{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found existing installation: torch 2.6.0+cu124\n",
      "Uninstalling torch-2.6.0+cu124:\n",
      "  Successfully uninstalled torch-2.6.0+cu124\n",
      "Found existing installation: torchvision 0.21.0+cu124\n",
      "Uninstalling torchvision-0.21.0+cu124:\n",
      "  Successfully uninstalled torchvision-0.21.0+cu124\n",
      "Found existing installation: torchaudio 2.6.0+cu124\n",
      "Uninstalling torchaudio-2.6.0+cu124:\n",
      "  Successfully uninstalled torchaudio-2.6.0+cu124\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Failed to remove contents in a temporary directory 'C:\\Users\\Develope\\anaconda3\\Lib\\site-packages\\~.rch'.\n",
      "You can safely remove it manually.\n"
     ]
    }
   ],
   "source": [
    "!pip uninstall -y torch torchvision torchaudio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Looking in indexes: https://download.pytorch.org/whl/cu124\n",
      "Collecting torch\n",
      "  Using cached https://download.pytorch.org/whl/cu124/torch-2.6.0%2Bcu124-cp312-cp312-win_amd64.whl.metadata (28 kB)\n",
      "Collecting torchvision\n",
      "  Using cached https://download.pytorch.org/whl/cu124/torchvision-0.21.0%2Bcu124-cp312-cp312-win_amd64.whl.metadata (6.3 kB)\n",
      "Collecting torchaudio\n",
      "  Using cached https://download.pytorch.org/whl/cu124/torchaudio-2.6.0%2Bcu124-cp312-cp312-win_amd64.whl.metadata (6.8 kB)\n",
      "Requirement already satisfied: filelock in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torch) (3.13.1)\n",
      "Requirement already satisfied: typing-extensions>=4.10.0 in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torch) (4.11.0)\n",
      "Requirement already satisfied: networkx in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torch) (3.3)\n",
      "Requirement already satisfied: jinja2 in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torch) (3.1.4)\n",
      "Requirement already satisfied: fsspec in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torch) (2024.6.1)\n",
      "Requirement already satisfied: setuptools in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torch) (75.1.0)\n",
      "Requirement already satisfied: sympy==1.13.1 in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torch) (1.13.1)\n",
      "Requirement already satisfied: mpmath<1.4,>=1.1.0 in c:\\users\\develope\\anaconda3\\lib\\site-packages (from sympy==1.13.1->torch) (1.3.0)\n",
      "Requirement already satisfied: numpy in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torchvision) (1.26.4)\n",
      "Requirement already satisfied: pillow!=8.3.*,>=5.3.0 in c:\\users\\develope\\anaconda3\\lib\\site-packages (from torchvision) (10.4.0)\n",
      "Requirement already satisfied: MarkupSafe>=2.0 in c:\\users\\develope\\anaconda3\\lib\\site-packages (from jinja2->torch) (2.1.3)\n",
      "Using cached https://download.pytorch.org/whl/cu124/torch-2.6.0%2Bcu124-cp312-cp312-win_amd64.whl (2532.3 MB)\n",
      "Using cached https://download.pytorch.org/whl/cu124/torchvision-0.21.0%2Bcu124-cp312-cp312-win_amd64.whl (6.1 MB)\n",
      "Using cached https://download.pytorch.org/whl/cu124/torchaudio-2.6.0%2Bcu124-cp312-cp312-win_amd64.whl (4.2 MB)\n",
      "Installing collected packages: torch, torchvision, torchaudio\n",
      "Successfully installed torch-2.6.0+cu124 torchaudio-2.6.0+cu124 torchvision-0.21.0+cu124\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu124"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 3817,
     "status": "ok",
     "timestamp": 1763643726663,
     "user": {
      "displayName": "sgoo k",
      "userId": "16229481617978281248"
     },
     "user_tz": -540
    },
    "id": "4iUv2CFDWZI8",
    "outputId": "a71b3653-8099-4803-864d-7617fbee8d0a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['.ipynb_checkpoints', 'Added_Trajectoy_Data_20251223_001953.json', 'Added_Trajectoy_Data_20251223_001953.pkl', 'Added_Trajectoy_Data_20251223_002907.json', 'Added_Trajectoy_Data_20251223_002907.pkl', 'Added_Trajectoy_Data_20251223_003826.json', 'Added_Trajectoy_Data_20251223_003826.pkl', 'Added_Trajectoy_Data_20251223_004809.json', 'Added_Trajectoy_Data_20251223_004809.pkl', 'Added_Trajectoy_Data_20251223_005831.json', 'Added_Trajectoy_Data_20251223_005831.pkl', 'Added_Trajectoy_Data_20251223_010905.json', 'Added_Trajectoy_Data_20251223_010905.pkl', 'Added_Trajectoy_Data_20251223_012013.json', 'Added_Trajectoy_Data_20251223_012013.pkl', 'Added_Trajectoy_Data_20251223_013126.json', 'Added_Trajectoy_Data_20251223_013126.pkl', 'Added_Trajectoy_Data_20251223_014310.json', 'Added_Trajectoy_Data_20251223_014310.pkl', 'Added_Trajectoy_Data_20251223_015544.json', 'Added_Trajectoy_Data_20251223_015544.pkl', 'Added_Trajectoy_Data_20251223_020824.json', 'Added_Trajectoy_Data_20251223_020824.pkl', 'Added_Trajectoy_Data_20251223_022148.json', 'Added_Trajectoy_Data_20251223_022148.pkl', 'Added_Trajectoy_Data_20251223_023555.json', 'Added_Trajectoy_Data_20251223_023555.pkl', 'Added_Trajectoy_Data_20251223_025008.json', 'Added_Trajectoy_Data_20251223_025008.pkl', 'Added_Trajectoy_Data_20251223_030449.json', 'Added_Trajectoy_Data_20251223_030449.pkl', 'Added_Trajectoy_Data_20251223_031925.json', 'Added_Trajectoy_Data_20251223_031925.pkl', 'Added_Trajectoy_Data_20251223_033436.json', 'Added_Trajectoy_Data_20251223_033436.pkl', 'Added_Trajectoy_Data_20251223_035010.json', 'Added_Trajectoy_Data_20251223_035010.pkl', 'Added_Trajectoy_Data_20251223_040519.json', 'Added_Trajectoy_Data_20251223_040519.pkl', 'Added_Trajectoy_Data_20251223_042046.json', 'Added_Trajectoy_Data_20251223_042046.pkl', 'Added_Trajectoy_Data_20251223_043520.json', 'Added_Trajectoy_Data_20251223_043520.pkl', 'Added_Trajectoy_Data_20251223_044930.json', 'Added_Trajectoy_Data_20251223_044930.pkl', 'Added_Trajectoy_Data_20251223_050416.json', 'Added_Trajectoy_Data_20251223_050416.pkl', 'Added_Trajectoy_Data_20251223_051958.json', 'Added_Trajectoy_Data_20251223_051958.pkl', 'Added_Trajectoy_Data_20251223_053535.json', 'Added_Trajectoy_Data_20251223_053535.pkl', 'Added_Trajectoy_Data_20251223_055153.json', 'Added_Trajectoy_Data_20251223_055153.pkl', 'Added_Trajectoy_Data_20251223_060925.json', 'Added_Trajectoy_Data_20251223_060925.pkl', 