Whisper-vs-Sensevoice-Small / requirements.txt
Luigi's picture
add tiktoken for whisper
c680485
# Gradio UI
gradio>=3.39.0
# Core ASR
torch>=2.0.0
torchaudio
transformers>=4.35.0
# FunASR SenseVoice
funasr>=1.0.14
# Audio handling
pydub>=0.25.1
ffmpeg-python>=0.2.0 # wrapper for ffmpeg; you’ll still need system ffmpeg installed
# Speaker Diarization
pyannote.audio>=2.1.1
huggingface-hub>=0.18.0 # for pyannote model download/auth
# (Optional) if you want GPU‐accelerated pipelines outside of HF Spaces
# accelerate>=0.20.0
opencc-python-reimplemented
ctranslate2==4.5.0
faster-whisper>=1.1.1
termcolor
nvidia-cublas-cu12
nvidia-cudnn-cu12
whisper
transformers_stream_generator
accelerate
tiktoken