Remove cancel generation feature as it didn't work

Restore proper cancel functionality

Test stopping criteria effectiveness

Improve cancel generation responsiveness

Fix cancel generation to gracefully stop ongoing response generation

remove qwen3 30b-a3b and qwen3 next 80b-a3b

Remove AOT indication from UI duration estimate

Remove AOT compilation completely and enable use_cache

Fix dynamic_shapes kwargs to match inputs structure for AOT compilation

Fix AOT compilation dynamic_shapes to match expected arg names for torch.export.export

Update requirements.txt

Update requirements.txt

Add dynamic GPU time estimate indicator to UI

Improve model size detection: replace ad-hoc string parsing with reliable params_b field in MODELS dict

Add qwen 80b-a3b

Set better defaults for free-tier users: Qwen3-1.7B model, 1024 max tokens, search disabled

Adjust duration estimation for H200 performance - reduce conservative estimates

Use actual parameter count for AOT decision instead of string matching

Make AOT compilation conditional for models >= 2B parameters to optimize free tier usage

Add AOT compilation optimization for ZeroGPU acceleration

add 4 20b+ models after enabling dynamic gpu duration

Add dynamic duration calculation for ZeroGPU acceleration

make qwen-4b default

disable two models that cannot run or too run too slowly on hf spaces with zerogpu

Update app.py

Update app.py

Update app.py

add original apriel 15b

use apriel 8bit

run Apriel on 4bit

Add Apriel-1.5-15b-Thinker

Update app.py

Update app.py

Update requirements.txt

Update requirements.txt

Update requirements.txt

Update app.py

add 4 models from qwen3 family

add qwen3 32b awq

Update app.py

Update README.md

Update requirements.txt

Update app.py

Update app.py

Update app.py

Update app.py

Update app.py

feat(models): add Granite-4.0-Micro and Qwen3-4B-Instruct-2507 to MODELS registry

feat(models): Added three new models

add 5 models from liquid ai