Gal Cohen (galco) issues

Repositories
Issues
Comments

Results 4 issues of


                                            Gal Cohen (galco)

FasterTransformer

Hi, I saw in mpt model card that the models could run with FasterTransformer I didn't find any details about that anywhere can you guys share the conversion scripts or...

Finetuning with llm-foundry (composer format)

Hi, Is there any option to convert HF ckpt to composer and FT with the llm-foundry scripts? Thanks!

Support CPU quantization

Hi is there a way to run EETQ without accelerator ? at least for the quantization process thanks

fix: dont change quant storage dtype in case of fsdp

# Description ## Motivation and Context Currently jamba storage dtype changes if not prequant in case of fsdp training and it should act as in dsz3 ## How has this...