Face recognition improvements (#16034)

* Start working on bird processor

* Initial setup for bird processing

* Improvements to handling

* Get classification working

* Cleanup classification

* Add classification config

* Update sort

.cspell/frigate-dictionary.txt

@@ -2,6 +2,7 @@ aarch
 absdiff
 airockchip
 Alloc
+alpr
 Amcrest
 amdgpu
 analyzeduration
@@ -12,6 +13,7 @@ argmax
 argmin
 argpartition
 ascontiguousarray
+astype
 authelia
 authentik
 autodetected
@@ -42,6 +44,7 @@ codeproject
 colormap
 colorspace
 comms
+coro
 ctypeslib
 CUDA
 Cuvid
@@ -59,6 +62,8 @@ dsize
 dtype
 ECONNRESET
 edgetpu
+facenet
+fastapi
 faststart
 fflags
 ffprobe
@@ -111,6 +116,8 @@ itemsize
 Jellyfin
 jetson
 jetsons
+jina
+jinaai
 joserfc
 jsmpeg
 jsonify
@@ -184,6 +191,7 @@ openai
 opencv
 openvino
 OWASP
+paddleocr
 paho
 passwordless
 popleft
@@ -193,6 +201,7 @@ poweroff
 preexec
 probesize
 protobuf
+pstate
 psutil
 pubkey
 putenv
@@ -212,6 +221,7 @@ rcond
 RDONLY
 rebranded
 referer
+reindex
 Reolink
 restream
 restreamed
@@ -236,6 +246,7 @@ sleeptime
 SNDMORE
 socs
 sqliteq
+sqlitevecq
 ssdlite
 statm
 stimeout
@@ -270,9 +281,11 @@ unraid
 unreviewed
 userdata
 usermod
+uvicorn
 vaapi
 vainfo
 variations
+vbios
 vconcat
 vitb
 vstream
@@ -300,4 +313,4 @@ yolo
 yolonas
 yolox
 zeep
-zerolatency
+zerolatency

.devcontainer/post_create.sh

@@ -3,10 +3,12 @@
 set -euxo pipefail
 
 # Cleanup the old github host key
-sed -i -e '/AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31\/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi\/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==/d' ~/.ssh/known_hosts
-# Add new github host key
-curl -L https://api.github.com/meta | jq -r '.ssh_keys | .[]' | \
-  sed -e 's/^/github.com /' >> ~/.ssh/known_hosts
+if [[ -f ~/.ssh/known_hosts ]]; then
+  # Add new github host key
+  sed -i -e '/AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31\/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi\/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==/d' ~/.ssh/known_hosts
+  curl -L https://api.github.com/meta | jq -r '.ssh_keys | .[]' | \
+    sed -e 's/^/github.com /' >> ~/.ssh/known_hosts
+fi
 
 # Frigate normal container runs as root, so it have permission to create
 # the folders. But the devcontainer runs as the host user, so we need to

.github/DISCUSSION_TEMPLATE/detector-support.yml

@@ -74,19 +74,6 @@ body:
         - CPU (no coral)
     validations:
       required: true
-  - type: dropdown
-    id: object-detector
-    attributes:
-      label: Object Detector
-      options:
-        - Coral
-        - OpenVino
-        - TensorRT
-        - RKNN
-        - Other
-        - CPU (no coral)
-    validations:
-      required: true
   - type: textarea
     id: screenshots
     attributes:

.github/DISCUSSION_TEMPLATE/general-support.yml

@@ -102,19 +102,6 @@ body:
         - CPU (no coral)
     validations:
       required: true
-  - type: dropdown
-    id: object-detector
-    attributes:
-      label: Object Detector
-      options:
-        - Coral
-        - OpenVino
-        - TensorRT
-        - RKNN
-        - Other
-        - CPU (no coral)
-    validations:
-      required: true
   - type: dropdown
     id: network
     attributes:

.github/pull_request_template.md

@@ -13,6 +13,7 @@
 - [ ] New feature
 - [ ] Breaking change (fix/feature causing existing functionality to break)
 - [ ] Code quality improvements to existing code
+- [ ] Documentation Update
 
 ## Additional information
 

.github/workflows/ci.yml

@@ -6,6 +6,8 @@ on:
     branches:
       - dev
       - master
+    paths-ignore:
+      - "docs/**"
 
 # only run the latest commit to avoid cache overwrites
 concurrency:
@@ -22,6 +24,8 @@ jobs:
     steps:
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - name: Set up QEMU and Buildx
         id: setup
         uses: ./.github/actions/setup
@@ -43,6 +47,8 @@ jobs:
     steps:
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - name: Set up QEMU and Buildx
         id: setup
         uses: ./.github/actions/setup
@@ -69,21 +75,14 @@ jobs:
             rpi.tags=${{ steps.setup.outputs.image-name }}-rpi
             *.cache-from=type=registry,ref=${{ steps.setup.outputs.cache-name }}-arm64
             *.cache-to=type=registry,ref=${{ steps.setup.outputs.cache-name }}-arm64,mode=max
-      - name: Build and push Rockchip build
-        uses: docker/bake-action@v3
-        with:
-          push: true
-          targets: rk
-          files: docker/rockchip/rk.hcl
-          set: |
-            rk.tags=${{ steps.setup.outputs.image-name }}-rk
-            *.cache-from=type=gha
   jetson_jp4_build:
     runs-on: ubuntu-latest
     name: Jetson Jetpack 4
     steps:
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - name: Set up QEMU and Buildx
         id: setup
         uses: ./.github/actions/setup
@@ -110,6 +109,8 @@ jobs:
     steps:
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - name: Set up QEMU and Buildx
         id: setup
         uses: ./.github/actions/setup
@@ -138,6 +139,8 @@ jobs:
     steps:
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - name: Set up QEMU and Buildx
         id: setup
         uses: ./.github/actions/setup
@@ -155,6 +158,30 @@ jobs:
             tensorrt.tags=${{ steps.setup.outputs.image-name }}-tensorrt
             *.cache-from=type=registry,ref=${{ steps.setup.outputs.cache-name }}-amd64
             *.cache-to=type=registry,ref=${{ steps.setup.outputs.cache-name }}-amd64,mode=max
+  arm64_extra_builds:
+    runs-on: ubuntu-latest
+    name: ARM Extra Build
+    needs:
+      - arm64_build
+    steps:
+      - name: Check out code
+        uses: actions/checkout@v4
+        with:
+          persist-credentials: false
+      - name: Set up QEMU and Buildx
+        id: setup
+        uses: ./.github/actions/setup
+        with:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      - name: Build and push Rockchip build
+        uses: docker/bake-action@v3
+        with:
+          push: true
+          targets: rk
+          files: docker/rockchip/rk.hcl
+          set: |
+            rk.tags=${{ steps.setup.outputs.image-name }}-rk
+            *.cache-from=type=gha
   combined_extra_builds:
     runs-on: ubuntu-latest
     name: Combined Extra Builds
@@ -164,6 +191,8 @@ jobs:
     steps:
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - name: Set up QEMU and Buildx
         id: setup
         uses: ./.github/actions/setup

.github/workflows/dependabot-auto-merge.yaml

@@ -1,24 +0,0 @@
-name: dependabot-auto-merge
-on: pull_request
-
-permissions:
-  contents: write
-
-jobs:
-  dependabot-auto-merge:
-    runs-on: ubuntu-latest
-    if: github.actor == 'dependabot[bot]'
-    steps:
-      - name: Get Dependabot metadata
-        id: metadata
-        uses: dependabot/fetch-metadata@v2
-        with:
-          github-token: ${{ secrets.GITHUB_TOKEN }}
-      - name: Enable auto-merge for Dependabot PRs
-        if: steps.metadata.outputs.dependency-type == 'direct:development' && (steps.metadata.outputs.update-type == 'version-update:semver-minor' || steps.metadata.outputs.update-type == 'version-update:semver-patch')
-        run: |
-          gh pr review --approve "$PR_URL"
-          gh pr merge --auto --squash "$PR_URL"
-        env:
-          PR_URL: ${{ github.event.pull_request.html_url }}
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/pull_request.yml

@@ -1,9 +1,12 @@
 name: On pull request
 
-on: pull_request
+on:
+  pull_request:
+    paths-ignore:
+      - "docs/**"
 
 env:
-  DEFAULT_PYTHON: 3.9
+  DEFAULT_PYTHON: 3.11
 
 jobs:
   build_devcontainer:
@@ -16,6 +19,8 @@ jobs:
       DOCKER_BUILDKIT: "1"
     steps:
       - uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - uses: actions/setup-node@master
         with:
           node-version: 16.x
@@ -35,6 +40,8 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - uses: actions/setup-node@master
         with:
           node-version: 16.x
@@ -49,6 +56,8 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - uses: actions/setup-node@master
         with:
           node-version: 20.x
@@ -64,8 +73,10 @@ jobs:
     steps:
       - name: Check out the repository
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - name: Set up Python ${{ env.DEFAULT_PYTHON }}
-        uses: actions/setup-python@v5.1.0
+        uses: actions/setup-python@v5.3.0
         with:
           python-version: ${{ env.DEFAULT_PYTHON }}
       - name: Install requirements
@@ -85,6 +96,8 @@ jobs:
     steps:
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - uses: actions/setup-node@master
         with:
           node-version: 16.x

.github/workflows/release.yml

@@ -11,6 +11,8 @@ jobs:
 
     steps:
       - uses: actions/checkout@v4
+        with:
+          persist-credentials: false
       - id: lowercaseRepo
         uses: ASzc/change-string-case-action@v6
         with:
@@ -22,10 +24,13 @@ jobs:
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
       - name: Create tag variables
+        env:
+          TAG: ${{ github.ref_name }}
+          LOWERCASE_REPO: ${{ steps.lowercaseRepo.outputs.lowercase }}
         run: |
-          BUILD_TYPE=$([[ "${{ github.ref_name }}" =~ ^v[0-9]+\.[0-9]+\.[0-9]+$ ]] && echo "stable" || echo "beta")
+          BUILD_TYPE=$([[ "${TAG}" =~ ^v[0-9]+\.[0-9]+\.[0-9]+$ ]] && echo "stable" || echo "beta")
           echo "BUILD_TYPE=${BUILD_TYPE}" >> $GITHUB_ENV
-          echo "BASE=ghcr.io/${{ steps.lowercaseRepo.outputs.lowercase }}" >> $GITHUB_ENV
+          echo "BASE=ghcr.io/${LOWERCASE_REPO}" >> $GITHUB_ENV
           echo "BUILD_TAG=${GITHUB_SHA::7}" >> $GITHUB_ENV
           echo "CLEAN_VERSION=$(echo ${GITHUB_REF##*/} | tr '[:upper:]' '[:lower:]' | sed 's/^[v]//')" >> $GITHUB_ENV
       - name: Tag and push the main image
@@ -34,14 +39,14 @@ jobs:
           STABLE_TAG=${BASE}:stable
           PULL_TAG=${BASE}:${BUILD_TAG}
           docker run --rm -v $HOME/.docker/config.json:/config.json quay.io/skopeo/stable:latest copy --authfile /config.json --multi-arch all docker://${PULL_TAG} docker://${VERSION_TAG}
-          for variant in standard-arm64 tensorrt tensorrt-jp4 tensorrt-jp5 rk; do
+          for variant in standard-arm64 tensorrt tensorrt-jp4 tensorrt-jp5 rk h8l rocm; do
             docker run --rm -v $HOME/.docker/config.json:/config.json quay.io/skopeo/stable:latest copy --authfile /config.json --multi-arch all docker://${PULL_TAG}-${variant} docker://${VERSION_TAG}-${variant}
           done
 
           # stable tag
           if [[ "${BUILD_TYPE}" == "stable" ]]; then
             docker run --rm -v $HOME/.docker/config.json:/config.json quay.io/skopeo/stable:latest copy --authfile /config.json --multi-arch all docker://${PULL_TAG} docker://${STABLE_TAG}
-            for variant in standard-arm64 tensorrt tensorrt-jp4 tensorrt-jp5 rk; do
+            for variant in standard-arm64 tensorrt tensorrt-jp4 tensorrt-jp5 rk h8l rocm; do
               docker run --rm -v $HOME/.docker/config.json:/config.json quay.io/skopeo/stable:latest copy --authfile /config.json --multi-arch all docker://${PULL_TAG}-${variant} docker://${STABLE_TAG}-${variant}
             done
           fi

.github/workflows/stale.yml

@@ -23,7 +23,9 @@ jobs:
           exempt-pr-labels: "pinned,security,dependencies"
           operations-per-run: 120
       - name: Print outputs
-        run: echo ${{ join(steps.stale.outputs.*, ',') }}
+        env:
+          STALE_OUTPUT: ${{ join(steps.stale.outputs.*, ',') }}
+        run: echo "$STALE_OUTPUT"
 
   # clean_ghcr:
   #   name: Delete outdated dev container images
@@ -38,4 +40,3 @@ jobs:
   #         account-type: personal
   #         token: ${{ secrets.GITHUB_TOKEN }}
   #         token-type: github-token
-

Makefile

@@ -1,7 +1,7 @@
 default_target: local
 
 COMMIT_HASH := $(shell git log -1 --pretty=format:"%h"|tail -1)
-VERSION = 0.15.0
+VERSION = 0.16.0
 IMAGE_REPO ?= ghcr.io/blakeblackshear/frigate
 GITHUB_REF_NAME ?= $(shell git rev-parse --abbrev-ref HEAD)
 BOARDS= #Initialized empty

benchmark.py

@@ -61,7 +61,7 @@ def start(id, num_detections, detection_queue, event):
     object_detector.cleanup()
     print(f"{id} - Processed for {duration:.2f} seconds.")
     print(f"{id} - FPS: {object_detector.fps.eps():.2f}")
-    print(f"{id} - Average frame processing time: {mean(frame_times)*1000:.2f}ms")
+    print(f"{id} - Average frame processing time: {mean(frame_times) * 1000:.2f}ms")
 
 
 ######

docker-compose.yml

@@ -23,7 +23,7 @@ services:
     #                     count: 1
     #                     capabilities: [gpu]
     environment:
-      YOLO_MODELS: yolov7-320
+      YOLO_MODELS: ""
     devices:
       - /dev/bus/usb:/dev/bus/usb
       # - /dev/dri:/dev/dri # for intel hwaccel, needs to be updated for your hardware

docker/hailo8l/Dockerfile

@@ -5,6 +5,7 @@ ARG DEBIAN_FRONTEND=noninteractive
 # Build Python wheels
 FROM wheels AS h8l-wheels
 
+RUN python3 -m pip config set global.break-system-packages true
 COPY docker/main/requirements-wheels.txt /requirements-wheels.txt
 COPY docker/hailo8l/requirements-wheels-h8l.txt /requirements-wheels-h8l.txt
 
@@ -16,89 +17,26 @@ RUN mkdir /h8l-wheels
 # Build the wheels
 RUN pip3 wheel --wheel-dir=/h8l-wheels -c /requirements-wheels.txt -r /requirements-wheels-h8l.txt
 
-# Build HailoRT and create wheel
-FROM wheels AS build-hailort
+FROM wget AS hailort
 ARG TARGETARCH
-
-SHELL ["/bin/bash", "-c"]
-
-# Install necessary APT packages
-RUN apt-get -qq update \
-    && apt-get -qq install -y \
-    apt-transport-https \
-    gnupg \
-    wget \
-    # the key fingerprint can be obtained from https://ftp-master.debian.org/keys.html
-    && wget -qO- "https://keyserver.ubuntu.com/pks/lookup?op=get&search=0xA4285295FC7B1A81600062A9605C66F00D6C9793" | \
-    gpg --dearmor > /usr/share/keyrings/debian-archive-bullseye-stable.gpg \
-    && echo "deb [signed-by=/usr/share/keyrings/debian-archive-bullseye-stable.gpg] http://deb.debian.org/debian bullseye main contrib non-free" | \
-    tee /etc/apt/sources.list.d/debian-bullseye-nonfree.list \
-    && apt-get -qq update \
-    && apt-get -qq install -y \
-    python3.9 \
-    python3.9-dev \
-    build-essential cmake git \
-    && rm -rf /var/lib/apt/lists/*
-
-# Extract Python version and set environment variables
-RUN PYTHON_VERSION=$(python3 --version 2>&1 | awk '{print $2}' | cut -d. -f1,2) && \
-    PYTHON_VERSION_NO_DOT=$(echo $PYTHON_VERSION | sed 's/\.//') && \
-    echo "PYTHON_VERSION=$PYTHON_VERSION" > /etc/environment && \
-    echo "PYTHON_VERSION_NO_DOT=$PYTHON_VERSION_NO_DOT" >> /etc/environment
-
-# Clone and build HailoRT
-RUN . /etc/environment && \
-    git clone https://github.com/hailo-ai/hailort.git /opt/hailort && \
-    cd /opt/hailort && \
-    git checkout v4.18.0 && \
-    cmake -H. -Bbuild -DCMAKE_BUILD_TYPE=Release -DHAILO_BUILD_PYBIND=1 -DPYBIND11_PYTHON_VERSION=${PYTHON_VERSION} && \
-    cmake --build build --config release --target libhailort && \
-    cmake --build build --config release --target _pyhailort && \
-    cp build/hailort/libhailort/bindings/python/src/_pyhailort.cpython-${PYTHON_VERSION_NO_DOT}-$(if [ $TARGETARCH == "amd64" ]; then echo 'x86_64'; else echo 'aarch64'; fi )-linux-gnu.so hailort/libhailort/bindings/python/platform/hailo_platform/pyhailort/ && \
-    cp build/hailort/libhailort/src/libhailort.so hailort/libhailort/bindings/python/platform/hailo_platform/pyhailort/
-
-RUN ls -ahl /opt/hailort/build/hailort/libhailort/src/
-RUN ls -ahl /opt/hailort/hailort/libhailort/bindings/python/platform/hailo_platform/pyhailort/
-
-# Remove the existing setup.py if it exists in the target directory
-RUN rm -f /opt/hailort/hailort/libhailort/bindings/python/platform/setup.py
-
-# Copy generate_wheel_conf.py and setup.py
-COPY docker/hailo8l/pyhailort_build_scripts/generate_wheel_conf.py /opt/hailort/hailort/libhailort/bindings/python/platform/generate_wheel_conf.py
-COPY docker/hailo8l/pyhailort_build_scripts/setup.py /opt/hailort/hailort/libhailort/bindings/python/platform/setup.py
-
-# Run the generate_wheel_conf.py script
-RUN python3 /opt/hailort/hailort/libhailort/bindings/python/platform/generate_wheel_conf.py
-
-# Create a wheel file using pip3 wheel
-RUN cd /opt/hailort/hailort/libhailort/bindings/python/platform && \
-    python3 setup.py bdist_wheel --dist-dir /hailo-wheels
+RUN --mount=type=bind,source=docker/hailo8l/install_hailort.sh,target=/deps/install_hailort.sh \
+    /deps/install_hailort.sh
 
 # Use deps as the base image
 FROM deps AS h8l-frigate
 
 # Copy the wheels from the wheels stage
 COPY --from=h8l-wheels /h8l-wheels /deps/h8l-wheels
-COPY --from=build-hailort /hailo-wheels /deps/hailo-wheels
-COPY --from=build-hailort /etc/environment /etc/environment
-RUN CC=$(python3 -c "import sysconfig; import shlex; cc = sysconfig.get_config_var('CC'); cc_cmd = shlex.split(cc)[0]; print(cc_cmd[:-4] if cc_cmd.endswith('-gcc') else cc_cmd)") && \
-    echo "CC=$CC" >> /etc/environment
+COPY --from=hailort /hailo-wheels /deps/hailo-wheels
+COPY --from=hailort /rootfs/ /
 
