Merge acb17a7b50 into 97b29d177a

x

.github/workflows/ci.yml

@@ -225,3 +225,29 @@ jobs:
           sources: |
             ghcr.io/${{ steps.lowercaseRepo.outputs.lowercase }}:${{ env.SHORT_SHA }}-amd64
             ghcr.io/${{ steps.lowercaseRepo.outputs.lowercase }}:${{ env.SHORT_SHA }}-rpi
+  axera_build:
+    runs-on: ubuntu-22.04
+    name: AXERA Build
+    needs:
+      - amd64_build
+      - arm64_build
+    steps:
+      - name: Check out code
+        uses: actions/checkout@v5
+        with:
+          persist-credentials: false
+      - name: Set up QEMU and Buildx
+        id: setup
+        uses: ./.github/actions/setup
+        with:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      - name: Build and push Axera build
+        uses: docker/bake-action@v6
+        with:
+          source: .
+          push: true
+          targets: axcl
+          files: docker/axcl/axcl.hcl
+          set: |
+            axcl.tags=${{ steps.setup.outputs.image-name }}-axcl
+            *.cache-from=type=gha

docker/axcl/Dockerfile

@@ -0,0 +1,55 @@
+# syntax=docker/dockerfile:1.6
+
+# https://askubuntu.com/questions/972516/debian-frontend-environment-variable
+ARG DEBIAN_FRONTEND=noninteractive
+
+# Globally set pip break-system-packages option to avoid having to specify it every time
+ARG PIP_BREAK_SYSTEM_PACKAGES=1
+
+
+FROM frigate AS frigate-axcl
+ARG TARGETARCH
+ARG PIP_BREAK_SYSTEM_PACKAGES
+
+# Install axpyengine
+RUN wget https://github.com/AXERA-TECH/pyaxengine/releases/download/0.1.3.rc1/axengine-0.1.3-py3-none-any.whl -O /axengine-0.1.3-py3-none-any.whl
+RUN pip3 install -i https://mirrors.aliyun.com/pypi/simple/ /axengine-0.1.3-py3-none-any.whl \
+    && rm /axengine-0.1.3-py3-none-any.whl
+
+# Install axcl
+RUN if [ "$TARGETARCH" = "amd64" ]; then \
+        echo "Installing x86_64 version of axcl"; \
+        wget https://github.com/ivanshi1108/assets/releases/download/v0.16.2/axcl_host_x86_64_V3.6.5_20250908154509_NO4973.deb -O /axcl.deb; \
+    else \
+        echo "Installing aarch64 version of axcl"; \
+        wget https://github.com/ivanshi1108/assets/releases/download/v0.16.2/axcl_host_aarch64_V3.6.5_20250908154509_NO4973.deb -O /axcl.deb; \
+    fi
+
+RUN mkdir /unpack_axcl && \
+    dpkg-deb -x /axcl.deb /unpack_axcl && \
+    cp -R /unpack_axcl/usr/bin/axcl /usr/bin/ && \
+    cp -R /unpack_axcl/usr/lib/axcl /usr/lib/ && \
+    rm -rf /unpack_axcl /axcl.deb
+
+
+# Install axcl ffmpeg
+RUN mkdir -p /usr/lib/ffmpeg/axcl
+
+RUN if [ "$TARGETARCH" = "amd64" ]; then \
+        wget https://github.com/ivanshi1108/assets/releases/download/v0.16.2/ffmpeg-x64 -O /usr/lib/ffmpeg/axcl/ffmpeg && \
+        wget https://github.com/ivanshi1108/assets/releases/download/v0.16.2/ffprobe-x64 -O /usr/lib/ffmpeg/axcl/ffprobe; \
+    else \
+        wget https://github.com/ivanshi1108/assets/releases/download/v0.16.2/ffmpeg-aarch64 -O /usr/lib/ffmpeg/axcl/ffmpeg && \
+        wget https://github.com/ivanshi1108/assets/releases/download/v0.16.2/ffprobe-aarch64 -O /usr/lib/ffmpeg/axcl/ffprobe; \
+    fi
+
+RUN chmod +x /usr/lib/ffmpeg/axcl/ffmpeg /usr/lib/ffmpeg/axcl/ffprobe
+
+# Set ldconfig path
+RUN echo "/usr/lib/axcl" > /etc/ld.so.conf.d/ax.conf
+
+# Set env
+ENV PATH="$PATH:/usr/bin/axcl"
+ENV LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/lib/axcl"
+
+ENTRYPOINT ["sh", "-c", "ldconfig && exec /init"]

docker/axcl/axcl.hcl

@@ -0,0 +1,13 @@
+target frigate {
+  dockerfile = "docker/main/Dockerfile"
+  platforms = ["linux/amd64", "linux/arm64"]
+  target = "frigate"
+}
+
+target axcl {
+  dockerfile = "docker/axcl/Dockerfile"
+  contexts = {
+    frigate = "target:frigate",
+  }
+  platforms = ["linux/amd64", "linux/arm64"]
+}

docker/axcl/axcl.mk

@@ -0,0 +1,15 @@
+BOARDS += axcl
+
+local-axcl: version
+	docker buildx bake --file=docker/axcl/axcl.hcl axcl \
+		--set axcl.tags=frigate:latest-axcl \
+		--load
+
+build-axcl: version
+	docker buildx bake --file=docker/axcl/axcl.hcl axcl \
+		--set axcl.tags=$(IMAGE_REPO):${GITHUB_REF_NAME}-$(COMMIT_HASH)-axcl
+
+push-axcl: build-axcl
+	docker buildx bake --file=docker/axcl/axcl.hcl axcl \
+		--set axcl.tags=$(IMAGE_REPO):${GITHUB_REF_NAME}-$(COMMIT_HASH)-axcl \
+		--push

