git clone https://github.com/meituan-longcat/LongCat-Videocd LongCat-Video
# create conda environmentconda create -n longcat-video python=3.10conda activate longcat-video# install torch (configure according to your CUDA version)pip install torch==2.6.0+cu124 torchvision==0.21.0+cu124 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu124# install flash-attn-2pip install ninja pip install psutil pip install packaging pip install flash_attn==2.7.4.post1# install other requirementspip install -r requirements.txt
# Single-GPU inferencetorchrun run_demo_text_to_video.py --checkpoint_dir=./weights/LongCat-Video --enable_compile# Multi-GPU inferencetorchrun --nproc_per_node=2 run_demo_text_to_video.py --context_parallel_size=2 --checkpoint_dir=./weights/LongCat-Video --enable_compile
# Single-GPU inferencetorchrun run_demo_image_to_video.py --checkpoint_dir=./weights/LongCat-Video --enable_compile# Multi-GPU inferencetorchrun --nproc_per_node=2 run_demo_image_to_video.py --context_parallel_size=2 --checkpoint_dir=./weights/LongCat-Video --enable_compile
# Single-GPU inferencetorchrun run_demo_video_continuation.py --checkpoint_dir=./weights/LongCat-Video --enable_compile# Multi-GPU inferencetorchrun --nproc_per_node=2 run_demo_video_continuation.py --context_parallel_size=2 --checkpoint_dir=./weights/LongCat-Video --enable_compile
Single-GPU inferencetorchrun run_demo_long_video.py --checkpoint_dir=./weights/LongCat-Video --enable_compile# Multi-GPU inferencetorchrun --nproc_per_node=2 run_demo_long_video.py --context_parallel_size=2 --checkpoint_dir=./weights/LongCat-Video --enable_compile
美团的AI技术发展路线,LongCat-Flash-Chat、LongCat-Thinking、再到LongCat-Video,再到最新 LongCat-Flash-Omni.
这条技术演进路线,从速度最快的对话交互,到专业深度的多模态理解,再到如今全面融合各类场景的音视频智能。
可以看到, 它家LongCat系每一次升级,都是AI向更强“多模态智能”迈进的一步。AI进化越来越快了,这也挺好,我们的生活也更方便了。
以上全文3729字,16张图和视频。若可以,给我个三连击:点赞、转发和在看。若可以再给我加个⭐️,谢谢你看我的文章,我们下篇再见。

