atonamy/wan-alpha 🔢📝❓✓ → 🖼️

▶️ 63 runs 📅 Oct 2025 ⚙️ Cog 0.16.8 🔗 GitHub

alpha-channel text-to-video transparent-background transparent-backgrounds

About

Wan-Alpha text-to-video generation with alpha channel, optimized for Replicate deployment. WebM/WebP output with transparency support.

Example Output

Prompt:

"Medium shot. A little girl holds a bubble wand and blows out colorful bubbles that float and pop in the air. The background of this video is transparent. Realistic style."

Output

Performance Metrics

110.68s Prediction Time

444.07s Total Time

All Input Parameters

{
  "fps": 16,
  "seed": -1,
  "prompt": "Medium shot. A little girl holds a bubble wand and blows out colorful bubbles that float and pop in the air. The background of this video is transparent. Realistic style.",
  "solver": "unipc",
  "num_frames": 81,
  "resolution": "832*480",
  "sample_steps": 4,
  "output_format": "webm",
  "guidance_scale": 1,
  "output_quality": 85,
  "negative_prompt": "",
  "remove_last_frame": false,
  "remove_first_frame": false
}

Input Parameters

fps Type: integerDefault: 16Range: 5 - 30: Output framerate (5-30 fps).
seed Type: integerDefault: -1: Random seed (-1 for random).
prompt (required) Type: string: Text prompt for video generation.
solver Default: unipc: Sampling solver.
num_frames Type: integerDefault: 81Range: 5 - 121: Desired number of frames (5-121). Internally adjusted to valid 4n+1 values (min 81, max 121).
resolution Default: 832*480: Video resolution formatted as width*height.
sample_steps Type: integerDefault: 4: Number of diffusion sampling steps.
output_format Default: webm: Output video format with transparency support.
guidance_scale Type: numberDefault: 1: Classifier-free guidance scale.
output_quality Type: integerDefault: 85Range: 1 - 100: Output quality (1-100, higher is better).
negative_prompt Type: stringDefault:: Optional negative prompt. If blank, uses the model default.
remove_last_frame Type: booleanDefault: false: Remove the last frame from the final output video.
remove_first_frame Type: booleanDefault: false: Remove the first frame from the final output video.

Output Schema

Output

Type: string • Format: uri

Example Execution Logs

/src/wan/text2video_dora_lightx2v.py:220: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(dtype=self.param_dtype), torch.no_grad(), no_sync():
  0%|          | 0/4 [00:00<?, ?it/s]/src/wan/modules/model_dora_lightx2v.py:618: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(dtype=torch.float32):
/src/wan/modules/model_dora_lightx2v.py:299: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(dtype=torch.float32):
/src/wan/modules/model_dora_lightx2v.py:307: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(dtype=torch.float32):
/src/wan/modules/model_dora_lightx2v.py:314: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(dtype=torch.float32):
/src/wan/modules/model_dora_lightx2v.py:346: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(dtype=torch.float32):
 25%|██▌       | 1/4 [00:09<00:27,  9.20s/it]
 50%|█████     | 2/4 [00:18<00:18,  9.15s/it]
 75%|███████▌  | 3/4 [00:27<00:09,  9.16s/it]
100%|██████████| 4/4 [00:36<00:00,  9.15s/it]
100%|██████████| 4/4 [00:36<00:00,  9.16s/it]
/src/wan/modules/vae.py:735: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(dtype=self.dtype):

Version Details

Version ID: 9e5a38f8658c6f9d18d1876ef8b65982a9962c705f08600d23b12dc6e2992d23
Version Created: October 7, 2025

Run on Replicate →