docker/axcl/user_installation.sh

@@ -0,0 +1,83 @@
+#!/bin/bash
+
+# Update package list and install dependencies
+sudo apt-get update
+sudo apt-get install -y build-essential cmake git wget pciutils kmod udev
+
+# Check if gcc-12 is needed
+current_gcc_version=$(gcc --version | head -n1 | awk '{print $NF}')
+gcc_major_version=$(echo $current_gcc_version | cut -d'.' -f1)
+
+if [[ $gcc_major_version -lt 12 ]]; then
+    echo "Current GCC version ($current_gcc_version) is lower than 12, installing gcc-12..."
+    sudo apt-get install -y gcc-12
+    sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-12 12
+    echo "GCC-12 installed and set as default"
+else
+    echo "Current GCC version ($current_gcc_version) is sufficient, skipping GCC installation"
+fi
+
+# Determine architecture
+arch=$(uname -m)
+download_url=""
+
+if [[ $arch == "x86_64" ]]; then
+    download_url="https://github.com/ivanshi1108/assets/releases/download/v0.16.2/axcl_host_x86_64_V3.6.5_20250908154509_NO4973.deb"
+    deb_file="axcl_host_x86_64_V3.6.5_20250908154509_NO4973.deb"
+elif [[ $arch == "aarch64" ]]; then
+    download_url="https://github.com/ivanshi1108/assets/releases/download/v0.16.2/axcl_host_aarch64_V3.6.5_20250908154509_NO4973.deb"
+    deb_file="axcl_host_aarch64_V3.6.5_20250908154509_NO4973.deb"
+else
+    echo "Unsupported architecture: $arch"
+    exit 1
+fi
+
+# Download AXCL driver
+echo "Downloading AXCL driver for $arch..."
+wget "$download_url" -O "$deb_file"
+
+if [ $? -ne 0 ]; then
+    echo "Failed to download AXCL driver"
+    exit 1
+fi
+
+# Install AXCL driver
+echo "Installing AXCL driver..."
+sudo dpkg -i "$deb_file"
+
+if [ $? -ne 0 ]; then
+    echo "Failed to install AXCL driver, attempting to fix dependencies..."
+    sudo apt-get install -f -y
+    sudo dpkg -i "$deb_file"
+
+    if [ $? -ne 0 ]; then
+        echo "AXCL driver installation failed"
+        exit 1
+    fi
+fi
+
+# Update environment
+echo "Updating environment..."
+source /etc/profile
+
+# Verify installation
+echo "Verifying AXCL installation..."
+if command -v axcl-smi &> /dev/null; then
+    echo "AXCL driver detected, checking AI accelerator status..."
+
+    axcl_output=$(axcl-smi 2>&1)
+    axcl_exit_code=$?
+
+    echo "$axcl_output"
+
+    if [ $axcl_exit_code -eq 0 ]; then
+        echo "AXCL driver installation completed successfully!"
+    else
+        echo "AXCL driver installed but no AI accelerator detected or communication failed."
+        echo "Please check if the AI accelerator is properly connected and powered on."
+        exit 1
+    fi
+else
+    echo "axcl-smi command not found. AXCL driver installation may have failed."
+    exit 1
+fi

docs/docs/configuration/audio_detectors.md

@@ -157,3 +157,19 @@ Only one `speech` event may be transcribed at a time. Frigate does not automatic
 :::
 
 Recorded `speech` events will always use a `whisper` model, regardless of the `model_size` config setting. Without a supported Nvidia GPU, generating transcriptions for longer `speech` events may take a fair amount of time, so be patient.
+
+#### FAQ
+
+1. Why doesn't Frigate automatically transcribe all `speech` events?
+
+   Frigate does not implement a queue mechanism for speech transcription, and adding one is not trivial. A proper queue would need backpressure, prioritization, memory/disk buffering, retry logic, crash recovery, and safeguards to prevent unbounded growth when events outpace processing. That’s a significant amount of complexity for a feature that, in most real-world environments, would mostly just churn through low-value noise.
+
+   Because transcription is **serialized (one event at a time)** and speech events can be generated far faster than they can be processed, an auto-transcribe toggle would very quickly create an ever-growing backlog and degrade core functionality. For the amount of engineering and risk involved, it adds **very little practical value** for the majority of deployments, which are often on low-powered, edge hardware.
+
+   If you hear speech that’s actually important and worth saving/indexing for the future, **just press the transcribe button in Explore** on that specific `speech` event - that keeps things explicit, reliable, and under your control.
+
+2. Why don't you save live transcription text and use that for `speech` events?
+
+   There’s no guarantee that a `speech` event is even created from the exact audio that went through the transcription model. Live transcription and `speech` event creation are **separate, asynchronous processes**. Even when both are correctly configured, trying to align the **precise start and end time of a speech event** with whatever audio the model happened to be processing at that moment is unreliable.
+
+   Automatically persisting that data would often result in **misaligned, partial, or irrelevant transcripts**, while still incurring all of the CPU, storage, and privacy costs of transcription. That’s why Frigate treats transcription as an **explicit, user-initiated action** rather than an automatic side-effect of every `speech` event.

docs/docs/configuration/object_detectors.md

@@ -49,6 +49,11 @@ Frigate supports multiple different detectors that work on different types of ha
 
 - [Synaptics](#synaptics): synap models can run on Synaptics devices(e.g astra machina) with included NPUs.
 
