tutorials/TrainData_biofounction_context.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "e48c5f86-f0aa-4963-8913-604443fa0a3c",
   "metadata": {},
   "source": [
    "# Tutorial: Train single cell data using biofounction conext method"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5aac073f-ef61-49f9-bf3e-13b2a3137121",
   "metadata": {},
   "source": [
    "## Introduction\n",
    "\n",
    "This tutorial demonstrates how to use **scGenAI** to train a model for predicting cell types from single-cell RNA sequencing (scRNA-seq) data. In this example, we apply the **biofounction context** method to generate input sequences.\n",
    "\n",
    "**Note:**\n",
    "1. The dataset used for this tutorial consists of **40 cells**, selected for demonstration purposes. For practical applications, it is recommended to use a larger, more representative dataset.\n",
    "2. In this tutorial, we will load the model template directly from the cache folder.\n",
    "3. While the model in this example is trained to predict cell types, it can be used to predict any feature in practical applications.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8294154a-4036-4788-b358-0e91342ac224",
   "metadata": {},
   "source": [
    "## Train the Model Using Biofounction Context (Multiple GPU)\n",
    "\n",
    "The training process can be easily initiated by calling the `run_training_from_config` function in **scGenAI**, using the input files specified in a configuration YAML file. Example template YAML files are available in the `scgenai/examples/config_templates` folder of the package. These templates provide the necessary structure for defining training parameters, model configurations, and input data locations.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2989a022-6bea-4f9c-861d-452b0d85e11f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scGenAI.training.train import run_training_from_config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5a140f4c-6d3c-4a18-9f15-87d148179ac8",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model_dir: /home/ubuntu/scgenai/examples/testrun/test1/\n",
      "train_file: /home/ubuntu/scgenai/examples/data/example_train.h5ad\n",
      "val_file: /home/ubuntu/scgenai/examples/data/example_val.h5ad\n",
      "gmtfile: /home/ubuntu/scgenai/examples/data/c8.BONE_MARROW.v2023.2.Hs.symbols.gmt\n",
      "log_dir: /home/ubuntu/scgenai/examples/logs/\n",
      "output_dir: /home/ubuntu/scgenai/examples/output\n",
      "seed: 1314521\n",
      "cache_dir: /home/ubuntu/scgenai/examples/tmp/cache\n",
      "mode: Train\n",
      "min_cells: 2\n",
      "target_feature: ct\n",
      "multiomics: No\n",
      "keepIntermediateFiles: No\n",
      "savelog: Yes\n",
      "savesetting: Yes\n",
      "model_backbone_name: llama\n",
      "model_backbone_size: small\n",
      "context_method: biofounction\n",
      "num_bins: 10\n",
      "optimizer: AdamW\n",
      "weight_decay: 0.01\n",
      "depth: 2\n",
      "max_length: 1024\n",
      "batch_size: 1\n",
      "learning_rate: 1e-05\n",
      "num_epochs: 20\n",
      "world_size: 1\n",
      "evaluate_during_training: True\n",
      "hidden_size: 256\n",
      "num_layers: 6\n",
      "num_heads: 4\n",
      "intermediate_size: 1024\n",
      "stride: 512\n",
      "Processed 40/40 cells\r"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-10-15 21:38:52.693 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 0, Total GPUs used: 4\n",
