PKU-YuanGroup / Video-LLaVA

【EMNLP 2024🔥】Video-LLaVA: Learning United Visual Representation by Alignment Before Projection
https://arxiv.org/pdf/2311.10122.pdf
Apache License 2.0
2.88k stars 207 forks source link

extremely slow with transformers #167

Open RaulKite opened 3 months ago

RaulKite commented 3 months ago

I have this warning using Video-llava and it spends about 4 minutes to process a video. In previous versions it took about 30 seconds.

I have this warning that maybe is the reason:

UserWarning: Creating a tensor from a list of numpy.ndarrays is extremely slow. Please consider converting the list to a single numpy.ndarray with numpy.array() before converting to a tensor. (Triggered internally at ../torch/csrc/utils/tensor_new.cpp:245.) return torch.tensor(value)

Do you know how solve it?

Thanks