Closed nokados closed 6 months ago
Thank you for opening your first issue in this project! Engagement like this is essential for open source projects! :hugs:
If you haven't done so already, check out Jupyter's Code of Conduct. Also, please try to follow the issue template as it helps other other community members to contribute more effectively.
You can meet the other Jovyans by joining our Discourse forum. There is also an intro thread there where you can stop by and say Hi! :wave:
Welcome to the Jupyter community! :tada:
Met the same problem. And I think I'm already get access to this model through "https://huggingface.co/bigcode/starcoder"
I have the same problem with different models as well (including mistralai/Mistral-7B-Instruct-v0.2 and meta-llama/Meta-Llama-3-8B)
I have the same problem with different models as well (including mistralai/Mistral-7B-Instruct-v0.2 and meta-llama/Meta-Llama-3-8B)
I can solve it by updating the libraries:langchain and huggingface-hub
according to https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/discussions/115, changing huggingface_hub.interface_api file on 152 line i replaced api_url as as self.api_url = f"{INFERENCE_ENDPOINT}/models/{repo_id}" it works for me.
But it seems that the length of answer is limitted. I don't know why :(
Thank you all for reporting this issue! We just merged a PR to fix this. It will be included in the next release, tentatively scheduled for early next week.
Note: this PR will not be backported to 1.x as JupyterLab 3 reached end-of-maintenance yesterday. See https://github.com/jupyterlab/jupyter-ai/issues/761.
Description
Hugging Face Hub Integration does not work. I can access HF Inference API from curl, but jupyter-ai chat fails with the
ValueError: Error raised by inference API: Cannot override task for LLM models
(see Context section for the full traceback). I've tried different open models: mistralai/Mistral-7B-Instruct-v0.2, bigcode/starcoder2-3b. None of them works.I guess, the reason is that you set
task
argument explicitly, while langchain does not set it. Moreover, Langchain uses InferenceClient instead of deprecated InferenceApi.Anyway, the best solution, in my opinion, is to replace langchain's HuggingFaceHub with HuggingFaceEndpoint, because the first is deprecated in langchain.
Reproduce
Expected behavior
Any answer from the model served on HF Inference API
Context
Traceback From Chat
Troubleshoot Output
Command Line Output