 # Install the wheels
+RUN python3 -m pip config set global.break-system-packages true
 RUN pip3 install -U /deps/h8l-wheels/*.whl
 RUN pip3 install -U /deps/hailo-wheels/*.whl
 
-RUN . /etc/environment && \
-    mv /usr/local/lib/python${PYTHON_VERSION}/dist-packages/hailo_platform/pyhailort/libhailort.so /usr/lib/${CC} && \
-    cd /usr/lib/${CC}/ && \
-    ln -s libhailort.so libhailort.so.4.18.0
-
 # Copy base files from the rootfs stage
 COPY --from=rootfs / /
 
-# Set environment variables for Hailo SDK
-ENV PATH="/opt/hailort/bin:${PATH}"
-ENV LD_LIBRARY_PATH="/usr/lib/$(if [ $TARGETARCH == "amd64" ]; then echo 'x86_64'; else echo 'aarch64'; fi )-linux-gnu:${LD_LIBRARY_PATH}"
-
 # Set workdir
 WORKDIR /opt/frigate/

docker/hailo8l/h8l.hcl

@@ -1,3 +1,9 @@
+target wget {
+  dockerfile = "docker/main/Dockerfile"
+  platforms = ["linux/arm64","linux/amd64"]
+  target = "wget"
+}
+
 target wheels {
   dockerfile = "docker/main/Dockerfile"
   platforms = ["linux/arm64","linux/amd64"]
@@ -19,6 +25,7 @@ target rootfs {
 target h8l {
   dockerfile = "docker/hailo8l/Dockerfile"
   contexts = {
+    wget = "target:wget"
     wheels = "target:wheels"
     deps = "target:deps"
     rootfs = "target:rootfs"

docker/hailo8l/install_hailort.sh

@@ -0,0 +1,19 @@
+#!/bin/bash
+
+set -euxo pipefail
+
+hailo_version="4.20.0"
+
+if [[ "${TARGETARCH}" == "amd64" ]]; then
+    arch="x86_64"
+elif [[ "${TARGETARCH}" == "arm64" ]]; then
+    arch="aarch64"
+fi
+
+wget -qO- "https://github.com/frigate-nvr/hailort/releases/download/v${hailo_version}/hailort-${TARGETARCH}.tar.gz" |
+    tar -C / -xzf -
+
+mkdir -p /hailo-wheels
+
+wget -P /hailo-wheels/ "https://github.com/frigate-nvr/hailort/releases/download/v${hailo_version}/hailort-${hailo_version}-cp311-cp311-linux_${arch}.whl"
+

docker/hailo8l/pyhailort_build_scripts/generate_wheel_conf.py

@@ -1,67 +0,0 @@
-import json
-import os
-import platform
-import sys
-import sysconfig
-
-
-def extract_toolchain_info(compiler):
-    # Remove the "-gcc" or "-g++" suffix if present
-    if compiler.endswith("-gcc") or compiler.endswith("-g++"):
-        compiler = compiler.rsplit("-", 1)[0]
-
-    # Extract the toolchain and ABI part (e.g., "gnu")
-    toolchain_parts = compiler.split("-")
-    abi_conventions = next(
-        (part for part in toolchain_parts if part in ["gnu", "musl", "eabi", "uclibc"]),
-        "",
-    )
-
-    return abi_conventions
-
-
-def generate_wheel_conf():
-    conf_file_path = os.path.join(
-        os.path.abspath(os.path.dirname(__file__)), "wheel_conf.json"
-    )
-
-    # Extract current system and Python version information
-    py_version = f"cp{sys.version_info.major}{sys.version_info.minor}"
-    arch = platform.machine()
-    system = platform.system().lower()
-    libc_version = platform.libc_ver()[1]
-
-    # Get the compiler information
-    compiler = sysconfig.get_config_var("CC")
-    abi_conventions = extract_toolchain_info(compiler)
-
-    # Create the new configuration data
-    new_conf_data = {
-        "py_version": py_version,
-        "arch": arch,
-        "system": system,
-        "libc_version": libc_version,
-        "abi": abi_conventions,
-        "extension": {
-            "posix": "so",
-            "nt": "pyd",  # Windows
-        }[os.name],
-    }
-
-    # If the file exists, load the existing data
-    if os.path.isfile(conf_file_path):
-        with open(conf_file_path, "r") as conf_file:
-            conf_data = json.load(conf_file)
-        # Update the existing data with the new data
-        conf_data.update(new_conf_data)
-    else:
-        # If the file does not exist, use the new data
-        conf_data = new_conf_data
-
-    # Write the updated data to the file
-    with open(conf_file_path, "w") as conf_file:
-        json.dump(conf_data, conf_file, indent=4)
-
-
-if __name__ == "__main__":
-    generate_wheel_conf()

docker/hailo8l/pyhailort_build_scripts/setup.py

@@ -1,111 +0,0 @@
-import json
-import os
-
-from setuptools import find_packages, setup
-from wheel.bdist_wheel import bdist_wheel as orig_bdist_wheel
-
-
-class NonPurePythonBDistWheel(orig_bdist_wheel):
-    """Makes the wheel platform-dependent so it can be based on the _pyhailort architecture"""
-
-    def finalize_options(self):
-        orig_bdist_wheel.finalize_options(self)
-        self.root_is_pure = False
-
-
-def _get_hailort_lib_path():
-    lib_filename = "libhailort.so"
-    lib_path = os.path.join(
-        os.path.abspath(os.path.dirname(__file__)),
-        f"hailo_platform/pyhailort/{lib_filename}",
-    )
-    if os.path.exists(lib_path):
-        print(f"Found libhailort shared library at: {lib_path}")
-    else:
-        print(f"Error: libhailort shared library not found at: {lib_path}")
-        raise FileNotFoundError(f"libhailort shared library not found at: {lib_path}")
-    return lib_path
-
-
-def _get_pyhailort_lib_path():
-    conf_file_path = os.path.join(
-        os.path.abspath(os.path.dirname(__file__)), "wheel_conf.json"
-    )
-    if not os.path.isfile(conf_file_path):
-        raise FileNotFoundError(f"Configuration file not found: {conf_file_path}")
-
-    with open(conf_file_path, "r") as conf_file:
-        content = json.load(conf_file)
-        py_version = content["py_version"]
-        arch = content["arch"]
-        system = content["system"]
-        extension = content["extension"]
-        abi = content["abi"]
-
-        # Construct the filename directly
-        lib_filename = f"_pyhailort.cpython-{py_version.split('cp')[1]}-{arch}-{system}-{abi}.{extension}"
-        lib_path = os.path.join(
-            os.path.abspath(os.path.dirname(__file__)),
-            f"hailo_platform/pyhailort/{lib_filename}",
-        )
-
-        if os.path.exists(lib_path):
-            print(f"Found _pyhailort shared library at: {lib_path}")
-        else:
-            print(f"Error: _pyhailort shared library not found at: {lib_path}")
-            raise FileNotFoundError(
-                f"_pyhailort shared library not found at: {lib_path}"
-            )
-
-        return lib_path
-
-
-def _get_package_paths():
-    packages = []
-    pyhailort_lib = _get_pyhailort_lib_path()
-    hailort_lib = _get_hailort_lib_path()
-    if pyhailort_lib:
-        packages.append(pyhailort_lib)
-    if hailort_lib:
-        packages.append(hailort_lib)
-    packages.append(os.path.abspath("hailo_tutorials/notebooks/*"))
-    packages.append(os.path.abspath("hailo_tutorials/hefs/*"))
-    return packages
-
-
-if __name__ == "__main__":
-    setup(
-        author="Hailo team",
-        author_email="contact@hailo.ai",
-        cmdclass={
-            "bdist_wheel": NonPurePythonBDistWheel,
-        },
-        description="HailoRT",
-        entry_points={
-            "console_scripts": [
-                "hailo=hailo_platform.tools.hailocli.main:main",
-            ]
-        },
-        install_requires=[
-            "argcomplete",
-            "contextlib2",
-            "future",
-            "netaddr",
-            "netifaces",
-            "verboselogs",
-            "numpy==1.23.3",
-        ],
-        name="hailort",
-        package_data={
-            "hailo_platform": _get_package_paths(),
-        },
-        packages=find_packages(),
-        platforms=[
-            "linux_x86_64",
-            "linux_aarch64",
-            "win_amd64",
-        ],
-        url="https://hailo.ai/",
-        version="4.17.0",
-        zip_safe=False,
-    )

docker/hailo8l/requirements-wheels-h8l.txt

@@ -1,12 +1,12 @@
-appdirs==1.4.4
-argcomplete==2.0.0
-contextlib2==0.6.0.post1
-distlib==0.3.6
-filelock==3.8.0
-future==0.18.2
-importlib-metadata==5.1.0
-importlib-resources==5.1.2
-netaddr==0.8.0
-netifaces==0.10.9
-verboselogs==1.7
-virtualenv==20.17.0
+appdirs==1.4.*
+argcomplete==2.0.*
+contextlib2==0.6.*
+distlib==0.3.*
+filelock==3.8.*
+future==0.18.*
+importlib-metadata==5.1.*
+importlib-resources==5.1.*
+netaddr==0.8.*
+netifaces==0.10.*
+verboselogs==1.7.*
+virtualenv==20.17.*

docker/hailo8l/user_installation.sh

@@ -4,6 +4,7 @@
 sudo apt-get update
 sudo apt-get install -y build-essential cmake git wget
 
+hailo_version="4.20.0"
 arch=$(uname -m)
 
 if [[ $arch == "x86_64" ]]; then
@@ -13,7 +14,7 @@ else
 fi
 
 # Clone the HailoRT driver repository
-git clone --depth 1 --branch v4.18.0 https://github.com/hailo-ai/hailort-drivers.git
+git clone --depth 1 --branch v${hailo_version} https://github.com/hailo-ai/hailort-drivers.git
 
 # Build and install the HailoRT driver
 cd hailort-drivers/linux/pcie
@@ -38,7 +39,7 @@ cd ../../
 if [ ! -d /lib/firmware/hailo ]; then
   sudo mkdir /lib/firmware/hailo
 fi
-sudo mv hailo8_fw.4.17.0.bin /lib/firmware/hailo/hailo8_fw.bin
+sudo mv hailo8_fw.*.bin /lib/firmware/hailo/hailo8_fw.bin
 
 # Install udev rules
 sudo cp ./linux/pcie/51-hailo-udev.rules /etc/udev/rules.d/

docker/main/Dockerfile

@@ -3,12 +3,12 @@
 # https://askubuntu.com/questions/972516/debian-frontend-environment-variable
 ARG DEBIAN_FRONTEND=noninteractive
 
-ARG BASE_IMAGE=debian:11
-ARG SLIM_BASE=debian:11-slim
+ARG BASE_IMAGE=debian:12
+ARG SLIM_BASE=debian:12-slim
 
 FROM ${BASE_IMAGE} AS base
 
-FROM --platform=${BUILDPLATFORM} debian:11 AS base_host
+FROM --platform=${BUILDPLATFORM} debian:12 AS base_host
 
 FROM ${SLIM_BASE} AS slim-base
 
@@ -66,8 +66,8 @@ COPY docker/main/requirements-ov.txt /requirements-ov.txt
 RUN apt-get -qq update \
     && apt-get -qq install -y wget python3 python3-dev python3-distutils gcc pkg-config libhdf5-dev \
     && wget -q https://bootstrap.pypa.io/get-pip.py -O get-pip.py \
-    && python3 get-pip.py "pip" \
-    && pip install -r /requirements-ov.txt
+    && python3 get-pip.py "pip" --break-system-packages \
+    && pip install --break-system-packages -r /requirements-ov.txt
 
 # Get OpenVino Model
 RUN --mount=type=bind,source=docker/main/build_ov_model.py,target=/build_ov_model.py \
@@ -139,24 +139,17 @@ ARG TARGETARCH
 # Use a separate container to build wheels to prevent build dependencies in final image
 RUN apt-get -qq update \
     && apt-get -qq install -y \
-    apt-transport-https \
-    gnupg \
-    wget \
-    # the key fingerprint can be obtained from https://ftp-master.debian.org/keys.html
-    && wget -qO- "https://keyserver.ubuntu.com/pks/lookup?op=get&search=0xA4285295FC7B1A81600062A9605C66F00D6C9793" | \
-    gpg --dearmor > /usr/share/keyrings/debian-archive-bullseye-stable.gpg \
-    && echo "deb [signed-by=/usr/share/keyrings/debian-archive-bullseye-stable.gpg] http://deb.debian.org/debian bullseye main contrib non-free" | \
-    tee /etc/apt/sources.list.d/debian-bullseye-nonfree.list \
+    apt-transport-https wget \
     && apt-get -qq update \
     && apt-get -qq install -y \
-    python3.9 \
-    python3.9-dev \
+    python3 \
+    python3-dev \
     # opencv dependencies
     build-essential cmake git pkg-config libgtk-3-dev \
     libavcodec-dev libavformat-dev libswscale-dev libv4l-dev \
     libxvidcore-dev libx264-dev libjpeg-dev libpng-dev libtiff-dev \
     gfortran openexr libatlas-base-dev libssl-dev\
-    libtbb2 libtbb-dev libdc1394-22-dev libopenexr-dev \
+    libtbbmalloc2 libtbb-dev libdc1394-dev libopenexr-dev \
     libgstreamer-plugins-base1.0-dev libgstreamer1.0-dev \
     # sqlite3 dependencies
     tclsh \
@@ -164,14 +157,11 @@ RUN apt-get -qq update \
     gcc gfortran libopenblas-dev liblapack-dev && \
     rm -rf /var/lib/apt/lists/*
 
-# Ensure python3 defaults to python3.9
-RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.9 1
-
 RUN wget -q https://bootstrap.pypa.io/get-pip.py -O get-pip.py \
-    && python3 get-pip.py "pip"
+    && python3 get-pip.py "pip" --break-system-packages
 
 COPY docker/main/requirements.txt /requirements.txt
-RUN pip3 install -r /requirements.txt
+RUN pip3 install -r /requirements.txt --break-system-packages
 
 # Build pysqlite3 from source
 COPY docker/main/build_pysqlite3.sh /build_pysqlite3.sh
@@ -180,9 +170,6 @@ RUN /build_pysqlite3.sh
 COPY docker/main/requirements-wheels.txt /requirements-wheels.txt
 RUN pip3 wheel --wheel-dir=/wheels -r /requirements-wheels.txt
 
-COPY docker/main/requirements-wheels-post.txt /requirements-wheels-post.txt
-RUN pip3 wheel --no-deps --wheel-dir=/wheels-post -r /requirements-wheels-post.txt
-
 
 # Collect deps in a single layer
 FROM scratch AS deps-rootfs
@@ -214,6 +201,9 @@ ENV TOKENIZERS_PARALLELISM=true
 # https://github.com/huggingface/transformers/issues/27214
 ENV TRANSFORMERS_NO_ADVISORY_WARNINGS=1
 
+# Set OpenCV ffmpeg loglevel to fatal: https://ffmpeg.org/doxygen/trunk/log_8h.html
+ENV OPENCV_FFMPEG_LOGLEVEL=8
+
 ENV PATH="/usr/local/go2rtc/bin:/usr/local/tempio/bin:/usr/local/nginx/sbin:${PATH}"
 ENV LIBAVFORMAT_VERSION_MAJOR=60
 
@@ -222,16 +212,8 @@ RUN --mount=type=bind,source=docker/main/install_deps.sh,target=/deps/install_de
     /deps/install_deps.sh
 
 RUN --mount=type=bind,from=wheels,source=/wheels,target=/deps/wheels \
-    python3 -m pip install --upgrade pip && \
-    pip3 install -U /deps/wheels/*.whl
-
-# We have to uninstall this dependency specifically
-# as it will break onnxruntime-openvino
-RUN pip3 uninstall -y onnxruntime
-
-RUN --mount=type=bind,from=wheels,source=/wheels-post,target=/deps/wheels \
-    python3 -m pip install --upgrade pip && \
-    pip3 install -U /deps/wheels/*.whl
+    python3 -m pip install --upgrade pip --break-system-packages && \
+    pip3 install -U /deps/wheels/*.whl --break-system-packages
 
 COPY --from=deps-rootfs / /
 
@@ -278,7 +260,7 @@ RUN apt-get update \
     && rm -rf /var/lib/apt/lists/*
 
 RUN --mount=type=bind,source=./docker/main/requirements-dev.txt,target=/workspace/frigate/requirements-dev.txt \
-    pip3 install -r requirements-dev.txt
+    pip3 install -r requirements-dev.txt --break-system-packages
 
 HEALTHCHECK NONE
 

docker/main/build_nginx.sh

@@ -8,8 +8,7 @@ SECURE_TOKEN_MODULE_VERSION="1.5"
 SET_MISC_MODULE_VERSION="v0.33"
 NGX_DEVEL_KIT_VERSION="v0.3.3"
 
-cp /etc/apt/sources.list /etc/apt/sources.list.d/sources-src.list
-sed -i 's|deb http|deb-src http|g' /etc/apt/sources.list.d/sources-src.list
+sed -i '/^Types:/s/deb/& deb-src/' /etc/apt/sources.list.d/debian.sources
 apt-get update
 
 apt-get -yqq build-dep nginx

docker/main/build_ov_model.py

@@ -4,7 +4,7 @@ from openvino.tools import mo
 ov_model = mo.convert_model(
     "/models/ssdlite_mobilenet_v2_coco_2018_05_09/frozen_inference_graph.pb",
     compress_to_fp16=True,
-    transformations_config="/usr/local/lib/python3.9/dist-packages/openvino/tools/mo/front/tf/ssd_v2_support.json",
+    transformations_config="/usr/local/lib/python3.11/dist-packages/openvino/tools/mo/front/tf/ssd_v2_support.json",
     tensorflow_object_detection_api_pipeline_config="/models/ssdlite_mobilenet_v2_coco_2018_05_09/pipeline.config",
     reverse_input_channels=True,
 )

docker/main/build_sqlite_vec.sh

@@ -4,8 +4,7 @@ set -euxo pipefail
 
 SQLITE_VEC_VERSION="0.1.3"
 
-cp /etc/apt/sources.list /etc/apt/sources.list.d/sources-src.list
-sed -i 's|deb http|deb-src http|g' /etc/apt/sources.list.d/sources-src.list
+sed -i '/^Types:/s/deb/& deb-src/' /etc/apt/sources.list.d/debian.sources
 apt-get update
 apt-get -yqq build-dep sqlite3 gettext git
 

docker/main/install_deps.sh

@@ -8,35 +8,37 @@ apt-get -qq install --no-install-recommends -y \
     apt-transport-https \
     gnupg \
     wget \
+    lbzip2 \
     procps vainfo \
     unzip locales tzdata libxml2 xz-utils \
-    python3.9 \
+    python3 \
     python3-pip \
     curl \
     lsof \
     jq \
-    nethogs
-
-# ensure python3 defaults to python3.9
-update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.9 1
+    nethogs \
+    libgl1 \
+    libglib2.0-0 \
+    libusb-1.0.0
 
 mkdir -p -m 600 /root/.gnupg
 
-# add coral repo
-curl -fsSLo - https://packages.cloud.google.com/apt/doc/apt-key.gpg | \
-    gpg --dearmor -o /etc/apt/trusted.gpg.d/google-cloud-packages-archive-keyring.gpg
-echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | tee /etc/apt/sources.list.d/coral-edgetpu.list
-echo "libedgetpu1-max libedgetpu/accepted-eula select true" | debconf-set-selections
+# install coral runtime
+wget -q -O /tmp/libedgetpu1-max.deb "https://github.com/feranick/libedgetpu/releases/download/16.0TF2.17.0-1/libedgetpu1-max_16.0tf2.17.0-1.bookworm_${TARGETARCH}.deb"
+unset DEBIAN_FRONTEND
+yes | dpkg -i /tmp/libedgetpu1-max.deb && export DEBIAN_FRONTEND=noninteractive
+rm /tmp/libedgetpu1-max.deb
 
