diff options
Diffstat (limited to 'train_colab.ipynb')
| -rw-r--r-- | train_colab.ipynb | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/train_colab.ipynb b/train_colab.ipynb index 2a092d7..66b244e 100644 --- a/train_colab.ipynb +++ b/train_colab.ipynb @@ -65,6 +65,13 @@ "outputs": [] }, { + "cell_type": "code", + "source": "import matplotlib.pyplot as plt\n\ndf = pd.read_csv(\"blazing_ppo_2p_log.csv\")\n\nfig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 4))\n\nax1.plot(df[\"episode\"], df[\"vs_greedy_wr\"] * 100, \"o-\", color=\"tab:blue\", markersize=3)\nax1.axhline(50, color=\"gray\", linestyle=\"--\", alpha=0.5, label=\"random baseline\")\nax1.set_xlabel(\"Episode\")\nax1.set_ylabel(\"Win Rate vs Greedy (%)\")\nax1.set_title(\"Win Rate\")\nax1.legend()\nax1.grid(alpha=0.3)\n\nax2.plot(df[\"episode\"], df[\"avg_len\"], \"o-\", color=\"tab:orange\", markersize=3)\nax2.set_xlabel(\"Episode\")\nax2.set_ylabel(\"Avg Game Length (steps)\")\nax2.set_title(\"Game Length\")\nax2.grid(alpha=0.3)\n\nfig.tight_layout()\nplt.show()", + "metadata": {}, + "execution_count": null, + "outputs": [] + }, + { "cell_type": "markdown", "source": "## 3. Download model & logs", "metadata": {} |