'Added_Trajectoy_Data_20251223_063002.json', 'Added_Trajectoy_Data_20251223_063002.pkl', 'Added_Trajectoy_Data_20251223_065328.json', 'Added_Trajectoy_Data_20251223_065328.pkl', 'Added_Trajectoy_Data_20251223_072347.json', 'Added_Trajectoy_Data_20251223_072347.pkl', 'Added_Trajectoy_Data_20251223_080649.json', 'Added_Trajectoy_Data_20251223_080649.pkl', 'Backup', 'benchmark_plot.py', 'chart_visualize-Copy1.py', 'chart_visualize.py', 'convert_added_json_to_pickle.py', 'convert_filtered.py', 'convert_smart_fire_to_pickle.py', 'convert_to_pickle.py', 'dataset_dt.py', 'DicisionTransformer.ipynb', 'DT_BC_100.pth', 'DT_C_10.pth', 'DT_C_5.pth', 'dt_model_new_trained_V10.onnx', 'dt_model_new_trained_V12.onnx', 'dt_model_new_trained_V12.pth', 'dt_model_new_trained_V12_50000.onnx', 'dt_model_new_trained_V12_50000.pth', 'dt_model_new_trained_V12_50000_int32.onnx', 'dt_model_new_trained_V12_int32.onnx', 'DT_SC_10.pth', 'DT_SC_5.pth', 'DT_S_100.pth', 'Experiment_Plan_V13.ipynb', 'export_onnx.py', 'export_onnx_v2.py', 'E_1_DT_BC_100.pth', 'E_1_DT_C_10.onnx', 'E_1_DT_C_10.pth', 'E_1_DT_C_5.onnx', 'E_1_DT_C_5.pth', 'E_1_DT_SC_10.onnx', 'E_1_DT_SC_10.pth', 'E_1_DT_SC_5.onnx', 'E_1_DT_SC_5.pth', 'E_1_DT_SC_5_F.onnx', 'E_1_DT_S_100.onnx', 'E_1_DT_S_100.pth', 'E_2_DT_BC_100.pth', 'E_2_DT_C_10.onnx', 'E_2_DT_C_10.pth', 'E_2_DT_C_5.onnx', 'E_2_DT_C_5.pth', 'E_2_DT_SC_10.onnx', 'E_2_DT_SC_10.pth', 'E_2_DT_SC_5.onnx', 'E_2_DT_SC_5.pth', 'E_2_DT_S_100.onnx', 'E_2_DT_S_100.pth', 'E_2_DT_S_100_F.onnx', 'E_3_DT_BC_100.pth', 'E_3_DT_C_10.onnx', 'E_3_DT_C_10.pth', 'E_3_DT_C_5.onnx', 'E_3_DT_C_5.pth', 'E_3_DT_SC_10.onnx', 'E_3_DT_SC_10.pth', 'E_3_DT_SC_5.onnx', 'E_3_DT_SC_5.pth', 'E_3_DT_S_100.onnx', 'E_3_DT_S_100.pth', 'Fig1_Model_Comparison.png', 'Fig2_RTG_Analysis.png', 'Fig2_RTG_Analysis_Revised.png', 'Fig3_RTG_Boxplot.png', 'Fig7_RTG_Analysis_Revised.png', 'Fig8_RTG_Boxplot.png', 'FINAL_RESULT', 'FINAL_RESULT.zip', 'FINAL_RESULT_BACKUP_SUBMIT1', 'FINAL_RESULT_For_Revision', 'FineTuning.py', 'FineTuningV.py', 'finetuning_ext_RLStep.py', 'generate_dataset.py', 'inference_pipeline.py', 'LogMergerTool', 'logs', 'Models', 'model_dt.py', 'new_train_sequential.py', 'nonmun.code-workspace', 'paper_plots_final.py', 'Performance Analysis Contingent upon Return-to-Go .ipynb', 'plot_rtg_paper.py', 'plot_statistical_results.py', 'README.md', 'restore_notebook.py', 'split_json.py', 'statistical_tests.py', 'train_best.py', 'train_perception_model.py', 'train_sequential.py', 'train_sequential_ext_RLStep.py', 'train_sequential_ext_RLStep_For_BC.py', 'trajectory_data_part_0.pkl', 'trajectory_data_part_1.pkl', 'trajectory_data_part_10.pkl', 'trajectory_data_part_11.pkl', 'trajectory_data_part_12.pkl', 'trajectory_data_part_2.pkl', 'trajectory_data_part_3.pkl', 'trajectory_data_part_4.pkl', 'trajectory_data_part_5.pkl', 'trajectory_data_part_6.pkl', 'trajectory_data_part_7.pkl', 'trajectory_data_part_8.pkl', 'trajectory_data_part_9.pkl', 'Trajectory_For_firetooneonlywhenhit', 'trajectory_log.json', 'UnityScript', '__pycache__']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<module 'convert_to_pickle' from 'C:\\\\Users\\\\Develope\\\\DT\\\\convert_to_pickle.py'>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "print(os.listdir())\n",
    "import importlib\n",
    "import model_dt\n",
    "import dataset_dt\n",
    "import convert_to_pickle\n",
    "\n",
    "from dataset_dt import TrajectoryDataset\n",
    "from model_dt import DecisionTransformer  # 모델 정의가 들어있는 파일\n",
    "importlib.reload(dataset_dt)\n",
    "importlib.reload(model_dt)\n",
    "importlib.reload(convert_to_pickle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] Processing trajectory_log.json...\n",
      "[INFO] File size: 1234.41 MB\n",
      "Progress: 8.0% | Total Steps: 192265\n",
      "[INFO] Saving part 0 to trajectory_data_part_0.pkl (200000 steps)...\n",
      "Progress: 16.0% | Total Steps: 387873\n",
      "[INFO] Saving part 1 to trajectory_data_part_1.pkl (200000 steps)...\n",
      "Progress: 24.0% | Total Steps: 584194\n",
      "[INFO] Saving part 2 to trajectory_data_part_2.pkl (200000 steps)...\n",
      "Progress: 32.0% | Total Steps: 780717\n",
      "[INFO] Saving part 3 to trajectory_data_part_3.pkl (200000 steps)...\n",
      "Progress: 40.0% | Total Steps: 977204\n",
      "[INFO] Saving part 4 to trajectory_data_part_4.pkl (200000 steps)...\n",
      "Progress: 49.0% | Total Steps: 1193064\n",
      "[INFO] Saving part 5 to trajectory_data_part_5.pkl (200000 steps)...\n",
      "Progress: 57.0% | Total Steps: 1387467\n",
      "[INFO] Saving part 6 to trajectory_data_part_6.pkl (200000 steps)...\n",
      "Progress: 65.0% | Total Steps: 1579118\n",
      "[INFO] Saving part 7 to trajectory_data_part_7.pkl (200000 steps)...\n",
      "Progress: 74.0% | Total Steps: 1795344\n",
      "[INFO] Saving part 8 to trajectory_data_part_8.pkl (200000 steps)...\n",
      "Progress: 82.0% | Total Steps: 1990089\n",
      "[INFO] Saving part 9 to trajectory_data_part_9.pkl (200000 steps)...\n",
      "Progress: 90.0% | Total Steps: 2186254\n",
      "[INFO] Saving part 10 to trajectory_data_part_10.pkl (200000 steps)...\n",
      "Progress: 98.0% | Total Steps: 2382831\n",
      "[INFO] Saving part 11 to trajectory_data_part_11.pkl (200000 steps)...\n",
      "Progress: 99.0% | Total Steps: 2407383\n",
      "[INFO] Saving final part 12 to trajectory_data_part_12.pkl (31933 steps)...\n",
      "\n",
      "[INFO] Parsing complete. Total steps: 2431933\n",
      "[INFO] Data saved in 13 parts.\n"
     ]
    }
   ],
   "source": [
    "## 강화학습에서 추출한 Log를 pkl로 저장\n",
    "from convert_to_pickle import convert_json_to_pickle\n",
    "convert_json_to_pickle();\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] Using device: cuda\n",
      "[INFO] Found 13 split files matching 'trajectory_data_part_'.\n",
      "[INFO] Peeking at first file for dimensions...\n",
      "[INFO] Loading specific file: trajectory_data_part_0.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_0.pkl.\n",
      "[INFO] Processed 1458 episodes. Total 200000 samples.\n",
      "[INFO] Obs Dim: 9, Act Dim: 3\n",
      "[INFO] Starting FRESH training with RL STEP LIMIT: 1000000 (Pre-loading to Memory)\n",
