EXOKERN/skill-forge-peginsert-v0.1.1

EXOKERN Skill v0.1.1 - Robust Peg Insertion Under Domain Randomization

skill-forge-peginsert-v0.1.1 is the domain-randomized reference model release in the EXOKERN catalog. It is trained on EXOKERN ContactBench v0.1.1 and ships the same paired comparison structure as v0:

• full_ft_best_model.pt: primary checkpoint with 22D observations, including force/torque input
• no_ft_best_model.pt: ablation checkpoint with the same architecture and 16D state-only observations

This release should be read as a robustness benchmark first. Both policies remain successful under severe domain randomization, and the repo is valuable precisely because it makes the mixed result on force reduction explicit.

Quick Facts

Item	Value
Task	Peg insertion in simulation under domain randomization
Dataset	EXOKERN/contactbench-forge-peginsert-v0.1.1
Simulator	NVIDIA Isaac Lab (Isaac Sim 4.5)
Robot	Franka FR3
Architecture	TemporalUNet1D diffusion policy
Parameters	71.3M
Observation horizon	10 frames
Prediction / execution horizon	16 / 8 actions
Seeds evaluated	42, 123, 7
Total rollouts reported	600

Benchmark Summary

The Hub metadata for this repo tracks the primary full_ft checkpoint. The full repo includes the paired no_ft ablation for comparison.

Checkpoint	Success Rate	Avg Contact Force (N)	Peak Contact Force (N)	Avg Episode Time (s)
full_ft	100.0	3.67 +/- 0.45	10.63	25.63
no_ft	100.0	3.37 +/- 0.06	10.33	25.73

Figure: multi-seed benchmark summary built from the published eval_seed42/123/7.json artifacts.

Per-seed results:

Seed	Condition	Success Rate	Avg Force (N)	Peak Force (N)	Avg Time (s)
42	full_ft	100.0	3.24	10.44	25.61
42	no_ft	100.0	3.38	10.38	25.73
123	full_ft	100.0	4.12	10.57	25.74
123	no_ft	100.0	3.34	10.32	25.79
7	full_ft	100.0	3.69	10.93	25.54
7	no_ft	100.0	3.37	10.31	25.68

Interpretation:

• This release demonstrates robust task completion under a much harder collection regime than v0.
• On this particular peg-in-hole setup, domain randomization largely closed the force gap between full_ft and no_ft.
• That does not prove force/torque is unnecessary in general. It shows that this release is best used as a robust benchmark and an honest reference point for harder future tasks.

What Changed Compared To v0

Topic	v0	v0.1.1
Dataset regime	Mostly fixed conditions	Multi-layer domain randomization
Dataset size	2,221 episodes / 330,929 frames	5,000 episodes / 745,000 frames
Robot	Franka Emika Panda	Franka FR3
Force reduction takeaway	Clear F/T advantage	Inconclusive on this task
Best use	Clean baseline	Robustness benchmark

Architecture

This release uses the same 1D Temporal U-Net diffusion policy family as v0.

Component	Value
Action dimension	7
Observation dimensions	22 (full_ft) / 16 (no_ft)
Diffusion training steps	100
DDIM inference steps	16
Base channels	256
Channel multipliers	(1, 2, 4)
Normalization	Min-max to [-1, 1]

Repository Contents

File	Description
full_ft_best_model.pt	Best checkpoint with force/torque input
no_ft_best_model.pt	Ablation checkpoint without force/torque input
inference.py	Self-contained inference helper and model definition
config.yaml	Training, dataset, and environment configuration
eval_seed42.json	Seed 42 evaluation artifact
eval_seed123.json	Seed 123 evaluation artifact
eval_seed7.json	Seed 7 evaluation artifact
training_curve_full_ft_seed42.png	Training curve for full_ft, seed 42
training_curve_full_ft_seed123.png	Training curve for full_ft, seed 123
training_curve_full_ft_seed7.png	Training curve for full_ft, seed 7
training_curve_no_ft_seed42.png	Training curve for no_ft, seed 42
training_curve_no_ft_seed123.png	Training curve for no_ft, seed 123
training_curve_no_ft_seed7.png	Training curve for no_ft, seed 7

Usage

Reproduce evaluation with exokern-eval

pip install exokern-eval

wget https://huggingface.co/EXOKERN/skill-forge-peginsert-v0.1.1/resolve/main/full_ft_best_model.pt

exokern-eval \
  --policy full_ft_best_model.pt \
  --env Isaac-Forge-PegInsert-Direct-v0 \
  --episodes 100

Load the repo helper locally

import os
import sys

from huggingface_hub import snapshot_download

repo_dir = snapshot_download(
    repo_id="EXOKERN/skill-forge-peginsert-v0.1.1",
    allow_patterns=["*.pt", "inference.py"],
)
sys.path.insert(0, repo_dir)

from inference import DiffusionPolicyInference

policy = DiffusionPolicyInference(
    os.path.join(repo_dir, "full_ft_best_model.pt"),
    device="cpu",
)

policy.add_observation([0.0] * 22)
actions = policy.get_actions()
print(len(actions))

Training And Evaluation Setup

Item	Value
Train / val split	85% / 15% by episode
Epochs	300
Batch size	256
Optimizer	AdamW, lr=1e-4, weight_decay=1e-4
LR schedule	Cosine annealing to 1e-6
EMA decay	0.995
Physics rate	120 Hz
Control rate	15 Hz
Domain randomization	Enabled in the training dataset

Related Work

• FORGE: Force-Guided Exploration for Robust Contact-Rich Manipulation under Uncertainty
• Diffusion Policy: Visuomotor Policy Learning via Action Diffusion
• Factory: Fast Contact for Robotic Assembly

Citation

@misc{exokern_skill_peginsert_v011_2026,
  title        = {EXOKERN Skill v0.1.1: Robust Peg Insertion Under Domain Randomization},
  author       = {{EXOKERN}},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/EXOKERN/skill-forge-peginsert-v0.1.1}},
  note         = {Paired full_ft and no_ft diffusion-policy checkpoints}
}

Security Note

The checkpoints in this repo are PyTorch pickles. Load them only in a trusted or isolated environment after reviewing the repository contents.

Limitations

• Simulation only. This release does not claim real-robot readiness.
• Reported robustness is specific to the peg-in-hole task and the randomization ranges documented in the paired dataset card.
• The ablation result is mixed: use this repo to study robustness, not to overclaim a universal force/torque effect.
• The repo exposes paired checkpoints for research comparison; the intended production-style reference in this repo is full_ft_best_model.pt.

Related Resources

• Dataset: EXOKERN/contactbench-forge-peginsert-v0.1.1
• Baseline predecessor: EXOKERN/skill-forge-peginsert-v0
• Evaluation CLI: github.com/Exokern/exokern_eval
• Organization page: huggingface.co/EXOKERN