Commit History

Enable 8-bit quantization option
a6b24a3
verified

Alovestocode commited on

Load model during startup for GPU reservation
6a8403c
verified

Alovestocode commited on

Expose @spaces.GPU decorator with shim
00b5731
verified

Alovestocode commited on

Use spaces.GPU decorator
abe3cd7
verified

Alovestocode commited on

Add root healthcheck and spaces GPU decorator
e551bf6
verified

Alovestocode commited on

Restore spaces.GPU usage
405a7ef
verified

Alovestocode commited on

Replace Gradio with FastAPI HTML console
ee25577
verified

Alovestocode commited on

Serve Gradio UI at /gradio path
74309f5
verified

Alovestocode commited on

Remove manual uvicorn loop
62eb658
verified

Alovestocode commited on

Retry uvicorn with fallback ports
a003115
verified

Alovestocode commited on

Run uvicorn with default PORT
3e313d0
verified

Alovestocode commited on

Remove unavailable Llama fallback
49afe36
verified

Alovestocode commited on

Add tokenizer fallback cascade
82dcfd3
verified

Alovestocode commited on

Default to smaller Llama checkpoint for faster init
f5c6fe4
verified

Alovestocode commited on

Mount Gradio at root with queued blocks
534388e
verified

Alovestocode commited on

Serve Gradio UI at /gradio path
4be9bb4
verified

Alovestocode commited on

Queue Gradio blocks before mount
5e3cb8e
verified

Alovestocode commited on

Revert to Spaces-managed server startup
3498ceb
verified

Alovestocode commited on

Fix uvicorn mount reference
6411bc4
verified

Alovestocode commited on

Start uvicorn server in Space entrypoint
b5f9b82
verified

Alovestocode commited on

Use merged Qwen checkpoint by default
335d4f5
verified

Alovestocode commited on

Initial ZeroGPU router backend
35d68ae
verified

Alovestocode commited on