      "[INFO] Pre-loading chunk 1/13: trajectory_data_part_0.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_0.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_0.pkl.\n",
      "[INFO] Processed 1458 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 200000 / 1000000\n",
      "[INFO] Pre-loading chunk 2/13: trajectory_data_part_1.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_1.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_1.pkl.\n",
      "[INFO] Processed 1735 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 400000 / 1000000\n",
      "[INFO] Pre-loading chunk 3/13: trajectory_data_part_10.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_10.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_10.pkl.\n",
      "[INFO] Processed 704 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 600000 / 1000000\n",
      "[INFO] Pre-loading chunk 4/13: trajectory_data_part_11.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_11.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_11.pkl.\n",
      "[INFO] Processed 712 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 800000 / 1000000\n",
      "[INFO] Pre-loading chunk 5/13: trajectory_data_part_12.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_12.pkl...\n",
      "[INFO] Loaded 31933 steps from trajectory_data_part_12.pkl.\n",
      "[INFO] Processed 106 episodes. Total 31933 samples.\n",
      "  [PROGRESS] Memory Buffer: 831933 / 1000000\n",
      "[INFO] Pre-loading chunk 6/13: trajectory_data_part_2.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_2.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_2.pkl.\n",
      "[INFO] Processed 1511 episodes. Total 200000 samples.\n",
      "  [LIMIT] Trimming chunk to 168067 samples.\n",
      "[INFO] Data Pre-loading Complete. Starting Training on 6 chunks.\n",
      "\n",
      "=== Epoch 1/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0709\n",
      "  Chunk 2 Finished. Avg Loss: 0.0431\n",
      "  Chunk 3 Finished. Avg Loss: 0.0405\n",
      "  Chunk 4 Finished. Avg Loss: 0.0378\n",
      "  Chunk 5 Finished. Avg Loss: 0.0330\n",
      "  Chunk 6 Finished. Avg Loss: 0.0380\n",
      "Epoch 1 completed in 1016.35s.\n",
      "[INFO] Saved Epoch 1 Model to: E_1_DT_S_100.pth\n",
      "\n",
      "=== Epoch 2/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0595\n",
      "  Chunk 2 Finished. Avg Loss: 0.0368\n",
      "  Chunk 3 Finished. Avg Loss: 0.0342\n",
      "  Chunk 4 Finished. Avg Loss: 0.0334\n",
      "  Chunk 5 Finished. Avg Loss: 0.0286\n",
      "  Chunk 6 Finished. Avg Loss: 0.0356\n",
      "Epoch 2 completed in 1018.93s.\n",
      "[INFO] Saved Epoch 2 Model to: E_2_DT_S_100.pth\n",
      "\n",
      "=== Epoch 3/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0557\n",
      "  Chunk 2 Finished. Avg Loss: 0.0350\n",
      "  Chunk 3 Finished. Avg Loss: 0.0315\n",
      "  Chunk 4 Finished. Avg Loss: 0.0307\n",
      "  Chunk 5 Finished. Avg Loss: 0.0258\n",
      "  Chunk 6 Finished. Avg Loss: 0.0340\n",
      "Epoch 3 completed in 1018.62s.\n",
      "[INFO] Saved Epoch 3 Model to: E_3_DT_S_100.pth\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed (cached): 1000000\n",
      "  Total Gradient Steps: 93753\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed: 1000000\n",
      "  Total Gradient Steps: 93753\n",
      "[INFO] Saved Final Model to: DT_S_100.pth\n",
      "[INFO] Using device: cuda\n",
      "[INFO] Found 31 split files matching 'Added_Trajectoy_Data_'.\n",
      "[INFO] Peeking at first file for dimensions...\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "[INFO] Obs Dim: 9, Act Dim: 3\n",
      "[INFO] Starting FRESH training with RL STEP LIMIT: 50000 (Pre-loading to Memory)\n",
      "[INFO] Pre-loading chunk 1/31: Added_Trajectoy_Data_20251223_001953.pkl\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "  [LIMIT] Trimming chunk to 50000 samples.\n",
      "[INFO] Data Pre-loading Complete. Starting Training on 1 chunks.\n",
      "\n",
      "=== Epoch 1/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0892\n",
      "Epoch 1 completed in 49.93s.\n",
      "[INFO] Saved Epoch 1 Model to: E_1_DT_C_5.pth\n",
      "\n",
      "=== Epoch 2/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0531\n",
      "Epoch 2 completed in 50.05s.\n",
      "[INFO] Saved Epoch 2 Model to: E_2_DT_C_5.pth\n",
      "\n",
      "=== Epoch 3/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0442\n",
      "Epoch 3 completed in 50.20s.\n",
      "[INFO] Saved Epoch 3 Model to: E_3_DT_C_5.pth\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed (cached): 50000\n",
      "  Total Gradient Steps: 4689\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed: 50000\n",
      "  Total Gradient Steps: 4689\n",
      "[INFO] Saved Final Model to: DT_C_5.pth\n",
      "[INFO] Using device: cuda\n",
      "[INFO] Found 31 split files matching 'Added_Trajectoy_Data_'.\n",
      "[INFO] Peeking at first file for dimensions...\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "[INFO] Obs Dim: 9, Act Dim: 3\n",
      "[INFO] Starting FRESH training with RL STEP LIMIT: 100000 (Pre-loading to Memory)\n",
      "[INFO] Pre-loading chunk 1/31: Added_Trajectoy_Data_20251223_001953.pkl\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "  [LIMIT] Trimming chunk to 100000 samples.\n",
      "[INFO] Data Pre-loading Complete. Starting Training on 1 chunks.\n",
      "\n",
      "=== Epoch 1/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0747\n",
      "Epoch 1 completed in 100.00s.\n",
      "[INFO] Saved Epoch 1 Model to: E_1_DT_C_10.pth\n",
      "\n",
      "=== Epoch 2/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0474\n",
      "Epoch 2 completed in 100.50s.\n",
      "[INFO] Saved Epoch 2 Model to: E_2_DT_C_10.pth\n",
      "\n",
      "=== Epoch 3/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0404\n",
      "Epoch 3 completed in 101.71s.\n",
      "[INFO] Saved Epoch 3 Model to: E_3_DT_C_10.pth\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed (cached): 100000\n",
      "  Total Gradient Steps: 9375\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed: 100000\n",
      "  Total Gradient Steps: 9375\n",
      "[INFO] Saved Final Model to: DT_C_10.pth\n",
