-
Point clone URL at mohameddalii/translation-audio
-
Do not preload cuDNN 12 when running Cohere/PyTorch
-
Add Cohere Transcribe Arabic backend for Whisper vs Cohere compare
-
Batched Whisper, replica pool for concurrent users, SSE streaming
-
Add Locust UI load test for /translate
-
Fix CUDA lib discovery for namespace nvidia package
-
Preload CUDA 12 cublas for Vast CUDA 13.2 + RTX 5090
-
Audio translation layer for RTX 5090 / Vast.ai
-
initial commit