      "2024-10-15 21:38:52.700 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 1, Total GPUs used: 4\n",
      "2024-10-15 21:38:52.707 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 2, Total GPUs used: 4\n",
      "2024-10-15 21:38:52.708 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 3, Total GPUs used: 4\n",
      "2024-10-15 21:38:54.957 | INFO     | scGenAI.training.train:model_train_and_eval:214 - Total number of trainable parameters: 74945796; Max Sequence Legth: 512 genes\n",
      "Epoch 1/20: 100%|██████████| 46/46 [00:09<00:00,  4.78it/s, GPU_ID=0, train_accuracy=0.587, train_loss=1.05]]\n",
      "Epoch 1/20: 100%|██████████| 46/46 [00:09<00:00,  4.78it/s, GPU_ID=1, train_accuracy=0.565, train_loss=0.898]\n",
      "\n",
      "Epoch 1/20: 100%|██████████| 46/46 [00:09<00:00,  4.78it/s, GPU_ID=2, train_accuracy=0.543, train_loss=1.16]\n",
      "Evaluating Epoch 1/20: 100%|██████████| 184/184 [00:04<00:00, 38.78it/s]\n",
      "Evaluating Epoch 1/20: 100%|██████████| 184/184 [00:04<00:00, 38.69it/s]\n",
      "Evaluating Epoch 1/20: 100%|██████████| 184/184 [00:04<00:00, 38.69it/s]\n",
      "2024-10-15 21:39:09.901 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 1/20, Train Loss: 1.0489, Train Accuracy: 0.5870, Context-level Val Loss: 0.8362, Context-level Val Accuracy: 0.6467\n",
      "Evaluating Epoch 1/20: 100%|██████████| 184/184 [00:04<00:00, 38.30it/s]\n",
      "Epoch 2/20:   0%|          | 0/46 [00:00<?, ?it/s]2024-10-15 21:39:10.474 | INFO     | scGenAI.utils.distributed:savemodel:54 - Model state saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model/scGenAI_model.pt\n",
      "2024-10-15 21:39:10.625 | INFO     | scGenAI.utils.distributed:savemodel:59 - Tokenizer saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model\n",
      "2024-10-15 21:39:10.628 | INFO     | scGenAI.utils.distributed:savemodel:75 - Custom vocab and label encoder saved\n",
      "2024-10-15 21:39:10.629 | INFO     | scGenAI.utils.distributed:savemodel:81 - Model configuration saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model\n",
      "Epoch 2/20: 100%|██████████| 46/46 [00:09<00:00,  4.71it/s, GPU_ID=1, train_accuracy=0.761, train_loss=0.595]\n",
      "Epoch 2/20: 100%|██████████| 46/46 [00:09<00:00,  4.73it/s, GPU_ID=3, train_accuracy=0.696, train_loss=0.708]\n",
      "Epoch 2/20: 100%|██████████| 46/46 [00:09<00:00,  5.05it/s, GPU_ID=0, train_accuracy=0.739, train_loss=0.769]\n",
      "Epoch 2/20: 100%|██████████| 46/46 [00:09<00:00,  4.70it/s, GPU_ID=2, train_accuracy=0.63, train_loss=0.906]\n",
      "Evaluating Epoch 2/20: 100%|██████████| 184/184 [00:04<00:00, 38.68it/s]\n",
      "2024-10-15 21:39:24.492 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 2/20, Train Loss: 0.7693, Train Accuracy: 0.7391, Context-level Val Loss: 0.8166, Context-level Val Accuracy: 0.6467\n",
      "Evaluating Epoch 2/20: 100%|██████████| 184/184 [00:04<00:00, 38.63it/s]\n",
      "Evaluating Epoch 2/20: 100%|██████████| 184/184 [00:04<00:00, 38.47it/s]\n",
      "Evaluating Epoch 2/20: 100%|██████████| 184/184 [00:04<00:00, 38.36it/s]\n",
      "Epoch 3/20:   0%|          | 0/46 [00:00<?, ?it/s]2024-10-15 21:39:26.313 | INFO     | scGenAI.utils.distributed:savemodel:54 - Model state saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model/scGenAI_model.pt\n",
      "2024-10-15 21:39:26.473 | INFO     | scGenAI.utils.distributed:savemodel:59 - Tokenizer saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model\n",