      "[INFO] Using device: cuda\n",
      "[INFO] Found 31 split files matching 'Added_Trajectoy_Data_'.\n",
      "[INFO] Peeking at first file for dimensions...\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "[INFO] Obs Dim: 9, Act Dim: 3\n",
      "[INFO] Loading checkpoint: DT_S_100.pth\n",
      "[INFO] Model loaded successfully.\n",
      "[INFO] Learning rate: 1e-05\n",
      "[INFO] Starting FINE-TUNING with RL STEP LIMIT: 50000 (Pre-loading to Memory)\n",
      "[INFO] Pre-loading chunk 1/31: Added_Trajectoy_Data_20251223_001953.pkl\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "  [LIMIT] Trimming chunk to 50000 samples.\n",
      "[INFO] Data Pre-loading Complete. Starting Fine-Tuning on 1 chunks.\n",
      "\n",
      "=== Fine-Tuning Epoch 1/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0499\n",
      "Epoch 1 completed in 49.71s.\n",
      "[INFO] Saved Epoch 1 Checkpoint to: E_1_DT_SC_5.pth\n",
      "\n",
      "=== Fine-Tuning Epoch 2/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0355\n",
      "Epoch 2 completed in 49.89s.\n",
      "[INFO] Saved Epoch 2 Checkpoint to: E_2_DT_SC_5.pth\n",
      "\n",
      "=== Fine-Tuning Epoch 3/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0328\n",
      "Epoch 3 completed in 50.03s.\n",
      "[INFO] Saved Epoch 3 Checkpoint to: E_3_DT_SC_5.pth\n",
      "\n",
      "[DONE] Fine-Tuning Finished.\n",
      "  Total RL Steps Processed (cached): 50000\n",
      "  Total Gradient Steps: 4689\n",
      "[INFO] Saved Final Model to: DT_SC_5.pth\n",
      "[INFO] Using device: cuda\n",
      "[INFO] Found 31 split files matching 'Added_Trajectoy_Data_'.\n",
      "[INFO] Peeking at first file for dimensions...\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "[INFO] Obs Dim: 9, Act Dim: 3\n",
      "[INFO] Loading checkpoint: DT_S_100.pth\n",
      "[INFO] Model loaded successfully.\n",
      "[INFO] Learning rate: 1e-05\n",
      "[INFO] Starting FINE-TUNING with RL STEP LIMIT: 100000 (Pre-loading to Memory)\n",
      "[INFO] Pre-loading chunk 1/31: Added_Trajectoy_Data_20251223_001953.pkl\n",
      "[INFO] Loading specific file: Added_Trajectoy_Data_20251223_001953.pkl...\n",
      "[INFO] Loaded 258606 steps from Added_Trajectoy_Data_20251223_001953.pkl.\n",
      "[INFO] Processed 1000 episodes. Total 258606 samples.\n",
      "  [LIMIT] Trimming chunk to 100000 samples.\n",
      "[INFO] Data Pre-loading Complete. Starting Fine-Tuning on 1 chunks.\n",
      "\n",
      "=== Fine-Tuning Epoch 1/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0454\n",
      "Epoch 1 completed in 99.60s.\n",
      "[INFO] Saved Epoch 1 Checkpoint to: E_1_DT_SC_10.pth\n",
      "\n",
      "=== Fine-Tuning Epoch 2/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0359\n",
      "Epoch 2 completed in 100.08s.\n",
      "[INFO] Saved Epoch 2 Checkpoint to: E_2_DT_SC_10.pth\n",
      "\n",
      "=== Fine-Tuning Epoch 3/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0339\n",
      "Epoch 3 completed in 101.30s.\n",
      "[INFO] Saved Epoch 3 Checkpoint to: E_3_DT_SC_10.pth\n",
      "\n",
      "[DONE] Fine-Tuning Finished.\n",
      "  Total RL Steps Processed (cached): 100000\n",
      "  Total Gradient Steps: 9375\n",
      "[INFO] Saved Final Model to: DT_SC_10.pth\n"
     ]
    }
   ],
   "source": [
    "import importlib\n",
    "import train_sequential_ext_RLStep\n",
    "importlib.reload(train_sequential_ext_RLStep)\n",
    "from train_sequential_ext_RLStep import train_sequential_rl_steps\n",
    "\n",
    "train_sequential_rl_steps(\n",
    "    data_prefix=\"trajectory_data_part_\", \n",
    "    output_model=\"DT_S_100.pth\",\n",
    "    target_rl_steps = 1000000)\n",
    "\n",
    "train_sequential_rl_steps(\n",
    "    data_prefix=\"Added_Trajectoy_Data_\", \n",
    "    output_model=\"DT_C_5.pth\",\n",
    "    target_rl_steps = 50000)\n",
    "\n",
    "train_sequential_rl_steps(\n",
    "    data_prefix=\"Added_Trajectoy_Data_\", \n",
    "    output_model=\"DT_C_10.pth\",\n",
    "    target_rl_steps = 100000)\n",
    "\n",
    "import finetuning_ext_RLStep\n",
    "importlib.reload(finetuning_ext_RLStep)\n",
    "from finetuning_ext_RLStep import finetuning_rl_steps\n",
    "\n",
    "finetuning_rl_steps(\n",
    "    load_model = \"DT_S_100.pth\",\n",
    "    data_prefix = \"Added_Trajectoy_Data_\",\n",
    "    target_rl_steps = 50000,\n",
    "    epochs = 3,\n",
    "    output_model = \"DT_SC_5.pth\");\n",
    "\n",
    "finetuning_rl_steps(\n",
    "    load_model = \"DT_S_100.pth\",\n",
    "    data_prefix = \"Added_Trajectoy_Data_\",\n",
    "    target_rl_steps = 100000,\n",
    "    epochs = 3,\n",
    "    output_model = \"DT_SC_10.pth\");\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] Using device: cuda\n",
      "[INFO] Found 13 split files matching 'trajectory_data_part_'.\n",
      "[INFO] Peeking at first file for dimensions...\n",
      "[INFO] Loading specific file: trajectory_data_part_0.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_0.pkl.\n",
      "[INFO] Processed 1458 episodes. Total 200000 samples.\n",
      "[INFO] Obs Dim: 9, Act Dim: 3\n",
      "[INFO] Starting FRESH training with RL STEP LIMIT: 1000000 (Pre-loading to Memory)\n",
      "[INFO] Pre-loading chunk 1/13: trajectory_data_part_0.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_0.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_0.pkl.\n",
      "[INFO] Processed 1458 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 200000 / 1000000\n",
      "[INFO] Pre-loading chunk 2/13: trajectory_data_part_1.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_1.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_1.pkl.\n",
      "[INFO] Processed 1735 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 400000 / 1000000\n",
      "[INFO] Pre-loading chunk 3/13: trajectory_data_part_10.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_10.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_10.pkl.\n",
      "[INFO] Processed 704 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 600000 / 1000000\n",
      "[INFO] Pre-loading chunk 4/13: trajectory_data_part_11.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_11.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_11.pkl.\n",
