Instructions to use wfen/Cosmos3-Nano-NVFP4-Blockwise with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use wfen/Cosmos3-Nano-NVFP4-Blockwise with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("wfen/Cosmos3-Nano-NVFP4-Blockwise", 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
Add Action-tab example conditioning inputs (FP8 parity)
Browse filesAdds the action-demo conditioning INPUTS the WebUI Action 'Run demo' reads (agibotworld first-frame image for policy/forward_dynamics; av clips for inverse_dynamics), which the NVFP4 assets/ lacked vs the FP8 checkpoint. Weights/config unchanged. AM-S5 follow-up.
assets/example_action_fd_agibotworld_first_frame.png
ADDED
|
Git LFS Details
|
assets/example_action_id_av_0_input.mp4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ff205f86ae169031c0e12c8c6bc8aabe24aeb44b84a16cba977669b29912f5fc
|
| 3 |
+
size 1125838
|
assets/example_action_id_av_1_input.mp4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:169e65cee76ef7c6366987a8e4c8e4ec6f803f5e624ba2ef39ba1f0c672925aa
|
| 3 |
+
size 1639407
|