      "2024-10-15 21:39:26.476 | INFO     | scGenAI.utils.distributed:savemodel:75 - Custom vocab and label encoder saved\n",
      "2024-10-15 21:39:26.477 | INFO     | scGenAI.utils.distributed:savemodel:81 - Model configuration saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model\n",
      "Epoch 3/20: 100%|██████████| 46/46 [00:11<00:00,  4.18it/s, GPU_ID=1, train_accuracy=0.804, train_loss=0.527]\n",
      "\n",
      "Epoch 3/20: 100%|██████████| 46/46 [00:10<00:00,  4.18it/s, GPU_ID=3, train_accuracy=0.739, train_loss=0.654]\n",
      "Epoch 3/20: 100%|██████████| 46/46 [00:11<00:00,  4.17it/s, GPU_ID=2, train_accuracy=0.696, train_loss=0.809]\n",
      "Evaluating Epoch 3/20: 100%|██████████| 184/184 [00:04<00:00, 38.68it/s]\n",
      "2024-10-15 21:39:40.347 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 3/20, Train Loss: 0.7049, Train Accuracy: 0.7174, Context-level Val Loss: 0.8217, Context-level Val Accuracy: 0.6685\n",
      "Evaluating Epoch 3/20: 100%|██████████| 184/184 [00:04<00:00, 38.53it/s]\n",
      "Evaluating Epoch 3/20: 100%|██████████| 184/184 [00:04<00:00, 38.38it/s]\n",
      "Evaluating Epoch 3/20: 100%|██████████| 184/184 [00:04<00:00, 37.97it/s]\n",
      "Epoch 4/20:   0%|          | 0/46 [00:00<?, ?it/s]2024-10-15 21:39:42.180 | INFO     | scGenAI.utils.distributed:savemodel:54 - Model state saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model/scGenAI_model.pt\n",
      "2024-10-15 21:39:42.335 | INFO     | scGenAI.utils.distributed:savemodel:59 - Tokenizer saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model\n",
      "2024-10-15 21:39:42.341 | INFO     | scGenAI.utils.distributed:savemodel:75 - Custom vocab and label encoder saved\n",
      "2024-10-15 21:39:42.342 | INFO     | scGenAI.utils.distributed:savemodel:81 - Model configuration saved to /home/ubuntu/scgenai/examples/testrun/test1/best_model\n",
      "Epoch 4/20: 100%|██████████| 46/46 [00:09<00:00,  5.05it/s, GPU_ID=0, train_accuracy=0.739, train_loss=0.643]\n",
      "Epoch 4/20: 100%|██████████| 46/46 [00:11<00:00,  4.18it/s, GPU_ID=1, train_accuracy=0.848, train_loss=0.428]\n",
      "Epoch 4/20: 100%|██████████| 46/46 [00:10<00:00,  4.20it/s, GPU_ID=3, train_accuracy=0.761, train_loss=0.57]\n",
      "Epoch 4/20: 100%|██████████| 46/46 [00:11<00:00,  4.17it/s, GPU_ID=2, train_accuracy=0.761, train_loss=0.714]\n",
      "Evaluating Epoch 4/20: 100%|██████████| 184/184 [00:04<00:00, 38.59it/s]\n",
      "2024-10-15 21:39:56.223 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 4/20, Train Loss: 0.6435, Train Accuracy: 0.7391, Context-level Val Loss: 0.8327, Context-level Val Accuracy: 0.6467\n",
      "Evaluating Epoch 4/20: 100%|██████████| 184/184 [00:04<00:00, 38.41it/s]\n",
      "Evaluating Epoch 4/20: 100%|██████████| 184/184 [00:04<00:00, 38.24it/s]\n",
      "Evaluating Epoch 4/20: 100%|██████████| 184/184 [00:04<00:00, 37.84it/s]\n",
      "Epoch 5/20: 100%|██████████| 46/46 [00:09<00:00,  5.01it/s, GPU_ID=1, train_accuracy=0.848, train_loss=0.305]\n",
      "Epoch 5/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=0, train_accuracy=0.804, train_loss=0.532]\n",
      "Epoch 5/20: 100%|██████████| 46/46 [00:09<00:00,  5.04it/s, GPU_ID=3, train_accuracy=0.848, train_loss=0.448]\n",
      "Epoch 5/20: 100%|██████████| 46/46 [00:09<00:00,  5.00it/s, GPU_ID=2, train_accuracy=0.826, train_loss=0.581]\n",
      "Evaluating Epoch 5/20: 100%|██████████| 184/184 [00:04<00:00, 38.47it/s]\n",
