第六十篇-ComfyUI+V100-32G+运行Wan2.2-图生视频

环境

复制代码

系统：CentOS-7
CPU : E5-2680V4 14核28线程
内存：DDR4 2133 32G * 2
显卡：Tesla V100-32G【PG503】 (水冷)
驱动: 535
CUDA: 12.2
ComfyUI version: 0.4.0
ComfyUI frontend version: 1.34.8

系统软件信息

复制代码

系统信息
OS
linux
Python Version
3.12.12 | packaged by Anaconda, Inc. | (main, Oct 21 2025, 20:16:04) [GCC 11.2.0]
Embedded Python
false
Pytorch Version
2.9.1+cu128
Arguments
main.py --listen --port 8188 --cuda-malloc --lowvram
RAM Total
62.68 GB
RAM Free
60.25 GB

启动

bash 复制代码

python main.py --listen --port 8188 --cuda-malloc --lowvram

模版

Wan 2.2 14B图生视频

bash 复制代码

https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors
https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/loras/wan2.2_i2v_lightx2v_4steps_lora_v1_high_noise.safetensors
https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/loras/wan2.2_i2v_lightx2v_4steps_lora_v1_low_noise.safetensors
https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors
https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/diffusion_models/wan2.2_i2v_high_noise_14B_fp8_scaled.safetensors
https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/resolve/main/split_files/diffusion_models/wan2.2_i2v_low_noise_14B_fp8_scaled.safetensors

配置参数

可以使用迅雷单独下载文件

Setp2 选择文件上传，上传一张图片

点击运行

运行数据

视频尺寸调整 720*1280

GPU

bash 复制代码

Mon Dec 15 23:45:38 2025
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.129.03             Driver Version: 535.129.03   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  Tesla PG503-216                On  | 00000000:04:00.0 Off |                    0 |
| N/A   38C    P0             234W / 250W |  23024MiB / 32768MiB |    100%      Default ||                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+

时间

bash 复制代码

loaded completely; 95367431640625005117571072.00 MB usable, 6419.49 MB loaded, full load: True
CLIP/text encoder model load device: cpu, offload device: cpu, current: cpu, dtype: torch.float16
Requested to load WanVAE
loaded completely; 21176.38 MB usable, 242.03 MB loaded, full load: True
Found quantization metadata version 1
Detected mixed precision quantization
Using mixed precision operations
model weight dtype torch.float16, manual cast: torch.float16
model_type FLOW
unet unexpected: ['scaled_fp8']
Requested to load WAN21
loaded completely; 17787.26 MB usable, 13631.43 MB loaded, full load: True
100%|███████████████████████████████████████████████████████████████████████| 2/2 [05:55<00:00, 177.72s/it]
Found quantization metadata version 1
Detected mixed precision quantization
Using mixed precision operations
model weight dtype torch.float16, manual cast: torch.float16
model_type FLOW
unet unexpected: ['scaled_fp8']
Requested to load WAN21
loaded completely; 17787.26 MB usable, 13631.43 MB loaded, full load: True
100%|███████████████████████████████████████████████████████████████████████| 2/2 [05:55<00:00, 177.85s/it]
Requested to load WanVAE
loaded completely; 5784.79 MB usable, 242.03 MB loaded, full load: True
Prompt executed in 00:16:35

效果

动态的5秒视频

总结

复制代码

V100-32G Wan2.2 生成1280*720的5秒视频用时16分钟，最大24G显存
效果还是挺好，生成时间长了点其他还好。