Can the 4090 Run It?
I ran a sample
python generate.py --task i2v-14B --size 1280*720 --ckpt_dir ./Wan2.1-I2V-14B-720P --image examples/i2v_input.JPG --prompt "Summer beach vacation style, a white cat wearing sunglasses sits on a surfboard. The fluffy-furred feline gazes directly at the camera with a relaxed expression. Blurred beach scenery forms the background featuring crystal-clear waters, distant green hills, and a blue sky dotted with white clouds. The cat assumes a naturally relaxed posture, as if savoring the sea breeze and warm sunlight. A close-up shot highlights the feline's intricate details and the refreshing atmosphere of the seaside."
but the progress bar didn't move at all.
try 480p version
From github:
DiffSynth-Studio provides more support for Wan, including video-to-video, FP8 quantization, VRAM optimization, LoRA training, and more. Please refer to their examples.
is working for me Thanks
Will we get FP4 for Blackwell soon (or not soon)? That should be faster then, right?
FYI
I tried this:
python generate.py --task t2v-1.3B --size 832*480 --ckpt_dir C:\Users\userx\Python_Scripts\Wan2.1-T2V-1.3B --prompt "A white cat wearing sunglasses on a surfboard at the beach"
and it took 58 minutes on a 4090
I am also using a 13th Gen Intel(R) Core(TM) i9-13900KF, 3000 Mhz, 24 Core(s), 32 Logical Processor(s) cpu
Note, I set it to 5 steps. While 28/64GB was used and 8/24GB GPU was used, the cpu usage had two jobs running at about 5% for at least 10 minutes before I wandered off to do something else. There was no progress bar at this point . Once the progress bar started moving it looks like it moved fast.
Question: What is it doing during this time???? Is there a way to get more logging? What is it doing this time?
[2025-05-04 12:06:03,090] INFO: Input prompt: A white cat wearing sunglasses on a surfboard at the beach
[2025-05-04 12:06:03,090] INFO: Creating WanT2V pipeline.
[2025-05-04 12:06:59,441] INFO: loading C:\Users\userX\Python_Scripts\Wan2.1-T2V-1.3B\models_t5_umt5-xxl-enc-bf16.pth
[2025-05-04 12:07:05,652] INFO: loading C:\Users\userX\Python_Scripts\Wan2.1-T2V-1.3B\Wan2.1_VAE.pth
[2025-05-04 12:07:06,286] INFO: Creating WanModel from C:\Users\userX\Python_Scripts\Wan2.1-T2V-1.3B
[2025-05-04 12:07:10,393] INFO: Generating video ...
100%|████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:25<00:00, 5.17s/it]
[2025-05-04 13:05:32,369] INFO: Saving generated video to t2v-1.3B_832x480_1_1_A_white_cat_wearing_sunglasses_on_a_surfboard_at_t_20250504_130532.mp4
[2025-05-04 13:05:32,990] INFO: Finished....
Note sorted by CPU and GPU highest to lowest. only 5% cpu is used and 37% gpu, 44% of memory used.
Removing the --offload_model True --t5_cpu - much faster...
[2025-05-04 17:33:29,195] INFO: Creating WanT2V pipeline.
[2025-05-04 17:34:23,852] INFO: loading C:\Users\vivek\Python_Scripts\Wan2.1-T2V-1.3B\models_t5_umt5-xxl-enc-bf16.pth
[2025-05-04 17:34:30,424] INFO: loading C:\Users\vivek\Python_Scripts\Wan2.1-T2V-1.3B\Wan2.1_VAE.pth
[2025-05-04 17:34:31,033] INFO: Creating WanModel from C:\Users\vivek\Python_Scripts\Wan2.1-T2V-1.3B
[2025-05-04 17:34:35,172] INFO: Generating video ...
100%|██████████████████████████████████████████████████████████████████████████████████| 50/50 [04:20<00:00, 5.21s/it]
[2025-05-04 17:39:11,471] INFO: Saving generated video to t2v-1.3B_832x480_1_1_Sexy_nude_20_year_old_muscular_woman_laying_on_top_20250504_173911.mp4
[2025-05-04 17:39:11,856] INFO: Finished.
Thanks,
V