Hi Team, Thanks for your work but I am not able to run the 14B model using the step given.
I have 8 gpu with 24GB ram.
Cmd used:
torchrun --standalone --nproc_per_node=8 examples/wanvideo/acceleration/unified_sequence_parallel.py
Code used:
import torch
from PIL import Image
from diffsynth import save_video, VideoData
from diffsynth.pipelines.wan_video_new import WanVideoPipeline, ModelConfig
import torch.distributed as dist
pipe = WanVideoPipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
use_usp=True,
skip_download = True,
redirect_common_files = False,
model_configs=[
ModelConfig(path = [
"/data/akshay/wan/Wan2.1-T2V-14B/diffusion_pytorch_model-00001-of-00006.safetensors",
"/data/akshay/wan/Wan2.1-T2V-14B/diffusion_pytorch_model-00002-of-00006.safetensors",
"/data/akshay/wan/Wan2.1-T2V-14B/diffusion_pytorch_model-00003-of-00006.safetensors",
"/data/akshay/wan/Wan2.1-T2V-14B/diffusion_pytorch_model-00004-of-00006.safetensors",
"/data/akshay/wan/Wan2.1-T2V-14B/diffusion_pytorch_model-00005-of-00006.safetensors",
"/data/akshay/wan/Wan2.1-T2V-14B/diffusion_pytorch_model-00006-of-00006.safetensors",
], offload_device="cpu"),
ModelConfig(path = "/data/akshay/wan/Wan2.1-T2V-14B/models_t5_umt5-xxl-enc-bf16.pth", offload_device="cpu"),
ModelConfig(path = "/data/akshay/wan/Wan2.1-T2V-14B/Wan2.1_VAE.pth", offload_device="cpu"),
],
)
print('loading complete')
pipe.enable_vram_management()
video = pipe(
prompt="一名宇航员身穿太空服，面朝镜头骑着一匹机械马在火星表面驰骋。红色的荒凉地表延伸至远方，点缀着巨大的陨石坑和奇特的岩石结构。机械马的步伐稳健，扬起微弱的尘埃，展现出未来科技与原始探索的完美结合。宇航员手持操控装置，目光坚定，仿佛正在开辟人类的新疆域。背景是深邃的宇宙和蔚蓝的地球，画面既科幻又充满希望，让人不禁畅想未来的星际生活。",
negative_prompt="色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走",
seed=0, tiled=True,
)
if dist.get_rank() == 0:
save_video(video, "video1.mp4", fps=15, quality=5)
Error:
Traceback (most recent call last):
[rank5]: File "/data/akshay/wan/DiffSynth-Studio/examples/wanvideo/acceleration/unified_sequence_parallel.py", line 51, in <module>
[rank5]: pipe.enable_vram_management()
[rank5]: File "/data/akshay/wan/DiffSynth-Studio/diffsynth/pipelines/wan_video_new.py", line 278, in enable_vram_management
[rank5]: vram_limit = self.get_vram()
[rank5]: ^^^^^^^^^^^^^^^
[rank5]: File "/data/akshay/wan/DiffSynth-Studio/diffsynth/pipelines/wan_video_new.py", line 147, in get_vram
[rank5]: return torch.cuda.mem_get_info(self.device)[1] / (1024 ** 3)
[rank5]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[rank5]: File "/opt/conda/lib/python3.11/site-packages/torch/cuda/memory.py", line 684, in mem_get_info
[rank5]: device = _get_device_index(device)
[rank5]: ^^^^^^^^^^^^^^^^^^^^^^^^^
[rank5]: File "/opt/conda/lib/python3.11/site-packages/torch/cuda/_utils.py", line 38, in _get_device_index
[rank5]: return _torch_get_device_index(device, optional, allow_cpu)
[rank5]: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[rank5]: File "/opt/conda/lib/python3.11/site-packages/torch/_utils.py", line 803, in _get_device_index
[rank5]: raise ValueError(
[rank5]: ValueError: Expected a torch.device with a specified index or an integer, but got:cuda
Can you please help me with this?
Thanks
Hi Team, Thanks for your work but I am not able to run the 14B model using the step given.
I have 8 gpu with 24GB ram.
Cmd used:
torchrun --standalone --nproc_per_node=8 examples/wanvideo/acceleration/unified_sequence_parallel.pyCode used:
Error:
Can you please help me with this?
Thanks