erm

re added num_beams

Added EOS toke stuff increased new token limit and added QOL features to frontent

Added it back to gpt2

Removed beam thing

Added max new tokens value

Create tiny.py

Create mamba.py

Create fast.py

Create llama2.py

Fuck padding

Attention mask shi

Fixed GPT2 and made the API POST and GET

Added gpt2 and fixed shi

Added hermes model

Pls be fixed

idk

Testing

Should be faster and work