Hugging Face's logo

Spaces:
polats
/
ZeroGPU-LLM-Inference
Running on Zero

App Files Community
ZeroGPU-LLM-Inference

Commit History

add bot avatar image
9f8aeaf
Running

polats commited on

suppress asyncio cleanup warnings from DDGS library
c9a4c50

polats Claude Opus 4.5 commited on

fix infinite page expansion on mobile
936637d

polats Claude Opus 4.5 commited on

fix mobile chat input being cut off
115cffb

polats Claude Opus 4.5 commited on

update ui for mobile
51a9e4a

polats commited on

keep audio component visible
c6a3e9f

polats commited on

simplify UI
8b6384b

polats commited on

Restore buttons param and update Gradio SDK to 6.3.0
26998d4

polats commited on

Remove unsupported buttons parameter from Chatbot for HF compatibility
2ed4599

polats commited on

Use pocket-tts from PyPI instead of local directory
5c81660

polats commited on

add pocket-tts voice
65be932

polats commited on

Add comprehensive changelog for UI/UX improvement session
31c47f7

Luigi commited on

Add UI/UX improvements summary documentation
a68593c

Luigi commited on

Add comprehensive documentation and user guide
c1bc514

Luigi commited on

Major UI/UX improvements for better user experience
df40b1d

Luigi commited on

Fix UI not resetting after cancel - add reset handler
b7e5000

Luigi commited on

Fix GeneratorExit handling to prevent runtime error
c49f312

Luigi commited on

Fix cancel generation feature by removing GeneratorExit handler
9466288

Luigi commited on

Handle GeneratorExit by yielding UI reset and returning cleanly
a430811

Luigi commited on

Simplify GeneratorExit handling - let it propagate and rely on finally block
6dc4bdb

Luigi commited on

Fix GeneratorExit handling by re-raising after UI reset
6ff484e

Luigi commited on

Fix UI state reset after generation cancellation
fcb2aaa

Luigi commited on

Fix GeneratorExit exception in cancel generation by adding proper exception handling
4bc617b

Luigi commited on

bugfix
fd70375

Luigi commited on

Improve cancel generation with robust UI state management and orchestrator pattern
9ac7f36

Luigi commited on

Restore cancel generation feature with improved UI integration
a94befb

Luigi commited on

Remove cancel generation feature as it didn't work
efa3ae6

Luigi commited on

Restore proper cancel functionality
afa1066

Luigi commited on

Test stopping criteria effectiveness
f0d2ab7

Luigi commited on

Improve cancel generation responsiveness
48e50b8

Luigi commited on

Fix cancel generation to gracefully stop ongoing response generation
a73d8f4

Luigi commited on

remove qwen3 30b-a3b and qwen3 next 80b-a3b
4af617b

Luigi commited on

Remove AOT indication from UI duration estimate
f97cbfc

Luigi commited on

Remove AOT compilation completely and enable use_cache
426163f

Luigi commited on

Fix dynamic_shapes kwargs to match inputs structure for AOT compilation
273acf8

Luigi commited on

Fix AOT compilation dynamic_shapes to match expected arg names for torch.export.export
0a99dfc

Luigi commited on

Update requirements.txt
4d2c362
verified

Luigi commited on

Update requirements.txt
48d806a
verified

Luigi commited on

Add dynamic GPU time estimate indicator to UI
fc989b4

Luigi commited on

Improve model size detection: replace ad-hoc string parsing with reliable params_b field in MODELS dict
ab92e0d

Luigi commited on

Add qwen 80b-a3b
8cdf3e1

Luigi commited on

Set better defaults for free-tier users: Qwen3-1.7B model, 1024 max tokens, search disabled
2cae073

Luigi commited on

Adjust duration estimation for H200 performance - reduce conservative estimates
de766da

Luigi commited on

Use actual parameter count for AOT decision instead of string matching
e3e334f

Luigi commited on

Make AOT compilation conditional for models >= 2B parameters to optimize free tier usage
4500f92

Luigi commited on

Add AOT compilation optimization for ZeroGPU acceleration
a7866ff

Luigi commited on

add 4 20b+ models after enabling dynamic gpu duration
fea2910
verified

Luigi commited on

Add dynamic duration calculation for ZeroGPU acceleration
6073cc2

Luigi commited on

make qwen-4b default
d3726c6
verified

Luigi commited on

disable two models that cannot run or too run too slowly on hf spaces with zerogpu
3dc7ced

Luigi commited on