Spaces:
Running
on
Zero
Apply for community grant: Academic project (gpu and storage)
AI Tamil Nadu (https://aitamilnadu.org/) is a non-profit hyperlocal community that aims to democratize AI specially in the Tamil Nadu, India. We are a bunch of volunteers spending our free time to run technical events and impart AI knowledge to the common public.
We're collaborating with CIP (https://cip.org/) to build a constitution around Tamil LLM by involving the people from the AITN community (15-20 people to start with). We want to convene a small Tamil-speaking group to run a Pol.is, to generate principles for a collective constitution, prompt a default Tamil model (HF space: https://huggingface.co/spaces/aitamilnadu/LLaMa_Advisor_Tamil) based on it, and assess how well it behaves. The bot will be designed to be an lifestyle/realtionship advisor for the Tamil community where people can prompt the model and get advice.
The objective here is to first run a MVP to get the opinion from the Tamil community about the model's actual behavior and supposedly expected behavior and build a constitution around it and use it to build better models for Tamil. This activity is run as a open-source initiative. It would be really great if we could get GPU and persitent storage support for this activity from HuggingFace for the duration of the planned project (approximately a month or two for the MVP and extend it to a larger timeframe where we open this to the general Tamil public). Awaiting a positive response.
Regards,
Abinaya Mahendiran
On behalf of AI Tamil Nadu
As this Space seems to be a modified version of the llama 2 Space, I thought it should be able to run on Zero as is, but I'm seeing this error:
Traceback (most recent call last):
File "/home/user/app/app.py", line 115, in <module>
gr.Examples(examples=examples, inputs=[msg], label="எடுத்துக்காட்டுகள் / Examples")
File "/usr/local/lib/python3.10/site-packages/gradio/helpers.py", line 56, in create_examples
examples_obj = Examples(
File "/usr/local/lib/python3.10/site-packages/gradio/helpers.py", line 154, in __init__
raise ValueError("If caching examples, `fn` and `outputs` must be provided")
ValueError: If caching examples, `fn` and `outputs` must be provided
IMPORTANT: You are using gradio version 4.27.0, however version 4.29.0 is available, please upgrade.
I think you can fix this by simply upgrading the gradio to 4.29.0. You can change the version in the README.md. https://huggingface.co/spaces/aitamilnadu/tamil-ai-assistant/blob/45634489e48314354f4feb9fef9506265073f885/README.md?code=true#L7
BTW, on Spaces, you can use 50GB of ephemeral storage https://huggingface.co/docs/hub/en/spaces-storage#persistent-storage-specs and most of the Spaces don't need the persistent storage, but let us know if you actually need one.