Apply for community grant: Academic project (gpu)

#1
by AnwenHu - opened
mPLUG org

mPLUG-DocOwl 1.5 is an OCR-free Multimodal Large Language Model with 8B parameters. It supports 1) document/webpage parsing; 2) table to markdown; 3) chart to markdown; 4) natural image parsing; 5) multi-grained text recognition; 6)multi-grained text grounding; 7) VQA with detailed explanations; 8) VQA with concise answers; 9) information extraction; 10) image captioning; 11) natural language inference

Hi @AnwenHu Looks like this Space is already using ZeroGPU. So, didn't ZeroGPU work for your Space?
But the following error is in the log and it doesn't seem to be related to ZeroGPU.

2024-04-03 06:53:44 | ERROR | stderr | Traceback (most recent call last):
2024-04-03 06:53:44 | ERROR | stderr |   File "/home/user/app/app.py", line 28, in <module>
2024-04-03 06:53:44 | ERROR | stderr |     no_change_btn = gr.Button.update()
2024-04-03 06:53:44 | ERROR | stderr | AttributeError: type object 'Button' has no attribute 'update'

If ZeroGPU doesn't work, we can still assign a normal GPU as a grant, but as we recently started using ZeroGPU as the default hardware for grants, it would be nice if you could check if it's possible to run on it.

mPLUG org

Hi, @hysts , we have successfully run the demo with ZeroGPU, thanks for your reply.

Awesome!!

Oh, BTW, with the latest gradio, you can use multimodal chatbot. Maybe your Space can use the feature.
You can check this Space as an example.

Sign up or log in to comment