metadata
language:
- en
tags:
- text generation
- pytorch
- causal-lm
- gpt_neox
license: mit
datasets:
- hoskinson-center/proof-pile
ProofGPT-v0.1
Model Description
ProofGPT-v0.1 is a 1.3B parameter language model based on the GPT-NeoX architecture and trained on the proof-pile (v1.1). The model is initialized with pythia-1.3b weights. ProofGPT-v0.1's Weights & Biases training log is viewable here.
Detailed evaluations coming soon :)
Note: Commit 9695b51
updated the tokenizer to have bos, eos, and unk tokens.