      "[INFO] Processed 712 episodes. Total 200000 samples.\n",
      "  [PROGRESS] Memory Buffer: 800000 / 1000000\n",
      "[INFO] Pre-loading chunk 5/13: trajectory_data_part_12.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_12.pkl...\n",
      "[INFO] Loaded 31933 steps from trajectory_data_part_12.pkl.\n",
      "[INFO] Processed 106 episodes. Total 31933 samples.\n",
      "  [PROGRESS] Memory Buffer: 831933 / 1000000\n",
      "[INFO] Pre-loading chunk 6/13: trajectory_data_part_2.pkl\n",
      "[INFO] Loading specific file: trajectory_data_part_2.pkl...\n",
      "[INFO] Loaded 200000 steps from trajectory_data_part_2.pkl.\n",
      "[INFO] Processed 1511 episodes. Total 200000 samples.\n",
      "  [LIMIT] Trimming chunk to 168067 samples.\n",
      "[INFO] Data Pre-loading Complete. Starting Training on 6 chunks.\n",
      "\n",
      "=== Epoch 1/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0708\n",
      "  Chunk 2 Finished. Avg Loss: 0.0417\n",
      "  Chunk 3 Finished. Avg Loss: 0.0392\n",
      "  Chunk 4 Finished. Avg Loss: 0.0369\n",
      "  Chunk 5 Finished. Avg Loss: 0.0321\n",
      "  Chunk 6 Finished. Avg Loss: 0.0372\n",
      "Epoch 1 completed in 1044.00s.\n",
      "[INFO] Saved Epoch 1 Model to: E_1_DT_BC_100.pth\n",
      "\n",
      "=== Epoch 2/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0592\n",
      "  Chunk 2 Finished. Avg Loss: 0.0364\n",
      "  Chunk 3 Finished. Avg Loss: 0.0336\n",
      "  Chunk 4 Finished. Avg Loss: 0.0329\n",
      "  Chunk 5 Finished. Avg Loss: 0.0280\n",
      "  Chunk 6 Finished. Avg Loss: 0.0350\n",
      "Epoch 2 completed in 1046.82s.\n",
      "[INFO] Saved Epoch 2 Model to: E_2_DT_BC_100.pth\n",
      "\n",
      "=== Epoch 3/3 ===\n",
      "  Chunk 1 Finished. Avg Loss: 0.0551\n",
      "  Chunk 2 Finished. Avg Loss: 0.0347\n",
      "  Chunk 3 Finished. Avg Loss: 0.0311\n",
      "  Chunk 4 Finished. Avg Loss: 0.0305\n",
      "  Chunk 5 Finished. Avg Loss: 0.0254\n",
      "  Chunk 6 Finished. Avg Loss: 0.0337\n",
      "Epoch 3 completed in 1054.18s.\n",
      "[INFO] Saved Epoch 3 Model to: E_3_DT_BC_100.pth\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed (cached): 1000000\n",
      "  Total Gradient Steps: 93753\n",
      "\n",
      "[DONE] Training Finished.\n",
      "  Total RL Steps Processed: 1000000\n",
      "  Total Gradient Steps: 93753\n",
      "[INFO] Saved Final Model to: DT_BC_100.pth\n"
     ]
    }
   ],
   "source": [
    "import importlib\n",
    "import train_sequential_ext_RLStep_For_BC\n",
    "importlib.reload(train_sequential_ext_RLStep_For_BC)\n",
    "from train_sequential_ext_RLStep_For_BC import train_sequential_rl_steps_BC\n",
    "\n",
    "train_sequential_rl_steps_BC(\n",
    "    data_prefix=\"trajectory_data_part_\", \n",
    "    output_model=\"DT_BC_100.pth\",\n",
    "    target_rl_steps = 1000000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "executionInfo": {
     "elapsed": 70,
     "status": "error",
     "timestamp": 1763477753862,
     "user": {
      "displayName": "sgoo k",
      "userId": "16229481617978281248"
     },
     "user_tz": -540
    },
    "id": "eZRuDC32cwli",
    "outputId": "1256c6fe-8a6d-4700-8f62-932cd70b8ab6"
   },
   "outputs": [],
   "source": [
    "## 모델 정보를 확인\n",
    "import torch\n",
    "model = DecisionTransformer(obs_dim=9, act_dim=3, hidden=256, n_layers=4, n_heads=4, max_len = 4096)\n",
    "model.load_state_dict(torch.load(\"dt_model_new_trained_V12.pth\"))\n",
    "model.eval()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 7884,
     "status": "ok",
     "timestamp": 1763478051931,
     "user": {
      "displayName": "sgoo k",
      "userId": "16229481617978281248"
     },
     "user_tz": -540
    },
    "id": "0IyKB8bGe3dR",
    "outputId": "eb4f0182-cf91-4a68-add6-fc095688a12d"
   },
   "outputs": [],
   "source": [
    "## ONNX 변환 준비\n",
    "!pip install onnx\n",
    "!pip install onnxscript"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] ONNX export complete: E_1_DT_S_100.onnx\n",
      "[INFO] ONNX export complete: E_2_DT_S_100.onnx\n",
      "[INFO] ONNX export complete: E_3_DT_S_100.onnx\n",
      "[INFO] ONNX export complete: E_1_DT_C_5.onnx\n",
      "[INFO] ONNX export complete: E_2_DT_C_5.onnx\n",
      "[INFO] ONNX export complete: E_3_DT_C_5.onnx\n",
      "[INFO] ONNX export complete: E_1_DT_C_10.onnx\n",
      "[INFO] ONNX export complete: E_2_DT_C_10.onnx\n",
      "[INFO] ONNX export complete: E_3_DT_C_10.onnx\n",
      "[INFO] ONNX export complete: E_1_DT_SC_5.onnx\n",
      "[INFO] ONNX export complete: E_2_DT_SC_5.onnx\n",
      "[INFO] ONNX export complete: E_3_DT_SC_5.onnx\n",
      "[INFO] ONNX export complete: E_1_DT_SC_10.onnx\n",
      "[INFO] ONNX export complete: E_2_DT_SC_10.onnx\n",
      "[INFO] ONNX export complete: E_3_DT_SC_10.onnx\n"
     ]
    }
   ],
   "source": [
    "#ONNX 로 변환\n",
    "# 모델 초기화 (학습 시 사용한 차원과 동일하게)\n",
    "import torch\n",
    "\n",
    "#\"dt_model_new_trained_V12.pth\"\n",
    "def CreateONNX(filePath):\n",
    "    obs_dim = 9\n",
    "    act_dim = 3\n",
    "    hidden = 256\n",
    "    seq_len = 32\n",
    "    \n",
    "    model = DecisionTransformer(obs_dim=9, act_dim=3, hidden=256, n_layers=4, n_heads=4, max_len = 4096)\n",
    "    model.load_state_dict(torch.load(filePath))\n",
    "    model.eval()\n",
    "    \n",
    "    # 더미 입력 생성 (batch=1, seq_len=32)\n",
    "    dummy_obs = torch.randn(1, seq_len, obs_dim)\n",
    "    dummy_act = torch.randn(1, seq_len, act_dim)\n",
    "    dummy_rtg = torch.randn(1, seq_len, 1)\n",
    "    dummy_ts = torch.arange(seq_len).unsqueeze(0)  # (1, seq_len)\n",
    "    \n",
    "    onnx_filename = filePath.replace(\".pth\", \".onnx\")\n",
    "    # ONNX로 내보내기\n",
    "    torch.onnx.export(\n",
    "        model,\n",
    "        (dummy_obs, dummy_act, dummy_rtg, dummy_ts),\n",
    "        onnx_filename,\n",
    "        input_names=[\"observations\", \"actions\", \"returns_to_go\", \"timesteps\"],\n",
    "        output_names=[\"predicted_actions\"],\n",
    "        dynamic_axes={\n",