+**AXERA** <CommunityBadge />
+
+- [AXEngine](#axera): axmodels can run on AXERA AI acceleration.
+
+
 **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.
@@ -1438,6 +1443,41 @@ model:
   input_pixel_format: rgb/bgr # look at the model.json to figure out which to put here
 ```
 
+## AXERA
+
+Hardware accelerated object detection is supported on the following SoCs:
+
+- AX650N
+- AX8850N
+
+This implementation uses the [AXera Pulsar2 Toolchain](https://huggingface.co/AXERA-TECH/Pulsar2).
+
+See the [installation docs](../frigate/installation.md#axera) for information on configuring the AXEngine hardware.
+
+### Configuration
+
+When configuring the AXEngine detector, you have to specify the model name.
+
+#### yolov9
+
+A yolov9 model is provided in the container at /axmodels and is used by this detector type by default.
+
+Use the model configuration shown below when using the axengine detector with the default axmodel:
+
+```yaml
+detectors:
+  axengine:
+    type: axengine
+
+model:
+  path: frigate-yolov9-tiny
+  model_type: yolo-generic
+  width: 320
+  height: 320
+  tensor_format: bgr
+  labelmap_path: /labelmap/coco-80.txt
+```
+
 # Models
 
 Some model types are not included in Frigate by default.

docs/docs/frigate/hardware.md

@@ -104,6 +104,10 @@ Frigate supports multiple different detectors that work on different types of ha
 
 - [Synaptics](#synaptics): synap models can run on Synaptics devices(e.g astra machina) with included NPUs to provide efficient object detection.
 
+**AXERA** <CommunityBadge />
+
+- [AXEngine](#axera): axera models can run on AXERA NPUs via AXEngine, delivering highly efficient object detection.
+
 :::
 
 ### Hailo-8
@@ -287,6 +291,14 @@ The inference time of a rk3588 with all 3 cores enabled is typically 25-30 ms fo
 | ssd mobilenet | ~ 25 ms                         |
 | yolov5m       | ~ 118 ms                        |
 
+### AXERA
+
+- **AXEngine** Default model is **yolov9**
+
+| Name             | AXERA AX650N/AX8850N Inference Time |
+| ---------------- | ----------------------------------- |
+| yolov9-tiny      | ~ 4 ms                              |
+
 ## What does Frigate use the CPU for and what does it use a detector for? (ELI5 Version)
 
 This is taken from a [user question on reddit](https://www.reddit.com/r/homeassistant/comments/q8mgau/comment/hgqbxh5/?utm_source=share&utm_medium=web2x&context=3). Modified slightly for clarity.
@@ -307,4 +319,4 @@ Basically - When you increase the resolution and/or the frame rate of the stream
 
 YES! The Coral does not help with decoding video streams.
 
-Decompressing video streams takes a significant amount of CPU power. Video compression uses key frames (also known as I-frames) to send a full frame in the video stream. The following frames only include the difference from the key frame, and the CPU has to compile each frame by merging the differences with the key frame. [More detailed explanation](https://support.video.ibm.com/hc/en-us/articles/18106203580316-Keyframes-InterFrame-Video-Compression). Higher resolutions and frame rates mean more processing power is needed to decode the video stream, so try and set them on the camera to avoid unnecessary decoding work.
+Decompressing video streams takes a significant amount of CPU power. Video compression uses key frames (also known as I-frames) to send a full frame in the video stream. The following frames only include the difference from the key frame, and the CPU has to compile each frame by merging the differences with the key frame. [More detailed explanation](https://support.video.ibm.com/hc/en-us/articles/18106203580316-Keyframes-InterFrame-Video-Compression). Higher resolutions and frame rates mean more processing power is needed to decode the video stream, so try and set them on the camera to avoid unnecessary decoding work.

docs/docs/frigate/installation.md

@@ -287,6 +287,42 @@ or add these options to your `docker run` command:
 
 Next, you should configure [hardware object detection](/configuration/object_detectors#synaptics) and [hardware video processing](/configuration/hardware_acceleration_video#synaptics).
 
+### AXERA
+
+<details>
+<summary>AXERA accelerators</summary>
+AXERA accelerators are available in an M.2 form factor, compatible with both Raspberry Pi and Orange Pi. This form factor has also been successfully tested on x86 platforms, making it a versatile choice for various computing environments.
+
+#### Installation
+
+Using AXERA accelerators requires the installation of the AXCL driver. We provide a convenient Linux script to complete this installation.
+
+Follow these steps for installation:
+
+1. Copy or download [this script](https://github.com/ivanshi1108/assets/releases/download/v0.16.2/user_installation.sh).
+2. Ensure it has execution permissions with `sudo chmod +x user_installation.sh`
+3. Run the script with `./user_installation.sh`
+
+#### Setup
+
+To set up Frigate, follow the default installation instructions, for example: `ghcr.io/blakeblackshear/frigate:stable`
+
+Next, grant Docker permissions to access your hardware by adding the following lines to your `docker-compose.yml` file:
+
+```yaml
+devices:
+  - /dev/axcl_host
+  - /dev/ax_mmb_dev
+  - /dev/msg_userdev
+```
+
+If you are using `docker run`, add this option to your command `--device /dev/axcl_host --device /dev/ax_mmb_dev --device /dev/msg_userdev`
+
+#### Configuration
+
+Finally, configure [hardware object detection](/configuration/object_detectors#axera) to complete the setup.
+</details>
+
 ## Docker
 
 Running through Docker with Docker Compose is the recommended install method.

frigate/data_processing/real_time/custom_classification.py

@@ -99,6 +99,42 @@ class CustomStateClassificationProcessor(RealTimeProcessorApi):
         if self.inference_speed:
             self.inference_speed.update(duration)
 
+    def _should_save_image(
+        self, camera: str, detected_state: str, score: float = 1.0
+    ) -> bool:
+        """
+        Determine if we should save the image for training.
+        Save when:
+        - State is changing or being verified (regardless of score)
+        - Score is less than 100% (even if state matches, useful for training)
+        Don't save when:
+        - State is stable (matches current_state) AND score is 100%
+        """
+        if camera not in self.state_history:
+            # First detection for this camera, save it
+            return True
+
+        verification = self.state_history[camera]
+        current_state = verification.get("current_state")
+        pending_state = verification.get("pending_state")
+
