{"spec_id":"line-loss-training","library":"seaborn","language":"python","code":"\"\"\" anyplot.ai\nline-loss-training: Training Loss Curve\nLibrary: seaborn 0.13.2 | Python 3.13.13\nQuality: 91/100 | Updated: 2026-05-14\n\"\"\"\n\nimport os\n\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\n\n\n# Theme tokens\nTHEME = os.getenv(\"ANYPLOT_THEME\", \"light\")\nPAGE_BG = \"#FAF8F1\" if THEME == \"light\" else \"#1A1A17\"\nELEVATED_BG = \"#FFFDF6\" if THEME == \"light\" else \"#242420\"\nINK = \"#1A1A17\" if THEME == \"light\" else \"#F0EFE8\"\nINK_SOFT = \"#4A4A44\" if THEME == \"light\" else \"#B8B7B0\"\n\n# Okabe-Ito palette\nBRAND = \"#009E73\"  # Training loss - first series\nACCENT = \"#C475FD\"  # Validation loss - second series\n\n# Data - simulate realistic neural network training loss curves\nnp.random.seed(42)\nepochs = np.arange(1, 101)\n\n# Training loss: exponential decay with some noise\ntrain_loss = 2.5 * np.exp(-0.04 * epochs) + 0.15 + np.random.normal(0, 0.03, len(epochs))\ntrain_loss = np.clip(train_loss, 0.1, 3.0)\n\n# Validation loss: similar decay but plateaus earlier and shows slight overfitting\nval_loss = 2.5 * np.exp(-0.035 * epochs) + 0.25 + np.random.normal(0, 0.04, len(epochs))\n# Add slight overfitting after epoch 70\nval_loss[69:] = val_loss[69:] + 0.002 * (epochs[69:] - 70)\nval_loss = np.clip(val_loss, 0.15, 3.0)\n\n# Find optimal epoch (minimum validation loss)\noptimal_epoch = epochs[np.argmin(val_loss)]\noptimal_val_loss = val_loss.min()\n\n# Create DataFrame for seaborn\ndf = pd.DataFrame(\n    {\n        \"Epoch\": np.tile(epochs, 2),\n        \"Loss\": np.concatenate([train_loss, val_loss]),\n        \"Type\": [\"Training Loss\"] * len(epochs) + [\"Validation Loss\"] * len(epochs),\n    }\n)\n\n# Set theme and style\nsns.set_theme(\n    style=\"ticks\",\n    rc={\n        \"figure.facecolor\": PAGE_BG,\n        \"axes.facecolor\": PAGE_BG,\n        \"axes.edgecolor\": INK_SOFT,\n        \"axes.labelcolor\": INK,\n        \"text.color\": INK,\n        \"xtick.color\": INK_SOFT,\n        \"ytick.color\": INK_SOFT,\n        \"grid.color\": INK,\n        \"grid.alpha\": 0.10,\n        \"legend.facecolor\": ELEVATED_BG,\n        \"legend.edgecolor\": INK_SOFT,\n    },\n)\n\n# Plot\nfig, ax = plt.subplots(figsize=(16, 9))\n\n# Use seaborn lineplot with hue for dual curves\nsns.lineplot(data=df, x=\"Epoch\", y=\"Loss\", hue=\"Type\", palette=[BRAND, ACCENT], linewidth=3, ax=ax)\n\n# Mark optimal stopping point\nax.axvline(x=optimal_epoch, color=INK_SOFT, linestyle=\"--\", linewidth=2, alpha=0.5)\nax.scatter([optimal_epoch], [optimal_val_loss], s=200, color=ACCENT, zorder=5, edgecolor=INK_SOFT, linewidth=2)\nax.annotate(\n    f\"Optimal: Epoch {optimal_epoch}\",\n    xy=(optimal_epoch, optimal_val_loss),\n    xytext=(optimal_epoch + 8, optimal_val_loss + 0.15),\n    fontsize=16,\n    color=INK,\n    arrowprops={\"arrowstyle\": \"->\", \"color\": INK_SOFT, \"lw\": 2},\n)\n\n# Style\nax.set_xlabel(\"Epoch\", fontsize=20, color=INK)\nax.set_ylabel(\"Cross-Entropy Loss\", fontsize=20, color=INK)\nax.set_title(\"line-loss-training · seaborn · anyplot.ai\", fontsize=24, color=INK)\nax.tick_params(axis=\"both\", labelsize=16, colors=INK_SOFT)\n\n# Remove top and right spines\nax.spines[\"top\"].set_visible(False)\nax.spines[\"right\"].set_visible(False)\nfor spine in (\"left\", \"bottom\"):\n    ax.spines[spine].set_color(INK_SOFT)\n\n# Grid - y-axis only\nax.yaxis.grid(True, alpha=0.10, linewidth=0.8)\nax.xaxis.grid(False)\n\n# Customize legend\nax.legend(fontsize=16, loc=\"upper right\", framealpha=0.95)\n\n# Set axis limits\nax.set_xlim(0, 105)\nax.set_ylim(0, 2.8)\n\nplt.tight_layout()\nplt.savefig(f\"plot-{THEME}.png\", dpi=300, bbox_inches=\"tight\", facecolor=PAGE_BG)\n"}