      "2024-10-15 21:40:10.294 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 5/20, Train Loss: 0.5315, Train Accuracy: 0.8043, Context-level Val Loss: 0.8533, Context-level Val Accuracy: 0.6087\n",
      "Evaluating Epoch 5/20: 100%|██████████| 184/184 [00:04<00:00, 38.33it/s]\n",
      "Evaluating Epoch 5/20: 100%|██████████| 184/184 [00:04<00:00, 38.16it/s]\n",
      "Evaluating Epoch 5/20: 100%|██████████| 184/184 [00:04<00:00, 37.72it/s]\n",
      "Epoch 6/20: 100%|██████████| 46/46 [00:09<00:00,  5.04it/s, GPU_ID=3, train_accuracy=0.935, train_loss=0.323] \n",
      "Epoch 6/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=0, train_accuracy=0.891, train_loss=0.35]\n",
      "Epoch 6/20: 100%|██████████| 46/46 [00:09<00:00,  5.01it/s, GPU_ID=1, train_accuracy=0.978, train_loss=0.181]\n",
      "Epoch 6/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=2, train_accuracy=0.87, train_loss=0.427]\n",
      "Evaluating Epoch 6/20: 100%|██████████| 184/184 [00:04<00:00, 38.45it/s]\n",
      "2024-10-15 21:40:24.371 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 6/20, Train Loss: 0.3499, Train Accuracy: 0.8913, Context-level Val Loss: 0.8995, Context-level Val Accuracy: 0.6033\n",
      "Evaluating Epoch 6/20: 100%|██████████| 184/184 [00:04<00:00, 38.20it/s]\n",
      "Evaluating Epoch 6/20: 100%|██████████| 184/184 [00:04<00:00, 38.09it/s]\n",
      "Evaluating Epoch 6/20: 100%|██████████| 184/184 [00:04<00:00, 37.56it/s]\n",
      "Epoch 7/20: 100%|██████████| 46/46 [00:09<00:00,  5.00it/s, GPU_ID=1, train_accuracy=0.978, train_loss=0.112] \n",
      "Epoch 7/20: 100%|██████████| 46/46 [00:09<00:00,  5.04it/s, GPU_ID=3, train_accuracy=0.935, train_loss=0.227]\n",
      "\n",
      "Epoch 7/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=2, train_accuracy=0.935, train_loss=0.275]\n",
      "Evaluating Epoch 7/20: 100%|██████████| 184/184 [00:04<00:00, 38.31it/s]\n",
      "2024-10-15 21:40:38.480 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 7/20, Train Loss: 0.1792, Train Accuracy: 0.9783, Context-level Val Loss: 0.9348, Context-level Val Accuracy: 0.5815\n",
      "Evaluating Epoch 7/20: 100%|██████████| 184/184 [00:04<00:00, 38.17it/s]\n",
      "Evaluating Epoch 7/20: 100%|██████████| 184/184 [00:04<00:00, 38.06it/s]\n",
      "Evaluating Epoch 7/20: 100%|██████████| 184/184 [00:04<00:00, 37.49it/s]\n",
      "Epoch 8/20: 100%|██████████| 46/46 [00:09<00:00,  5.00it/s, GPU_ID=1, train_accuracy=1, train_loss=0.0875]47] \n",
      "Epoch 8/20: 100%|██████████| 46/46 [00:09<00:00,  5.04it/s, GPU_ID=3, train_accuracy=0.957, train_loss=0.147]\n",
      "Epoch 8/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=0, train_accuracy=1, train_loss=0.0914]\n",
      "Epoch 8/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=2, train_accuracy=1, train_loss=0.155]\n",
      "Evaluating Epoch 8/20: 100%|██████████| 184/184 [00:04<00:00, 38.33it/s]\n",
      "2024-10-15 21:40:52.582 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 8/20, Train Loss: 0.0914, Train Accuracy: 1.0000, Context-level Val Loss: 0.9503, Context-level Val Accuracy: 0.6196\n",
      "Evaluating Epoch 8/20: 100%|██████████| 184/184 [00:04<00:00, 38.13it/s]\n",
      "Evaluating Epoch 8/20: 100%|██████████| 184/184 [00:04<00:00, 37.98it/s]\n",
      "Evaluating Epoch 8/20: 100%|██████████| 184/184 [00:04<00:00, 37.47it/s]\n",
      "Epoch 9/20: 100%|██████████| 46/46 [00:09<00:00,  5.00it/s, GPU_ID=1, train_accuracy=1, train_loss=0.0563] \n",