+        # Save if there's a pending state change being verified
+        if pending_state is not None:
+            return True
+
+        # Save if the detected state differs from the current verified state
+        # (state is changing)
+        if current_state is not None and detected_state != current_state:
+            return True
+
+        # If score is less than 100%, save even if state matches
+        # (useful for training to improve confidence)
+        if score < 1.0:
+            return True
+
+        # Don't save if state is stable (detected_state == current_state) AND score is 100%
+        return False
+
     def verify_state_change(self, camera: str, detected_state: str) -> str | None:
         """
         Verify state change requires 3 consecutive identical states before publishing.
@@ -212,14 +248,16 @@ class CustomStateClassificationProcessor(RealTimeProcessorApi):
                 return
 
         if self.interpreter is None:
-            write_classification_attempt(
-                self.train_dir,
-                cv2.cvtColor(frame, cv2.COLOR_RGB2BGR),
-                "none-none",
-                now,
-                "unknown",
-                0.0,
-            )
+            # When interpreter is None, always save (score is 0.0, which is < 1.0)
+            if self._should_save_image(camera, "unknown", 0.0):
+                write_classification_attempt(
+                    self.train_dir,
+                    cv2.cvtColor(frame, cv2.COLOR_RGB2BGR),
+                    "none-none",
+                    now,
+                    "unknown",
+                    0.0,
+                )
             return
 
         input = np.expand_dims(resized_frame, axis=0)
@@ -236,14 +274,17 @@ class CustomStateClassificationProcessor(RealTimeProcessorApi):
         score = round(probs[best_id], 2)
         self.__update_metrics(datetime.datetime.now().timestamp() - now)
 
-        write_classification_attempt(
-            self.train_dir,
-            cv2.cvtColor(frame, cv2.COLOR_RGB2BGR),
-            "none-none",
-            now,
-            self.labelmap[best_id],
-            score,
-        )
+        detected_state = self.labelmap[best_id]
+
+        if self._should_save_image(camera, detected_state, score):
+            write_classification_attempt(
+                self.train_dir,
+                cv2.cvtColor(frame, cv2.COLOR_RGB2BGR),
+                "none-none",
+                now,
+                detected_state,
+                score,
+            )
 
         if score < self.model_config.threshold:
             logger.debug(
@@ -251,7 +292,6 @@ class CustomStateClassificationProcessor(RealTimeProcessorApi):
             )
             return
 
-        detected_state = self.labelmap[best_id]
         verified_state = self.verify_state_change(camera, detected_state)
 
         if verified_state is not None:

frigate/detectors/plugins/axengine.py

@@ -0,0 +1,86 @@
+import logging
+import os.path
+import re
+import urllib.request
+from typing import Literal
+
+import axengine as axe
+
+from frigate.const import MODEL_CACHE_DIR
+from frigate.detectors.detection_api import DetectionApi
+from frigate.detectors.detector_config import BaseDetectorConfig, ModelTypeEnum
+from frigate.util.model import post_process_yolo
+
+logger = logging.getLogger(__name__)
+
+DETECTOR_KEY = "axengine"
+
+supported_models = {
+    ModelTypeEnum.yologeneric: "frigate-yolov9-.*$",
+}
+
+model_cache_dir = os.path.join(MODEL_CACHE_DIR, "axengine_cache/")
+
+
+class AxengineDetectorConfig(BaseDetectorConfig):
+    type: Literal[DETECTOR_KEY]
+
+
+class Axengine(DetectionApi):
+    type_key = DETECTOR_KEY
+
+    def __init__(self, config: AxengineDetectorConfig):
+        logger.info("__init__ axengine")
+        super().__init__(config)
+        self.height = config.model.height
+        self.width = config.model.width
+        model_path = config.model.path or "frigate-yolov9-tiny"
+        model_props = self.parse_model_input(model_path)
+        self.session = axe.InferenceSession(model_props["path"])
+
+    def __del__(self):
+        pass
+
+    def parse_model_input(self, model_path):
+        model_props = {}
+        model_props["preset"] = True
+
+        model_matched = False
+
+        for model_type, pattern in supported_models.items():
+            if re.match(pattern, model_path):
+                model_matched = True
+                model_props["model_type"] = model_type
+
+        if model_matched:
+            model_props["filename"] = model_path + ".axmodel"
+            model_props["path"] = model_cache_dir + model_props["filename"]
+
+            if not os.path.isfile(model_props["path"]):
+                self.download_model(model_props["filename"])
+        else:
+            supported_models_str = ", ".join(model[1:-1] for model in supported_models)
+            raise Exception(
+                f"Model {model_path} is unsupported. Provide your own model or choose one of the following: {supported_models_str}"
+            )
+        return model_props
+
+    def download_model(self, filename):
+        if not os.path.isdir(model_cache_dir):
+            os.mkdir(model_cache_dir)
+
+        GITHUB_ENDPOINT = os.environ.get("GITHUB_ENDPOINT", "https://github.com")
+        urllib.request.urlretrieve(
+            f"{GITHUB_ENDPOINT}/ivanshi1108/assets/releases/download/v0.16.2/{filename}",
+            model_cache_dir + filename,
+        )
+
+    def detect_raw(self, tensor_input):
+        results = None
+        results = self.session.run(None, {"images": tensor_input})
+        if self.detector_config.model.model_type == ModelTypeEnum.yologeneric:
+            return post_process_yolo(results, self.width, self.height)
+        else:
+            raise ValueError(
+                f'Model type "{self.detector_config.model.model_type}" is currently not supported.'
+            )

frigate/ptz/onvif.py

@@ -190,7 +190,11 @@ class OnvifController:
         ptz: ONVIFService = await onvif.create_ptz_service()
         self.cams[camera_name]["ptz"] = ptz
 
-        imaging: ONVIFService = await onvif.create_imaging_service()
+        try:
+            imaging: ONVIFService = await onvif.create_imaging_service()
+        except (Fault, ONVIFError, TransportError, Exception) as e:
+            logger.debug(f"Imaging service not supported for {camera_name}: {e}")
+            imaging = None
         self.cams[camera_name]["imaging"] = imaging
         try:
             video_sources = await media.GetVideoSources()
@@ -381,7 +385,10 @@ class OnvifController:
                             f"Disabling autotracking zooming for {camera_name}: Absolute zoom not supported. Exception: {e}"
                         )
 
-        if self.cams[camera_name]["video_source_token"] is not None:
+        if (
+            self.cams[camera_name]["video_source_token"] is not None
+            and imaging is not None
+        ):
             try:
                 imaging_capabilities = await imaging.GetImagingSettings(
                     {"VideoSourceToken": self.cams[camera_name]["video_source_token"]}
@@ -421,6 +428,7 @@ class OnvifController:
         if (
             "focus" in self.cams[camera_name]["features"]
             and self.cams[camera_name]["video_source_token"]
+            and self.cams[camera_name]["imaging"] is not None
         ):
             try:
                 stop_request = self.cams[camera_name]["imaging"].create_type("Stop")
@@ -648,6 +656,7 @@ class OnvifController:
         if (
             "focus" not in self.cams[camera_name]["features"]
             or not self.cams[camera_name]["video_source_token"]
+            or self.cams[camera_name]["imaging"] is None
         ):
             logger.error(f"{camera_name} does not support ONVIF continuous focus.")
             return

frigate/video.py

@@ -124,45 +124,50 @@ def capture_frames(
         config_subscriber.check_for_updates()
         return config.enabled
 
-    while not stop_event.is_set():
-        if not get_enabled_state():
-            logger.debug(f"Stopping capture thread for disabled {config.name}")
-            break
-
-        fps.value = frame_rate.eps()
-        skipped_fps.value = skipped_eps.eps()
-        current_frame.value = datetime.now().timestamp()
-        frame_name = f"{config.name}_frame{frame_index}"
-        frame_buffer = frame_manager.write(frame_name)
-        try:
-            frame_buffer[:] = ffmpeg_process.stdout.read(frame_size)
-        except Exception:
-            # shutdown has been initiated
-            if stop_event.is_set():
+    try:
+        while not stop_event.is_set():
+            if not get_enabled_state():
+                logger.debug(f"Stopping capture thread for disabled {config.name}")
                 break
 
-            logger.error(f"{config.name}: Unable to read frames from ffmpeg process.")
+            fps.value = frame_rate.eps()
+            skipped_fps.value = skipped_eps.eps()
+            current_frame.value = datetime.now().timestamp()
+            frame_name = f"{config.name}_frame{frame_index}"
+            frame_buffer = frame_manager.write(frame_name)
+            try:
+                frame_buffer[:] = ffmpeg_process.stdout.read(frame_size)
+            except Exception:
+                # shutdown has been initiated
+                if stop_event.is_set():
+                    break
 
-            if ffmpeg_process.poll() is not None:
                 logger.error(
-                    f"{config.name}: ffmpeg process is not running. exiting capture thread..."
+                    f"{config.name}: Unable to read frames from ffmpeg process."
                 )
-                break
 
-            continue
+                if ffmpeg_process.poll() is not None:
+                    logger.error(
+                        f"{config.name}: ffmpeg process is not running. exiting capture thread..."
+                    )
+                    break
 
-        frame_rate.update()
+                continue
 
-        # don't lock the queue to check, just try since it should rarely be full
-        try:
-            # add to the queue
-            frame_queue.put((frame_name, current_frame.value), False)
-            frame_manager.close(frame_name)
-        except queue.Full:
-            # if the queue is full, skip this frame
-            skipped_eps.update()
+            frame_rate.update()
 
-        frame_index = 0 if frame_index == shm_frame_count - 1 else frame_index + 1
+            # don't lock the queue to check, just try since it should rarely be full
+            try:
+                # add to the queue
+                frame_queue.put((frame_name, current_frame.value), False)
+                frame_manager.close(frame_name)
+            except queue.Full:
+                # if the queue is full, skip this frame
+                skipped_eps.update()
+
+            frame_index = 0 if frame_index == shm_frame_count - 1 else frame_index + 1
+    finally:
+        config_subscriber.stop()
 
 
 class CameraWatchdog(threading.Thread):
@@ -234,6 +239,16 @@ class CameraWatchdog(threading.Thread):
                 else:
                     self.ffmpeg_detect_process.wait()
 
+        # Wait for old capture thread to fully exit before starting a new one
+        if self.capture_thread is not None and self.capture_thread.is_alive():
+            self.logger.info("Waiting for capture thread to exit...")
+            self.capture_thread.join(timeout=5)
+
+            if self.capture_thread.is_alive():
+                self.logger.warning(
+                    f"Capture thread for {self.config.name} did not exit in time"
+                )
+
         self.logger.error(
             "The following ffmpeg logs include the last 100 lines prior to exit."
         )

web/src/components/classification/ClassificationModelEditDialog.tsx

@@ -37,7 +37,7 @@ import { useForm } from "react-hook-form";
 import { useTranslation } from "react-i18next";
 import { LuPlus, LuX } from "react-icons/lu";
 import { toast } from "sonner";
-import useSWR from "swr";
+import useSWR, { mutate } from "swr";
 import { z } from "zod";
 
 type ClassificationModelEditDialogProps = {
@@ -240,15 +240,61 @@ export default function ClassificationModelEditDialog({
             position: "top-center",
           });
         } else {
-          // State model - update classes
-          // Note: For state models, updating classes requires renaming categories
-          // which is handled through the dataset API, not the config API
-          // We'll need to implement this by calling the rename endpoint for each class
-          // For now, we just show a message that this requires retraining
+          const stateData = data as StateFormData;
+          const newClasses = stateData.classes.filter(
+            (c) => c.trim().length > 0,
+          );
+          const oldClasses = dataset?.categories
+            ? Object.keys(dataset.categories).filter((key) => key !== "none")
+            : [];
 
-          toast.info(t("edit.stateClassesInfo"), {
-            position: "top-center",
-          });
+          const renameMap = new Map<string, string>();
+          const maxLength = Math.max(oldClasses.length, newClasses.length);
