LoreonLabsPlatform

Overview

Intelligence

Markets
Builders
Research
Narratives
Ecosystems
Launchpads

Discover

Search
Sources

Python

Megatron-LM-perl

MikaStars39/Megatron-LM-perl

PythonEmerging

Stars

—

Forks

—

Contributors

8

Last push

1mo ago

Recent commits

Latest commits.

Merge pull request #6 from MikaStars39/rooplus
6ad4036ChenYXxxx2mo ago
fix: avoid mutating live SOAP optimizer state in sharded_state_dict
f283ee6keming.32mo ago
fix: handle ShardedTensorFactory in SOAP preconditioner sharding
9103b4ckeming.32mo ago
fix: propagate prepend_axis_num for SOAP preconditioner sharding
33831a0MikaStars392mo ago
fix: use 1D stacking for SOAP preconditioner sharding under TP
4068aa1MikaStars392mo ago

fix: use temporary config mutation for SOAP Adam path

9fad900MikaStars392mo ago

fix: guard init_state_fn for None in sharded_state_dict

2cd8f8eMikaStars392mo ago

fix: avoid mutating shared config and ensure fp32 dtype in SOAP init

f6569daMikaStars392mo ago

Top contributors

Builders behind this project.