Megatron-LM icon indicating copy to clipboard operation
Megatron-LM copied to clipboard

Published 20 hours ago • verified publisher icon NVIDIA

Fix wrong key for output_layer_init_method

Open zhipeng93 opened this issue 11 months ago • 1 comments

Given that we aim to use mcore to do the training, we have a function to parse the args from Megatron-LM to mcore.

Howover, the key of output_layer_init_method is incorrect. [1]

This PR fix this.

[1] https://github.com/NVIDIA/Megatron-LM/blob/main/megatron/core/transformer/transformer_config.py#L95

zhipeng93 avatar Feb 28 '24 10:02 zhipeng93

Marking as stale. No activity in 60 days.

github-actions[bot] avatar Apr 28 '24 18:04 github-actions[bot]