nxdev-org/nemotron-speech-streaming-en-0.6b

Changing title

Clean up HuggingFace Hub settings

Reduce logging verbosity

Add fair usage disclaimer for 20s recording limit

Fix reconnection loop after idle timeout

Add connection management, idle timeout, and server stats

Add connection management, health monitoring, and circuit breaker

Fix: Use last interim transcript as fallback when no final transcript received

feat: Add WER evaluation mode with LibriSpeech

feat: UI improvements and NVCF robustness fixes

Add per-session Triton client and 20s recording limit

Switch from NVCF gRPC to Triton Inference Server

Add package-lock.json

Added the default

Dummy change - Revert2

Dummy change - Revert

Dummy change

Changed icon and title

Added NVIDIA logo -3

Added NVIDIA logo -2

Added NVIDIA logo -1

Added NVIDIA logo

Adding logs

Added the new parameters at inference time

wss fix

Fixed the NVIDIA logo issue

First Commit

initial commit