BORIStheBLADE
Getting comfortable
Stupid question... Does it record the video clip still but label it audio instead of person or dog?
Blue Iris does not support the Sound Classifier (TensorFlow) module yet. I will be sending Ken an email to add the support.Stupid question... Does it record the video clip still but label it audio instead of person or dog?
I will have to look into what it will take to retrain the model so it can detect Glass Breaking.Can Glass Breaking be added to the sounds?
At this time the module does not support GPU.@MikeLud1 Will the sound classifier use GPU if available? If so, which GPU devices are supported?
It needs to be a 16 bit wav fileIs the new sound classifier able to take a video clip and analyze the sound or would it have to be a sound file?
In other words, could someone export a video clip and run it thru it outside of BI?
New Release CodeProject.AI v2.5-RC4
v2.5-RC4 Updates
- Dynamic Explorer UI: Each module now supplies its own UI for the explorer
- Support for sound audition modules in the explorer
- Improved dashboard
- Updated module settings schema that includes module author and original project acknowledgement
- Installer fixes
- Improved Jetson support
- Windows on Arm support
CodeProject.AI Server: AI the easy way.
Version 2.6.5. Our fast, free, self-hosted Artificial Intelligence Server for any platform, any languagewww.codeproject.com
version: '3.3'
services:
ai-server:
network_mode: host
# ports:
# - '32168:32168'
container_name: CodeProject.AI-Server
volumes:
- '/opt/senseai/ai:/etc/codeproject/ai'
- '/opt/senseai/modules:/app/AnalysisLayer/ObjectDetectionYolo/custom-models'
image: 'codeproject/ai-server:cuda12_2-2.5.0'
restart: always
deploy:
resources:
reservations:
devices:
- driver: nvidia
capabilities: [gpu]
What version of Blue Iris are you on? The current version does not need to find the custom models in a folder it will pull the list from CodeProject.AI APIHi folks, I'm having trouble getting the delivery and package models to work.
Here are my settings, what am I missing here?
View attachment 182995View attachment 182996View attachment 182998
View from inside the AI docker container showing where the .pt files are:
View attachment 182997
Docker compose for AI
YAML:version: '3.3' services: ai-server: network_mode: host # ports: # - '32168:32168' container_name: CodeProject.AI-Server volumes: - '/opt/senseai/ai:/etc/codeproject/ai' - '/opt/senseai/modules:/app/AnalysisLayer/ObjectDetectionYolo/custom-models' image: 'codeproject/ai-server:cuda12_2-2.5.0' restart: always deploy: resources: reservations: devices: - driver: nvidia capabilities: [gpu]
I've restarted both AI and BI and then testing with a recent clip where a DHL truck was in frame I only get tags for 'ipcam-general' and 'ipcam-animal' with no attempt for delivery.
View attachment 182999
Is my config incorrect somewhere?
What version of Blue Iris are you on? The current version does not need to find the custom models in a folder it will pull the list from CodeProject.AI API
Try restarting Blue Iris service to see if it pulls the custom model list.I'm on 5.8.4.2....
going to be 5.8.4.3 in a few minutes because I live on the edge
I removed the file path, but no change in what model BI is using
View attachment 183008View attachment 183009
Also what custom models show in ExplorerI'm on 5.8.4.2....
going to be 5.8.4.3 in a few minutes because I live on the edge
I removed the file path, but no change in what model BI is using
View attachment 183008View attachment 183009
One other thing make sure all the model names in you docker container are lowercase.Try restarting Blue Iris service to see if it pulls the custom model list.
version: '3.3'
services:
ai-server:
network_mode: host
# ports:
# - '32168:32168'
container_name: CodeProject.AI-Server
volumes:
- '/opt/senseai/ai:/etc/codeproject/ai'
- '/opt/senseai/modules:/app/preinstalled-modules/ObjectDetectionYOLOv5-6.2/custom-models'
image: 'codeproject/ai-server:cuda12_2-2.5.0'
restart: always
deploy:
resources:
reservations:
devices:
- driver: nvidia
capabilities: [gpu]
/vision
but that's not working for me