[Detector Support]: Hailo 8L and the Frigate Plus model support. #15397
-
Describe the problem you are havingI am trying to use the Hailo 8L with my Frigate Plus model. I am getting an error on all of the different model options that they are not supported. I am assuming that I need to convert this to Hailo's HEF format. I am trying to use the hailo DFC program but am having issues with those conversion commands. Version15 Beta Frigate config filemqtt:
enabled: false
host: CCCCCCCCC
port: 1883
topic_prefix: frigate
client_id: frigate
user: CCCCCCCCC
password: CCCCCCCCC
ffmpeg:
hwaccel_args: preset-vaapi
#input_args: preset-rtsp-udp
#output_args: preset-record-generic-audio-copy
detectors:
hailo8l:
type: hailo8l
device: PCIe
model:
path: /config/model_cache/h8l_cache/ssd_mobilenet_v1.hef
#path: plus://5da8db58f6b07967ee161de9c004fe52
model:
width: 300
height: 300
input_tensor: nhwc
input_pixel_format: bgr
model_type: ssd
objects:
# Optional: list of objects to track from labelmap.txt (default: shown below)
track:
- person
- car
- cat
- motorcycle
- bicycle
- boat
- deer
- dog
- horse
- deer
- bird
- raccoon
- fox
- squirrel
- rabbit
- package
- umbrella
- face
- fedex
- amazon
- usps
- dhl
- license_plate
- package
- person
- ups
snapshots:
# Optional: Enable writing jpg snapshot to /media/frigate/clips (default: shown below)
enabled: true
# Optional: save a clean PNG copy of the snapshot image (default: shown below)
clean_copy: true
# Optional: print a timestamp on the snapshots (default: shown below)
timestamp: false
# Optional: draw bounding box on the snapshots (default: shown below)
bounding_box: true
# Optional: crop the snapshot (default: shown below)
crop: false
# Optional: height to resize the snapshot to (default: original size)
height: 175
# Optional: Restrict snapshots to objects that entered any of the listed zones (default: no required zones)
#required_zones: []
# Optional: Camera override for retention settings (default: global values)
retain:
# Required: Default retention days (default: shown below)
default: 60
# Optional: Per object retention days
objects:
person: 100
# Optional: quality of the encoded jpeg, 0-100 (default: shown below)
quality: 80
cameras:
logan_north_front:
ffmpeg:
inputs:
- path: rtsp://CCCCCCCCC
roles:
- record
- detect
detect:
width: 2592 # <- optional, by default Frigate tries to automatically detect resolution
height: 1944
record:
enabled: true
# retain:
# days: 60
# mode: motion
# events:
# Optional: Number of seconds before the event to include (default: shown below)
# pre_capture: 10
# Optional: Number of seconds after the event to include (default: shown below)
# post_capture: 10
# Optional: Objects to save recordings for. (default: all tracked objects)
# retain:
# default: 100
# mode: motion
motion:
mask: 0.628,0.306,0.269,0.31,0.103,0.353,0.101,0.216,0.217,0.004,0.695,0.002
threshold: 32
contour_area: 8
improve_contrast: true
zones:
North_Gate:
coordinates:
0.003,0.547,0.587,0.574,0.666,0.567,0.675,0.469,0.998,0.63,0.995,0.993,0.005,0.991
loitering_time: 0
review:
alerts:
required_zones: North_Gate
logan_door_front:
ffmpeg:
inputs:
- path: rtsp://CCCCCCCCC
roles:
- record
- detect
record:
enabled: true
# retain:
# days: 60
# mode: motion
# events:
# Optional: Number of seconds before the event to include (default: shown below)
# pre_capture: 10
# Optional: Number of seconds after the event to include (default: shown below)
# post_capture: 10
# Optional: Objects to save recordings for. (default: all tracked objects)
# retain:
# default: 100
# mode: motion
zones:
Entry:
coordinates: 0,0,0.268,0,0.242,0.349,0.439,0.472,0.516,0.544,0.543,1,0,1
loitering_time: 0
inertia: 3
objects:
- package
- person
motion:
mask:
0.313,0.142,0.405,0.229,0.448,0.327,0.457,0.406,0.427,0.49,0.379,0.549,0.253,0.621,0.245,0.523,0.268,0.135
review:
alerts:
required_zones: Entry
detect:
annotation_offset: 0
logan_south_front:
ffmpeg:
inputs:
- path: rtsp://CCCCCCCCC
