forresti

6 ms·

1. 1.▲
   LongVU – New Video LLM from Meta(huggingface.co)

   2 points by forresti 2y ago|1 comments
2. 2.▲
   by forresti 2y ago

   VideoLLM from Meta! LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding Paper: https://huggingface.co/papers/2410.17434 Code: https://github.com/Vision-CAIR/LongVU P
3. 3.▲
   by forresti 8y ago

   Heh, I think the part of the point of PlaidML is to avoid the speed/efficiency limitations of TensorFlow in deployment/production.