-# enable non-free repo in Debian
-if grep -q "Debian" /etc/issue; then
-    sed -i -e's/ main/ main contrib non-free/g' /etc/apt/sources.list
+# install python3 & tflite runtime
+if [[ "${TARGETARCH}" == "amd64" ]]; then
+    pip3 install --break-system-packages https://github.com/feranick/TFlite-builds/releases/download/v2.17.0/tflite_runtime-2.17.0-cp311-cp311-linux_x86_64.whl
+    pip3 install --break-system-packages https://github.com/feranick/pycoral/releases/download/2.0.2TF2.17.0/pycoral-2.0.2-cp311-cp311-linux_x86_64.whl
 fi
 
-# coral drivers
-apt-get -qq update
-apt-get -qq install --no-install-recommends --no-install-suggests -y \
-    libedgetpu1-max python3-tflite-runtime python3-pycoral
+if [[ "${TARGETARCH}" == "arm64" ]]; then
+    pip3 install --break-system-packages https://github.com/feranick/TFlite-builds/releases/download/v2.17.0/tflite_runtime-2.17.0-cp311-cp311-linux_aarch64.whl
+    pip3 install --break-system-packages https://github.com/feranick/pycoral/releases/download/2.0.2TF2.17.0/pycoral-2.0.2-cp311-cp311-linux_aarch64.whl
+fi
 
 # btbn-ffmpeg -> amd64
 if [[ "${TARGETARCH}" == "amd64" ]]; then
@@ -45,7 +47,7 @@ if [[ "${TARGETARCH}" == "amd64" ]]; then
     wget -qO btbn-ffmpeg.tar.xz "https://github.com/NickM-27/FFmpeg-Builds/releases/download/autobuild-2022-07-31-12-37/ffmpeg-n5.1-2-g915ef932a3-linux64-gpl-5.1.tar.xz"
     tar -xf btbn-ffmpeg.tar.xz -C /usr/lib/ffmpeg/5.0 --strip-components 1
     rm -rf btbn-ffmpeg.tar.xz /usr/lib/ffmpeg/5.0/doc /usr/lib/ffmpeg/5.0/bin/ffplay
-    wget -qO btbn-ffmpeg.tar.xz "https://github.com/BtbN/FFmpeg-Builds/releases/download/autobuild-2024-09-30-15-36/ffmpeg-n7.1-linux64-gpl-7.1.tar.xz"
+    wget -qO btbn-ffmpeg.tar.xz "https://github.com/NickM-27/FFmpeg-Builds/releases/download/autobuild-2024-09-19-12-51/ffmpeg-n7.0.2-18-g3e6cec1286-linux64-gpl-7.0.tar.xz"
     tar -xf btbn-ffmpeg.tar.xz -C /usr/lib/ffmpeg/7.0 --strip-components 1
     rm -rf btbn-ffmpeg.tar.xz /usr/lib/ffmpeg/7.0/doc /usr/lib/ffmpeg/7.0/bin/ffplay
 fi
@@ -57,34 +59,29 @@ if [[ "${TARGETARCH}" == "arm64" ]]; then
     wget -qO btbn-ffmpeg.tar.xz "https://github.com/NickM-27/FFmpeg-Builds/releases/download/autobuild-2022-07-31-12-37/ffmpeg-n5.1-2-g915ef932a3-linuxarm64-gpl-5.1.tar.xz"
     tar -xf btbn-ffmpeg.tar.xz -C /usr/lib/ffmpeg/5.0 --strip-components 1
     rm -rf btbn-ffmpeg.tar.xz /usr/lib/ffmpeg/5.0/doc /usr/lib/ffmpeg/5.0/bin/ffplay
-    wget -qO btbn-ffmpeg.tar.xz "https://github.com/BtbN/FFmpeg-Builds/releases/download/autobuild-2024-09-30-15-36/ffmpeg-n7.1-linuxarm64-gpl-7.1.tar.xz"
+    wget -qO btbn-ffmpeg.tar.xz "https://github.com/NickM-27/FFmpeg-Builds/releases/download/autobuild-2024-09-19-12-51/ffmpeg-n7.0.2-18-g3e6cec1286-linuxarm64-gpl-7.0.tar.xz"
     tar -xf btbn-ffmpeg.tar.xz -C /usr/lib/ffmpeg/7.0 --strip-components 1
     rm -rf btbn-ffmpeg.tar.xz /usr/lib/ffmpeg/7.0/doc /usr/lib/ffmpeg/7.0/bin/ffplay
 fi
 
 # arch specific packages
 if [[ "${TARGETARCH}" == "amd64" ]]; then
-    # use debian bookworm for amd / intel-i965 driver packages
-    echo 'deb https://deb.debian.org/debian bookworm main contrib non-free' >/etc/apt/sources.list.d/debian-bookworm.list
-    apt-get -qq update
+    # install amd / intel-i965 driver packages
     apt-get -qq install --no-install-recommends --no-install-suggests -y \
         i965-va-driver intel-gpu-tools onevpl-tools \
         libva-drm2 \
         mesa-va-drivers radeontop
 
-    # something about this dependency requires it to be installed in a separate call rather than in the line above
-    apt-get -qq install --no-install-recommends --no-install-suggests -y \
-        i965-va-driver-shaders
-
-    rm -f /etc/apt/sources.list.d/debian-bookworm.list
+    # intel packages use zst compression so we need to update dpkg
+    apt-get install -y dpkg
 
     # use intel apt intel packages
     wget -qO - https://repositories.intel.com/gpu/intel-graphics.key | gpg --yes --dearmor --output /usr/share/keyrings/intel-graphics.gpg
     echo "deb [arch=amd64 signed-by=/usr/share/keyrings/intel-graphics.gpg] https://repositories.intel.com/gpu/ubuntu jammy client" | tee /etc/apt/sources.list.d/intel-gpu-jammy.list
     apt-get -qq update
     apt-get -qq install --no-install-recommends --no-install-suggests -y \
-        intel-opencl-icd intel-level-zero-gpu intel-media-va-driver-non-free \
-        libmfx1 libmfxgen1 libvpl2
+        intel-opencl-icd=24.35.30872.31-996~22.04 intel-level-zero-gpu=1.3.29735.27-914~22.04 intel-media-va-driver-non-free=24.3.3-996~22.04 \
+        libmfx1=23.2.2-880~22.04 libmfxgen1=24.2.4-914~22.04 libvpl2=1:2.13.0.0-996~22.04
 
     rm -f /usr/share/keyrings/intel-graphics.gpg
     rm -f /etc/apt/sources.list.d/intel-gpu-jammy.list

docker/main/requirements-wheels-post.txt

@@ -1,3 +0,0 @@
-# ONNX
-onnxruntime-openvino == 1.19.* ; platform_machine == 'x86_64'
-onnxruntime == 1.19.* ; platform_machine == 'aarch64'

docker/main/requirements-wheels.txt

@@ -1,40 +1,47 @@
 click == 8.1.*
 # FastAPI
+aiohttp == 3.11.2
+starlette == 0.41.2
 starlette-context == 0.3.6
-fastapi == 0.115.0
+fastapi == 0.115.*
 uvicorn == 0.30.*
-slowapi == 0.1.9
+slowapi == 0.1.*
 imutils == 0.5.*
 joserfc == 1.0.*
 pathvalidate == 3.2.*
 markupsafe == 2.1.*
+python-multipart == 0.0.12
+# General
 mypy == 1.6.1
-numpy == 1.26.*
-onvif_zeep == 0.2.12
-opencv-python-headless == 4.9.0.*
+onvif-zeep-async == 3.1.*
 paho-mqtt == 2.1.*
 pandas == 2.2.*
 peewee == 3.17.*
 peewee_migrate == 1.13.*
-psutil == 5.9.*
+psutil == 6.1.*
 pydantic == 2.8.*
 git+https://github.com/fbcotter/py3nvml#egg=py3nvml
-pytz == 2024.1
+pytz == 2024.*
 pyzmq == 26.2.*
 ruamel.yaml == 0.18.*
 tzlocal == 5.2
 requests == 2.32.*
 types-requests == 2.32.*
-scipy == 1.13.*
 norfair == 2.2.*
 setproctitle == 1.3.*
 ws4py == 0.5.*
 unidecode == 1.3.*
-# OpenVino (ONNX installed in wheels-post)
-openvino == 2024.3.*
+# Image Manipulation
+numpy == 1.26.*
+opencv-python-headless == 4.10.0.*
+opencv-contrib-python == 4.9.0.*
+scipy == 1.14.*
+# OpenVino & ONNX
+openvino == 2024.4.*
+onnxruntime-openvino == 1.20.* ; platform_machine == 'x86_64'
+onnxruntime == 1.20.* ; platform_machine == 'aarch64'
 # Embeddings
 transformers == 4.45.*
-onnx_clip == 4.0.*
 # Generative AI
 google-generativeai == 0.8.*
 ollama == 0.3.*
@@ -42,3 +49,6 @@ openai == 1.51.*
 # push notifications
 py-vapid == 1.9.*
 pywebpush == 2.0.*
+# alpr
+pyclipper == 1.3.*
+shapely == 2.0.*

docker/main/requirements.txt

@@ -1,2 +1,2 @@
-scikit-build == 0.17.*
+scikit-build == 0.18.*
 nvidia-pyindex

docker/main/rootfs/usr/local/go2rtc/create_config.py

@@ -165,7 +165,7 @@ if config.get("birdseye", {}).get("restream", False):
     birdseye: dict[str, any] = config.get("birdseye")
 
     input = f"-f rawvideo -pix_fmt yuv420p -video_size {birdseye.get('width', 1280)}x{birdseye.get('height', 720)} -r 10 -i {BIRDSEYE_PIPE}"
-    ffmpeg_cmd = f"exec:{parse_preset_hardware_acceleration_encode(ffmpeg_path, config.get('ffmpeg', {}).get('hwaccel_args'), input, '-rtsp_transport tcp -f rtsp {output}')}"
+    ffmpeg_cmd = f"exec:{parse_preset_hardware_acceleration_encode(ffmpeg_path, config.get('ffmpeg', {}).get('hwaccel_args', ''), input, '-rtsp_transport tcp -f rtsp {output}')}"
 
     if go2rtc_config.get("streams"):
         go2rtc_config["streams"]["birdseye"] = ffmpeg_cmd

docker/main/rootfs/usr/local/nginx/conf/nginx.conf

@@ -81,6 +81,9 @@ http {
         open_file_cache_errors on;
         aio on;
 
+        # file upload size
+        client_max_body_size 10M;
+
         # https://github.com/kaltura/nginx-vod-module#vod_open_file_thread_pool
         vod_open_file_thread_pool default;
 

docker/rockchip/COCO/coco_subset_20.txt

@@ -0,0 +1,20 @@
+./subset/000000005001.jpg
+./subset/000000038829.jpg
+./subset/000000052891.jpg
+./subset/000000075612.jpg
+./subset/000000098261.jpg
+./subset/000000181542.jpg
+./subset/000000215245.jpg
+./subset/000000277005.jpg
+./subset/000000288685.jpg
+./subset/000000301421.jpg
+./subset/000000334371.jpg
+./subset/000000348481.jpg
+./subset/000000373353.jpg
+./subset/000000397681.jpg
+./subset/000000414673.jpg
+./subset/000000419312.jpg
+./subset/000000465822.jpg
+./subset/000000475732.jpg
+./subset/000000559707.jpg
+./subset/000000574315.jpg

docker/rockchip/Dockerfile

@@ -7,21 +7,26 @@ FROM wheels as rk-wheels
 COPY docker/main/requirements-wheels.txt /requirements-wheels.txt
 COPY docker/rockchip/requirements-wheels-rk.txt /requirements-wheels-rk.txt
 RUN sed -i "/https:\/\//d" /requirements-wheels.txt
+RUN sed -i "/onnxruntime/d" /requirements-wheels.txt
+RUN python3 -m pip config set global.break-system-packages true
 RUN pip3 wheel --wheel-dir=/rk-wheels -c /requirements-wheels.txt -r /requirements-wheels-rk.txt
+RUN rm -rf /rk-wheels/opencv_python-*
 
 FROM deps AS rk-frigate
 ARG TARGETARCH
 
 RUN --mount=type=bind,from=rk-wheels,source=/rk-wheels,target=/deps/rk-wheels \
-    pip3 install -U /deps/rk-wheels/*.whl
+    pip3 install --no-deps -U /deps/rk-wheels/*.whl --break-system-packages
 
 WORKDIR /opt/frigate/
 COPY --from=rootfs / /
+COPY docker/rockchip/COCO /COCO
+COPY docker/rockchip/conv2rknn.py /opt/conv2rknn.py
 
-ADD https://github.com/MarcA711/rknn-toolkit2/releases/download/v2.0.0/librknnrt.so /usr/lib/
+ADD https://github.com/MarcA711/rknn-toolkit2/releases/download/v2.3.0/librknnrt.so /usr/lib/
 
 RUN rm -rf /usr/lib/btbn-ffmpeg/bin/ffmpeg
 RUN rm -rf /usr/lib/btbn-ffmpeg/bin/ffprobe
-ADD --chmod=111 https://github.com/MarcA711/Rockchip-FFmpeg-Builds/releases/download/6.1-5/ffmpeg /usr/lib/ffmpeg/6.0/bin/
-ADD --chmod=111 https://github.com/MarcA711/Rockchip-FFmpeg-Builds/releases/download/6.1-5/ffprobe /usr/lib/ffmpeg/6.0/bin/
+ADD --chmod=111 https://github.com/MarcA711/Rockchip-FFmpeg-Builds/releases/download/6.1-6/ffmpeg /usr/lib/ffmpeg/6.0/bin/
+ADD --chmod=111 https://github.com/MarcA711/Rockchip-FFmpeg-Builds/releases/download/6.1-6/ffprobe /usr/lib/ffmpeg/6.0/bin/
 ENV PATH="/usr/lib/ffmpeg/6.0/bin/:${PATH}"

docker/rockchip/conv2rknn.py

@@ -0,0 +1,82 @@
+import os
+
+import rknn
+import yaml
+from rknn.api import RKNN
+
+try:
+    with open(rknn.__path__[0] + "/VERSION") as file:
+        tk_version = file.read().strip()
+except FileNotFoundError:
+    pass
+
+try:
+    with open("/config/conv2rknn.yaml", "r") as config_file:
+        configuration = yaml.safe_load(config_file)
+except FileNotFoundError:
+    raise Exception("Please place a config.yaml file in /config/conv2rknn.yaml")
+
+if configuration["config"] != None:
+    rknn_config = configuration["config"]
+else:
+    rknn_config = {}
+
+if not os.path.isdir("/config/model_cache/rknn_cache/onnx"):
+    raise Exception(
+        "Place the onnx models you want to convert to rknn format in /config/model_cache/rknn_cache/onnx"
+    )
+
+if "soc" not in configuration:
+    try:
+        with open("/proc/device-tree/compatible") as file:
+            soc = file.read().split(",")[-1].strip("\x00")
+    except FileNotFoundError:
+        raise Exception("Make sure to run docker in privileged mode.")
+
+    configuration["soc"] = [
+        soc,
+    ]
+
+if "quantization" not in configuration:
+    configuration["quantization"] = False
+
+if "output_name" not in configuration:
+    configuration["output_name"] = "{{input_basename}}"
+
+for input_filename in os.listdir("/config/model_cache/rknn_cache/onnx"):
+    for soc in configuration["soc"]:
+        quant = "i8" if configuration["quantization"] else "fp16"
+
+        input_path = "/config/model_cache/rknn_cache/onnx/" + input_filename
+        input_basename = input_filename[: input_filename.rfind(".")]
+
+        output_filename = (
+            configuration["output_name"].format(
+                quant=quant,
+                input_basename=input_basename,
+                soc=soc,
+                tk_version=tk_version,
+            )
+            + ".rknn"
+        )
+        output_path = "/config/model_cache/rknn_cache/" + output_filename
+
+        rknn_config["target_platform"] = soc
+
+        rknn = RKNN(verbose=True)
+        rknn.config(**rknn_config)
+
+        if rknn.load_onnx(model=input_path) != 0:
+            raise Exception("Error loading model.")
+
+        if (
+            rknn.build(
+                do_quantization=configuration["quantization"],
+                dataset="/COCO/coco_subset_20.txt",
+            )
+            != 0
+        ):
+            raise Exception("Error building model.")
+
+        if rknn.export_rknn(output_path) != 0:
+            raise Exception("Error exporting rknn model.")

docker/rockchip/requirements-wheels-rk.txt

@@ -1 +1,2 @@
-rknn-toolkit-lite2 @ https://github.com/MarcA711/rknn-toolkit2/releases/download/v2.0.0/rknn_toolkit_lite2-2.0.0b0-cp39-cp39-linux_aarch64.whl
+rknn-toolkit2 == 2.3.0
+rknn-toolkit-lite2 == 2.3.0

docker/rocm/Dockerfile

@@ -34,7 +34,7 @@ RUN mkdir -p /opt/rocm-dist/etc/ld.so.conf.d/
 RUN echo /opt/rocm/lib|tee /opt/rocm-dist/etc/ld.so.conf.d/rocm.conf
 
 #######################################################################
-FROM --platform=linux/amd64 debian:11 as debian-base
+FROM --platform=linux/amd64 debian:12 as debian-base
 
 RUN apt-get update && apt-get -y upgrade
 RUN apt-get -y install --no-install-recommends libelf1 libdrm2 libdrm-amdgpu1 libnuma1 kmod
@@ -51,7 +51,7 @@ COPY --from=rocm /opt/rocm-$ROCM /opt/rocm-$ROCM
 RUN ln -s /opt/rocm-$ROCM /opt/rocm
 
 RUN apt-get -y install g++ cmake
-RUN apt-get -y install python3-pybind11 python3.9-distutils python3-dev
+RUN apt-get -y install python3-pybind11 python3-distutils python3-dev
 
 WORKDIR /opt/build
 
@@ -70,10 +70,11 @@ RUN apt-get -y install libnuma1
 WORKDIR /opt/frigate/
 COPY --from=rootfs / /
 
-COPY docker/rocm/requirements-wheels-rocm.txt /requirements.txt
-RUN python3 -m pip install --upgrade pip \
-    && pip3 uninstall -y onnxruntime-openvino \
-    && pip3 install -r /requirements.txt
+# Temporarily disabled to see if a new wheel can be built to support py3.11
+#COPY docker/rocm/requirements-wheels-rocm.txt /requirements.txt
+#RUN python3 -m pip install --upgrade pip \
+#    && pip3 uninstall -y onnxruntime-openvino \
+#    && pip3 install -r /requirements.txt
 
 #######################################################################
 FROM scratch AS rocm-dist
@@ -86,12 +87,12 @@ COPY --from=rocm /opt/rocm-$ROCM/share/miopen/db/*$AMDGPU* /opt/rocm-$ROCM/share
 COPY --from=rocm /opt/rocm-$ROCM/share/miopen/db/*gfx908* /opt/rocm-$ROCM/share/miopen/db/
 COPY --from=rocm /opt/rocm-$ROCM/lib/rocblas/library/*$AMDGPU* /opt/rocm-$ROCM/lib/rocblas/library/
 COPY --from=rocm /opt/rocm-dist/ /
-COPY --from=debian-build /opt/rocm/lib/migraphx.cpython-39-x86_64-linux-gnu.so /opt/rocm-$ROCM/lib/
+COPY --from=debian-build /opt/rocm/lib/migraphx.cpython-311-x86_64-linux-gnu.so /opt/rocm-$ROCM/lib/
 