roles:
- record
- detect
record:
enabled: true
# retain:
# days: 60
# mode: motion
# events:
# Optional: Number of seconds before the event to include (default: shown below)
# pre_capture: 15
# Optional: Number of seconds after the event to include (default: shown below)
# post_capture: 15
# Optional: Objects to save recordings for. (default: all tracked objects)
# retain:
# default: 100
# mode: motion
zones:
South_Gate:
coordinates: 0.115,0.384,0.998,0.402,0.997,0.996,0.003,0.996,0.005,0.561
inertia: 3
loitering_time: 0
review:
alerts:
required_zones: South_Gate
version: 0.15-0
semantic_search:
enabled: false
reindex: false
model_size: small
camera_groups:
Main:
order: 1
icon: LuActivity
cameras: birdseye docker-compose file or Docker CLI commandversion: "3.9"
services:
frigate:
container_name: frigate
privileged: true # this may not be necessary for all setups
restart: unless-stopped
image: ghcr.io/blakeblackshear/frigate:d0cc8cb-h8l
#image: ghcr.io/blakeblackshear/frigate:0.14.0-beta4
shm_size: "1024mb" # 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
#- /dev/dri/card1 # for intel hwaccel, needs to be updated for your hardware
#- /dev/fb0
- /dev/hailo0
volumes:
- /etc/localtime:/etc/localtime:ro
- /mnt/Video/Frigate/config/:/config
- /mnt/Video/Frigate/media:/media/frigate
- type: tmpfs # Optional: 1GB of memory, reduces SSD/SD Card wear
target: /tmp/cache
tmpfs:
size: 1000000000
ports:
- "5000:5000"
- "8554:8554" # RTSP feeds
- "8555:8555/tcp" # WebRTC over tcp
- "8555:8555/udp" # WebRTC over udp
environment:
FRIGATE_RTSP_PASSWORD: "cccccc"
PLUS_API_KEY: "cccccc"
LIBVA_DRIVER_NAME: "radeonsi" Relevant Frigate log output2024-12-07 22:52:55.874153370 2024/12/07 22:52:55 [error] 172#172: *2 auth request unexpected status: 502 while sending to client, client: 127.0.0.1, server: , request: "GET /api/version HTTP/1.1", host: "127.0.0.1:5000"
2024-12-07 22:52:56.842034795 [2024-12-07 22:52:56] frigate.config.config WARNING : Customizing more than a detector model path is unsupported.
2024-12-07 22:52:56.842789278 *************************************************************
2024-12-07 22:52:56.842794287 *************************************************************
2024-12-07 22:52:56.842798445 *** Your config file is not valid! ***
2024-12-07 22:52:56.842804987 *** Please check the docs at ***
2024-12-07 22:52:56.842807391 *** https://docs.frigate.video/configuration/ ***
2024-12-07 22:52:56.842834522 *************************************************************
2024-12-07 22:52:56.842836886 *************************************************************
2024-12-07 22:52:56.842839160 *** Config Validation Errors ***
2024-12-07 22:52:56.842841294 *************************************************************
2024-12-07 22:52:56.842916324 : Value error, Model does not support detector type of hailo8l
2024-12-07 22:52:56.842920762 *************************************************************
2024-12-07 22:52:56.842923497 *** End Config Validation Errors ***
2024-12-07 22:52:56.842926202 *************************************************************
2024-12-07 22:52:57.285537606 [INFO] Service Frigate exited with code 1 (by signal 0)
s6-rc: info: service legacy-services: stopping Install methodDocker Compose Object DetectorOther Screenshots of the Frigate UI's System metrics pagesN/A Any other information that may be helpfulNo response |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 8 replies
-
Is it planned that the frigate plus service will support a model for the Hailo 8L in the future? Can someone point me to how to get these converted? |
Beta Was this translation helpful? Give feedback.
-
I would like to support it in the future, but I haven't been successful in converting either the mobiledet or yolonas models to HEF yet. |
Beta Was this translation helpful? Give feedback.
It's not in the supported list yet: https://community.hailo.ai/t/yolo-nas-onnx-to-hef/3168