+
+          for (let i = 0; i < maxLength; i++) {
+            const oldClass = oldClasses[i];
+            const newClass = newClasses[i];
+
+            if (oldClass && newClass && oldClass !== newClass) {
+              renameMap.set(oldClass, newClass);
+            }
+          }
+
+          const renamePromises = Array.from(renameMap.entries()).map(
+            async ([oldName, newName]) => {
+              try {
+                await axios.put(
+                  `/classification/${model.name}/dataset/${oldName}/rename`,
+                  {
+                    new_category: newName,
+                  },
+                );
+              } catch (err) {
+                const error = err as {
+                  response?: { data?: { message?: string; detail?: string } };
+                };
+                const errorMessage =
+                  error.response?.data?.message ||
+                  error.response?.data?.detail ||
+                  "Unknown error";
+                throw new Error(
+                  `Failed to rename ${oldName} to ${newName}: ${errorMessage}`,
+                );
+              }
+            },
+          );
+
+          if (renamePromises.length > 0) {
+            await Promise.all(renamePromises);
+            await mutate(`classification/${model.name}/dataset`);
+            toast.success(t("toast.success.updatedModel"), {
+              position: "top-center",
+            });
+          } else {
+            toast.info(t("edit.stateClassesInfo"), {
+              position: "top-center",
+            });
+          }
         }
 
         onSuccess();
@@ -256,8 +302,10 @@ export default function ClassificationModelEditDialog({
       } catch (err) {
         const error = err as {
           response?: { data?: { message?: string; detail?: string } };
+          message?: string;
         };
         const errorMessage =
+          error.message ||
           error.response?.data?.message ||
           error.response?.data?.detail ||
           "Unknown error";
@@ -268,7 +316,7 @@ export default function ClassificationModelEditDialog({
         setIsSaving(false);
       }
     },
-    [isObjectModel, model, t, onSuccess, onClose],
+    [isObjectModel, model, dataset, t, onSuccess, onClose],
   );
 
   const handleCancel = useCallback(() => {

web/src/components/menu/LiveContextMenu.tsx

@@ -48,6 +48,7 @@ import { useTranslation } from "react-i18next";
 import { useDateLocale } from "@/hooks/use-date-locale";
 import { useIsAdmin } from "@/hooks/use-is-admin";
 import { CameraNameLabel } from "../camera/FriendlyNameLabel";
+import { LiveStreamMetadata } from "@/types/live";
 
 type LiveContextMenuProps = {
   className?: string;
@@ -68,6 +69,7 @@ type LiveContextMenuProps = {
   resetPreferredLiveMode: () => void;
   config?: FrigateConfig;
   children?: ReactNode;
+  streamMetadata?: { [key: string]: LiveStreamMetadata };
 };
 export default function LiveContextMenu({
   className,
@@ -88,6 +90,7 @@ export default function LiveContextMenu({
   resetPreferredLiveMode,
   config,
   children,
+  streamMetadata,
 }: LiveContextMenuProps) {
   const { t } = useTranslation("views/live");
   const [showSettings, setShowSettings] = useState(false);
@@ -558,6 +561,7 @@ export default function LiveContextMenu({
           setGroupStreamingSettings={setGroupStreamingSettings}
           setIsDialogOpen={setShowSettings}
           onSave={onSave}
+          streamMetadata={streamMetadata}
         />
       </Dialog>
     </div>

web/src/components/settings/CameraStreamingDialog.tsx

@@ -38,6 +38,7 @@ import { useCameraFriendlyName } from "@/hooks/use-camera-friendly-name";
 type CameraStreamingDialogProps = {
   camera: string;
   groupStreamingSettings: GroupStreamingSettings;
+  streamMetadata?: { [key: string]: LiveStreamMetadata };
   setGroupStreamingSettings: React.Dispatch<
     React.SetStateAction<GroupStreamingSettings>
   >;
@@ -48,6 +49,7 @@ type CameraStreamingDialogProps = {
 export function CameraStreamingDialog({
   camera,
   groupStreamingSettings,
+  streamMetadata,
   setGroupStreamingSettings,
   setIsDialogOpen,
   onSave,
@@ -76,12 +78,7 @@ export function CameraStreamingDialog({
     [config, streamName],
   );
 
-  const { data: cameraMetadata } = useSWR<LiveStreamMetadata>(
-    isRestreamed ? `go2rtc/streams/${streamName}` : null,
-    {
-      revalidateOnFocus: false,
-    },
-  );
+  const cameraMetadata = streamName ? streamMetadata?.[streamName] : undefined;
 
   const supportsAudioOutput = useMemo(() => {
     if (!cameraMetadata) {

web/src/hooks/use-camera-live-mode.ts

@@ -1,8 +1,8 @@
-import { baseUrl } from "@/api/baseUrl";
 import { CameraConfig, FrigateConfig } from "@/types/frigateConfig";
 import { useCallback, useEffect, useState, useMemo } from "react";
 import useSWR from "swr";
-import { LivePlayerMode, LiveStreamMetadata } from "@/types/live";
+import { LivePlayerMode } from "@/types/live";
+import useDeferredStreamMetadata from "./use-deferred-stream-metadata";
 
 export default function useCameraLiveMode(
   cameras: CameraConfig[],
@@ -11,9 +11,9 @@ export default function useCameraLiveMode(
 ) {
   const { data: config } = useSWR<FrigateConfig>("config");
 
-  // Get comma-separated list of restreamed stream names for SWR key
-  const restreamedStreamsKey = useMemo(() => {
-    if (!cameras || !config) return null;
+  // Compute which streams need metadata (restreamed streams only)
+  const restreamedStreamNames = useMemo(() => {
+    if (!cameras || !config) return [];
 
     const streamNames = new Set<string>();
     cameras.forEach((camera) => {
@@ -32,56 +32,13 @@ export default function useCameraLiveMode(
       }
     });
 
-    return streamNames.size > 0
-      ? Array.from(streamNames).sort().join(",")
-      : null;
+    return Array.from(streamNames);
   }, [cameras, config, activeStreams]);
 
-  const streamsFetcher = useCallback(async (key: string) => {
-    const streamNames = key.split(",");
-
-    const metadataPromises = streamNames.map(async (streamName) => {
-      try {
-        const response = await fetch(
-          `${baseUrl}api/go2rtc/streams/${streamName}`,
-          {
-            priority: "low",
-          },
-        );
-
-        if (response.ok) {
-          const data = await response.json();
-          return { streamName, data };
-        }
-        return { streamName, data: null };
-      } catch (error) {
-        // eslint-disable-next-line no-console
-        console.error(`Failed to fetch metadata for ${streamName}:`, error);
-        return { streamName, data: null };