 #######################################################################
 FROM deps-prelim AS rocm-prelim-hsa-override0
-
-ENV HSA_ENABLE_SDMA=0
+\
+    ENV HSA_ENABLE_SDMA=0
 
 COPY --from=rocm-dist / /
 

docker/rpi/install_deps.sh

@@ -24,7 +24,7 @@ sed -i -e's/ main/ main contrib non-free/g' /etc/apt/sources.list
 if [[ "${TARGETARCH}" == "arm64" ]]; then
     # add raspberry pi repo
     gpg --no-default-keyring --keyring /usr/share/keyrings/raspbian.gpg --keyserver keyserver.ubuntu.com --recv-keys 82B129927FA3303E
-    echo "deb [signed-by=/usr/share/keyrings/raspbian.gpg] https://archive.raspberrypi.org/debian/ bullseye main" | tee /etc/apt/sources.list.d/raspi.list
+    echo "deb [signed-by=/usr/share/keyrings/raspbian.gpg] https://archive.raspberrypi.org/debian/ bookworm main" | tee /etc/apt/sources.list.d/raspi.list
     apt-get -qq update
     apt-get -qq install --no-install-recommends --no-install-suggests -y ffmpeg
 fi

docker/tensorrt/Dockerfile.amd64

@@ -7,33 +7,19 @@ ARG DEBIAN_FRONTEND=noninteractive
 FROM wheels as trt-wheels
 ARG DEBIAN_FRONTEND
 ARG TARGETARCH
+RUN python3 -m pip config set global.break-system-packages true
 
 # Add TensorRT wheels to another folder
 COPY docker/tensorrt/requirements-amd64.txt /requirements-tensorrt.txt
 RUN mkdir -p /trt-wheels && pip3 wheel --wheel-dir=/trt-wheels -r /requirements-tensorrt.txt
 
-# Build CuDNN
-FROM wget AS cudnn-deps
-
-ARG COMPUTE_LEVEL
-
-RUN apt-get update \
-    && apt-get install -y git build-essential
-
-RUN wget https://developer.download.nvidia.com/compute/cuda/repos/debian11/x86_64/cuda-keyring_1.1-1_all.deb \
-&& dpkg -i cuda-keyring_1.1-1_all.deb \
-&& apt-get update \
-&& apt-get -y install cuda-toolkit \
-&& rm -rf /var/lib/apt/lists/*
-
 FROM tensorrt-base AS frigate-tensorrt
-ENV TRT_VER=8.5.3
+ENV TRT_VER=8.6.1
+RUN python3 -m pip config set global.break-system-packages true
 RUN --mount=type=bind,from=trt-wheels,source=/trt-wheels,target=/deps/trt-wheels \
-    pip3 install -U /deps/trt-wheels/*.whl && \
+    pip3 install -U /deps/trt-wheels/*.whl --break-system-packages && \
     ldconfig
-COPY --from=cudnn-deps /usr/local/cuda-12.6 /usr/local/cuda
 
-ENV LD_LIBRARY_PATH=/usr/local/lib/python3.9/dist-packages/tensorrt:/usr/local/cuda/lib64:/usr/local/lib/python3.9/dist-packages/nvidia/cufft/lib
 WORKDIR /opt/frigate/
 COPY --from=rootfs / /
 
@@ -42,8 +28,8 @@ FROM devcontainer AS devcontainer-trt
 
 COPY --from=trt-deps /usr/local/lib/libyolo_layer.so /usr/local/lib/libyolo_layer.so
 COPY --from=trt-deps /usr/local/src/tensorrt_demos /usr/local/src/tensorrt_demos
-COPY --from=cudnn-deps /usr/local/cuda-12.6 /usr/local/cuda
+COPY --from=trt-deps /usr/local/cuda-12.1 /usr/local/cuda
 COPY docker/tensorrt/detector/rootfs/ /
 COPY --from=trt-deps /usr/local/lib/libyolo_layer.so /usr/local/lib/libyolo_layer.so
 RUN --mount=type=bind,from=trt-wheels,source=/trt-wheels,target=/deps/trt-wheels \
-    pip3 install -U /deps/trt-wheels/*.whl
+    pip3 install -U /deps/trt-wheels/*.whl --break-system-packages

docker/tensorrt/Dockerfile.arm64

@@ -10,8 +10,8 @@ ARG DEBIAN_FRONTEND
 # Use a separate container to build wheels to prevent build dependencies in final image
 RUN apt-get -qq update \
     && apt-get -qq install -y --no-install-recommends \
-       python3.9 python3.9-dev \
-       wget build-essential cmake git \
+    python3.9 python3.9-dev \
+    wget build-essential cmake git \
     && rm -rf /var/lib/apt/lists/*
 
 # Ensure python3 defaults to python3.9
@@ -41,7 +41,11 @@ RUN --mount=type=bind,source=docker/tensorrt/detector/build_python_tensorrt.sh,t
     && TENSORRT_VER=$(cat /etc/TENSORRT_VER) /deps/build_python_tensorrt.sh
 
 COPY docker/tensorrt/requirements-arm64.txt /requirements-tensorrt.txt
-RUN pip3 wheel --wheel-dir=/trt-wheels -r /requirements-tensorrt.txt
+ADD https://nvidia.box.com/shared/static/psl23iw3bh7hlgku0mjo1xekxpego3e3.whl /tmp/onnxruntime_gpu-1.15.1-cp311-cp311-linux_aarch64.whl
+
+RUN pip3 uninstall -y onnxruntime-openvino \
+    && pip3 wheel --wheel-dir=/trt-wheels -r /requirements-tensorrt.txt \
+    && pip3 install --no-deps /tmp/onnxruntime_gpu-1.15.1-cp311-cp311-linux_aarch64.whl
 
 FROM build-wheels AS trt-model-wheels
 ARG DEBIAN_FRONTEND

docker/tensorrt/Dockerfile.base

@@ -3,7 +3,7 @@
 # https://askubuntu.com/questions/972516/debian-frontend-environment-variable
 ARG DEBIAN_FRONTEND=noninteractive
 
-ARG TRT_BASE=nvcr.io/nvidia/tensorrt:23.03-py3
+ARG TRT_BASE=nvcr.io/nvidia/tensorrt:23.12-py3
 
 # Build TensorRT-specific library
 FROM ${TRT_BASE} AS trt-deps
@@ -24,8 +24,9 @@ ENV S6_CMD_WAIT_FOR_SERVICES_MAXTIME=0
 
 COPY --from=trt-deps /usr/local/lib/libyolo_layer.so /usr/local/lib/libyolo_layer.so
 COPY --from=trt-deps /usr/local/src/tensorrt_demos /usr/local/src/tensorrt_demos
+COPY --from=trt-deps /usr/local/cuda-12.* /usr/local/cuda
 COPY docker/tensorrt/detector/rootfs/ /
-ENV YOLO_MODELS="yolov7-320"
+ENV YOLO_MODELS=""
 
 HEALTHCHECK --start-period=600s --start-interval=5s --interval=15s --timeout=5s --retries=3 \
     CMD curl --fail --silent --show-error http://127.0.0.1:5000/api/version || exit 1

docker/tensorrt/detector/rootfs/etc/ld.so.conf.d/cuda_tensorrt.conf

@@ -1,6 +1,8 @@
 /usr/local/lib
-/usr/local/lib/python3.9/dist-packages/nvidia/cudnn/lib
-/usr/local/lib/python3.9/dist-packages/nvidia/cuda_runtime/lib
-/usr/local/lib/python3.9/dist-packages/nvidia/cublas/lib
-/usr/local/lib/python3.9/dist-packages/nvidia/cuda_nvrtc/lib
-/usr/local/lib/python3.9/dist-packages/tensorrt
+/usr/local/cuda/lib64
+/usr/local/lib/python3.11/dist-packages/nvidia/cudnn/lib
+/usr/local/lib/python3.11/dist-packages/nvidia/cuda_runtime/lib
+/usr/local/lib/python3.11/dist-packages/nvidia/cublas/lib
+/usr/local/lib/python3.11/dist-packages/nvidia/cuda_nvrtc/lib
+/usr/local/lib/python3.11/dist-packages/tensorrt
+/usr/local/lib/python3.11/dist-packages/nvidia/cufft/lib

docker/tensorrt/detector/rootfs/etc/s6-overlay/s6-rc.d/trt-model-prepare/run

@@ -11,6 +11,7 @@ set -o errexit -o nounset -o pipefail
 MODEL_CACHE_DIR=${MODEL_CACHE_DIR:-"/config/model_cache/tensorrt"}
 TRT_VER=${TRT_VER:-$(cat /etc/TENSORRT_VER)}
 OUTPUT_FOLDER="${MODEL_CACHE_DIR}/${TRT_VER}"
+YOLO_MODELS=${YOLO_MODELS:-""}
 
 # Create output folder
 mkdir -p ${OUTPUT_FOLDER}
@@ -19,6 +20,11 @@ FIRST_MODEL=true
 MODEL_DOWNLOAD=""
 MODEL_CONVERT=""
 
+if [ -z "$YOLO_MODELS"]; then
+    echo "tensorrt model preparation disabled"
+    exit 0
+fi
+
 for model in ${YOLO_MODELS//,/ }
 do
     # Remove old link in case path/version changed

docker/tensorrt/requirements-amd64.txt

@@ -1,14 +1,14 @@
 # NVidia TensorRT Support (amd64 only)
 --extra-index-url 'https://pypi.nvidia.com'
 numpy < 1.24; platform_machine == 'x86_64'
-tensorrt == 8.5.3.*; platform_machine == 'x86_64'
-cuda-python == 11.8; platform_machine == 'x86_64'
-cython == 0.29.*; platform_machine == 'x86_64'
+tensorrt == 8.6.1.*; platform_machine == 'x86_64'
+cuda-python == 11.8.*; platform_machine == 'x86_64'
+cython == 3.0.*; platform_machine == 'x86_64'
 nvidia-cuda-runtime-cu12 == 12.1.*; platform_machine == 'x86_64'
 nvidia-cuda-runtime-cu11 == 11.8.*; platform_machine == 'x86_64'
 nvidia-cublas-cu11 == 11.11.3.6; platform_machine == 'x86_64'
 nvidia-cudnn-cu11 == 8.6.0.*; platform_machine == 'x86_64'
 nvidia-cufft-cu11==10.*; platform_machine == 'x86_64'
-onnx==1.14.0; platform_machine == 'x86_64'
-onnxruntime-gpu==1.17.*; platform_machine == 'x86_64'
+onnx==1.16.*; platform_machine == 'x86_64'
+onnxruntime-gpu==1.18.*; platform_machine == 'x86_64'
 protobuf==3.20.3; platform_machine == 'x86_64'

docker/tensorrt/requirements-arm64.txt

@@ -1 +1 @@
-cuda-python == 11.7; platform_machine == 'aarch64'
+cuda-python == 11.7; platform_machine == 'aarch64'

docs/docs/configuration/advanced.md

@@ -174,7 +174,7 @@ NOTE: The folder that is set for the config needs to be the folder that contains
 
 ### Custom go2rtc version
 
-Frigate currently includes go2rtc v1.9.4, there may be certain cases where you want to run a different version of go2rtc.
+Frigate currently includes go2rtc v1.9.2, there may be certain cases where you want to run a different version of go2rtc.
 
 To do this:
 

docs/docs/configuration/autotracking.md

@@ -41,6 +41,7 @@ cameras:
     ...
     onvif:
       # Required: host of the camera being connected to.
+      # NOTE: HTTP is assumed by default; HTTPS is supported if you specify the scheme, ex: "https://0.0.0.0".
       host: 0.0.0.0
       # Optional: ONVIF port for device (default: shown below).
       port: 8000
@@ -49,6 +50,8 @@ cameras:
       user: admin
       # Optional: password for login.
       password: admin
+      # Optional: Skip TLS verification from the ONVIF server (default: shown below)
+      tls_insecure: False
       # Optional: PTZ camera object autotracking. Keeps a moving object in
       # the center of the frame by automatically moving the PTZ camera.
       autotracking:

docs/docs/configuration/camera_specific.md

@@ -67,14 +67,15 @@ ffmpeg:
 
 ### Annke C800
 
-This camera is H.265 only. To be able to play clips on some devices (like MacOs or iPhone) the H.265 stream has to be repackaged and the audio stream has to be converted to aac. Unfortunately direct playback of in the browser is not working (yet), but the downloaded clip can be played locally.
+This camera is H.265 only. To be able to play clips on some devices (like MacOs or iPhone) the H.265 stream has to be adjusted using the `apple_compatibility` config.
 
 ```yaml
 cameras:
   annkec800: # <------ Name the camera
     ffmpeg:
+      apple_compatibility: true  # <- Adds compatibility with MacOS and iPhone
       output_args:
-        record: -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c:v copy -tag:v hvc1 -bsf:v hevc_mp4toannexb -c:a aac
+        record: preset-record-generic-audio-aac
 
       inputs:
         - path: rtsp://user:password@camera-ip:554/H264/ch1/main/av_stream # <----- Update for your camera
@@ -156,7 +157,9 @@ cameras:
 
 #### Reolink Doorbell
 
-The reolink doorbell supports 2-way audio via go2rtc and other applications. It is important that the http-flv stream is still used for stability, a secondary rtsp stream can be added that will be using for the two way audio only.
+The reolink doorbell supports two way audio via go2rtc and other applications. It is important that the http-flv stream is still used for stability, a secondary rtsp stream can be added that will be using for the two way audio only.
+
+Ensure HTTP is enabled in the camera's advanced network settings. To use two way talk with Frigate, see the [Live view documentation](/configuration/live#two-way-talk).
 
 ```yaml
 go2rtc:
@@ -181,7 +184,7 @@ go2rtc:
       - rtspx://192.168.1.1:7441/abcdefghijk
 ```
 
-[See the go2rtc docs for more information](https://github.com/AlexxIT/go2rtc/tree/v1.9.4#source-rtsp)
+[See the go2rtc docs for more information](https://github.com/AlexxIT/go2rtc/tree/v1.9.2#source-rtsp)
 
 In the Unifi 2.0 update Unifi Protect Cameras had a change in audio sample rate which causes issues for ffmpeg. The input rate needs to be set for record if used directly with unifi protect.
 

docs/docs/configuration/cameras.md

@@ -109,7 +109,7 @@ This list of working and non-working PTZ cameras is based on user feedback.
 | Reolink E1 Zoom              |      ✅      |      ❌      |                                                                                                                                                 |
 | Reolink RLC-823A 16x         |      ✅      |      ❌      |                                                                                                                                                 |
 | Speco O8P32X                 |      ✅      |      ❌      |                                                                                                                                                 |
-| Sunba 405-D20X               |      ✅      |      ❌      |                                                                                                                                                 |
+| Sunba 405-D20X               |      ✅      |      ❌      | Incomplete ONVIF support reported on original, and 4k models. All models are suspected incompatable.                                            |
 | Tapo                         |      ✅      |      ❌      | Many models supported, ONVIF Service Port: 2020                                                                                                 |
 | Uniview IPC672LR-AX4DUPK     |      ✅      |      ❌      | Firmware says FOV relative movement is supported, but camera doesn't actually move when sending ONVIF commands                                  |
 | Uniview IPC6612SR-X33-VG     |      ✅      |      ✅      | Leave `calibrate_on_startup` as `False`. A user has reported that zooming with `absolute` is working.                                           |

docs/docs/configuration/face_recognition.md

@@ -0,0 +1,35 @@
+---
+id: face_recognition
+title: Face Recognition
+---
+
+Face recognition allows people to be assigned names and when their face is recognized Frigate will assign the person's name as a sub label. This information is included in the UI, filters, as well as in notifications.
+
+Frigate has support for FaceNet to create face embeddings, which runs locally. Embeddings are then saved to Frigate's database.
+
+## Minimum System Requirements
+
+Face recognition works by running a large AI model locally on your system. Systems without a GPU will not run Face Recognition reliably or at all.
+
+## Configuration
+
+Face recognition is disabled by default and requires semantic search to be enabled, face recognition must be enabled in your config file before it can be used. Semantic Search and face recognition are global configuration settings.
+
+```yaml
+face_recognition:
+  enabled: true
+```
+
+## Dataset
+
+The number of images needed for a sufficient training set for face recognition varies depending on several factors:
+
+- Complexity of the task: A simple task like recognizing faces of known individuals may require fewer images than a complex task like identifying unknown individuals in a large crowd.
+- Diversity of the dataset: A dataset with diverse images, including variations in lighting, pose, and facial expressions, will require fewer images per person than a less diverse dataset.
+- Desired accuracy: The higher the desired accuracy, the more images are typically needed.
+
+However, here are some general guidelines:
+
+- Minimum: For basic face recognition tasks, a minimum of 10-20 images per person is often recommended.
+- Recommended: For more robust and accurate systems, 30-50 images per person is a good starting point.
+- Ideal: For optimal performance, especially in challenging conditions, 100 or more images per person can be beneficial.

docs/docs/configuration/genai.md

@@ -3,9 +3,15 @@ id: genai
 title: Generative AI
 ---
 
-Generative AI can be used to automatically generate descriptions based on the thumbnails of your tracked objects. This helps with [Semantic Search](/configuration/semantic_search) in Frigate by providing detailed text descriptions as a basis of the search query.
+Generative AI can be used to automatically generate descriptive text based on the thumbnails of your tracked objects. This helps with [Semantic Search](/configuration/semantic_search) in Frigate to provide more context about your tracked objects. Descriptions are accessed via the _Explore_ view in the Frigate UI by clicking on a tracked object's thumbnail.
 
-Semantic Search must be enabled to use Generative AI. Descriptions are accessed via the _Explore_ view in the Frigate UI by clicking on a tracked object's thumbnail.
+Requests for a description are sent off automatically to your AI provider at the end of the tracked object's lifecycle. Descriptions can also be regenerated manually via the Frigate UI.
+
+:::info
+
+Semantic Search must be enabled to use Generative AI.
+
+:::
 
 ## Configuration
 
@@ -29,11 +35,21 @@ cameras:
 
 ## Ollama
 
-[Ollama](https://ollama.com/) allows you to self-host large language models and keep everything running locally. It provides a nice API over [llama.cpp](https://github.com/ggerganov/llama.cpp). It is highly recommended to host this server on a machine with an Nvidia graphics card, or on a Apple silicon Mac for best performance. Most of the 7b parameter 4-bit vision models will fit inside 8GB of VRAM. There is also a [docker container](https://hub.docker.com/r/ollama/ollama) available.
+:::warning
+
+Using Ollama on CPU is not recommended, high inference times make using Generative AI impractical.
+
+:::
+
+[Ollama](https://ollama.com/) allows you to self-host large language models and keep everything running locally. It provides a nice API over [llama.cpp](https://github.com/ggerganov/llama.cpp). It is highly recommended to host this server on a machine with an Nvidia graphics card, or on a Apple silicon Mac for best performance.
+
+Most of the 7b parameter 4-bit vision models will fit inside 8GB of VRAM. There is also a [Docker container](https://hub.docker.com/r/ollama/ollama) available.
+
+Parallel requests also come with some caveats. You will need to set `OLLAMA_NUM_PARALLEL=1` and choose a `OLLAMA_MAX_QUEUE` and `OLLAMA_MAX_LOADED_MODELS` values that are appropriate for your hardware and preferences. See the [Ollama documentation](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-does-ollama-handle-concurrent-requests).
 
 ### Supported Models
 
-You must use a vision capable model with Frigate. Current model variants can be found [in their model library](https://ollama.com/library). At the time of writing, this includes `llava`, `llava-llama3`, `llava-phi3`, and `moondream`.
+You must use a vision capable model with Frigate. Current model variants can be found [in their model library](https://ollama.com/library). At the time of writing, this includes `llava`, `llava-llama3`, `llava-phi3`, and `moondream`. Note that Frigate will not automatically download the model you specify in your config, you must download the model to your local instance of Ollama first i.e. by running `ollama pull llava:7b` on your Ollama server/Docker container. Note that the model specified in Frigate's config must match the downloaded model tag.
 
 :::note
 
@@ -48,7 +64,7 @@ genai:
   enabled: True
   provider: ollama
   base_url: http://localhost:11434
-  model: llava
+  model: llava:7b
 ```
 
