Fetching metadata from the HF Docker repository... Yesterday
-
Remove visible quota note; preserve Budget Guard logic
2 days ago
-
Preserve ZeroGPU platform errors using session queue events instead of null call errors
-
Expose sanitized quota readiness and preserve precise budget errors before queue
-
Remove public account/model headers and visitor OAuth checks; keep server quota guard with retries
-
Add coding workspace, text/code references, per-model profiles and Continue reply
-
Allow 4096-token replies and 110-second generation with 180-second quota reserve
-
Pin Astheral chat to tested V3 weights; preserve UI and quota guard
-
Keep selected GGUF identity in responses and reset missing temporary models
-
Register the GGUF ZeroGPU handler during application startup
-
Add private GGUF upload, model selection and quota-guarded CUDA inference
-
Serve llama-ui routes directly without Gradio SSR interception
-
Preserve Hugging Face visitor headers through the Gradio bridge
-
Install real llama-ui with streaming Astheral ZeroGPU adapter and existing quota guard
-
Pause GPU generation before final three minutes of included quota
-
Correct Gradio examples for response settings inputs
-
Deploy private Astheral demo with real weights on ZeroGPU
-
initial commit
