Spaces:
Running on Zero
Running on Zero
Commit History
-
Smooth playback and guard visitor role Running
4310ff6 -
Stream guarded stable audio prefixes
f705b22 -
Ground Lyra and drain completed replies
f937854 -
Measure RoleForge latency and streaming equivalence
3e17fa6 -
Allow longer bounded LFM story replies
bcd4aeb -
Extend bounded LFM story replies
c9db6ec -
Reject repeated generic LFM openings
a52c7f3
-
Allow complete long LFM replies
e970e6c -
Keep LFM text and audio aligned
c07452a -
Reduce coherent LFM turn latency
a5e1103 -
Honor the official LFM interleaved contract
ce33838 -
Make live dialogue quality-first and server-gated
cd521aa -
Restore sample-exact browser playback
b285539 -
Ground Lyra and regenerate early repeats
40a62e3 -
Stabilize PCM playback and barge-in
b0aaa04 -
Attach PCM routes to Gradio server
888fca5 -
Replace Gradio audio with WebSocket PCM transport
aa0f74f -
Smooth live audio streaming
cadb025 -
Start live session from microphone recording
022bba5 -
Repair live microphone session routing
d64559d -
Add interruptible live LFM conversation
8a350d4 -
Share LFM context across ZeroGPU workers
cecd236 -
Retain bounded LFM conversation context
1aa8308 -
Migrate RoleForge feasibility lab to LFM2.5-Audio
fd9ccf8 -
Apply TURN relay to both WebRTC peers
b68901c -
Buffer ICE without changing WebRTC component identity
3ed8155 -
Revert "Buffer early WebRTC ICE candidates"
This reverts commit 2aceb54c89f857be0750a4cabfa1eaf51fd30625.5576640 -
Buffer early WebRTC ICE candidates
2aceb54 -
Warm PersonaPlex under live inference mode
a4a16d2 -
Relay Space WebRTC through short-lived TURN
e9d083e -
Spike PersonaPlex WebRTC streaming on ZeroGPU
0382c60 -
Repair microphone capture frontend
e670013 -
Use ZeroGPU supported torch version syntax
e4204c7 -
Repair PersonaPlex for Blackwell ZeroGPU
cbc0e90 -
Pin Gradio-compatible web stack
8010b00 -
Pin Pydantic for Gradio schema compatibility
0ac1780
-
Align Gradio with PersonaPlex dependencies
baf3f23 -
Use platform-managed ZeroGPU runtime
bf9dd83 -
Integrate generated Hugging Face Space scaffold
c179431 -
initial commit
Verified signature15da56e -
Use scoped OAuth for gated model access
45d6972 -
Build RoleForge voice NPC feasibility lab
900c6ab
-
Add support for audio input upload in demo interface
1b0acf6
-
Enhance audio processing in generate_response by adding prepend and append silence to improve model response timing
b840b20 -
Add warmup function to initialize CUDA graphs and improve performance in get_models
d96db07 -
Refine sine wave token usage in generate_response function to simulate user silence with official PersonaPlex constants
d3a0030 -
Reset streaming state within context in generate_response function
4b1918b -
Enhance response generation by adding silence continuation and simulating user input with sine wave tokens
7e8f6df -
Reorganize imports for moshi models and streamline text tokenizer loading in get_models function
a3b50ab -
Implement lazy loading for models in get_models function to optimize GPU resource usage
8c0b009