      "Epoch 9/20: 100%|██████████| 46/46 [00:09<00:00,  5.03it/s, GPU_ID=3, train_accuracy=1, train_loss=0.0839]\n",
      "Epoch 9/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=0, train_accuracy=1, train_loss=0.0475]\n",
      "Epoch 9/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=2, train_accuracy=1, train_loss=0.0697]\n",
      "Evaluating Epoch 9/20: 100%|██████████| 184/184 [00:04<00:00, 38.29it/s]\n",
      "2024-10-15 21:41:06.700 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 9/20, Train Loss: 0.0475, Train Accuracy: 1.0000, Context-level Val Loss: 0.9796, Context-level Val Accuracy: 0.6467\n",
      "Evaluating Epoch 9/20: 100%|██████████| 184/184 [00:04<00:00, 38.12it/s]\n",
      "Evaluating Epoch 9/20: 100%|██████████| 184/184 [00:04<00:00, 37.95it/s]\n",
      "Evaluating Epoch 9/20: 100%|██████████| 184/184 [00:04<00:00, 37.29it/s]\n",
      "Epoch 10/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=1, train_accuracy=1, train_loss=0.0513]693]\n",
      "Epoch 10/20: 100%|██████████| 46/46 [00:09<00:00,  4.96it/s, GPU_ID=0, train_accuracy=1, train_loss=0.0324]693]\n",
      "\n",
      "Epoch 10/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=2, train_accuracy=1, train_loss=0.0457]\n",
      "Evaluating Epoch 10/20: 100%|██████████| 184/184 [00:04<00:00, 38.22it/s]\n",
      "2024-10-15 21:41:20.850 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 10/20, Train Loss: 0.0324, Train Accuracy: 1.0000, Context-level Val Loss: 0.9768, Context-level Val Accuracy: 0.6304\n",
      "Evaluating Epoch 10/20: 100%|██████████| 184/184 [00:04<00:00, 38.03it/s]\n",
      "Evaluating Epoch 10/20: 100%|██████████| 184/184 [00:04<00:00, 37.90it/s]\n",
      "Evaluating Epoch 10/20: 100%|██████████| 184/184 [00:04<00:00, 37.31it/s]\n",
      "Epoch 11/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=1, train_accuracy=1, train_loss=0.0197] \n",
      "Epoch 11/20: 100%|██████████| 46/46 [00:09<00:00,  5.03it/s, GPU_ID=3, train_accuracy=1, train_loss=0.0353]\n",
      "Epoch 11/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=0, train_accuracy=1, train_loss=0.0209]\n",
      "Epoch 11/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=2, train_accuracy=1, train_loss=0.0309]\n",
      "Evaluating Epoch 11/20: 100%|██████████| 184/184 [00:04<00:00, 38.18it/s]\n",
      "2024-10-15 21:41:34.998 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 11/20, Train Loss: 0.0209, Train Accuracy: 1.0000, Context-level Val Loss: 1.0232, Context-level Val Accuracy: 0.6522\n",
      "Evaluating Epoch 11/20: 100%|██████████| 184/184 [00:04<00:00, 38.04it/s]\n",
      "Evaluating Epoch 11/20: 100%|██████████| 184/184 [00:04<00:00, 37.92it/s]\n",
      "Evaluating Epoch 11/20: 100%|██████████| 184/184 [00:04<00:00, 37.31it/s]\n",
      "Epoch 12/20: 100%|██████████| 46/46 [00:09<00:00,  5.03it/s, GPU_ID=3, train_accuracy=1, train_loss=0.0154] \n",
      "Epoch 12/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=1, train_accuracy=1, train_loss=0.0135]\n",
      "Epoch 12/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=0, train_accuracy=1, train_loss=0.0173]\n",
      "Epoch 12/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=2, train_accuracy=1, train_loss=0.0192]\n",
      "Evaluating Epoch 12/20: 100%|██████████| 184/184 [00:04<00:00, 38.24it/s]\n",