-      }
-    });
-
-    const results = await Promise.allSettled(metadataPromises);
-
-    const metadata: { [key: string]: LiveStreamMetadata } = {};
-    results.forEach((result) => {
-      if (result.status === "fulfilled" && result.value.data) {
-        metadata[result.value.streamName] = result.value.data;
-      }
-    });
-
-    return metadata;
-  }, []);
-
-  const { data: allStreamMetadata = {} } = useSWR<{
-    [key: string]: LiveStreamMetadata;
-  }>(restreamedStreamsKey, streamsFetcher, {
-    revalidateOnFocus: false,
-    revalidateOnReconnect: false,
-    revalidateIfStale: false,
-    dedupingInterval: 60000,
-  });
+  // Fetch stream metadata with deferred loading (doesn't block initial render)
+  const streamMetadata = useDeferredStreamMetadata(restreamedStreamNames);
 
+  // Compute live mode states
   const [preferredLiveModes, setPreferredLiveModes] = useState<{
     [key: string]: LivePlayerMode;
   }>({});
@@ -122,10 +79,10 @@ export default function useCameraLiveMode(
         newPreferredLiveModes[camera.name] = isRestreamed ? "mse" : "jsmpeg";
       }
 
-      // check each stream for audio support
+      // Check each stream for audio support
       if (isRestreamed) {
         Object.values(camera.live.streams).forEach((streamName) => {
-          const metadata = allStreamMetadata?.[streamName];
+          const metadata = streamMetadata[streamName];
           newSupportsAudioOutputStates[streamName] = {
             supportsAudio: metadata
               ? metadata.producers.find(
@@ -150,7 +107,7 @@ export default function useCameraLiveMode(
     setPreferredLiveModes(newPreferredLiveModes);
     setIsRestreamedStates(newIsRestreamedStates);
     setSupportsAudioOutputStates(newSupportsAudioOutputStates);
-  }, [cameras, config, windowVisible, allStreamMetadata]);
+  }, [cameras, config, windowVisible, streamMetadata]);
 
   const resetPreferredLiveMode = useCallback(
     (cameraName: string) => {
@@ -180,5 +137,6 @@ export default function useCameraLiveMode(
     resetPreferredLiveMode,
     isRestreamedStates,
     supportsAudioOutputStates,
+    streamMetadata,
   };
 }

web/src/hooks/use-deferred-stream-metadata.ts

@@ -0,0 +1,90 @@
+import { baseUrl } from "@/api/baseUrl";
+import { useCallback, useEffect, useState, useMemo } from "react";
+import useSWR from "swr";
+import { LiveStreamMetadata } from "@/types/live";
+
+const FETCH_TIMEOUT_MS = 10000;
+const DEFER_DELAY_MS = 2000;
+
+/**
+ * Hook that fetches go2rtc stream metadata with deferred loading.
+ *
+ * Metadata fetching is delayed to prevent blocking initial page load
+ * and camera image requests.
+ *
+ * @param streamNames - Array of stream names to fetch metadata for
+ * @returns Object containing stream metadata keyed by stream name
+ */
+export default function useDeferredStreamMetadata(streamNames: string[]) {
+  const [fetchEnabled, setFetchEnabled] = useState(false);
+
+  useEffect(() => {
+    const timeoutId = setTimeout(() => {
+      setFetchEnabled(true);
+    }, DEFER_DELAY_MS);
+
+    return () => clearTimeout(timeoutId);
+  }, []);
+
+  const swrKey = useMemo(() => {
+    if (!fetchEnabled || streamNames.length === 0) return null;
+    // Use spread to avoid mutating the original array
+    return `deferred-streams:${[...streamNames].sort().join(",")}`;
+  }, [fetchEnabled, streamNames]);
+
+  const fetcher = useCallback(async (key: string) => {
+    // Extract stream names from key (remove prefix)
+    const names = key.replace("deferred-streams:", "").split(",");
+
+    const promises = names.map(async (streamName) => {
+      const controller = new AbortController();
+      const timeoutId = setTimeout(() => controller.abort(), FETCH_TIMEOUT_MS);
+
+      try {
+        const response = await fetch(
+          `${baseUrl}api/go2rtc/streams/${streamName}`,
+          {
+            priority: "low",
+            signal: controller.signal,
+          },
+        );
+        clearTimeout(timeoutId);
+
+        if (response.ok) {
+          const data = await response.json();
+          return { streamName, data };
+        }
+        return { streamName, data: null };
+      } catch (error) {
+        clearTimeout(timeoutId);
+        if ((error as Error).name !== "AbortError") {
+          // eslint-disable-next-line no-console
+          console.error(`Failed to fetch metadata for ${streamName}:`, error);
+        }
+        return { streamName, data: null };
+      }
+    });
+
+    const results = await Promise.allSettled(promises);
+
+    const metadata: { [key: string]: LiveStreamMetadata } = {};
+    results.forEach((result) => {
+      if (result.status === "fulfilled" && result.value.data) {
+        metadata[result.value.streamName] = result.value.data;
+      }
+    });
+
+    return metadata;
+  }, []);
+
+  const { data: metadata = {} } = useSWR<{
+    [key: string]: LiveStreamMetadata;
+  }>(swrKey, fetcher, {
+    revalidateOnFocus: false,
+    revalidateOnReconnect: false,
+    revalidateIfStale: false,
+    dedupingInterval: 60000,
+  });
+
+  return metadata;
+}

web/src/views/live/DraggableGridLayout.tsx

@@ -24,6 +24,7 @@ import "react-resizable/css/styles.css";
 import {
   AudioState,
   LivePlayerMode,
+  LiveStreamMetadata,
   StatsState,
   VolumeState,
 } from "@/types/live";
@@ -47,7 +48,6 @@ import {
   TooltipContent,
 } from "@/components/ui/tooltip";
 import { Toaster } from "@/components/ui/sonner";
-import useCameraLiveMode from "@/hooks/use-camera-live-mode";
 import LiveContextMenu from "@/components/menu/LiveContextMenu";
 import { useStreamingSettings } from "@/context/streaming-settings-provider";
 import { useTranslation } from "react-i18next";
@@ -65,6 +65,16 @@ type DraggableGridLayoutProps = {
   setIsEditMode: React.Dispatch<React.SetStateAction<boolean>>;
   fullscreen: boolean;
   toggleFullscreen: () => void;
+  preferredLiveModes: { [key: string]: LivePlayerMode };
+  setPreferredLiveModes: React.Dispatch<
+    React.SetStateAction<{ [key: string]: LivePlayerMode }>
+  >;
+  resetPreferredLiveMode: (cameraName: string) => void;
+  isRestreamedStates: { [key: string]: boolean };
+  supportsAudioOutputStates: {
+    [key: string]: { supportsAudio: boolean; cameraName: string };
+  };
+  streamMetadata: { [key: string]: LiveStreamMetadata };
 };
 export default function DraggableGridLayout({
   cameras,
@@ -79,6 +89,12 @@ export default function DraggableGridLayout({
   setIsEditMode,
   fullscreen,
   toggleFullscreen,
+  preferredLiveModes,
+  setPreferredLiveModes,
+  resetPreferredLiveMode,
+  isRestreamedStates,
+  supportsAudioOutputStates,
