Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# CartPole-v1_ppo
|
| 2 |
+
|
| 3 |
+
Run: `test-run` — Env: `CartPole-v1` — Algo: `ppo`
|
| 4 |
+
|
| 5 |
+
This repository contains artifacts from a Gymnasium Solver training run.
|
| 6 |
+
|
| 7 |
+
## Contents
|
| 8 |
+
- Config: `artifacts/configs/config.json`
|
| 9 |
+
- Checkpoints: `artifacts/checkpoints/*.ckpt`
|
| 10 |
+
- Logs: `artifacts/logs/*.log`
|
| 11 |
+
|
| 12 |
+
## Config (excerpt)
|
| 13 |
+
```json
|
| 14 |
+
{
|
| 15 |
+
"env_id": "CartPole-v1",
|
| 16 |
+
"algo_id": "ppo",
|
| 17 |
+
"n_steps": 8,
|
| 18 |
+
"batch_size": 8,
|
| 19 |
+
"n_epochs": 1,
|
| 20 |
+
"n_timesteps": 8,
|
| 21 |
+
"seed": 42,
|
| 22 |
+
"n_envs": 1,
|
| 23 |
+
"obs_type": "rgb",
|
| 24 |
+
"policy": "MlpPolicy",
|
| 25 |
+
"learning_rate": 0.001,
|
| 26 |
+
"gamma": 0.98,
|
| 27 |
+
"gae_lambda": 0.8,
|
| 28 |
+
"ent_coef": 0.0,
|
| 29 |
+
"vf_coef": 0.5,
|
| 30 |
+
"clip_range": 0.2,
|
| 31 |
+
"normalize_advantages": "batch"
|
| 32 |
+
}
|
| 33 |
+
```
|