label-studio-ml-backend
label-studio-ml-backend copied to clipboard
HumanSignal
Configs and boilerplates for Label Studio's Machine Learning backend
docker runtime tip: `RuntimeError: Cannot re-initialize CUDA in forked subprocess. To use CUDA with multiprocessing, you must use the 'spawn' start method` `- DEVICE=cuda:0`
- Remove unused imports. - Update import for label_studio_sdk (to adapt new changes).
I'm experiencing a "Validation error" in Label Studio when sending predictions from my backend. **Steps to Reproduce:** 1. I'm sending a POST request to `/predict` endpoint with the following JSON...
I am currently developing a custom ML Backend for active learning with Label Studio's enterprise version. I have managed to link my ML backend through AWS ECS. However, I am...
I deployed label-studio 1.8.0 in kubernetes, and deployed an ML Backend from the Segment Anything example. I can connect to the ML Backend from the Label Studio, but I get...
Rust is required to build the pip requirements, this commit adds it to the container.
I have the same problem #234 1. clone the repo 2. cd into default_config folder 3. docker-compose up start the project 4. then meet this problem I've try python 3.8-slim/python...
When my model's `predict(self, tasks, **kwards)` function is called each task has a `data` field containing a dictionary and in my case, the dictionary has an `image` key with `gs://bucket-name/image-name.jpg`...
Metadata
Owner
Metadata
Configs and boilerplates for Label Studio's Machine Learning backend