Downloads · 30 days
0
nano-anchor-forcing-bench-100k-checkpoints
shuaishuaicdp/nano-anchor-forcing-bench-100k-checkpoints
nano-anchor-forcing-bench-100k-checkpoints is a machine learning model from shuaishuaicdp. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for pytorch. The card lists the license as other.
Base initialization and the latest saved checkpoints for the custom 100k prompt-transition experiments in Dongping-Chen/Nano-Anchor-Forcing.
Access
Public
Updated Oct 25, 2026
Repo size
81.3 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt89.7 GB · 100%
At a glance
- Library
- pytorch
- License
- other
- Access
- Public
- Created
- Jul 18, 2026
- Updated
- Oct 25, 2026
- SHA
- 7beaa283
Model card
From the Hugging Face model README
Nano Anchor Forcing Bench 100k Checkpoints
Base initialization and the latest saved checkpoints for the custom 100k
prompt-transition experiments in
Dongping-Chen/Nano-Anchor-Forcing.
Files
| File | State | SHA256 |
|---|---|---|
ckpt/base.pt | generator initialization | 10a2aa8fcf89c77d9033f4c117405412a690e289625766619d293f0c5a208ee7 |
logs/bench_100k_vae_r256/checkpoint_model_003125/model.pt | curriculum/fixed-implicit, complete step 3125 | 77f3fc905c9237c2752a634786ccec240f73d42c86e3b5241b5f7f8adf5a6bf3 |
logs/bench_100k_vae_r256_random/checkpoint_model_002100/model.pt | curriculum/random-routing, resume step 2100 | 2cf8338ba8655a48e28e6ee8d658718ccbdf64b1fc52332ee75452c1d913c305 |
logs/bench_100k_vae_r256_shuffle_random/checkpoint_model_001000/model.pt | seed-42-shuffle/random-routing, resume step 1000 | 55505388d93eee1b890b69caf1f2a704733cc658d8768597838b7788e8370ada |
Each training checkpoint contains generator and critic LoRA weights, both
optimizer states, and the saved step. Downloading with --local-dir . restores
the directory structure expected by auto_resume: True:
hf download shuaishuaicdp/nano-anchor-forcing-bench-100k-checkpoints \
ckpt/base.pt \
logs/bench_100k_vae_r256/checkpoint_model_003125/model.pt \
logs/bench_100k_vae_r256_random/checkpoint_model_002100/model.pt \
logs/bench_100k_vae_r256_shuffle_random/checkpoint_model_001000/model.pt \
--local-dir .
The Wan 1.3B generator and 14B teacher weights are not duplicated here. The
exact upstream revisions and complete restore instructions are documented in
BENCH_100K_MIGRATION.md in the GitHub repository.
Use is subject to the licenses and terms of the upstream Wan2.1 and Anchor Forcing projects.
- Library
- pytorch
- License
- other
- Created
- Jul 18, 2026
- Updated
- Oct 25, 2026