Lon
ryukn.bsky.social
Lon
@ryukn.bsky.social
Software Engineer specializing in Machine Learning in Tokyo, Japan.
Working in the field of AI for science.

Opinions are my own.
I was using HuggingFace Transformers library for fine-tuning purposes.
Unsloth seems like it could be a good solution for the memory issues.
August 16, 2025 at 5:23 AM
Let's reproduce GPT-2 (124M)
YouTube video by Andrej Karpathy
www.youtube.com
August 2, 2025 at 11:21 AM