Spaces:
Paused
Paused
Commit History
-
\Enable video file uploads (mp4, mkv, mov, avi)"
HARI PREETHAM2aa7be4 -
\Map language codes to BCP-47 for Azure Whisper"
HARI PREETHAMd7193bb -
\Fix Azure SDK set_service_property keyword argument error"
HARI PREETHAMbc37cd1 -
\Refine Azure region handling"
HARI PREETHAMbe0ccf8
-
\Add Sarvam AI translation endpoint"
HARI PREETHAMf795322 -
\Fix Sarvam AI nested transcript parsing and termination detection"
HARI PREETHAM788fb16 -
\Fix Sarvam AI extraction: WAV conversion, chunking, and field mapping"
HARI PREETHAMa3f3f5b -
\Add detailed logging for Sarvam AI debugging"
HARI PREETHAMad462ff -
\Integrate Sarvam AI STT engine"
HARI PREETHAMdff797e
-
\Enable Azure Whisper model support + prompt hints"
HARI PREETHAM44e7f68 -
\Add Azure Speech-to-Text endpoint"
HARI PREETHAMe7b11c5
-
\Revert Whisper to large-v2 for verified accuracy (large-v3 hallucinated), keeping NLLB-3.3B"
HARI PREETHAM4fb8564 -
\Force rebuild with pre-downloaded large-v3 model to fix runtime timeout"
HARI PREETHAM5172015 -
\Revert to large-v2 because large-v3 failed to download on free tier"
HARI PREETHAMd1790cc -
\Switch to large-v3 for maximum accuracy, disabling streaming mode"
HARI PREETHAMfcb143f -
\Upgrade to Medium model for accuracy, keeping VAD + Repetition Penalty"
HARI PREETHAMbc8957c -
\Revert to Beam Search (Beam 5) for accuracy, keeping Small + VAD + Repetition Penalty"
HARI PREETHAM52fd1aa -
\Switch to small + VAD + RepetitionPenalty to fix hallucinations and improve speed"
HARI PREETHAMf44dc7a -
\Switch to base model + Greedy Decoding (beam_size=1) for 5x speedup"
HARI PREETHAM494e435 -
\Emergency Downgrade to small model to fix unresponsiveness"
HARI PREETHAM162937d -
\Downgrade to medium model (1.5GB) for faster CPU streaming and avoid timeout"
HARI PREETHAMe80864e -
\Force rebuild backend to clear stuck queues for streaming"
HARI PREETHAMfc6ef49 -
Downgrade to large-v2 for speed
HARI PREETHAM3003280 -
Force rebuild for large-v3 model download
HARI PREETHAMa571917 -
Switch to large-v3 for maximum accuracy
HARI PREETHAM957ff91 -
Downgrade model to large-v2 for speed
HARI PREETHAM4107394
-
Migrate to faster-whisper (CTranslate2) for 4x CPU speedup
HARI PREETHAM3732708 -
Add language and prompt support
HARI PREETHAMdaf5837 -
Upgrade Whisper Space to large-v2 for max accuracy
HARI PREETHAM4c38d17 -
Upgrade Whisper model to 'medium' for better Hindi/Telugu accuracy
HARI PREETHAM39bf4c7 -
Initial Deploy of Whisper Engine
HARI PREETHAM838100c -
initial commit
Verified signature90908b0