      "2024-10-15 21:41:49.126 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 12/20, Train Loss: 0.0173, Train Accuracy: 1.0000, Context-level Val Loss: 1.0524, Context-level Val Accuracy: 0.6630\n",
      "Evaluating Epoch 12/20: 100%|██████████| 184/184 [00:04<00:00, 38.04it/s]\n",
      "Evaluating Epoch 12/20: 100%|██████████| 184/184 [00:04<00:00, 37.88it/s]\n",
      "Evaluating Epoch 12/20: 100%|██████████| 184/184 [00:04<00:00, 37.22it/s]\n",
      "Epoch 13/20: 100%|██████████| 46/46 [00:09<00:00,  5.04it/s, GPU_ID=3, train_accuracy=1, train_loss=0.0117] \n",
      "Epoch 13/20: 100%|██████████| 46/46 [00:09<00:00,  4.96it/s, GPU_ID=0, train_accuracy=1, train_loss=0.0122]\n",
      "Epoch 13/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=1, train_accuracy=1, train_loss=0.0106]\n",
      "Epoch 13/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=2, train_accuracy=1, train_loss=0.0132]\n",
      "Evaluating Epoch 13/20: 100%|██████████| 184/184 [00:04<00:00, 38.22it/s]\n",
      "2024-10-15 21:42:03.270 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 13/20, Train Loss: 0.0122, Train Accuracy: 1.0000, Context-level Val Loss: 1.0628, Context-level Val Accuracy: 0.6576\n",
      "Evaluating Epoch 13/20: 100%|██████████| 184/184 [00:04<00:00, 37.98it/s]\n",
      "Evaluating Epoch 13/20: 100%|██████████| 184/184 [00:04<00:00, 37.81it/s]\n",
      "Evaluating Epoch 13/20: 100%|██████████| 184/184 [00:04<00:00, 37.27it/s]\n",
      "Epoch 14/20: 100%|██████████| 46/46 [00:09<00:00,  4.96it/s, GPU_ID=0, train_accuracy=1, train_loss=0.0096]]\n",
      "\n",
      "Epoch 14/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=1, train_accuracy=1, train_loss=0.00871]\n",
      "Epoch 14/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=2, train_accuracy=1, train_loss=0.0103]\n",
      "Evaluating Epoch 14/20: 100%|██████████| 184/184 [00:04<00:00, 38.21it/s]\n",
      "2024-10-15 21:42:17.418 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 14/20, Train Loss: 0.0096, Train Accuracy: 1.0000, Context-level Val Loss: 1.0715, Context-level Val Accuracy: 0.6630\n",
      "Evaluating Epoch 14/20: 100%|██████████| 184/184 [00:04<00:00, 37.97it/s]\n",
      "Evaluating Epoch 14/20: 100%|██████████| 184/184 [00:04<00:00, 37.86it/s]\n",
      "Evaluating Epoch 14/20: 100%|██████████| 184/184 [00:04<00:00, 37.27it/s]\n",
      "Epoch 15/20: 100%|██████████| 46/46 [00:09<00:00,  5.03it/s, GPU_ID=3, train_accuracy=1, train_loss=0.00809]\n",
      "Epoch 15/20: 100%|██████████| 46/46 [00:09<00:00,  4.96it/s, GPU_ID=0, train_accuracy=1, train_loss=0.00826]\n",
      "Epoch 15/20: 100%|██████████| 46/46 [00:09<00:00,  4.99it/s, GPU_ID=1, train_accuracy=1, train_loss=0.00742]\n",
      "Epoch 15/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=2, train_accuracy=1, train_loss=0.00868]\n",
      "Evaluating Epoch 15/20: 100%|██████████| 184/184 [00:04<00:00, 38.13it/s]\n",
      "2024-10-15 21:42:31.579 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 15/20, Train Loss: 0.0083, Train Accuracy: 1.0000, Context-level Val Loss: 1.0860, Context-level Val Accuracy: 0.6630\n",
      "Evaluating Epoch 15/20: 100%|██████████| 184/184 [00:04<00:00, 38.00it/s]\n",
      "Evaluating Epoch 15/20: 100%|██████████| 184/184 [00:04<00:00, 37.86it/s]\n",
      "Evaluating Epoch 15/20: 100%|██████████| 184/184 [00:04<00:00, 37.25it/s]\n",
      "Epoch 16/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=1, train_accuracy=1, train_loss=0.00649]\n",
      "\n",
      "\n",