 ## Google Gemini
@@ -122,12 +138,22 @@ genai:
   api_key: "{FRIGATE_OPENAI_API_KEY}"
 ```
 
+## Usage and Best Practices
+
+Frigate's thumbnail search excels at identifying specific details about tracked objects – for example, using an "image caption" approach to find a "person wearing a yellow vest," "a white dog running across the lawn," or "a red car on a residential street." To enhance this further, Frigate’s default prompts are designed to ask your AI provider about the intent behind the object's actions, rather than just describing its appearance.
+
+While generating simple descriptions of detected objects is useful, understanding intent provides a deeper layer of insight. Instead of just recognizing "what" is in a scene, Frigate’s default prompts aim to infer "why" it might be there or "what" it could do next. Descriptions tell you what’s happening, but intent gives context. For instance, a person walking toward a door might seem like a visitor, but if they’re moving quickly after hours, you can infer a potential break-in attempt. Detecting a person loitering near a door at night can trigger an alert sooner than simply noting "a person standing by the door," helping you respond based on the situation’s context.
+
+### Using GenAI for notifications
+
+Frigate provides an [MQTT topic](/integrations/mqtt), `frigate/tracked_object_update`, that is updated with a JSON payload containing `event_id` and `description` when your AI provider returns a description for a tracked object. This description could be used directly in notifications, such as sending alerts to your phone or making audio announcements. If additional details from the tracked object are needed, you can query the [HTTP API](/integrations/api/event-events-event-id-get) using the `event_id`, eg: `http://frigate_ip:5000/api/events/<event_id>`.
+
 ## Custom Prompts
 
 Frigate sends multiple frames from the tracked object along with a prompt to your Generative AI provider asking it to generate a description. The default prompt is as follows:
 
 ```
-Describe the {label} in the sequence of images with as much detail as possible. Do not describe the background.
+Analyze the sequence of images containing the {label}. Focus on the likely intent or behavior of the {label} based on its actions and movement, rather than describing its appearance or the surroundings. Consider what the {label} is doing, why, and what it might do next.
 ```
 
 :::tip
@@ -144,25 +170,25 @@ genai:
   provider: ollama
   base_url: http://localhost:11434
   model: llava
-  prompt: "Describe the {label} in these images from the {camera} security camera."
+  prompt: "Analyze the {label} in these images from the {camera} security camera. Focus on the actions, behavior, and potential intent of the {label}, rather than just describing its appearance."
   object_prompts:
-    person: "Describe the main person in these images (gender, age, clothing, activity, etc). Do not include where the activity is occurring (sidewalk, concrete, driveway, etc)."
-    car: "Label the primary vehicle in these images with just the name of the company if it is a delivery vehicle, or the color make and model."
+    person: "Examine the main person in these images. What are they doing and what might their actions suggest about their intent (e.g., approaching a door, leaving an area, standing still)? Do not describe the surroundings or static details."
+    car: "Observe the primary vehicle in these images. Focus on its movement, direction, or purpose (e.g., parking, approaching, circling). If it's a delivery vehicle, mention the company."
 ```
 
 Prompts can also be overriden at the camera level to provide a more detailed prompt to the model about your specific camera, if you desire. By default, descriptions will be generated for all tracked objects and all zones. But you can also optionally specify `objects` and `required_zones` to only generate descriptions for certain tracked objects or zones.
 
-Optionally, you can generate the description using a snapshot (if enabled) by setting `use_snapshot` to `True`. By default, this is set to `False`, which sends the thumbnails collected over the object's lifetime to the model. Using a snapshot provides the AI with a higher-resolution image (typically downscaled by the AI itself), but the trade-off is that only a single image is used, which might limit the model's ability to determine object movement or direction.
+Optionally, you can generate the description using a snapshot (if enabled) by setting `use_snapshot` to `True`. By default, this is set to `False`, which sends the uncompressed images from the `detect` stream collected over the object's lifetime to the model. Once the object lifecycle ends, only a single compressed and cropped thumbnail is saved with the tracked object. Using a snapshot might be useful when you want to _regenerate_ a tracked object's description as it will provide the AI with a higher-quality image (typically downscaled by the AI itself) than the cropped/compressed thumbnail. Using a snapshot otherwise has a trade-off in that only a single image is sent to your provider, which will limit the model's ability to determine object movement or direction.
 
 ```yaml
 cameras:
   front_door:
     genai:
       use_snapshot: True
-      prompt: "Describe the {label} in these images from the {camera} security camera at the front door of a house, aimed outward toward the street."
+      prompt: "Analyze the {label} in these images from the {camera} security camera at the front door. Focus on the actions and potential intent of the {label}."
       object_prompts:
-        person: "Describe the main person in these images (gender, age, clothing, activity, etc). Do not include where the activity is occurring (sidewalk, concrete, driveway, etc). If delivering a package, include the company the package is from."
-        cat: "Describe the cat in these images (color, size, tail). Indicate whether or not the cat is by the flower pots. If the cat is chasing a mouse, make up a name for the mouse."
+        person: "Examine the person in these images. What are they doing, and how might their actions suggest their purpose (e.g., delivering something, approaching, leaving)? If they are carrying or interacting with a package, include details about its source or destination."
+        cat: "Observe the cat in these images. Focus on its movement and intent (e.g., wandering, hunting, interacting with objects). If the cat is near the flower pots or engaging in any specific actions, mention it."
       objects:
         - person
         - cat

docs/docs/configuration/hardware_acceleration.md

@@ -175,6 +175,16 @@ For more information on the various values across different distributions, see h
 
 Depending on your OS and kernel configuration, you may need to change the `/proc/sys/kernel/perf_event_paranoid` kernel tunable. You can test the change by running `sudo sh -c 'echo 2 >/proc/sys/kernel/perf_event_paranoid'` which will persist until a reboot. Make it permanent by running `sudo sh -c 'echo kernel.perf_event_paranoid=2 >> /etc/sysctl.d/local.conf'`
 
+#### Stats for SR-IOV devices
+
+When using virtualized GPUs via SR-IOV, additional args are needed for GPU stats to function. This can be enabled with the following config:
+
+```yaml
+telemetry:
+  stats:
+    sriov: True
+```
+
 ## AMD/ATI GPUs (Radeon HD 2000 and newer GPUs) via libva-mesa-driver
 
 VAAPI supports automatic profile selection so it will work automatically with both H.264 and H.265 streams.
@@ -231,28 +241,11 @@ docker run -d \
 
 ### Setup Decoder
 
-The decoder you need to pass in the `hwaccel_args` will depend on the input video.
-
-A list of supported codecs (you can use `ffmpeg -decoders | grep cuvid` in the container to get the ones your card supports)
-
-```
- V..... h263_cuvid           Nvidia CUVID H263 decoder (codec h263)
- V..... h264_cuvid           Nvidia CUVID H264 decoder (codec h264)
- V..... hevc_cuvid           Nvidia CUVID HEVC decoder (codec hevc)
- V..... mjpeg_cuvid          Nvidia CUVID MJPEG decoder (codec mjpeg)
- V..... mpeg1_cuvid          Nvidia CUVID MPEG1VIDEO decoder (codec mpeg1video)
- V..... mpeg2_cuvid          Nvidia CUVID MPEG2VIDEO decoder (codec mpeg2video)
- V..... mpeg4_cuvid          Nvidia CUVID MPEG4 decoder (codec mpeg4)
- V..... vc1_cuvid            Nvidia CUVID VC1 decoder (codec vc1)
- V..... vp8_cuvid            Nvidia CUVID VP8 decoder (codec vp8)
- V..... vp9_cuvid            Nvidia CUVID VP9 decoder (codec vp9)
-```
-
-For example, for H264 video, you'll select `preset-nvidia-h264`.
+Using `preset-nvidia` ffmpeg will automatically select the necessary profile for the incoming video, and will log an error if the profile is not supported by your GPU.
 
 ```yaml
 ffmpeg:
-  hwaccel_args: preset-nvidia-h264
+  hwaccel_args: preset-nvidia
 ```
 
 If everything is working correctly, you should see a significant improvement in performance.

docs/docs/configuration/index.md

@@ -203,14 +203,13 @@ detectors:
   ov:
     type: openvino
     device: AUTO
-    model:
-      path: /openvino-model/ssdlite_mobilenet_v2.xml
 
 model:
   width: 300
   height: 300
   input_tensor: nhwc
   input_pixel_format: bgr
+  path: /openvino-model/ssdlite_mobilenet_v2.xml
   labelmap_path: /openvino-model/coco_91cl_bkgr.txt
 
 record:

docs/docs/configuration/license_plate_recognition.md

@@ -0,0 +1,45 @@
+---
+id: license_plate_recognition
+title: License Plate Recognition (LPR)
+---
+
+Frigate can recognize license plates on vehicles and automatically add the detected characters as a `sub_label` to objects that are of type `car`. A common use case may be to read the license plates of cars pulling into a driveway or cars passing by on a street with a dedicated LPR camera.
+
+Users running a Frigate+ model should ensure that `license_plate` is added to the [list of objects to track](https://docs.frigate.video/plus/#available-label-types) either globally or for a specific camera. This will improve the accuracy and performance of the LPR model.
+
+LPR is most effective when the vehicle’s license plate is fully visible to the camera. For moving vehicles, Frigate will attempt to read the plate continuously, refining its detection and keeping the most confident result. LPR will not run on stationary vehicles.
+
+## Minimum System Requirements
+
+License plate recognition works by running AI models locally on your system. The models are relatively lightweight and run on your CPU. At least 4GB of RAM is required.
+
+## Configuration
+
+License plate recognition is disabled by default. Enable it in your config file:
+
+```yaml
+lpr:
+  enabled: true
+```
+
+## Advanced Configuration
+
+Several options are available to fine-tune the LPR feature. For example, you can adjust the `min_area` setting, which defines the minimum size in pixels a license plate must be before LPR runs. The default is 500 pixels.
+
+Additionally, you can define `known_plates` as strings or regular expressions, allowing Frigate to label tracked vehicles with custom sub_labels when a recognized plate is detected. This information is then accessible in the UI, filters, and notifications.
+
+```yaml
+lpr:
+  enabled: true
+  min_area: 500
+  known_plates:
+    Wife's Car:
+      - "ABC-1234"
+      - "ABC-I234"
+    Johnny:
+      - "J*N-*234" # Using wildcards for H/M and 1/I
+    Sally:
+      - "[S5]LL-1234" # Matches SLL-1234 and 5LL-1234
+```
+
+In this example, "Wife's Car" will appear as the label for any vehicle matching the plate "ABC-1234." The model might occasionally interpret the digit 1 as a capital I (e.g., "ABC-I234"), so both variations are listed. Similarly, multiple possible variations are specified for Johnny and Sally.

docs/docs/configuration/live.md

@@ -23,13 +23,13 @@ If you are using go2rtc, you should adjust the following settings in your camera
 
 - Video codec: **H.264** - provides the most compatible video codec with all Live view technologies and browsers. Avoid any kind of "smart codec" or "+" codec like _H.264+_ or _H.265+_. as these non-standard codecs remove keyframes (see below).
 - Audio codec: **AAC** - provides the most compatible audio codec with all Live view technologies and browsers that support audio.
-- I-frame interval (sometimes called the keyframe interval, the interframe space, or the GOP length): match your camera's frame rate, or choose "1x" (for interframe space on Reolink cameras). For example, if your stream outputs 20fps, your i-frame interval should be 20 (or 1x on Reolink). Values higher than the frame rate will cause the stream to take longer to begin playback. See [this page](https://gardinal.net/understanding-the-keyframe-interval/) for more on keyframes.
+- I-frame interval (sometimes called the keyframe interval, the interframe space, or the GOP length): match your camera's frame rate, or choose "1x" (for interframe space on Reolink cameras). For example, if your stream outputs 20fps, your i-frame interval should be 20 (or 1x on Reolink). Values higher than the frame rate will cause the stream to take longer to begin playback. See [this page](https://gardinal.net/understanding-the-keyframe-interval/) for more on keyframes. For many users this may not be an issue, but it should be noted that that a 1x i-frame interval will cause more storage utilization if you are using the stream for the `record` role as well.
 
 The default video and audio codec on your camera may not always be compatible with your browser, which is why setting them to H.264 and AAC is recommended. See the [go2rtc docs](https://github.com/AlexxIT/go2rtc?tab=readme-ov-file#codecs-madness) for codec support information.
 
 ### Audio Support
 
-MSE Requires AAC audio, WebRTC requires PCMU/PCMA, or opus audio. If you want to support both MSE and WebRTC then your restream config needs to make sure both are enabled.
+MSE Requires PCMA/PCMU or AAC audio, WebRTC requires PCMA/PCMU or opus audio. If you want to support both MSE and WebRTC then your restream config needs to make sure both are enabled.
 
 ```yaml
 go2rtc:
@@ -138,3 +138,13 @@ services:
 :::
 
 See [go2rtc WebRTC docs](https://github.com/AlexxIT/go2rtc/tree/v1.8.3#module-webrtc) for more information about this.
+
+### Two way talk
+
+For devices that support two way talk, Frigate can be configured to use the feature from the camera's Live view in the Web UI. You should:
+
+- Set up go2rtc with [WebRTC](#webrtc-extra-configuration).
+- Ensure you access Frigate via https (may require [opening port 8971](/frigate/installation/#ports)).
+- For the Home Assistant Frigate card, [follow the docs](https://github.com/dermotduffy/frigate-hass-card?tab=readme-ov-file#using-2-way-audio) for the correct source.
+
+To use the Reolink Doorbell with two way talk, you should use the [recommended Reolink configuration](/configuration/camera_specific#reolink-doorbell)

docs/docs/configuration/motion_detection.md

@@ -92,10 +92,16 @@ motion:
   lightning_threshold: 0.8
 ```
 
-:::tip
+:::warning
 
 Some cameras like doorbell cameras may have missed detections when someone walks directly in front of the camera and the lightning_threshold causes motion detection to be re-calibrated. In this case, it may be desirable to increase the `lightning_threshold` to ensure these objects are not missed.
 
 :::
 
+:::note
+
+Lightning threshold does not stop motion based recordings from being saved.
+
+:::
+
 Large changes in motion like PTZ moves and camera switches between Color and IR mode should result in no motion detection. This is done via the `lightning_threshold` configuration. It is defined as the percentage of the image used to detect lightning or other substantial changes where motion detection needs to recalibrate. Increasing this value will make motion detection more likely to consider lightning or IR mode changes as valid motion. Decreasing this value will make motion detection more likely to ignore large amounts of motion such as a person approaching a doorbell camera.

docs/docs/configuration/object_detectors.md

@@ -22,14 +22,14 @@ Frigate supports multiple different detectors that work on different types of ha
 - [ONNX](#onnx): OpenVINO will automatically be detected and used as a detector in the default Frigate image when a supported ONNX model is configured.
 
 **Nvidia**
-- [TensortRT](#nvidia-tensorrt-detector): TensorRT can run on Nvidia GPUs, using one of many default models.
-- [ONNX](#onnx): TensorRT will automatically be detected and used as a detector in the `-tensorrt` Frigate image when a supported ONNX model is configured.
+- [TensortRT](#nvidia-tensorrt-detector): TensorRT can run on Nvidia GPUs and Jetson devices, using one of many default models.
+- [ONNX](#onnx): TensorRT will automatically be detected and used as a detector in the `-tensorrt` or `-tensorrt-jp(4/5)` Frigate images when a supported ONNX model is configured.
 
 **Rockchip**
 - [RKNN](#rockchip-platform): RKNN models can run on Rockchip devices with included NPUs.
 
 **For Testing**
-- [CPU Detector (not recommended for actual use](#cpu-detector-not-recommended): Use a CPU to run tflite model, this is not recommended and in most cases OpenVINO can be used in CPU mode with better results. 
+- [CPU Detector (not recommended for actual use](#cpu-detector-not-recommended): Use a CPU to run tflite model, this is not recommended and in most cases OpenVINO can be used in CPU mode with better results.
 
 :::
 
@@ -144,7 +144,9 @@ detectors:
 
 #### SSDLite MobileNet v2
 
-An OpenVINO model is provided in the container at `/openvino-model/ssdlite_mobilenet_v2.xml` and is used by this detector type by default. The model comes from Intel's Open Model Zoo [SSDLite MobileNet V2](https://github.com/openvinotoolkit/open_model_zoo/tree/master/models/public/ssdlite_mobilenet_v2) and is converted to an FP16 precision IR model. Use the model configuration shown below when using the OpenVINO detector with the default model.
+An OpenVINO model is provided in the container at `/openvino-model/ssdlite_mobilenet_v2.xml` and is used by this detector type by default. The model comes from Intel's Open Model Zoo [SSDLite MobileNet V2](https://github.com/openvinotoolkit/open_model_zoo/tree/master/models/public/ssdlite_mobilenet_v2) and is converted to an FP16 precision IR model.
+
+Use the model configuration shown below when using the OpenVINO detector with the default OpenVINO model:
 
 ```yaml
 detectors:
@@ -223,7 +225,7 @@ The model used for TensorRT must be preprocessed on the same hardware platform t
 
 The Frigate image will generate model files during startup if the specified model is not found. Processed models are stored in the `/config/model_cache` folder. Typically the `/config` path is mapped to a directory on the host already and the `model_cache` does not need to be mapped separately unless the user wants to store it in a different location on the host.
 
-By default, the `yolov7-320` model will be generated, but this can be overridden by specifying the `YOLO_MODELS` environment variable in Docker. One or more models may be listed in a comma-separated format, and each one will be generated. To select no model generation, set the variable to an empty string, `YOLO_MODELS=""`. Models will only be generated if the corresponding `{model}.trt` file is not present in the `model_cache` folder, so you can force a model to be regenerated by deleting it from your Frigate data folder.
+By default, no models will be generated, but this can be overridden by specifying the `YOLO_MODELS` environment variable in Docker. One or more models may be listed in a comma-separated format, and each one will be generated. Models will only be generated if the corresponding `{model}.trt` file is not present in the `model_cache` folder, so you can force a model to be regenerated by deleting it from your Frigate data folder.
 
 If you have a Jetson device with DLAs (Xavier or Orin), you can generate a model that will run on the DLA by appending `-dla` to your model name, e.g. specify `YOLO_MODELS=yolov7-320-dla`. The model will run on DLA0 (Frigate does not currently support DLA1). DLA-incompatible layers will fall back to running on the GPU.
 
@@ -254,6 +256,7 @@ yolov4x-mish-640
 yolov7-tiny-288
 yolov7-tiny-416
 yolov7-640
+yolov7-416
 yolov7-320
 yolov7x-640
 yolov7x-320
@@ -264,7 +267,7 @@ An example `docker-compose.yml` fragment that converts the `yolov4-608` and `yol
 ```yml
 frigate:
   environment:
-    - YOLO_MODELS=yolov4-608,yolov7x-640
+    - YOLO_MODELS=yolov7-320,yolov7x-640
     - USE_FP16=false
 ```
 
@@ -282,6 +285,8 @@ The TensorRT detector can be selected by specifying `tensorrt` as the model type
 
 The TensorRT detector uses `.trt` model files that are located in `/config/model_cache/tensorrt` by default. These model path and dimensions used will depend on which model you have generated.
 
