BarunLM-35M: a parameter-efficient 35M base language model trained on 5.7B tokens
54stars3forksPython
language-modelmachine-learningpytorchsmall-language-modeltext-generation
Real data pulled from GitHub this week. The author's original repo lives upstream.
View on GitHub