llama-cpp-python>=0.2.86
gradio>=4.31
huggingface_hub>=0.24
tokenizers>=0.19 
tqdm>=4.66
g4f[all]