Downloads · 30 days
0
grpo_stable_reasoning_nolow_0722_075_1200steps_lora64_withlow_0723
PhoenixHu/grpo_stable_reasoning_nolow_0722_075_1200steps_lora64_withlow_0723
grpo_stable_reasoning_nolow_0722_075_1200steps_lora64_withlow_0723 is a reinforcement learning model from PhoenixHu. Use it for the reinforcement learning task on the model card, and read the license before you ship it in a product.
GRPO RL checkpoints for a SimLingo-style driving VLA (InternVL2-1B + continuous speed-waypoint head).
Access
Public
Updated Jul 24, 2026
Repo size
6.1 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt6.1 GB · 100%
At a glance
- Task
- Reinforcement Learning
- Access
- Public
- Created
- Jul 24, 2026
- Updated
- Jul 24, 2026
- SHA
- fa8e7b6c
Model card
From the Hugging Face model README
grpo_stable_reasoning_nolow_0722_075_1200steps_lora64_withlow_0723
GRPO RL checkpoints for a SimLingo-style driving VLA (InternVL2-1B + continuous speed-waypoint head).
Variant: nolow pool + low-speed samples re-included (withlow), ratio 0.75, 1200-step schedule, LLM LoRA rank 64.
- Each
checkpoints/stepNNN.ckpt/pytorch_model.pt= consolidated fp32state_dict. .hydra/config.yaml= the training config — required to rebuild the model architecture for eval.
| file | step |
|---|---|
checkpoints/step500.ckpt/pytorch_model.pt | 500 |
checkpoints/step600.ckpt/pytorch_model.pt | 600 |
checkpoints/step1000.ckpt/pytorch_model.pt | 1000 |
import torch
sd = torch.load("checkpoints/step1000.ckpt/pytorch_model.pt", map_location="cpu")
Only steps 500 / 600 / 1000 are published.
rollouts.jsonl(RL training telemetry) is not included.
- Task
- Reinforcement Learning
- Created
- Jul 24, 2026
- Updated
- Jul 24, 2026