+Use the config below to work with generated TRT models:
+
 ```yaml
 detectors:
   tensorrt:
@@ -415,6 +420,24 @@ Note that the labelmap uses a subset of the complete COCO label set that has onl
 
 ONNX is an open format for building machine learning models, Frigate supports running ONNX models on CPU, OpenVINO, and TensorRT. On startup Frigate will automatically try to use a GPU if one is available.
 
+:::info
+
+If the correct build is used for your GPU then the GPU will be detected and used automatically.
+
+- **AMD**
+
+  - ROCm will automatically be detected and used with the ONNX detector in the `-rocm` Frigate image.
+
+- **Intel**
+
+  - OpenVINO will automatically be detected and used with the ONNX detector in the default Frigate image.
+
+- **Nvidia**
+  - Nvidia GPUs will automatically be detected and used with the ONNX detector in the `-tensorrt` Frigate image.
+  - Jetson devices will automatically be detected and used with the ONNX detector in the `-tensorrt-jp(4/5)` Frigate image.
+
+:::
+
 :::tip
 
 When using many cameras one detector may not be enough to keep up. Multiple detectors can be defined assuming GPU resources are available. An example configuration would be:
@@ -457,6 +480,7 @@ model:
   width: 320 # <--- should match whatever was set in notebook
   height: 320 # <--- should match whatever was set in notebook
   input_pixel_format: bgr
+  input_tensor: nchw
   path: /config/yolo_nas_s.onnx
   labelmap_path: /labelmap/coco-80.txt
 ```
@@ -482,11 +506,12 @@ detectors:
   cpu1:
     type: cpu
     num_threads: 3
-    model:
-      path: "/custom_model.tflite"
   cpu2:
     type: cpu
     num_threads: 3
+
+model:
+  path: "/custom_model.tflite"
 ```
 
 When using CPU detectors, you can add one CPU detector per camera. Adding more detectors than the number of cameras should not improve performance.
@@ -525,7 +550,7 @@ Hardware accelerated object detection is supported on the following SoCs:
 - RK3576
 - RK3588
 
-This implementation uses the [Rockchip's RKNN-Toolkit2](https://github.com/airockchip/rknn-toolkit2/), version v2.0.0.beta0. Currently, only [Yolo-NAS](https://github.com/Deci-AI/super-gradients/blob/master/YOLONAS.md) is supported as object detection model.
+This implementation uses the [Rockchip's RKNN-Toolkit2](https://github.com/airockchip/rknn-toolkit2/), version v2.3.0. Currently, only [Yolo-NAS](https://github.com/Deci-AI/super-gradients/blob/master/YOLONAS.md) is supported as object detection model.
 
 ### Prerequisites
 
@@ -598,7 +623,41 @@ $ cat /sys/kernel/debug/rknpu/load
 :::
 
 - All models are automatically downloaded and stored in the folder `config/model_cache/rknn_cache`. After upgrading Frigate, you should remove older models to free up space.
-- You can also provide your own `.rknn` model. You should not save your own models in the `rknn_cache` folder, store them directly in the `model_cache` folder or another subfolder. To convert a model to `.rknn` format see the `rknn-toolkit2` (requires a x86 machine). Note, that there is only post-processing for the supported models.
+- You can also provide your own `.rknn` model. You should not save your own models in the `rknn_cache` folder, store them directly in the `model_cache` folder or another subfolder. To convert a model to `.rknn` format see the `rknn-toolkit2`. Note, that there is only post-processing for the supported models.
+
+### Converting your own onnx model to rknn format
+
+To convert a onnx model to the rknn format using the [rknn-toolkit2](https://github.com/airockchip/rknn-toolkit2/) you have to:
+
+- Place one ore more models in onnx format in the directory `config/model_cache/rknn_cache/onnx` on your docker host (this might require `sudo` privileges).
+- Save the configuration file under `config/conv2rknn.yaml` (see below for details).
+- Run `docker exec <frigate_container_id> python3 /opt/conv2rknn.py`. If the conversion was successful, the rknn models will be placed in `config/model_cache/rknn_cache`.
+
+This is an example configuration file that you need to adjust to your specific onnx model:
+
+```yaml
+soc: ["rk3562","rk3566", "rk3568", "rk3576", "rk3588"]
+quantization: false
+
+output_name: "{input_basename}"
+
+config:
+  mean_values: [[0, 0, 0]]
+  std_values: [[255, 255, 255]]
+  quant_img_rgb2bgr: true
+```
+
+Explanation of the paramters:
+
+- `soc`: A list of all SoCs you want to build the rknn model for. If you don't specify this parameter, the script tries to find out your SoC and builds the rknn model for this one.
+- `quantization`: true: 8 bit integer (i8) quantization, false: 16 bit float (fp16). Default: false.
+- `output_name`: The output name of the model. The following variables are available:
+  - `quant`: "i8" or "fp16" depending on the config
+  - `input_basename`: the basename of the input model (e.g. "my_model" if the input model is calles "my_model.onnx")
+  - `soc`: the SoC this model was build for (e.g. "rk3588")
+  - `tk_version`: Version of `rknn-toolkit2` (e.g. "2.3.0")
+  - **example**: Specifying `output_name = "frigate-{quant}-{input_basename}-{soc}-v{tk_version}"` could result in a model called `frigate-i8-my_model-rk3588-v2.3.0.rknn`.
+- `config`: Configuration passed to `rknn-toolkit2` for model conversion. For an explanation of all available parameters have a look at section "2.2. Model configuration" of [this manual](https://github.com/MarcA711/rknn-toolkit2/releases/download/v2.3.0/03_Rockchip_RKNPU_API_Reference_RKNN_Toolkit2_V2.3.0_EN.pdf).
 
 ## Hailo-8l
 
@@ -613,8 +672,6 @@ detectors:
   hailo8l:
     type: hailo8l
     device: PCIe
-    model:
-      path: /config/model_cache/h8l_cache/ssd_mobilenet_v1.hef
 
 model:
   width: 300
@@ -622,4 +679,5 @@ model:
   input_tensor: nhwc
   input_pixel_format: bgr
   model_type: ssd
+  path: /config/model_cache/h8l_cache/ssd_mobilenet_v1.hef
 ```

docs/docs/configuration/objects.md

@@ -5,7 +5,7 @@ title: Available Objects
 
 import labels from "../../../labelmap.txt";
 
-Frigate includes the object models listed below from the Google Coral test data.
+Frigate includes the object labels listed below from the Google Coral test data.
 
 Please note:
 

docs/docs/configuration/reference.md

@@ -52,7 +52,7 @@ detectors:
   # Required: name of the detector
   detector_name:
     # Required: type of the detector
-    # Frigate provided types include 'cpu', 'edgetpu', 'openvino' and 'tensorrt' (default: shown below)
+    # Frigate provides many types, see https://docs.frigate.video/configuration/object_detectors for more details (default: shown below)
     # Additional detector types can also be plugged in.
     # Detectors may require additional configuration.
     # Refer to the Detectors configuration page for more information.
@@ -117,25 +117,27 @@ auth:
   hash_iterations: 600000
 
 # Optional: model modifications
+# NOTE: The default values are for the EdgeTPU detector.
+# Other detectors will require the model config to be set.
 model:
-  # Optional: path to the model (default: automatic based on detector)
+  # Required: path to the model (default: automatic based on detector)
   path: /edgetpu_model.tflite
-  # Optional: path to the labelmap (default: shown below)
+  # Required: path to the labelmap (default: shown below)
   labelmap_path: /labelmap.txt
   # Required: Object detection model input width (default: shown below)
   width: 320
   # Required: Object detection model input height (default: shown below)
   height: 320
-  # Optional: Object detection model input colorspace
+  # Required: Object detection model input colorspace
   # Valid values are rgb, bgr, or yuv. (default: shown below)
   input_pixel_format: rgb
-  # Optional: Object detection model input tensor format
+  # Required: Object detection model input tensor format
   # Valid values are nhwc or nchw (default: shown below)
   input_tensor: nhwc
-  # Optional: Object detection model type, currently only used with the OpenVINO detector
+  # Required: Object detection model type, currently only used with the OpenVINO detector
   # Valid values are ssd, yolox, yolonas (default: shown below)
   model_type: ssd
-  # Optional: Label name modifications. These are merged into the standard labelmap.
+  # Required: Label name modifications. These are merged into the standard labelmap.
   labelmap:
     2: vehicle
   # Optional: Map of object labels to their attribute labels (default: depends on model)
@@ -242,6 +244,8 @@ ffmpeg:
   # If set too high, then if a ffmpeg crash or camera stream timeout occurs, you could potentially lose up to a maximum of retry_interval second(s) of footage
   # NOTE: this can be a useful setting for Wireless / Battery cameras to reduce how much footage is potentially lost during a connection timeout.
   retry_interval: 10
+  # Optional: Set tag on HEVC (H.265) recording stream to improve compatibility with Apple players. (default: shown below)
+  apple_compatibility: false
 
 # Optional: Detect configuration
 # NOTE: Can be overridden at the camera level
@@ -518,6 +522,17 @@ semantic_search:
   enabled: False
   # Optional: Re-index embeddings database from historical tracked objects (default: shown below)
   reindex: False
+  # Optional: Set the model size used for embeddings. (default: shown below)
+  # NOTE: small model runs on CPU and large model runs on GPU
+  model_size: "small"
+
+# Optional: Configuration for face recognition capability
+face_recognition:
+  # Optional: Enable semantic search (default: shown below)
+  enabled: False
+  # Optional: Set the model size used for embeddings. (default: shown below)
+  # NOTE: small model runs on CPU and large model runs on GPU
+  model_size: "small"
 
 # Optional: Configuration for AI generated tracked object descriptions
 # NOTE: Semantic Search must be enabled for this to do anything.
@@ -545,10 +560,12 @@ genai:
 # Uses https://github.com/AlexxIT/go2rtc (v1.9.2)
 go2rtc:
 
-# Optional: jsmpeg stream configuration for WebUI
+# Optional: Live stream configuration for WebUI.
+# NOTE: Can be overridden at the camera level
 live:
-  # Optional: Set the name of the stream that should be used for live view
-  # in frigate WebUI. (default: name of camera)
+  # Optional: Set the name of the stream configured in go2rtc
+  # that should be used for live view in frigate WebUI. (default: name of camera)
+  # NOTE: In most cases this should be set at the camera level only.
   stream_name: camera_name
   # Optional: Set the height of the jsmpeg stream. (default: 720)
   # This must be less than or equal to the height of the detect stream. Lower resolutions
@@ -681,6 +698,7 @@ cameras:
     # to enable PTZ controls.
     onvif:
       # Required: host of the camera being connected to.
+      # NOTE: HTTP is assumed by default; HTTPS is supported if you specify the scheme, ex: "https://0.0.0.0".
       host: 0.0.0.0
       # Optional: ONVIF port for device (default: shown below).
       port: 8000
@@ -689,6 +707,8 @@ cameras:
       user: admin
       # Optional: password for login.
       password: admin
+      # Optional: Skip TLS verification from the ONVIF server (default: shown below)
+      tls_insecure: False
       # Optional: Ignores time synchronization mismatches between the camera and the server during authentication.
       # Using NTP on both ends is recommended and this should only be set to True in a "safe" environment due to the security risk it represents.
       ignore_time_mismatch: False
@@ -752,6 +772,8 @@ cameras:
         - cat
       # Optional: Restrict generation to objects that entered any of the listed zones (default: none, all zones qualify)
       required_zones: []
+      # Optional: Save thumbnails sent to generative AI for review/debugging purposes (default: shown below)
+      debug_save_thumbnails: False
 
 # Optional
 ui:
@@ -793,11 +815,13 @@ telemetry:
     - lo
   # Optional: Configure system stats
   stats:
-    # Enable AMD GPU stats (default: shown below)
+    # Optional: Enable AMD GPU stats (default: shown below)
     amd_gpu_stats: True
-    # Enable Intel GPU stats (default: shown below)
+    # Optional: Enable Intel GPU stats (default: shown below)
     intel_gpu_stats: True
-    # Enable network bandwidth stats monitoring for camera ffmpeg processes, go2rtc, and object detectors. (default: shown below)
+    # Optional: Treat GPU as SR-IOV to fix GPU stats (default: shown below)
+    sriov: False
+    # Optional: Enable network bandwidth stats monitoring for camera ffmpeg processes, go2rtc, and object detectors. (default: shown below)
     # NOTE: The container must either be privileged or have cap_net_admin, cap_net_raw capabilities enabled.
     network_bandwidth: False
   # Optional: Enable the latest version outbound check (default: shown below)

docs/docs/configuration/restream.md

@@ -7,7 +7,7 @@ title: Restream
 
 Frigate can restream your video feed as an RTSP feed for other applications such as Home Assistant to utilize it at `rtsp://<frigate_host>:8554/<camera_name>`. Port 8554 must be open. [This allows you to use a video feed for detection in Frigate and Home Assistant live view at the same time without having to make two separate connections to the camera](#reduce-connections-to-camera). The video feed is copied from the original video feed directly to avoid re-encoding. This feed does not include any annotation by Frigate.
 
-Frigate uses [go2rtc](https://github.com/AlexxIT/go2rtc/tree/v1.9.4) to provide its restream and MSE/WebRTC capabilities. The go2rtc config is hosted at the `go2rtc` in the config, see [go2rtc docs](https://github.com/AlexxIT/go2rtc/tree/v1.9.4#configuration) for more advanced configurations and features.
+Frigate uses [go2rtc](https://github.com/AlexxIT/go2rtc/tree/v1.9.2) to provide its restream and MSE/WebRTC capabilities. The go2rtc config is hosted at the `go2rtc` in the config, see [go2rtc docs](https://github.com/AlexxIT/go2rtc/tree/v1.9.2#configuration) for more advanced configurations and features.
 
 :::note
 
@@ -132,9 +132,31 @@ cameras:
             - detect
 ```
 
+## Handling Complex Passwords
+
+go2rtc expects URL-encoded passwords in the config, [urlencoder.org](https://urlencoder.org) can be used for this purpose. 
+
+For example:
+
+```yaml
+go2rtc:
+  streams:
+    my_camera: rtsp://username:$@foo%@192.168.1.100
+```
+
+becomes
+
+```yaml
+go2rtc:
+  streams:
+    my_camera: rtsp://username:$%40foo%25@192.168.1.100
+```
+
+See [this comment(https://github.com/AlexxIT/go2rtc/issues/1217#issuecomment-2242296489) for more information.
+
 ## Advanced Restream Configurations
 
-The [exec](https://github.com/AlexxIT/go2rtc/tree/v1.9.4#source-exec) source in go2rtc can be used for custom ffmpeg commands. An example is below:
+The [exec](https://github.com/AlexxIT/go2rtc/tree/v1.9.2#source-exec) source in go2rtc can be used for custom ffmpeg commands. An example is below:
 
 NOTE: The output will need to be passed with two curly braces `{{output}}`
 

docs/docs/configuration/semantic_search.md

@@ -5,13 +5,21 @@ title: Using Semantic Search
 
 Semantic Search in Frigate allows you to find tracked objects within your review items using either the image itself, a user-defined text description, or an automatically generated one. This feature works by creating _embeddings_ — numerical vector representations — for both the images and text descriptions of your tracked objects. By comparing these embeddings, Frigate assesses their similarities to deliver relevant search results.
 
-Frigate has support for two models to create embeddings, both of which run locally: [OpenAI CLIP](https://openai.com/research/clip) and [all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2). Embeddings are then saved to Frigate's database.
+Frigate uses [Jina AI's CLIP model](https://huggingface.co/jinaai/jina-clip-v1) to create and save embeddings to Frigate's database. All of this runs locally.
 
 Semantic Search is accessed via the _Explore_ view in the Frigate UI.
 
+## Minimum System Requirements
+
+Semantic Search works by running a large AI model locally on your system. Small or underpowered systems like a Raspberry Pi will not run Semantic Search reliably or at all.
+
+A minimum of 8GB of RAM is required to use Semantic Search. A GPU is not strictly required but will provide a significant performance increase over CPU-only systems.
+
+For best performance, 16GB or more of RAM and a dedicated GPU are recommended.
+
 ## Configuration
 
-Semantic search is disabled by default, and must be enabled in your config file before it can be used. Semantic Search is a global configuration setting.
+Semantic Search is disabled by default, and must be enabled in your config file or in the UI's Settings page before it can be used. Semantic Search is a global configuration setting.
 
 ```yaml
 semantic_search:
@@ -21,24 +29,64 @@ semantic_search:
 
 :::tip
 
-The embeddings database can be re-indexed from the existing tracked objects in your database by adding `reindex: True` to your `semantic_search` configuration. Depending on the number of tracked objects you have, it can take a long while to complete and may max out your CPU while indexing. Make sure to set the config back to `False` before restarting Frigate again.
+The embeddings database can be re-indexed from the existing tracked objects in your database by adding `reindex: True` to your `semantic_search` configuration or by toggling the switch on the Search Settings page in the UI and restarting Frigate. Depending on the number of tracked objects you have, it can take a long while to complete and may max out your CPU while indexing. Make sure to turn the UI's switch off or set the config back to `False` before restarting Frigate again.
 
-If you are enabling the Search feature for the first time, be advised that Frigate does not automatically index older tracked objects. You will need to enable the `reindex` feature in order to do that.
+If you are enabling Semantic Search for the first time, be advised that Frigate does not automatically index older tracked objects. You will need to enable the `reindex` feature in order to do that.
 
 :::
 
-### OpenAI CLIP
+### Jina AI CLIP
 
-This model is able to embed both images and text into the same vector space, which allows `image -> image` and `text -> image` similarity searches. Frigate uses this model on tracked objects to encode the thumbnail image and store it in the database. When searching for tracked objects via text in the search box, Frigate will perform a `text -> image` similarity search against this embedding. When clicking "Find Similar" in the tracked object detail pane, Frigate will perform an `image -> image` similarity search to retrieve the closest matching thumbnails.
+The vision model is able to embed both images and text into the same vector space, which allows `image -> image` and `text -> image` similarity searches. Frigate uses this model on tracked objects to encode the thumbnail image and store it in the database. When searching for tracked objects via text in the search box, Frigate will perform a `text -> image` similarity search against this embedding. When clicking "Find Similar" in the tracked object detail pane, Frigate will perform an `image -> image` similarity search to retrieve the closest matching thumbnails.
 
-### all-MiniLM-L6-v2
+The text model is used to embed tracked object descriptions and perform searches against them. Descriptions can be created, viewed, and modified on the Explore page when clicking on thumbnail of a tracked object. See [the Generative AI docs](/configuration/genai.md) for more information on how to automatically generate tracked object descriptions.
 
