accelerate>=0.26.0,<=1.14.0
transformers[sentencepiece]>=4.35.2,<=5.13.0
sentence-transformers>=2.2.2,<=5.3.0
openvino-genai
# last supported commit f50a0b326df82d4b05c461a9184a97f616bfa63a
optimum-intel[nncf]>=1.19.0
pandas>=2.0.3,<=3.0.5
numpy>=1.23.5,<=2.5.1
tqdm>=4.66.1,<=4.70.0
diffusers>=0.22.0,<=0.39.0
datasets>=3.6.0,<=5.0.1
jinja2>=3.1.0,<=3.1.6
scipy>=1.3.2,<=1.18.0
# For video-to-text VLM sample/VideoGen
opencv-python<=5.0.0.93
imageio[pyav]<=2.37.4
imageio-ffmpeg<=0.6.0
lpips==0.1.4
scikit-image<=0.26.0
# - openbmb/MiniCPM-o-2_6
soundfile==0.14.0
librosa==0.11.0
vocos==0.1.0
vector-quantize-pytorch==1.31.0
torchaudio>=2.1.0,<=2.11.0
torchvision>=0.16,<=0.28.0
# For LoRA adapters
peft==0.20.0
# For OpenGVLab/VideoChat-Flash-Qwen2_5-7B_InternVideo2-1B
timm==1.0.28
decord==0.6.0; platform_system != "Darwin"
# For speech generation evaluation
jiwer>=3.0.0,<=4.0.0
pywhispercpp>=1.4.1,<1.6.0
speechbrain>=1.0.0,<=1.1.0