    "            \"observations\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"returns_to_go\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"timesteps\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"predicted_actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        },\n",
    "        opset_version=17\n",
    "    )\n",
    "\n",
    "    print(\"[INFO] ONNX export complete: \"+ onnx_filename)\n",
    "\n",
    "    import onnx\n",
    "    from onnx import helper, TensorProto\n",
    "    model = onnx.load(onnx_filename)\n",
    "    for input in model.graph.input:\n",
    "        if input.name == \"timesteps\":\n",
    "            input.type.tensor_type.elem_type = TensorProto.INT32\n",
    "    onnx.save(model, onnx_filename)\n",
    "    print(\"[INFO] changed Compleate : \"+ onnx_filename)\n",
    "##########################################\n",
    "CreateONNX(filePath = \"E_1_DT_S_100.pth\");\n",
    "CreateONNX(filePath = \"E_2_DT_S_100.pth\");\n",
    "CreateONNX(filePath = \"E_3_DT_S_100.pth\");\n",
    "\n",
    "CreateONNX(filePath = \"E_1_DT_C_5.pth\");\n",
    "CreateONNX(filePath = \"E_2_DT_C_5.pth\");\n",
    "CreateONNX(filePath = \"E_3_DT_C_5.pth\");\n",
    "\n",
    "CreateONNX(filePath = \"E_1_DT_C_10.pth\");\n",
    "CreateONNX(filePath = \"E_2_DT_C_10.pth\");\n",
    "CreateONNX(filePath = \"E_3_DT_C_10.pth\");\n",
    "\n",
    "CreateONNX(filePath = \"E_1_DT_SC_5.pth\");\n",
    "CreateONNX(filePath = \"E_2_DT_SC_5.pth\");\n",
    "CreateONNX(filePath = \"E_3_DT_SC_5.pth\");\n",
    "\n",
    "CreateONNX(filePath = \"E_1_DT_SC_10.pth\");\n",
    "CreateONNX(filePath = \"E_2_DT_SC_10.pth\");\n",
    "CreateONNX(filePath = \"E_3_DT_SC_10.pth\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] ONNX export complete: E_1_DT_BC_100.onnx\n",
      "[INFO] changed Compleate : E_1_DT_BC_100.onnx\n",
      "[INFO] ONNX export complete: E_2_DT_BC_100.onnx\n",
      "[INFO] changed Compleate : E_2_DT_BC_100.onnx\n",
      "[INFO] ONNX export complete: E_3_DT_BC_100.onnx\n",
      "[INFO] changed Compleate : E_3_DT_BC_100.onnx\n",
      "[INFO] ONNX export complete: DT_BC_100.onnx\n",
      "[INFO] changed Compleate : DT_BC_100.onnx\n"
     ]
    }
   ],
   "source": [
    "#ONNX 로 변환\n",
    "# 모델 초기화 (학습 시 사용한 차원과 동일하게)\n",
    "import torch\n",
    "\n",
    "#\"dt_model_new_trained_V12.pth\"\n",
    "def CreateONNX(filePath):\n",
    "    obs_dim = 9\n",
    "    act_dim = 3\n",
    "    hidden = 256\n",
    "    seq_len = 32\n",
    "    \n",
    "    model = DecisionTransformer(obs_dim=9, act_dim=3, hidden=256, n_layers=4, n_heads=4, max_len = 4096)\n",
    "    model.load_state_dict(torch.load(filePath))\n",
    "    model.eval()\n",
    "    \n",
    "    # 더미 입력 생성 (batch=1, seq_len=32)\n",
    "    dummy_obs = torch.randn(1, seq_len, obs_dim)\n",
    "    dummy_act = torch.randn(1, seq_len, act_dim)\n",
    "    dummy_rtg = torch.randn(1, seq_len, 1)\n",
    "    dummy_ts = torch.arange(seq_len).unsqueeze(0)  # (1, seq_len)\n",
    "    \n",
    "    onnx_filename = filePath.replace(\".pth\", \".onnx\")\n",
    "    # ONNX로 내보내기\n",
    "    torch.onnx.export(\n",
    "        model,\n",
    "        (dummy_obs, dummy_act, dummy_rtg, dummy_ts),\n",
    "        onnx_filename,\n",
    "        input_names=[\"observations\", \"actions\", \"returns_to_go\", \"timesteps\"],\n",
    "        output_names=[\"predicted_actions\"],\n",
    "        dynamic_axes={\n",
    "            \"observations\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"returns_to_go\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"timesteps\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "            \"predicted_actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        },\n",
    "        opset_version=17\n",
    "    )\n",
    "\n",
    "    print(\"[INFO] ONNX export complete: \"+ onnx_filename)\n",
    "\n",
    "    import onnx\n",
    "    from onnx import helper, TensorProto\n",
    "    model = onnx.load(onnx_filename)\n",
    "    for input in model.graph.input:\n",
    "        if input.name == \"timesteps\":\n",
    "            input.type.tensor_type.elem_type = TensorProto.INT32\n",
    "    onnx.save(model, onnx_filename)\n",
    "    print(\"[INFO] changed Compleate : \"+ onnx_filename)\n",
    "##########################################\n",
    "CreateONNX(filePath = \"E_1_DT_BC_100.pth\");\n",
    "CreateONNX(filePath = \"E_2_DT_BC_100.pth\");\n",
    "CreateONNX(filePath = \"E_3_DT_BC_100.pth\");\n",
    "CreateONNX(filePath = \"DT_BC_100.pth\");\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Timestpes 에러 변환\n",
    "import onnx\n",
    "from onnx import helper, TensorProto\n",
    "\n",
    "def\n",
    "model = onnx.load(\"dt_model_new_trained_V12.onnx\")\n",
    "for input in model.graph.input:\n",
    "    if input.name == \"timesteps\":\n",
    "        input.type.tensor_type.elem_type = TensorProto.INT32\n",
    "onnx.save(model, \"dt_model_new_trained_V12_int32.onnx\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import importlib\n",
    "import convert_added_json_to_pickle\n",
    "importlib.reload(convert_added_json_to_pickle)\n",
    "from convert_added_json_to_pickle import convert_added_json_to_pickle\n",
    "\n",
    "if __name__ == \"__main__\":\n",
    "    convert_added_json_to_pickle()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import importlib\n",
    "import FineTuning\n",
    "importlib.reload(FineTuning)\n",
    "from FineTuning import fine_tuning\n",
    "\n",
    "import train_sequential\n",
    "importlib.reload(train_sequential)\n",
    "from train_sequential import train_sequential\n",
    "\n",
    "\n",
    "#train_sequential(data_prefix=\"trajectory_data_part_\", output_model=\"dt_model_new_trained_V12.pth\")\n",
    "train_sequential(data_prefix=\"Added_Trajectoy_Data_\",output_model=\"dt_model_new_trained_V11_Only_50000.pth\",target_steps =50000)\n",
    "\n",
    "train_sequential(\n",
    "     data_prefix=\"Added_Trajectoy_Data_\", \n",
    "     output_model=\"dt_model_new_trained_V10_Elite_Only.pth\",\n",
    "     target_steps =100000)\n",
    "\n",
    "#train_sequential(\n",
    "#     data_prefix=\"Added_Trajectoy_Data_\", \n",
    "#     output_model=\"dt_model_new_trained_V10_Elite_150000.pth\",\n",
    "#     target_steps =150000)\n",