      "Epoch 16/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=2, train_accuracy=1, train_loss=0.0075]\n",
      "Evaluating Epoch 16/20: 100%|██████████| 184/184 [00:04<00:00, 38.14it/s]\n",
      "2024-10-15 21:42:45.741 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 16/20, Train Loss: 0.0072, Train Accuracy: 1.0000, Context-level Val Loss: 1.0990, Context-level Val Accuracy: 0.6576\n",
      "Evaluating Epoch 16/20: 100%|██████████| 184/184 [00:04<00:00, 37.97it/s]\n",
      "Evaluating Epoch 16/20: 100%|██████████| 184/184 [00:04<00:00, 37.88it/s]\n",
      "Evaluating Epoch 16/20: 100%|██████████| 184/184 [00:04<00:00, 37.29it/s]\n",
      "Epoch 17/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=0, train_accuracy=1, train_loss=0.00635] \n",
      "Epoch 17/20: 100%|██████████| 46/46 [00:09<00:00,  5.02it/s, GPU_ID=3, train_accuracy=1, train_loss=0.00621]\n",
      "\n",
      "Epoch 17/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=2, train_accuracy=1, train_loss=0.00657]\n",
      "Evaluating Epoch 17/20: 100%|██████████| 184/184 [00:04<00:00, 38.14it/s]\n",
      "2024-10-15 21:42:59.893 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 17/20, Train Loss: 0.0064, Train Accuracy: 1.0000, Context-level Val Loss: 1.1114, Context-level Val Accuracy: 0.6630\n",
      "Evaluating Epoch 17/20: 100%|██████████| 184/184 [00:04<00:00, 37.96it/s]\n",
      "Evaluating Epoch 17/20: 100%|██████████| 184/184 [00:04<00:00, 37.87it/s]\n",
      "Evaluating Epoch 17/20: 100%|██████████| 184/184 [00:04<00:00, 37.33it/s]\n",
      "Epoch 18/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=0, train_accuracy=1, train_loss=0.00565] \n",
      "Epoch 18/20: 100%|██████████| 46/46 [00:09<00:00,  5.03it/s, GPU_ID=3, train_accuracy=1, train_loss=0.00553]\n",
      "\n",
      "Epoch 18/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=2, train_accuracy=1, train_loss=0.00583]\n",
      "Evaluating Epoch 18/20: 100%|██████████| 184/184 [00:04<00:00, 38.10it/s]\n",
      "2024-10-15 21:43:14.043 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 18/20, Train Loss: 0.0057, Train Accuracy: 1.0000, Context-level Val Loss: 1.1233, Context-level Val Accuracy: 0.6576\n",
      "Evaluating Epoch 18/20: 100%|██████████| 184/184 [00:04<00:00, 37.96it/s]\n",
      "Evaluating Epoch 18/20: 100%|██████████| 184/184 [00:04<00:00, 37.86it/s]\n",
      "Evaluating Epoch 18/20: 100%|██████████| 184/184 [00:04<00:00, 37.24it/s]\n",
      "Epoch 19/20: 100%|██████████| 46/46 [00:09<00:00,  4.98it/s, GPU_ID=1, train_accuracy=1, train_loss=0.00465] \n",
      "Epoch 19/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=0, train_accuracy=1, train_loss=0.00507]\n",
      "Epoch 19/20: 100%|██████████| 46/46 [00:09<00:00,  5.03it/s, GPU_ID=3, train_accuracy=1, train_loss=0.00496]\n",
      "Epoch 19/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=2, train_accuracy=1, train_loss=0.00522]\n",
      "Evaluating Epoch 19/20: 100%|██████████| 184/184 [00:04<00:00, 38.15it/s]\n",
      "2024-10-15 21:43:28.193 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 19/20, Train Loss: 0.0051, Train Accuracy: 1.0000, Context-level Val Loss: 1.1346, Context-level Val Accuracy: 0.6467\n",
      "Evaluating Epoch 19/20: 100%|██████████| 184/184 [00:04<00:00, 37.96it/s]\n",
      "Evaluating Epoch 19/20: 100%|██████████| 184/184 [00:04<00:00, 37.85it/s]\n",
      "Evaluating Epoch 19/20: 100%|██████████| 184/184 [00:04<00:00, 37.13it/s]\n",