-This is a sentence embedding model that has been fine tuned on over 1 billion sentence pairs. This model is used to embed tracked object descriptions and perform searches against them. Descriptions can be created, viewed, and modified on the Search page when clicking on the gray tracked object chip at the top left of each review item. See [the Generative AI docs](/configuration/genai.md) for more information on how to automatically generate tracked object descriptions.
+Differently weighted versions of the Jina model are available and can be selected by setting the `model_size` config option as `small` or `large`:
 
-## Usage
+```yaml
+semantic_search:
+  enabled: True
+  model_size: small
+```
 
-1. Semantic search is used in conjunction with the other filters available on the Search page. Use a combination of traditional filtering and semantic search for the best results.
-2. The comparison between text and image embedding distances generally means that results matching `description` will appear first, even if a `thumbnail` embedding may be a better match. Play with the "Search Type" filter to help find what you are looking for.
-3. Make your search language and tone closely match your descriptions. If you are using thumbnail search, phrase your query as an image caption.
-4. Semantic search on thumbnails tends to return better results when matching large subjects that take up most of the frame. Small things like "cat" tend to not work well.
-5. Experiment! Find a tracked object you want to test and start typing keywords to see what works for you.
+- Configuring the `large` model employs the full Jina model and will automatically run on the GPU if applicable.
+- Configuring the `small` model employs a quantized version of the Jina model that uses less RAM and runs on CPU with a very negligible difference in embedding quality.
+
+### GPU Acceleration
+
+The CLIP models are downloaded in ONNX format, and the `large` model can be accelerated using GPU hardware, when available. This depends on the Docker build that is used.
+
+```yaml
+semantic_search:
+  enabled: True
+  model_size: large
+```
+
+:::info
+
+If the correct build is used for your GPU and the `large` model is configured, then the GPU will be detected and used automatically.
+
+**NOTE:** Object detection and Semantic Search are independent features. If you want to use your GPU with Semantic Search, you must choose the appropriate Frigate Docker image for your GPU.
+
+- **AMD**
+
+  - ROCm will automatically be detected and used for Semantic Search in the `-rocm` Frigate image.
+
+- **Intel**
+
+  - OpenVINO will automatically be detected and used for Semantic Search in the default Frigate image.
+
+- **Nvidia**
+  - Nvidia GPUs will automatically be detected and used for Semantic Search in the `-tensorrt` Frigate image.
+  - Jetson devices will automatically be detected and used for Semantic Search in the `-tensorrt-jp(4/5)` Frigate image.
+
+:::
+
+## Usage and Best Practices
+
+1. Semantic Search is used in conjunction with the other filters available on the Explore page. Use a combination of traditional filtering and Semantic Search for the best results.
+2. Use the thumbnail search type when searching for particular objects in the scene. Use the description search type when attempting to discern the intent of your object.
+3. Because of how the AI models Frigate uses have been trained, the comparison between text and image embedding distances generally means that with multi-modal (`thumbnail` and `description`) searches, results matching `description` will appear first, even if a `thumbnail` embedding may be a better match. Play with the "Search Type" setting to help find what you are looking for. Note that if you are generating descriptions for specific objects or zones only, this may cause search results to prioritize the objects with descriptions even if the the ones without them are more relevant.
+4. Make your search language and tone closely match exactly what you're looking for. If you are using thumbnail search, **phrase your query as an image caption**. Searching for "red car" may not work as well as "red sedan driving down a residential street on a sunny day".
+5. Semantic search on thumbnails tends to return better results when matching large subjects that take up most of the frame. Small things like "cat" tend to not work well.
+6. Experiment! Find a tracked object you want to test and start typing keywords and phrases to see what works for you.

docs/docs/frigate/camera_setup.md

@@ -28,7 +28,7 @@ For the Dahua/Loryta 5442 camera, I use the following settings:
 - Encode Mode: H.264
 - Resolution: 2688\*1520
 - Frame Rate(FPS): 15
-- I Frame Interval: 30
+- I Frame Interval: 30 (15 can also be used to prioritize streaming performance - see the [camera settings recommendations](../configuration/live) for more info)
 
 **Sub Stream (Detection)**
 

docs/docs/frigate/installation.md

@@ -81,15 +81,15 @@ You can calculate the **minimum** shm size for each camera with the following fo
 
 ```console
 # Replace <width> and <height>
-$ python -c 'print("{:.2f}MB".format((<width> * <height> * 1.5 * 10 + 270480) / 1048576))'
+$ python -c 'print("{:.2f}MB".format((<width> * <height> * 1.5 * 20 + 270480) / 1048576))'
 
-# Example for 1280x720
-$ python -c 'print("{:.2f}MB".format((1280 * 720 * 1.5 * 10 + 270480) / 1048576))'
-13.44MB
+# Example for 1280x720, including logs
+$ python -c 'print("{:.2f}MB".format((1280 * 720 * 1.5 * 20 + 270480) / 1048576)) + 40'
+46.63MB
 
 # Example for eight cameras detecting at 1280x720, including logs
-$ python -c 'print("{:.2f}MB".format(((1280 * 720 * 1.5 * 10 + 270480) / 1048576) * 8 + 40))'
-136.99MB
+$ python -c 'print("{:.2f}MB".format(((1280 * 720 * 1.5 * 20 + 270480) / 1048576) * 8 + 40))'
+253MB
 ```
 
 The shm size cannot be set per container for Home Assistant add-ons. However, this is probably not required since by default Home Assistant Supervisor allocates `/dev/shm` with half the size of your total memory. If your machine has 8GB of memory, chances are that Frigate will have access to up to 4GB without any additional configuration.
@@ -193,8 +193,9 @@ services:
     container_name: frigate
     privileged: true # this may not be necessary for all setups
     restart: unless-stopped
+    stop_grace_period: 30s # allow enough time to shut down the various services
     image: ghcr.io/blakeblackshear/frigate:stable
-    shm_size: "64mb" # update for your cameras based on calculation above
+    shm_size: "512mb" # update for your cameras based on calculation above
     devices:
       - /dev/bus/usb:/dev/bus/usb # Passes the USB Coral, needs to be modified for other versions
       - /dev/apex_0:/dev/apex_0 # Passes a PCIe Coral, follow driver instructions here https://coral.ai/docs/m2/get-started/#2a-on-linux
@@ -224,6 +225,7 @@ If you can't use docker compose, you can run the container with something simila
 docker run -d \
   --name frigate \
   --restart=unless-stopped \
+  --stop-timeout 30 \
   --mount type=tmpfs,target=/tmp/cache,tmpfs-size=1000000000 \
   --device /dev/bus/usb:/dev/bus/usb \
   --device /dev/dri/renderD128 \
@@ -250,10 +252,7 @@ The community supported docker image tags for the current stable version are:
 - `stable-tensorrt-jp5` - Frigate build optimized for nvidia Jetson devices running Jetpack 5
 - `stable-tensorrt-jp4` - Frigate build optimized for nvidia Jetson devices running Jetpack 4.6
 - `stable-rk` - Frigate build for SBCs with Rockchip SoC
-- `stable-rocm` - Frigate build for [AMD GPUs and iGPUs](../configuration/object_detectors.md#amdrocm-gpu-detector), all drivers
-  - `stable-rocm-gfx900` - AMD gfx900 driver only
-  - `stable-rocm-gfx1030` - AMD gfx1030 driver only
-  - `stable-rocm-gfx1100` - AMD gfx1100 driver only
+- `stable-rocm` - Frigate build for [AMD GPUs](../configuration/object_detectors.md#amdrocm-gpu-detector)
   - `stable-h8l` - Frigate build for the Hailo-8L M.2 PICe Raspberry Pi 5 hat
 
 ## Home Assistant Addon
@@ -306,8 +305,15 @@ To install make sure you have the [community app plugin here](https://forums.unr
 
 ## Proxmox
 
-It is recommended to run Frigate in LXC, rather than in a VM, for maximum performance. The setup can be complex so be prepared to read the Proxmox and LXC documentation. Suggestions include:
+[According to Proxmox documentation](https://pve.proxmox.com/pve-docs/pve-admin-guide.html#chapter_pct) it is recommended that you run application containers like Frigate inside a Proxmox QEMU VM. This will give you all the advantages of application containerization, while also providing the benefits that VMs offer, such as strong isolation from the host and the ability to live-migrate, which otherwise isn’t possible with containers.
 
+:::warning
+
+If you choose to run Frigate via LXC in Proxmox the setup can be complex so be prepared to read the Proxmox and LXC documentation, Frigate does not officially support running inside of an LXC.
+
+:::
+
+ Suggestions include:
 - For Intel-based hardware acceleration, to allow access to the `/dev/dri/renderD128` device with major number 226 and minor number 128, add the following lines to the `/etc/pve/lxc/<id>.conf` LXC configuration:
   - `lxc.cgroup2.devices.allow: c 226:128 rwm`
   - `lxc.mount.entry: /dev/dri/renderD128 dev/dri/renderD128 none bind,optional,create=file`

docs/docs/guides/configuring_go2rtc.md

@@ -13,7 +13,15 @@ Use of the bundled go2rtc is optional. You can still configure FFmpeg to connect
 
 # Setup a go2rtc stream
 
-First, you will want to configure go2rtc to connect to your camera stream by adding the stream you want to use for live view in your Frigate config file. For the best experience, you should set the stream name under go2rtc to match the name of your camera so that Frigate will automatically map it and be able to use better live view options for the camera. Avoid changing any other parts of your config at this step. Note that go2rtc supports [many different stream types](https://github.com/AlexxIT/go2rtc/tree/v1.9.4#module-streams), not just rtsp.
+First, you will want to configure go2rtc to connect to your camera stream by adding the stream you want to use for live view in your Frigate config file. Avoid changing any other parts of your config at this step. Note that go2rtc supports [many different stream types](https://github.com/AlexxIT/go2rtc/tree/v1.9.2#module-streams), not just rtsp.
+
+:::tip
+
+For the best experience, you should set the stream name under `go2rtc` to match the name of your camera so that Frigate will automatically map it and be able to use better live view options for the camera.
+
+See [the live view docs](../configuration/live.md#setting-stream-for-live-ui) for more information.
+
+:::
 
 ```yaml
 go2rtc:
@@ -39,8 +47,8 @@ After adding this to the config, restart Frigate and try to watch the live strea
 
 - Check Video Codec:
     - If the camera stream works in go2rtc but not in your browser, the video codec might be unsupported.
-    - If using H265, switch to H264. Refer to [video codec compatibility](https://github.com/AlexxIT/go2rtc/tree/v1.9.4#codecs-madness) in go2rtc documentation.
-    - If unable to switch from H265 to H264, or if the stream format is different (e.g., MJPEG), re-encode the video using [FFmpeg parameters](https://github.com/AlexxIT/go2rtc/tree/v1.9.4#source-ffmpeg). It supports rotating and resizing video feeds and hardware acceleration. Keep in mind that transcoding video from one format to another is a resource intensive task and you may be better off using the built-in jsmpeg view.
+    - If using H265, switch to H264. Refer to [video codec compatibility](https://github.com/AlexxIT/go2rtc/tree/v1.9.2#codecs-madness) in go2rtc documentation.
+    - If unable to switch from H265 to H264, or if the stream format is different (e.g., MJPEG), re-encode the video using [FFmpeg parameters](https://github.com/AlexxIT/go2rtc/tree/v1.9.2#source-ffmpeg). It supports rotating and resizing video feeds and hardware acceleration. Keep in mind that transcoding video from one format to another is a resource intensive task and you may be better off using the built-in jsmpeg view.
         ```yaml
         go2rtc:
           streams:

docs/docs/guides/getting_started.md

@@ -115,6 +115,7 @@ services:
   frigate:
     container_name: frigate
     restart: unless-stopped
+    stop_grace_period: 30s
     image: ghcr.io/blakeblackshear/frigate:stable
     volumes:
       - ./config:/config
@@ -306,7 +307,9 @@ By default, Frigate will retain video of all tracked objects for 10 days. The fu
 
 ### Step 7: Complete config
 
-At this point you have a complete config with basic functionality. You can see the [full config reference](../configuration/reference.md) for a complete list of configuration options.
+At this point you have a complete config with basic functionality. 
+- View [common configuration examples](../configuration/index.md#common-configuration-examples) for a list of common configuration examples.
+- View [full config reference](../configuration/reference.md) for a complete list of configuration options.
 
 ### Follow up
 

docs/docs/integrations/mqtt.md

@@ -94,6 +94,18 @@ Message published for each changed tracked object. The first message is publishe
 }
 ```
 
+### `frigate/tracked_object_update`
+
+Message published for updates to tracked object metadata, for example when GenAI runs and returns a tracked object description.
+
+```json
+{
+  "type": "description",
+  "id": "1607123955.475377-mxklsc",
+  "description": "The car is a red sedan moving away from the camera."
+}
+```
+
 ### `frigate/reviews`
 
 Message published for each changed review item. The first message is published when the `detection` or `alert` is initiated. When additional objects are detected or when a zone change occurs, it will publish a, `update` message with the same id. When the review activity has ended a final `end` message is published.

docs/docs/plus/first_model.md

@@ -5,7 +5,7 @@ title: Requesting your first model
 
 ## Step 1: Upload and annotate your images
 
-Before requesting your first model, you will need to upload at least 10 images to Frigate+. But for the best results, you should provide at least 100 verified images per camera. Keep in mind that varying conditions should be included. You will want images from cloudy days, sunny days, dawn, dusk, and night. Refer to the [integration docs](../integrations/plus.md#generate-an-api-key) for instructions on how to easily submit images to Frigate+ directly from Frigate.
+Before requesting your first model, you will need to upload and verify at least 1 image to Frigate+. The more images you upload, annotate, and verify the better your results will be. Most users start to see very good results once they have at least 100 verified images per camera. Keep in mind that varying conditions should be included. You will want images from cloudy days, sunny days, dawn, dusk, and night. Refer to the [integration docs](../integrations/plus.md#generate-an-api-key) for instructions on how to easily submit images to Frigate+ directly from Frigate.
 
 It is recommended to submit **both** true positives and false positives. This will help the model differentiate between what is and isn't correct. You should aim for a target of 80% true positive submissions and 20% false positives across all of your images. If you are experiencing false positives in a specific area, submitting true positives for any object type near that area in similar lighting conditions will help teach the model what that area looks like when no objects are present.
 
@@ -13,7 +13,7 @@ For more detailed recommendations, you can refer to the docs on [improving your
 
 ## Step 2: Submit a model request
 
-Once you have an initial set of verified images, you can request a model on the Models page. Each model request requires 1 of the 12 trainings that you receive with your annual subscription. This model will support all [label types available](./index.md#available-label-types) even if you do not submit any examples for those labels. Model creation can take up to 36 hours.
+Once you have an initial set of verified images, you can request a model on the Models page. For guidance on choosing a model type, refer to [this part of the documentation](./index.md#available-model-types). Each model request requires 1 of the 12 trainings that you receive with your annual subscription. This model will support all [label types available](./index.md#available-label-types) even if you do not submit any examples for those labels. Model creation can take up to 36 hours.
 ![Plus Models Page](/img/plus/plus-models.jpg)
 
 ## Step 3: Set your model id in the config

docs/docs/plus/improving_model.md

@@ -3,7 +3,7 @@ id: improving_model
 title: Improving your model
 ---
 
-You may find that Frigate+ models result in more false positives initially, but by submitting true and false positives, the model will improve. Because a limited number of users submitted images to Frigate+ prior to this launch, you may need to submit several hundred images per camera to see good results. With all the new images now being submitted, future base models will improve as more and more users (including you) submit examples to Frigate+. Note that only verified images will be used when training your model. Submitting an image from Frigate as a true or false positive will not verify the image. You still must verify the image in Frigate+ in order for it to be used in training.
+You may find that Frigate+ models result in more false positives initially, but by submitting true and false positives, the model will improve. With all the new images now being submitted by subscribers, future base models will improve as more and more examples are incorporated. Note that only images with at least one verified label will be used when training your model. Submitting an image from Frigate as a true or false positive will not verify the image. You still must verify the image in Frigate+ in order for it to be used in training.
 
 - **Submit both true positives and false positives**. This will help the model differentiate between what is and isn't correct. You should aim for a target of 80% true positive submissions and 20% false positives across all of your images. If you are experiencing false positives in a specific area, submitting true positives for any object type near that area in similar lighting conditions will help teach the model what that area looks like when no objects are present.
 - **Lower your thresholds a little in order to generate more false/true positives near the threshold value**. For example, if you have some false positives that are scoring at 68% and some true positives scoring at 72%, you can try lowering your threshold to 65% and submitting both true and false positives within that range. This will help the model learn and widen the gap between true and false positive scores.
@@ -36,18 +36,17 @@ Misidentified objects should have a correct label added. For example, if a perso
 
 ## Shortcuts for a faster workflow
 
-|Shortcut Key|Description|
-|-----|--------|
-|`?`|Show all keyboard shortcuts|
-|`w`|Add box|
-|`d`|Toggle difficult|
-|`s`|Switch to the next label|
-|`tab`|Select next largest box|
-|`del`|Delete current box|
-|`esc`|Deselect/Cancel|
-|`← ↑ → ↓`|Move box|
-|`Shift + ← ↑ → ↓`|Resize box|
-|`-`|Zoom out|
-|`=`|Zoom in|
-|`f`|Hide/show all but current box|
-|`spacebar`|Verify and save|
+| Shortcut Key      | Description                   |
+| ----------------- | ----------------------------- |
+| `?`               | Show all keyboard shortcuts   |
+| `w`               | Add box                       |
+| `d`               | Toggle difficult              |
+| `s`               | Switch to the next label      |
+| `tab`             | Select next largest box       |
+| `del`             | Delete current box            |
+| `esc`             | Deselect/Cancel               |
+| `← ↑ → ↓`         | Move box                      |
+| `Shift + ← ↑ → ↓` | Resize box                    |
+| `scrollwheel`     | Zoom in/out                   |
+| `f`               | Hide/show all but current box |
+| `spacebar`        | Verify and save               |

docs/docs/plus/index.md

@@ -15,17 +15,36 @@ With a subscription, 12 model trainings per year are included. If you cancel you
 
 Information on how to integrate Frigate+ with Frigate can be found in the [integration docs](../integrations/plus.md).
 
+## Available model types
+
+There are two model types offered in Frigate+: `mobiledet` and `yolonas`. Both of these models are object detection models and are trained to detect the same set of labels [listed below](#available-label-types).
+
+Not all model types are supported by all detectors, so it's important to choose a model type to match your detector as shown in the table under [supported detector types](#supported-detector-types).
+
+| Model Type  | Description                                                                                                                                  |
+| ----------- | -------------------------------------------------------------------------------------------------------------------------------------------- |
+| `mobiledet` | Based on the same architecture as the default model included with Frigate. Runs on Google Coral devices and CPUs.                            |
+| `yolonas`   | A newer architecture that offers slightly higher accuracy and improved detection of small objects. Runs on Intel, NVidia GPUs, and AMD GPUs. |
+
 ## Supported detector types
 
+Currently, Frigate+ models support CPU (`cpu`), Google Coral (`edgetpu`), OpenVino (`openvino`), ONNX (`onnx`), and ROCm (`rocm`) detectors.
+
 :::warning
 
-Frigate+ models are not supported for TensorRT or OpenVino yet.
+Using Frigate+ models with `onnx` and `rocm` is only available with Frigate 0.15, which is still under development.
 
 :::
 
-Currently, Frigate+ models only support CPU (`cpu`) and Coral (`edgetpu`) models. OpenVino is next in line to gain support.
+| Hardware                                                                                                                     | Recommended Detector Type | Recommended Model Type |
+| ---------------------------------------------------------------------------------------------------------------------------- | ------------------------- | ---------------------- |
+| [CPU](/configuration/object_detectors.md#cpu-detector-not-recommended)                                                       | `cpu`                     | `mobiledet`            |
+| [Coral (all form factors)](/configuration/object_detectors.md#edge-tpu-detector)                                             | `edgetpu`                 | `mobiledet`            |
+| [Intel](/configuration/object_detectors.md#openvino-detector)                                                                | `openvino`                | `yolonas`              |
+| [NVidia GPU](https://deploy-preview-13787--frigate-docs.netlify.app/configuration/object_detectors#onnx)\*                   | `onnx`                    | `yolonas`              |
+| [AMD ROCm GPU](https://deploy-preview-13787--frigate-docs.netlify.app/configuration/object_detectors#amdrocm-gpu-detector)\* | `rocm`                    | `yolonas`              |
 
-The models are created using the same MobileDet architecture as the default model. Additional architectures will be added in future releases as needed.
+_\* Requires Frigate 0.15_
 
 ## Available label types
 

docs/docs/troubleshooting/edgetpu.md

@@ -49,7 +49,10 @@ The USB Coral can become stuck and need to be restarted, this can happen for a n
 
 ## PCIe Coral Not Detected
 
-The most common reason for the PCIe coral not being detected is that the driver has not been installed. See [the coral docs](https://coral.ai/docs/m2/get-started/#2-install-the-pcie-driver-and-edge-tpu-runtime) for how to install the driver for the PCIe based coral.
+The most common reason for the PCIe Coral not being detected is that the driver has not been installed. This process varies based on what OS and kernel that is being run. 
+
+- In most cases [the Coral docs](https://coral.ai/docs/m2/get-started/#2-install-the-pcie-driver-and-edge-tpu-runtime) show how to install the driver for the PCIe based Coral.
+- For Ubuntu 22.04+ https://github.com/jnicolson/gasket-builder can be used to build and install the latest version of the driver.
 