    "#\n",
    "#train_sequential(\n",
    "#    data_prefix=\"Added_Trajectoy_Data_\", \n",
    "#    output_model=\"dt_model_new_trained_V10_Elite_200000.pth\",\n",
    "#    target_steps =200000)\n",
    "if __name__ == \"__main__\":\n",
    "    fine_tuning(2, \"dt_model_new_trained_V12.pth\", \"dt_model_new_trained_V12_add_5.pth\", \"Added_Trajectoy_Data_*.pkl\", 50000);\n",
    "\n",
    "if __name__ == \"__main__\":\n",
    "    fine_tuning(5, \"dt_model_new_trained_V12.pth\", \"dt_model_new_trained_V10_add_10.pth\", \"Added_Trajectoy_Data_*.pkl\", 100000);\n",
    "\n",
    "# if __name__ == \"__main__\":\n",
    "#     fine_tuning(5, \"dt_model_new_trained_V12.pth\", \"dt_model_new_trained_V10_add_15.pth\", \"Added_Trajectoy_Data_*.pkl\", 150000);\n",
    "\n",
    "# if __name__ == \"__main__\":\n",
    "#     fine_tuning(5, \"dt_model_new_trained_V12.pth\", \"dt_model_new_trained_V10_add_20.pth\", \"Added_Trajectoy_Data_*.pkl\", 200000);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#ONNX 로 변환\n",
    "# 모델 초기화 (학습 시 사용한 차원과 동일하게)\n",
    "obs_dim = 9\n",
    "act_dim = 3\n",
    "hidden = 256\n",
    "seq_len = 32\n",
    "\n",
    "model = DecisionTransformer(obs_dim=9, act_dim=3, hidden=256, n_layers=4, n_heads=4, max_len = 4096)\n",
    "model.load_state_dict(torch.load(\"dt_model_new_trained_V12.pth\"))\n",
    "#model.load_state_dict(torch.load(\"dt_model_epoch_2.pth\"))\n",
    "model.eval()\n",
    "\n",
    "# 더미 입력 생성 (batch=1, seq_len=32)\n",
    "dummy_obs = torch.randn(1, seq_len, obs_dim)\n",
    "dummy_act = torch.randn(1, seq_len, act_dim)\n",
    "dummy_rtg = torch.randn(1, seq_len, 1)\n",
    "dummy_ts = torch.arange(seq_len).unsqueeze(0)  # (1, seq_len)\n",
    "\n",
    "# ONNX로 내보내기\n",
    "torch.onnx.export(\n",
    "    model,\n",
    "    (dummy_obs, dummy_act, dummy_rtg, dummy_ts),\n",
    "    \"dt_model_new_trained_V12.onnx\",\n",
    "    input_names=[\"observations\", \"actions\", \"returns_to_go\", \"timesteps\"],\n",
    "    output_names=[\"predicted_actions\"],\n",
    "    dynamic_axes={\n",
    "        \"observations\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"returns_to_go\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"timesteps\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"predicted_actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "    },\n",
    "    opset_version=17\n",
    ")\n",
    "\n",
    "print(\"[INFO] ONNX export complete: dt_model_new_trained_V12.onnx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Timestpes 에러 변환\n",
    "import onnx\n",
    "from onnx import helper, TensorProto\n",
    "\n",
    "model = onnx.load(\"dt_model_new_trained_V12.onnx\")\n",
    "for input in model.graph.input:\n",
    "    if input.name == \"timesteps\":\n",
    "        input.type.tensor_type.elem_type = TensorProto.INT32\n",
    "onnx.save(model, \"dt_model_new_trained_V12_int.onnx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import importlib\n",
    "import FineTuningV\n",
    "importlib.reload(FineTuningV)\n",
    "from FineTuningV import fine_tuningV\n",
    "\n",
    "if __name__ == \"__main__\":\n",
    "    fine_tuningV()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#ONNX 로 변환\n",
    "# 모델 초기화 (학습 시 사용한 차원과 동일하게)\n",
    "obs_dim = 9\n",
    "act_dim = 3\n",
    "hidden = 128\n",
    "seq_len = 32\n",
    "\n",
    "model = DecisionTransformer(obs_dim=9, act_dim=3, hidden=128, n_layers=3, n_heads=4, max_len = 32)\n",
    "model.load_state_dict(torch.load(\"dt_model_Trained_OnlyV_V3.pth\"))\n",
    "#model.load_state_dict(torch.load(\"dt_model_epoch_2.pth\"))\n",
    "model.eval()\n",
    "\n",
    "# 더미 입력 생성 (batch=1, seq_len=32)\n",
    "dummy_obs = torch.randn(1, seq_len, obs_dim)\n",
    "dummy_act = torch.randn(1, seq_len, act_dim)\n",
    "dummy_rtg = torch.randn(1, seq_len, 1)\n",
    "dummy_ts = torch.arange(seq_len).unsqueeze(0)  # (1, seq_len)\n",
    "\n",
    "# ONNX로 내보내기\n",
    "torch.onnx.export(\n",
    "    model,\n",
    "    (dummy_obs, dummy_act, dummy_rtg, dummy_ts),\n",
    "    \"dt_model_Trained_OnlyV3.onnx\",\n",
    "    input_names=[\"observations\", \"actions\", \"returns_to_go\", \"timesteps\"],\n",
    "    output_names=[\"predicted_actions\"],\n",
    "    dynamic_axes={\n",
    "        \"observations\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"returns_to_go\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"timesteps\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "        \"predicted_actions\": {0: \"batch_size\", 1: \"seq_len\"},\n",
    "    },\n",
    "    opset_version=17\n",
    ")\n",
    "\n",
    "print(\"[INFO] ONNX export complete: dt_model_Trained_OnlyV3.onnx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Timestpes 에러 변환\n",
    "import onnx\n",
    "from onnx import helper, TensorProto\n",
    "\n",
    "model = onnx.load(\"dt_model_Trained_OnlyV3.onnx\")\n",
    "for input in model.graph.input:\n",
    "    if input.name == \"timesteps\":\n",
    "        input.type.tensor_type.elem_type = TensorProto.INT32\n",
    "onnx.save(model, \"dt_model_int32_v10_trained_AddedV.onnx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 487
    },
    "executionInfo": {
     "elapsed": 164,
     "status": "error",
     "timestamp": 1763481364612,
     "user": {
      "displayName": "sgoo k",
      "userId": "16229481617978281248"
     },
     "user_tz": -540
    },
    "id": "Vi6mQGXod5cg",
    "outputId": "8b017b28-d545-4010-8f45-90eb558d8493"
   },
   "outputs": [],
   "source": [
    "torch.onnx.export(\n",
    "    model,\n",
    "    dummy_input,\n",
    "    \"dt_model_12.onnx\",\n",
    "    input_names=[\"observations\", \"actions\", \"returns_to_go\", \"timesteps\"],\n",
    "    output_names=[\"predicted_actions\"],\n",
    "    dynamic_axes={\n",
    "        \"observations\": {1: \"seq_len\"},\n",
    "        \"actions\": {1: \"seq_len\"},\n",
    "        \"returns_to_go\": {1: \"seq_len\"},\n",
    "        \"timesteps\": {1: \"seq_len\"},\n",
    "        \"predicted_actions\": {1: \"seq_len\"}\n",
    "    },\n",
    "    opset_version=17,\n",
    "    verbose=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] Saved Fig1_Model_Comparison.png\n",
