Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

Update requirements.txt

Update app.py

Update app.py

lazy load the model

try to sovle the socket finding problem

push multiple feature to make this tool more serious

change the visibility of the second part

problem in caching solve -> correct visualization

don't use zero caching

enable multiple head edit at the sametime

change the port for exposing the automatic port

UI for changing the attetnion instead command text

add visual edition of attention map

add visual edition of attention map

change the share=true to the location of server

change the share=true to the location of server

share=True

fwd_hooks handled differently

share=True

change fwd_hooks location

remove the quantization of Gemma2-2b

handle how model load differently

use HF_TOKEN for downloading model

remove gr.Cache()

update the Gradio version in Repo Card

remove huggingface_hub HFfolder

update app

refactor :update Gradio to latest version

Add missing YAML configuration to README

Fix HF Space URL to Nicknam

Fix HF Space URL to Nicknam

Initial commit of Attention Surgeon