Downloads · 30 days
0
orbital-thruster-grpo-test
pixxel-phantom/orbital-thruster-grpo-test
orbital-thruster-grpo-test is a machine learning model from pixxel-phantom. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Access
Public
Updated May 15, 2026
Repo size
468 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt296 MB · 45%
At a glance
- Access
- Public
- Created
- May 15, 2026
- Updated
- May 15, 2026
- SHA
- f5622a9a
Model card
From the Hugging Face model README
OrbitalThrusterEnv — GRPO LoRA adapter
Base model: Qwen/Qwen2.5-1.5B-Instruct
Source env: https://huggingface.co/spaces/pixxel-phantom/orbital-thruster-env
Trained via TRL GRPOTrainer + Unsloth on OpenEnv OrbitalThrusterEnv flagship task mission_ops_long_horizon.
5 independent reward funcs (format, env-step, mode-match, anti-spam, fuel-discipline) for anti-reward-hacking.
Artifacts
trainer_output/qwen_grpo/— final LoRA adaptertrainer_output/qwen_sft/— SFT warm-start adapteroutputs/training/grpo_metrics.png— reward + loss curvesoutputs/eval_trained/trained_vs_baseline.png— trained vs baselines on 4 tasks
- Created
- May 15, 2026
- Updated
- May 15, 2026