Text Generation
Transformers
PyTorch
TensorBoard
Safetensors
bloom
Eval Results
text-generation-inference
Inference Endpoints

Bloom-176B on Google Colab

#63
by thisisanshgupta - opened

I was wondering if there's any way to use Bloom-176B on Google Colab?

BigScience Workshop org

No

thisisanshgupta changed discussion status to closed

It's too big model for Google Colab usage limits. Even for premium or professional plans.

If inference time is not a constraint, you can run BLOOM 176-B on a desktop computer, even if you don’t have a GPU, you just need at least 16GB of RAM. The following blog post contains instructions to do this:

https://towardsdatascience.com/run-bloom-the-largest-open-access-ai-model-on-your-desktop-computer-f48e1e2a9a32

In my computer (i5 11gen, 16GB RAM, 1TB SSD Samsung 980 pro), the generation takes 3 minutes per token using only the CPU, which is a little slow but manageable.

I was wondering if there's any way to use Bloom-176B on Google Colab?

Found this:
https://harishgarg.com/writing/generate-content-using-bloom-open-source-ai-model/
Not sure if this helps or not.

You can run it on Runpod using KoboldAI, here (I believe you need 335GB+ of VRAM though, which isn't cheap):

https://runpod.io?ref=ev1llooa

Sign up or log in to comment