Is the original template supposed to be working?

#75
by moilanep - opened

When I try the original template app in https://huggingface.co/spaces/agents-course/First_agent_template I get a bunch of these errors:

Error in generating model output:
422 Client Error: Unprocessable Entity for url: https://api-inference.huggingface.co/models/Qwen/Qwen2.5-Coder-32B-Instruct/v1/chat/completions (Request ID: tRbpf6nat7SP2sdvj_kpA)

Input validation error: inputs tokens + max_new_tokens must be <= 16000. Given: 123164 inputs tokens and 2096 max_new_tokens
Make sure 'text-generation' task is supported by the model.

Moreover, when I duplicate the space, the app looks really weird (screencapture attached)
Agent_app.png
. Has anybody else this problem?

Please restart the space

I am still getting the same issue even after restarting the space. appreciate any help in this regard.

I have the same issue! I am duplicating the space and changing the model, will come back if that works.

edit: That did nothing because now I'm just getting tons of 404, 401 errors and stuff.

Enabling 3rd party cookies should fix this issue.

Hi @ChakradharG , Thanks. It works for my after enabling the 3rd party cookies in chrome.

Enabling 3rd party cookies should fix this issue.

It didn't for me, I tried on both Chrome and Edge.
bot avatar
Step 1

💥 Error

Error in generating model output:
422 Client Error: Unprocessable Entity for url: https://api-inference.huggingface.co/models/Qwen/Qwen2.5-Coder-32B-Instruct/v1/chat/completions (Request ID: lEMcbfXlvoeIQ2Wqa0MPT)

Input validation error: inputs tokens + max_new_tokens must be <= 16000. Given: 242578 inputs tokens and 2096 max_new_tokens
Make sure 'text-generation' task is supported by the model.

Step 1 | Input-tokens:12,930 | Output-tokens:130 | Duration: 0.48

Can you paste a screenshot of the actual UI?

Can you paste a screenshot of the actual UI?

image.png

Most likely the model is overloaded, can you try changing the model endpoint in app.py?
image.png

Sign up or log in to comment