Vchitect / Latte

Latte: Latent Diffusion Transformer for Video Generation.
Apache License 2.0
1.45k stars 147 forks source link

Does Latte support multiple GPUs #29

Closed afezeriaWrnbbmm closed 4 months ago

afezeriaWrnbbmm commented 4 months ago

Does Latte support multiple GPUs, and if so, how should it be set up?

maxin-cn commented 4 months ago

Does Latte support multiple GPUs, and if so, how should it be set up?

Hi, I'm not sure what you mean by multi-GPU? Do you mean multi-GPU training or multi-GPU testing?

afezeriaWrnbbmm commented 4 months ago

I followed the steps in the readme to generate videos from text, but only one of my GPUs was used. (base) root@nvidia3090:/home/nvidia3090# nvidia-smi Sat Feb 24 07:13:27 2024 +---------------------------------------------------------------------------------------+ | NVIDIA-SMI 535.154.05 Driver Version: 535.154.05 CUDA Version: 12.2 | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 NVIDIA GeForce RTX 3090 Off | 00000000:04:00.0 Off | N/A | | 0% 24C P2 349W / 350W | 21275MiB / 24576MiB | 100% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 1 NVIDIA GeForce RTX 3090 Off | 00000000:06:00.0 Off | N/A | | 0% 5C P8 28W / 350W | 2788MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 2 NVIDIA GeForce RTX 3090 Off | 00000000:07:00.0 Off | N/A | | 0% 5C P8 24W / 350W | 2770MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 3 NVIDIA GeForce RTX 3090 Off | 00000000:0C:00.0 Off | N/A | | 0% 5C P8 20W / 370W | 2786MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 4 NVIDIA GeForce RTX 3090 Off | 00000000:0D:00.0 Off | N/A | | 0% 4C P8 20W / 350W | 2776MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 5 NVIDIA GeForce RTX 3090 Off | 00000000:0E:00.0 Off | N/A | | 0% 4C P8 24W / 420W | 2786MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+

+---------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=======================================================================================| | 0 N/A N/A 1210829 C python 2810MiB | | 0 N/A N/A 1355236 C python 18456MiB | | 1 N/A N/A 1210766 C python 2782MiB | | 2 N/A N/A 1210696 C python 2764MiB | | 3 N/A N/A 1210629 C python 2780MiB | | 4 N/A N/A 1210479 C python 2770MiB | | 5 N/A N/A 1210343 C python 2780MiB | +---------------------------------------------------------------------------------------+

maxin-cn commented 4 months ago

I followed the steps in the readme to generate videos from text, but only one of my GPUs was used. (base) root@nvidia3090:/home/nvidia3090# nvidia-smi Sat Feb 24 07:13:27 2024 +---------------------------------------------------------------------------------------+ | NVIDIA-SMI 535.154.05 Driver Version: 535.154.05 CUDA Version: 12.2 | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 NVIDIA GeForce RTX 3090 Off | 00000000:04:00.0 Off | N/A | | 0% 24C P2 349W / 350W | 21275MiB / 24576MiB | 100% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 1 NVIDIA GeForce RTX 3090 Off | 00000000:06:00.0 Off | N/A | | 0% 5C P8 28W / 350W | 2788MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 2 NVIDIA GeForce RTX 3090 Off | 00000000:07:00.0 Off | N/A | | 0% 5C P8 24W / 350W | 2770MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 3 NVIDIA GeForce RTX 3090 Off | 00000000:0C:00.0 Off | N/A | | 0% 5C P8 20W / 370W | 2786MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 4 NVIDIA GeForce RTX 3090 Off | 00000000:0D:00.0 Off | N/A | | 0% 4C P8 20W / 350W | 2776MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 5 NVIDIA GeForce RTX 3090 Off | 00000000:0E:00.0 Off | N/A | | 0% 4C P8 24W / 420W | 2786MiB / 24576MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+

+---------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=======================================================================================| | 0 N/A N/A 1210829 C python 2810MiB | | 0 N/A N/A 1355236 C python 18456MiB | | 1 N/A N/A 1210766 C python 2782MiB | | 2 N/A N/A 1210696 C python 2764MiB | | 3 N/A N/A 1210629 C python 2780MiB | | 4 N/A N/A 1210479 C python 2770MiB | | 5 N/A N/A 1210343 C python 2780MiB | +---------------------------------------------------------------------------------------+

The code that generates videos from texts does not currently support the multi-GPU parallel mode. If you want to generate video on multiple Gpus, please refer to sample_ddp.py to modify your code.

afezeriaWrnbbmm commented 4 months ago

Thanks for your answer