Hugging Face's logo
Nikhil Verma's picture
8

Nikhil Verma

vermanic

AI & ML interests

None yet

Organizations

Amazon Web Services's profile picture

vermanic's activity

New activity in HuggingFaceH4/starchat-beta 4 months ago

Incomplete Output even with max_new_tokens

6
#27 opened 9 months ago by vermanic
New activity in HuggingFaceH4/starchat-beta 8 months ago

SFT taking high memory with Transformers (>5x the amount it takes to load model checkpoint )
#29 opened 8 months ago by vermanic
New activity in codellama/CodeLlama-34b-Instruct-hf 9 months ago

What is right GPU to run this

4
#7 opened 9 months ago by Varunk29
New activity in HuggingFaceH4/starchat-beta 9 months ago

Every Inference gives the prompt as part of output also, any way to remove that?

1
#26 opened 9 months ago by vermanic
New activity in HuggingFaceH4/starchat-alpha 9 months ago

Incomplete Output even with max_new_tokens
#26 opened 9 months ago by vermanic

Model doesn't end/terminate generation: Need to modify EOS token

9
#2 opened about 1 year ago by eugenesiow
New activity in WizardLMTeam/WizardCoder-15B-V1.0 9 months ago

Parallel Inferences using GPU?

#38 opened 9 months ago by vermanic