numpy
openai
mem0ai

[all]
voicemem[clap,slm,stream]

[audio]
soundfile

[clap]
torch
laion-clap
soundfile
scipy

[demo]
voicemem[web]

[environment]
torch
transformers
soundfile
scipy

[finetune]
ms-swift==4.5.2
bitsandbytes

[omni]
voicemem[slm]

[slm]
torch
transformers==4.52.3
accelerate
qwen-omni-utils
soundfile
scipy
funasr
torchaudio
modelscope
sherpa-onnx
sentence-transformers
fastapi
uvicorn
pydantic
sounddevice
pywebrtc-audio
websockets>=14

[stream]
voicemem[wav]
funasr
torchaudio
modelscope

[text]
sentence-transformers

[tts_local]
piper-tts
voxcpm

[voiceprint]
sherpa-onnx
funasr
torchaudio
modelscope

[wav]
voicemem[text]
sherpa-onnx
soundfile
scipy
torch
transformers

[web]
voicemem[all]
fastapi
uvicorn
