Hugging Face
Models
Datasets
Spaces
Buckets
new
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
pottokao
/
H3-RotNVFP4-ComfyUI-Loader
Like
2
MiniMax H3
comfyui
nunchaku
nvfp4
fp4
video
blackwell
License:
apache-2.0
Model card
Files
Files and versions
xet
Community
1
Deploy
Copy to bucket
new
main
H3-RotNVFP4-ComfyUI-Loader
57.1 kB
Ctrl+K
Ctrl+K
1 contributor
History:
8 commits
pottokao
Add workflow setup notes (files, node deps, sampler rationale)
74c7c6d
verified
29 days ago
.gitattributes
Safe
1.52 kB
initial commit
about 2 months ago
METHODOLOGY.md
Safe
8.18 kB
METHODOLOGY: measured max clip length per config (dense 310f/124f; VSA 68f/OOM)
about 2 months ago
README.md
Safe
2.83 kB
H3RotNVFP4Loader node + methodology (device-bug fixed)
about 2 months ago
WORKFLOW.md
Safe
1.7 kB
Add workflow setup notes (files, node deps, sampler rationale)
29 days ago
__init__.py
Safe
131 Bytes
H3RotNVFP4Loader node + methodology (device-bug fixed)
about 2 months ago
nodes.py
Safe
38.9 kB
loader: build fp4 blocks directly (no ~72GB bf16 load transient); --gpu-only loads in 4s @10.57GB peak, DiT resident 12.777GB zero offload; output bit-identical (SSIM 1.0). Device-bug fix retained.
about 2 months ago
requirements.txt
Safe
835 Bytes
H3RotNVFP4Loader node + methodology (device-bug fixed)
about 2 months ago
workflow.json
Safe
2.91 kB
Add the workflow.json the README promised (h3_fasth3_T1.safetensors, 4 steps, euler)
29 days ago