+  streamMetadata,
 }: DraggableGridLayoutProps) {
   const { t } = useTranslation(["views/live"]);
   const { data: config } = useSWR<FrigateConfig>("config");
@@ -98,33 +114,6 @@ export default function DraggableGridLayout({
     }
   }, [allGroupsStreamingSettings, cameraGroup]);
 
-  const activeStreams = useMemo(() => {
-    const streams: { [cameraName: string]: string } = {};
-    cameras.forEach((camera) => {
-      const availableStreams = camera.live.streams || {};
-      const streamNameFromSettings =
-        currentGroupStreamingSettings?.[camera.name]?.streamName || "";
-      const streamExists =
-        streamNameFromSettings &&
-        Object.values(availableStreams).includes(streamNameFromSettings);
-
-      const streamName = streamExists
-        ? streamNameFromSettings
-        : Object.values(availableStreams)[0] || "";
-
-      streams[camera.name] = streamName;
-    });
-    return streams;
-  }, [cameras, currentGroupStreamingSettings]);
-
-  const {
-    preferredLiveModes,
-    setPreferredLiveModes,
-    resetPreferredLiveMode,
-    isRestreamedStates,
-    supportsAudioOutputStates,
-  } = useCameraLiveMode(cameras, windowVisible, activeStreams);
-
   // grid layout
 
   const ResponsiveGridLayout = useMemo(() => WidthProvider(Responsive), []);
@@ -624,6 +613,7 @@ export default function DraggableGridLayout({
                     resetPreferredLiveMode(camera.name)
                   }
                   config={config}
+                  streamMetadata={streamMetadata}
                 >
                   <LivePlayer
                     key={camera.name}
@@ -838,6 +828,7 @@ type GridLiveContextMenuProps = {
   unmuteAll: () => void;
   resetPreferredLiveMode: () => void;
   config?: FrigateConfig;
+  streamMetadata?: { [key: string]: LiveStreamMetadata };
 };
 
 const GridLiveContextMenu = React.forwardRef<
@@ -868,6 +859,7 @@ const GridLiveContextMenu = React.forwardRef<
       unmuteAll,
       resetPreferredLiveMode,
       config,
+      streamMetadata,
       ...props
     },
     ref,
@@ -899,6 +891,7 @@ const GridLiveContextMenu = React.forwardRef<
           unmuteAll={unmuteAll}
           resetPreferredLiveMode={resetPreferredLiveMode}
           config={config}
+          streamMetadata={streamMetadata}
         >
           {children}
         </LiveContextMenu>

web/src/views/live/LiveDashboardView.tsx

@@ -265,6 +265,7 @@ export default function LiveDashboardView({
     resetPreferredLiveMode,
     isRestreamedStates,
     supportsAudioOutputStates,
+    streamMetadata,
   } = useCameraLiveMode(cameras, windowVisible, activeStreams);
 
   const birdseyeConfig = useMemo(() => config?.birdseye, [config]);
@@ -650,6 +651,12 @@ export default function LiveDashboardView({
               setIsEditMode={setIsEditMode}
               fullscreen={fullscreen}
               toggleFullscreen={toggleFullscreen}
+              preferredLiveModes={preferredLiveModes}
+              setPreferredLiveModes={setPreferredLiveModes}
+              resetPreferredLiveMode={resetPreferredLiveMode}
+              isRestreamedStates={isRestreamedStates}
+              supportsAudioOutputStates={supportsAudioOutputStates}
+              streamMetadata={streamMetadata}
             />
           )}
         </>

web/src/views/settings/AuthenticationView.tsx

@@ -478,33 +478,32 @@ export default function AuthenticationView({
                       <TableCell className="text-right">
                         <TooltipProvider>
                           <div className="flex items-center justify-end gap-2">
-                            {user.username !== "admin" &&
-                              user.username !== "viewer" && (
-                                <Tooltip>
-                                  <TooltipTrigger asChild>
-                                    <Button
-                                      size="sm"
-                                      variant="outline"
-                                      className="h-8 px-2"
-                                      onClick={() => {
-                                        setSelectedUser(user.username);
-                                        setSelectedUserRole(
-                                          user.role || "viewer",
-                                        );
-                                        setShowRoleChange(true);
-                                      }}
-                                    >
-                                      <LuUserCog className="size-3.5" />
-                                      <span className="ml-1.5 hidden sm:inline-block">
-                                        {t("role.title", { ns: "common" })}
-                                      </span>
-                                    </Button>
-                                  </TooltipTrigger>
-                                  <TooltipContent>
-                                    <p>{t("users.table.changeRole")}</p>
-                                  </TooltipContent>
-                                </Tooltip>
-                              )}
+                            {user.username !== "admin" && (
+                              <Tooltip>
+                                <TooltipTrigger asChild>
+                                  <Button
+                                    size="sm"
+                                    variant="outline"
+                                    className="h-8 px-2"
+                                    onClick={() => {
+                                      setSelectedUser(user.username);
+                                      setSelectedUserRole(
+                                        user.role || "viewer",
+                                      );
+                                      setShowRoleChange(true);
+                                    }}
+                                  >
+                                    <LuUserCog className="size-3.5" />
+                                    <span className="ml-1.5 hidden sm:inline-block">
+                                      {t("role.title", { ns: "common" })}
+                                    </span>
+                                  </Button>
+                                </TooltipTrigger>
+                                <TooltipContent>
+                                  <p>{t("users.table.changeRole")}</p>
+                                </TooltipContent>
+                              </Tooltip>
+                            )}
 
                             <Tooltip>
                               <TooltipTrigger asChild>