Mohit Kumar

icon indicating a website link themohitkumar.github.io icon indicating an email [email protected]

icon company 5techlab icon location Dubai icon location • AI Engineer with 4+ years of experience in pretraining, fine-tuning, and alignment of Large Language Models (LLMs) and Visual Language Models (VLMs).

Results 16 comments of Mohit Kumar

TypeError: pad_sequence(): argument 'padding_value' (position 3) must be float, not NoneType

I am also getting the same error which base model you are using?

TypeError: pad_sequence(): argument 'padding_value' (position 3) must be float, not NoneType

I am also using Qwen-7b actually for llama unk_token is but qwen it is null. SO it is causing error. Pad sequence error. unk_token == null for Qwen unk_token"" If...

TypeError: pad_sequence(): argument 'padding_value' (position 3) must be float, not NoneType

@haotian-liu Could you please give your insights how we can train using qwen7b base model and solve the issue?

TypeError: pad_sequence(): argument 'padding_value' (position 3) must be float, not NoneType

Have you trained model by assigning pad_token_id to -100, could you please share config if feasible otherwise make comment over accuracy for the model you have trained.

remove GPU dependency for interference?

Hey Abhishek this is related to more your dependencies On Tue, 4 Jul, 2023, 11:47 Abhishek G, ***@***.***> wrote: > Hi Folks I admire the work of @iiLaurens > and...

Process hanged when using cpu offloading