      "[INFO] Saved Fig2_RTG_Analysis.png\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x800 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "# Set style for academic paper\n",
    "sns.set_theme(style=\"whitegrid\")\n",
    "plt.rcParams['font.family'] = 'serif' # Times New Roman style often preferred in papers\n",
    "plt.rcParams['font.size'] = 11\n",
    "\n",
    "def save_plot(filename):\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(filename, dpi=300, bbox_inches='tight')\n",
    "    print(f\"[INFO] Saved {filename}\")\n",
    "# ==========================================\n",
    "# Data Preparation\n",
    "# ==========================================\n",
    "# 1. Model Comparison Data\n",
    "models = ['V2-SC', 'V3-C', 'PPO', 'V1-S']\n",
    "win_rates = [68.0, 58.0, 46.0, .0]\n",
    "rewards = [49.33, 42.87, 49.28, 29.48]\n",
    "steps = [819, 838, 895, 915]\n",
    "accuracies = [1.13, 0.95, 1.09, 0.51]\n",
    "# 2. RTG Analysis Data\n",
    "rtg_values = ['RTG 55', 'RTG 30', 'RTG 20', 'RTG 10']\n",
    "rtg_wins = [68.0, 0.0, 0.0, 2.0]\n",
    "rtg_rewards = [50.34, 27.36, 14.38, 14.08]\n",
    "# ==========================================\n",
    "# Figure 1: Overall Performance Comparison\n",
    "# ==========================================\n",
    "def plot_model_comparison():\n",
    "    # We will create a 2x2 grid for 4 metrics\n",
    "    fig, axes = plt.subplots(2, 2, figsize=(10, 8))\n",
    "    \n",
    "    # Common Bar Kwargs\n",
    "    bar_width = 0.6\n",
    "    palette = sns.color_palette(\"viridis\", len(models))\n",
    "    # Highlight the best model (Index 0)\n",
    "    colors = [palette[0] if i==0 else palette[2] for i in range(len(models))]\n",
    "    colors[2] = 'gray' # PPO as baseline\n",
    "    \n",
    "    # A. Win Rate\n",
    "    ax = axes[0, 0]\n",
    "    bars = ax.bar(models, win_rates, color=colors, width=bar_width, edgecolor='black', alpha=0.9)\n",
    "    ax.set_title('(a) Success Rate (%)', fontweight='bold', pad=10)\n",
    "    ax.set_ylabel('Win Rate (%)')\n",
    "    ax.set_ylim(0, 80)\n",
    "    # Add values\n",
    "    for bar in bars:\n",
    "        ax.text(bar.get_x() + bar.get_width()/2, bar.get_height()+1, f'{int(bar.get_height())}%', ha='center', fontsize=9)\n",
    "    ax.tick_params(axis='x', rotation=20)\n",
    "    # B. Average Reward\n",
    "    ax = axes[0, 1]\n",
    "    bars = ax.bar(models, rewards, color=colors, width=bar_width, edgecolor='black', alpha=0.9)\n",
    "    ax.set_title('(b) Average Cumulative Reward', fontweight='bold', pad=10)\n",
    "    ax.set_ylabel('Reward')\n",
    "    ax.set_ylim(0, 60)\n",
    "    for bar in bars:\n",
    "        ax.text(bar.get_x() + bar.get_width()/2, bar.get_height()+1, f'{bar.get_height():.1f}', ha='center', fontsize=9)\n",
    "    ax.tick_params(axis='x', rotation=20)\n",
    "    # C. Steps to Kill (Lower is Better)\n",
    "    ax = axes[1, 0]\n",
    "    bars = ax.bar(models, steps, color=colors, width=bar_width, edgecolor='black', alpha=0.95)\n",
    "    ax.set_title('(c) Avg. Steps to Victory (Lower is Better)', fontweight='bold', pad=10)\n",
    "    ax.set_ylabel('Steps')\n",
    "    ax.set_ylim(700, 1000)\n",
    "    for bar in bars:\n",
    "        ax.text(bar.get_x() + bar.get_width()/2, bar.get_height()+10, f'{int(bar.get_height())}', ha='center', fontsize=9)\n",
    "    ax.tick_params(axis='x', rotation=20)\n",
    "    # D. Accuracy (Arbitrary Unit from Log)\n",
    "    ax = axes[1, 1]\n",
    "    bars = ax.bar(models, accuracies, color=colors, width=bar_width, edgecolor='black', alpha=0.9)\n",
    "    ax.set_title('(d) Aiming Accuracy Score', fontweight='bold', pad=10)\n",
    "    ax.set_ylabel('Score')\n",
    "    ax.set_ylim(0, 1.4)\n",
    "    for bar in bars:\n",
    "        ax.text(bar.get_x() + bar.get_width()/2, bar.get_height()+0.05, f'{bar.get_height():.2f}', ha='center', fontsize=9)\n",
    "    ax.tick_params(axis='x', rotation=20)\n",
    "    plt.suptitle(\"Figure 1. Comparison of DT Variants vs PPO Baseline\", fontsize=14, y=0.98)\n",
    "    save_plot('Fig1_Model_Comparison.png')\n",
    "# ==========================================\n",
    "# Figure 2: RTG Sensitivity\n",
    "# ==========================================\n",
    "def plot_rtg_analysis():\n",
    "    fig, ax1 = plt.subplots(figsize=(7, 5))\n",
    "    x = np.arange(len(rtg_values))\n",
    "    width = 0.5\n",
    "    \n",
    "    # Dual Axis Plot\n",
    "    # Bar for Reward\n",
    "    color1 = 'tab:blue'\n",
    "    bars = ax1.bar(x, rtg_rewards, width, label='Avg. Reward', color=color1, alpha=0.7, edgecolor='black')\n",
    "    ax1.set_xlabel('Target Return (RTG)', fontweight='bold')\n",
    "    ax1.set_ylabel('Average Reward', color=color1, fontweight='bold')\n",
    "    ax1.tick_params(axis='y', labelcolor=color1)\n",
    "    ax1.set_xticks(x)\n",
    "    ax1.set_xticklabels(rtg_values)\n",
    "    ax1.set_ylim(0, 60)\n",
    "    # Line for Win Rate\n",
    "    ax2 = ax1.twinx()\n",
    "    color2 = 'tab:red'\n",
    "    ax2.plot(x, rtg_wins, color=color2, marker='o', linewidth=3, label='Win Rate (%)')\n",
    "    ax2.set_ylabel('Win Rate (%)', color=color2, fontweight='bold')\n",
    "    ax2.tick_params(axis='y', labelcolor=color2)\n",
    "    ax2.set_ylim(-5, 80)\n",
    "    # Annotations\n",
    "    for i, v in enumerate(rtg_rewards):\n",
    "        ax1.text(i, v + 1, f'{v:.1f}', ha='center', color='black', fontsize=10)\n",
    "    \n",
    "    for i, v in enumerate(rtg_wins):\n",
    "        ax2.text(i, v + 3, f'{int(v)}%', ha='center', color=color2, fontweight='bold', fontsize=10)\n",
    "    plt.title(\"Figure 2. Impact of Returns-To-Go (RTG) on Performance\", fontsize=13, pad=15)\n",
    "    fig.tight_layout()\n",
    "    save_plot('Fig2_RTG_Analysis.png')\n",
    "if __name__ == \"__main__\":\n",
    "    plot_model_comparison()\n",
    "    plot_rtg_analysis()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "authorship_tag": "ABX9TyNglFYnIAAo1EQRA/3bNYVL",
   "gpuType": "T4",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python [conda env:base] *",
   "language": "python",
   "name": "conda-base-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}