🌍 A minimalist repository for training video world models based on diffusion-forcing. • 20 items • Updated • 7
NanoWM-B/2 · RT-1 · Ablation: pred_name = x
One of three checkpoints from the pred_target ablation on RT-1 fractal (x-prediction arm). Each arm runs in its native schedule environment — cosine + ZTSNR for v and x, linear + no-ZTSNR for epsilon — so the comparison isolates the prediction target rather than handicapping any one of them.
Run identity
- wandb: https://wandb.ai/better_guidance/nano-world-model-ablation/runs/cn20uh5d
- launcher:
src/scripts/ablation/pred_x.sh - collection: https://huggingface.co/collections/knightnemo/nano-world-model
Training setup
| Key | Value |
|---|---|
| Architecture | NanoWM-B/2 (12 layers, d=768, patch=2, 158.6M params) |
| Dataset | RT-1 fractal (lerobot/fractal20220817_data) |
| Frames × resolution | 4 × 256² → 4 × 32² latents (SD-VAE) |
| Context frames | 1 (sequential / self-forcing scheduling) |
| Action injection | additive (7-dim continuous) |
| Steps | 50,000 |
| Batch | 8/GPU × 8 × H20 = 64 effective |
| Optimizer | AdamW, lr 1e-4, wd 0.01, warmup 1000, grad clip 0.1 after 20k |
| Precision | bf16-mixed (params fp32), VAE fp32, torch.compile on |
| Seed | 3407 |
Diffusion setup
| Key | Value |
|---|---|
| pred_name | x |
| noise_schedule | squaredcos_cap_v2 (cosine) |
| zero_terminal_snr | true |
| timestep_sampling | logit_normal (SD3-style, μ=0, σ=1) |
| snr_gamma | 5.0 (Min-SNR loss weighting) |
| diffusion_steps | 1000 train · 250 DDIM sample |
| history_stabilization_level (inference) | 0.02 |
Loading
git clone git@github.com:knightnemo/nano-world-model.git
cd nano-world-model
huggingface-cli download knightnemo/nanowm-b2-rt1-abl-pred-x-50k --local-dir ./ckpt
import sys
from omegaconf import OmegaConf
from safetensors.torch import load_file
sys.path.insert(0, "src")
from models import get_models
cfg = OmegaConf.load("ckpt/config.yaml")
cfg.experiment.infra.compile = False
model = get_models(cfg).eval()
state_dict = load_file("ckpt/model.safetensors")
model.load_state_dict(state_dict, strict=True) # 0 missing / 0 unexpected
- Downloads last month
- 1
Safetensors
Model size
0.2B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support