Text-to-Video
Diffusers
Safetensors
MiniMaxH3ModularPipeline
video-generation
audio
fastvideo
nvfp4
pruned
Instructions to use FastVideo/FastVideo-FastH3-Trim-8-Step-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use FastVideo/FastVideo-FastH3-Trim-8-Step-NVFP4 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("FastVideo/FastVideo-FastH3-Trim-8-Step-NVFP4", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
Upload fastvideo_inference.json with huggingface_hub
Browse files- fastvideo_inference.json +4 -0
fastvideo_inference.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"schema_version": "fasth3-inference-contract-v1", "checkpoint_step": 300, "task": "t2av",
|
| 2 |
+
"dmd_denoising_steps": [999, 874, 749, 624, 500, 375, 250, 125], "num_inference_steps": 9,
|
| 3 |
+
"transformer_forwards": 8, "video_scheduler_shift": 10.0, "audio_scheduler_shift": 3.0, "guidance_scale": 1.0,
|
| 4 |
+
"attention_backend": "VIDEO_SPARSE_ATTN_H3", "vsa_sparsity": 0.8, "vsa_tile_size": 64}
|