PyPI page
Home page
Author:
vovaRL
License:
MIT
Summary:
Masked Diffusion Language Model (MDLM) pretraining in PyTorch with FSDP2, Muon optimizer, and Hugging Face Jobs support.
Latest version:
0.1.74
Required dependencies:
cut-cross-entropy
|
filelock
|
huggingface-hub
|
numpy
|
safetensors
|
torch
Optional dependencies:
build
|
flash-attn-4
|
pytest
|
twine
Downloads last day:
183
Downloads last week:
242
Downloads last month:
5,366
7-day smoothing