 ## Only One PCIe Coral Is Detected With Coral Dual EdgeTPU
 

docs/docs/troubleshooting/faqs.md

@@ -98,3 +98,11 @@ docker run -d \
   -p 8555:8555/udp \
   ghcr.io/blakeblackshear/frigate:stable
 ```
+
+### My RTSP stream works fine in VLC, but it does not work when I put the same URL in my Frigate config. Is this a bug?
+
+No. Frigate uses the TCP protocol to connect to your camera's RTSP URL. VLC automatically switches between UDP and TCP depending on network conditions and stream availability. So a stream that works in VLC but not in Frigate is likely due to VLC selecting UDP as the transfer protocol.
+
+TCP ensures that all data packets arrive in the correct order. This is crucial for video recording, decoding, and stream processing, which is why Frigate enforces a TCP connection. UDP is faster but less reliable, as it does not guarantee packet delivery or order, and VLC does not have the same requirements as Frigate.
+
+You can still configure Frigate to use UDP by using ffmpeg input args or the preset `preset-rtsp-udp`. See the [ffmpeg presets](/configuration/ffmpeg_presets) documentation.

docs/docs/troubleshooting/recordings.md

@@ -3,7 +3,15 @@ id: recordings
 title: Troubleshooting Recordings
 ---
 
-### WARNING : Unable to keep up with recording segments in cache for camera. Keeping the 5 most recent segments out of 6 and discarding the rest...
+## I have Frigate configured for motion recording only, but it still seems to be recording even with no motion. Why?
+
+You'll want to:
+
+- Make sure your camera's timestamp is masked out with a motion mask. Even if there is no motion occurring in your scene, your motion settings may be sensitive enough to count your timestamp as motion.
+- If you have audio detection enabled, keep in mind that audio that is heard above `min_volume` is considered motion.
+- [Tune your motion detection settings](/configuration/motion_detection) either by editing your config file or by using the UI's Motion Tuner.
+
+## I see the message: WARNING : Unable to keep up with recording segments in cache for camera. Keeping the 5 most recent segments out of 6 and discarding the rest...
 
 This error can be caused by a number of different issues. The first step in troubleshooting is to enable debug logging for recording. This will enable logging showing how long it takes for recordings to be moved from RAM cache to the disk.
 
@@ -40,6 +48,7 @@ On linux, some helpful tools/commands in diagnosing would be:
 On modern linux kernels, the system will utilize some swap if enabled. Setting vm.swappiness=1 no longer means that the kernel will only swap in order to avoid OOM. To prevent any swapping inside a container, set allocations memory and memory+swap to be the same and disable swapping by setting the following docker/podman run parameters:
 
 **Compose example**
+
 ```yaml
 version: "3.9"
 services:
@@ -54,6 +63,7 @@ services:
 ```
 
 **Run command example**
+
 ```
 --memory=<MAXRAM> --memory-swap=<MAXSWAP> --memory-swappiness=0
 ```

docs/package.json

@@ -17,15 +17,15 @@
     "write-heading-ids": "docusaurus write-heading-ids"
   },
   "dependencies": {
-    "@docusaurus/core": "^3.5.2",
-    "@docusaurus/preset-classic": "^3.5.2",
-    "@docusaurus/theme-mermaid": "^3.5.2",
-    "@docusaurus/plugin-content-docs": "^3.5.2",
-    "@mdx-js/react": "^3.0.1",
+    "@docusaurus/core": "^3.6.3",
+    "@docusaurus/preset-classic": "^3.6.3",
+    "@docusaurus/theme-mermaid": "^3.6.3",
+    "@docusaurus/plugin-content-docs": "^3.6.3",
+    "@mdx-js/react": "^3.1.0",
     "clsx": "^2.1.1",
-    "docusaurus-plugin-openapi-docs": "^4.1.0",
-    "docusaurus-theme-openapi-docs": "^4.1.0",
-    "prism-react-renderer": "^2.4.0",
+    "docusaurus-plugin-openapi-docs": "^4.3.1",
+    "docusaurus-theme-openapi-docs": "^4.3.1",
+    "prism-react-renderer": "^2.4.1",
     "raw-loader": "^4.0.2",
     "react": "^18.3.1",
     "react-dom": "^18.3.1"

docs/sidebars.ts

@@ -26,7 +26,7 @@ const sidebars: SidebarsConfig = {
         {
           type: 'link',
           label: 'Go2RTC Configuration Reference',
-          href: 'https://github.com/AlexxIT/go2rtc/tree/v1.9.4#configuration',
+          href: 'https://github.com/AlexxIT/go2rtc/tree/v1.9.2#configuration',
         } as PropSidebarItemLink,
       ],
       Detectors: [
@@ -36,6 +36,8 @@ const sidebars: SidebarsConfig = {
       'Semantic Search': [
         'configuration/semantic_search',
         'configuration/genai',
+        'configuration/face_recognition',
+        'configuration/license_plate_recognition',
       ],
       Cameras: [
         'configuration/cameras',

frigate/__main__.py

@@ -3,12 +3,15 @@ import faulthandler
 import signal
 import sys
 import threading
+from typing import Union
 
+import ruamel.yaml
 from pydantic import ValidationError
 
 from frigate.app import FrigateApp
 from frigate.config import FrigateConfig
 from frigate.log import setup_logging
+from frigate.util.config import find_config_file
 
 
 def main() -> None:
@@ -42,10 +45,51 @@ def main() -> None:
         print("*************************************************************")
         print("*************************************************************")
         print("***    Config Validation Errors                           ***")
-        print("*************************************************************")
+        print("*************************************************************\n")
+        # Attempt to get the original config file for line number tracking
+        config_path = find_config_file()
+        with open(config_path, "r") as f:
+            yaml_config = ruamel.yaml.YAML()
+            yaml_config.preserve_quotes = True
+            full_config = yaml_config.load(f)
+
         for error in e.errors():
-            location = ".".join(str(item) for item in error["loc"])
-            print(f"{location}: {error['msg']}")
+            error_path = error["loc"]
+
+            current = full_config
+            line_number = "Unknown"
+            last_line_number = "Unknown"
+
+            try:
+                for i, part in enumerate(error_path):
+                    key: Union[int, str] = (
+                        int(part) if isinstance(part, str) and part.isdigit() else part
+                    )
+
+                    if isinstance(current, ruamel.yaml.comments.CommentedMap):
+                        current = current[key]
+                    elif isinstance(current, list):
+                        if isinstance(key, int):
+                            current = current[key]
+
+                    if hasattr(current, "lc"):
+                        last_line_number = current.lc.line
+
+                    if i == len(error_path) - 1:
+                        if hasattr(current, "lc"):
+                            line_number = current.lc.line
+                        else:
+                            line_number = last_line_number
+
+            except Exception as traverse_error:
+                print(f"Could not determine exact line number: {traverse_error}")
+
+            if current != full_config:
+                print(f"Line #  : {line_number}")
+                print(f"Key     : {' -> '.join(map(str, error_path))}")
+                print(f"Value   : {error.get('input', '-')}")
+            print(f"Message : {error.get('msg', error.get('type', 'Unknown'))}\n")
+
         print("*************************************************************")
         print("***    End Config Validation Errors                       ***")
         print("*************************************************************")

frigate/api/app.py

@@ -7,27 +7,30 @@ import os
 import traceback
 from datetime import datetime, timedelta
 from functools import reduce
+from io import StringIO
 from typing import Any, Optional
 
 import requests
+import ruamel.yaml
 from fastapi import APIRouter, Body, Path, Request, Response
 from fastapi.encoders import jsonable_encoder
 from fastapi.params import Depends
 from fastapi.responses import JSONResponse, PlainTextResponse
 from markupsafe import escape
 from peewee import operator
+from pydantic import ValidationError
 
-from frigate.api.defs.app_body import AppConfigSetBody
-from frigate.api.defs.app_query_parameters import AppTimelineHourlyQueryParameters
+from frigate.api.defs.query.app_query_parameters import AppTimelineHourlyQueryParameters
+from frigate.api.defs.request.app_body import AppConfigSetBody
 from frigate.api.defs.tags import Tags
 from frigate.config import FrigateConfig
-from frigate.const import CONFIG_DIR
 from frigate.models import Event, Timeline
 from frigate.util.builtin import (
     clean_camera_user_pass,
     get_tz_modifiers,
     update_yaml_from_url,
 )
+from frigate.util.config import find_config_file
 from frigate.util.services import (
     ffprobe_stream,
     get_nvidia_driver_info,
@@ -134,9 +137,27 @@ def config(request: Request):
         for zone_name, zone in config_obj.cameras[camera_name].zones.items():
             camera_dict["zones"][zone_name]["color"] = zone.color
 
+    # remove go2rtc stream passwords
+    go2rtc: dict[str, any] = config_obj.go2rtc.model_dump(
+        mode="json", warnings="none", exclude_none=True
+    )
+    for stream_name, stream in go2rtc.get("streams", {}).items():
+        if stream is None:
+            continue
+        if isinstance(stream, str):
+            cleaned = clean_camera_user_pass(stream)
+        else:
+            cleaned = []
+
+            for item in stream:
+                cleaned.append(clean_camera_user_pass(item))
+
+        config["go2rtc"]["streams"][stream_name] = cleaned
+
     config["plus"] = {"enabled": request.app.frigate_config.plus_api.is_active()}
     config["model"]["colormap"] = config_obj.model.colormap
 
+    # use merged labelamp
     for detector_config in config["detectors"].values():
         detector_config["model"]["labelmap"] = (
             request.app.frigate_config.model.merged_labelmap
@@ -147,13 +168,7 @@ def config(request: Request):
 
 @router.get("/config/raw")
 def config_raw():
-    config_file = os.environ.get("CONFIG_FILE", "/config/config.yml")
-
-    # Check if we can use .yaml instead of .yml
-    config_file_yaml = config_file.replace(".yml", ".yaml")
-
-    if os.path.isfile(config_file_yaml):
-        config_file = config_file_yaml
+    config_file = find_config_file()
 
     if not os.path.isfile(config_file):
         return JSONResponse(
@@ -173,7 +188,6 @@ def config_raw():
 @router.post("/config/save")
 def config_save(save_option: str, body: Any = Body(media_type="text/plain")):
     new_config = body.decode()
-
     if not new_config:
         return JSONResponse(
             content=(
@@ -184,13 +198,64 @@ def config_save(save_option: str, body: Any = Body(media_type="text/plain")):
 
     # Validate the config schema
     try:
+        # Use ruamel to parse and preserve line numbers
+        yaml_config = ruamel.yaml.YAML()
+        yaml_config.preserve_quotes = True
+        full_config = yaml_config.load(StringIO(new_config))
+
         FrigateConfig.parse_yaml(new_config)
+
+    except ValidationError as e:
+        error_message = []
+
+        for error in e.errors():
+            error_path = error["loc"]
+            current = full_config
+            line_number = "Unknown"
+            last_line_number = "Unknown"
+
+            try:
+                for i, part in enumerate(error_path):
+                    key = int(part) if part.isdigit() else part
+
+                    if isinstance(current, ruamel.yaml.comments.CommentedMap):
+                        current = current[key]
+                    elif isinstance(current, list):
+                        current = current[key]
+
+                    if hasattr(current, "lc"):
+                        last_line_number = current.lc.line
+
+                    if i == len(error_path) - 1:
+                        if hasattr(current, "lc"):
+                            line_number = current.lc.line
+                        else:
+                            line_number = last_line_number
+
+            except Exception:
+                line_number = "Unable to determine"
+
+            error_message.append(
+                f"Line {line_number}: {' -> '.join(map(str, error_path))} - {error.get('msg', error.get('type', 'Unknown'))}"
+            )
+
+        return JSONResponse(
+            content=(
+                {
+                    "success": False,
+                    "message": "Your configuration is invalid.\nSee the official documentation at docs.frigate.video.\n\n"
+                    + "\n".join(error_message),
+                }
+            ),
+            status_code=400,
+        )
+
     except Exception:
         return JSONResponse(
             content=(
                 {
                     "success": False,
-                    "message": f"\nConfig Error:\n\n{escape(str(traceback.format_exc()))}",
+                    "message": f"\nYour configuration is invalid.\nSee the official documentation at docs.frigate.video.\n\n{escape(str(traceback.format_exc()))}",
                 }
             ),
             status_code=400,
@@ -198,13 +263,7 @@ def config_save(save_option: str, body: Any = Body(media_type="text/plain")):
 
     # Save the config to file
     try:
-        config_file = os.environ.get("CONFIG_FILE", "/config/config.yml")
-
-        # Check if we can use .yaml instead of .yml
-        config_file_yaml = config_file.replace(".yml", ".yaml")
-
-        if os.path.isfile(config_file_yaml):
-            config_file = config_file_yaml
+        config_file = find_config_file()
 
         with open(config_file, "w") as f:
             f.write(new_config)
@@ -253,13 +312,7 @@ def config_save(save_option: str, body: Any = Body(media_type="text/plain")):
 
 @router.put("/config/set")
 def config_set(request: Request, body: AppConfigSetBody):
-    config_file = os.environ.get("CONFIG_FILE", f"{CONFIG_DIR}/config.yml")
-
-    # Check if we can use .yaml instead of .yml
-    config_file_yaml = config_file.replace(".yml", ".yaml")
-
-    if os.path.isfile(config_file_yaml):
-        config_file = config_file_yaml
+    config_file = find_config_file()
 
     with open(config_file, "r") as f:
         old_raw_config = f.read()

frigate/api/auth.py

@@ -18,7 +18,7 @@ from joserfc import jwt
 from peewee import DoesNotExist
 from slowapi import Limiter
 
-from frigate.api.defs.app_body import (
+from frigate.api.defs.request.app_body import (
     AppPostLoginBody,
     AppPostUsersBody,
     AppPutPasswordBody,
@@ -85,7 +85,12 @@ def get_remote_addr(request: Request):
             return str(ip)
 
     # if there wasn't anything in the route, just return the default
-    return request.remote_addr or "127.0.0.1"
+    remote_addr = None
+
+    if hasattr(request, "remote_addr"):
+        remote_addr = request.remote_addr
+
+    return remote_addr or "127.0.0.1"
 
 
 def get_jwt_secret() -> str:
@@ -324,7 +329,7 @@ def login(request: Request, body: AppPostLoginBody):
     try:
         db_user: User = User.get_by_id(user)
     except DoesNotExist:
-        return JSONResponse(content={"message": "Login failed"}, status_code=400)
+        return JSONResponse(content={"message": "Login failed"}, status_code=401)
 
     password_hash = db_user.password_hash
     if verify_password(password, password_hash):
@@ -335,7 +340,7 @@ def login(request: Request, body: AppPostLoginBody):
             response, JWT_COOKIE_NAME, encoded_jwt, expiration, JWT_COOKIE_SECURE
         )
         return response
-    return JSONResponse(content={"message": "Login failed"}, status_code=400)
+    return JSONResponse(content={"message": "Login failed"}, status_code=401)
 
 
 @router.get("/users")
@@ -357,6 +362,7 @@ def create_user(request: Request, body: AppPostUsersBody):
         {
             User.username: body.username,
             User.password_hash: password_hash,
+            User.notification_tokens: [],
         }
     ).execute()
     return JSONResponse(content={"username": body.username})

frigate/api/classification.py

@@ -0,0 +1,127 @@
+"""Object classification APIs."""
+
+import logging
+import os
+import random
+import shutil
+import string
+
+from fastapi import APIRouter, Request, UploadFile
+from fastapi.responses import JSONResponse
+from pathvalidate import sanitize_filename
+
+from frigate.api.defs.tags import Tags
+from frigate.const import FACE_DIR
+from frigate.embeddings import EmbeddingsContext
+
+logger = logging.getLogger(__name__)
+
+router = APIRouter(tags=[Tags.events])
+
+
+@router.get("/faces")
+def get_faces():
+    face_dict: dict[str, list[str]] = {}
+
+    for name in os.listdir(FACE_DIR):
+        face_dir = os.path.join(FACE_DIR, name)
+
+        if not os.path.isdir(face_dir):
+            continue
+
+        face_dict[name] = []
+
+        for file in sorted(
+            os.listdir(face_dir),
+            key=lambda f: os.path.getctime(os.path.join(face_dir, f)),
+            reverse=True,
+        ):
+            face_dict[name].append(file)
+
+    return JSONResponse(status_code=200, content=face_dict)
+
+
+@router.post("/faces/{name}")
+async def register_face(request: Request, name: str, file: UploadFile):
+    if not request.app.frigate_config.face_recognition.enabled:
+        return JSONResponse(
+            status_code=400,
+            content={"message": "Face recognition is not enabled.", "success": False},
+        )
+
+    context: EmbeddingsContext = request.app.embeddings
+    result = context.register_face(name, await file.read())
+    return JSONResponse(
+        status_code=200 if result.get("success", True) else 400,
+        content=result,
+    )
+
+
+@router.post("/faces/train/{name}/classify")
+def train_face(request: Request, name: str, body: dict = None):
+    if not request.app.frigate_config.face_recognition.enabled:
+        return JSONResponse(
+            status_code=400,
+            content={"message": "Face recognition is not enabled.", "success": False},
+        )
+
+    json: dict[str, any] = body or {}
+    training_file = os.path.join(
+        FACE_DIR, f"train/{sanitize_filename(json.get('training_file', ''))}"
+    )
+
+    if not training_file or not os.path.isfile(training_file):
+        return JSONResponse(
+            content=(
+                {
+                    "success": False,
+                    "message": f"Invalid filename or no file exists: {training_file}",
+                }
+            ),
+            status_code=404,
+        )
+
+    rand_id = "".join(random.choices(string.ascii_lowercase + string.digits, k=6))
+    new_name = f"{name}-{rand_id}.webp"
+    new_file = os.path.join(FACE_DIR, f"{name}/{new_name}")
+    shutil.move(training_file, new_file)
+
+    context: EmbeddingsContext = request.app.embeddings
+    context.clear_face_classifier()
+
+    return JSONResponse(
+        content=(
+            {
+                "success": True,
+                "message": f"Successfully saved {training_file} as {new_name}.",
+            }
+        ),
+        status_code=200,
+    )
+
+
+@router.post("/faces/{name}/delete")
+def deregister_faces(request: Request, name: str, body: dict = None):
+    if not request.app.frigate_config.face_recognition.enabled:
+        return JSONResponse(
+            status_code=400,
+            content={"message": "Face recognition is not enabled.", "success": False},
+        )
+
+    json: dict[str, any] = body or {}
+    list_of_ids = json.get("ids", "")
+
+    if not list_of_ids or len(list_of_ids) == 0:
+        return JSONResponse(
+            content=({"success": False, "message": "Not a valid list of ids"}),
+            status_code=404,
+        )
+
+    context: EmbeddingsContext = request.app.embeddings
+    context.delete_face_ids(
+        name, map(lambda file: sanitize_filename(file), list_of_ids)
+    )
+    return JSONResponse(
+        content=({"success": True, "message": "Successfully deleted faces."}),
+        status_code=200,
+    )