      "Epoch 20/20: 100%|██████████| 46/46 [00:09<00:00,  4.97it/s, GPU_ID=1, train_accuracy=1, train_loss=0.00422] \n",
      "Epoch 20/20: 100%|██████████| 46/46 [00:09<00:00,  4.95it/s, GPU_ID=0, train_accuracy=1, train_loss=0.00458]\n",
      "Epoch 20/20: 100%|██████████| 46/46 [00:09<00:00,  5.02it/s, GPU_ID=3, train_accuracy=1, train_loss=0.00449]\n",
      "Epoch 20/20: 100%|██████████| 46/46 [00:09<00:00,  4.96it/s, GPU_ID=2, train_accuracy=1, train_loss=0.00471]\n",
      "Evaluating Epoch 20/20: 100%|██████████| 184/184 [00:04<00:00, 38.18it/s]\n",
      "2024-10-15 21:43:42.368 | INFO     | scGenAI.training.train:evaluate_and_save_predictions:328 - Epoch 20/20, Train Loss: 0.0046, Train Accuracy: 1.0000, Context-level Val Loss: 1.1455, Context-level Val Accuracy: 0.6413\n",
      "Evaluating Epoch 20/20: 100%|██████████| 184/184 [00:04<00:00, 38.00it/s]\n",
      "Evaluating Epoch 20/20: 100%|██████████| 184/184 [00:04<00:00, 37.83it/s]\n",
      "Evaluating Epoch 20/20: 100%|██████████| 184/184 [00:04<00:00, 37.07it/s]\n",
      "2024-10-15 21:43:42.943 | INFO     | scGenAI.utils.distributed:savemodel:54 - Model state saved to /home/ubuntu/scgenai/examples/testrun/test1/last_model/scGenAI_model.pt\n",
      "2024-10-15 21:43:43.086 | INFO     | scGenAI.utils.distributed:savemodel:59 - Tokenizer saved to /home/ubuntu/scgenai/examples/testrun/test1/last_model\n",
      "2024-10-15 21:43:43.088 | INFO     | scGenAI.utils.distributed:savemodel:75 - Custom vocab and label encoder saved\n",
      "2024-10-15 21:43:43.089 | INFO     | scGenAI.utils.distributed:savemodel:81 - Model configuration saved to /home/ubuntu/scgenai/examples/testrun/test1/last_model\n",
      "2024-10-15 21:43:43.090 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "2024-10-15 21:43:43.090 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "2024-10-15 21:43:43.090 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "2024-10-15 21:43:43.090 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "\u001b[32m2024-10-15 21:43:44.315\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 1\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.418\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 1 - Validation Accuracy: 0.4500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.423\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 2\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.515\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 2 - Validation Accuracy: 0.4250\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.521\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 3\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.611\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 3 - Validation Accuracy: 0.5000\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.616\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 4\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.706\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 4 - Validation Accuracy: 0.4750\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.711\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 5\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.800\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 5 - Validation Accuracy: 0.4500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.805\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 6\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.896\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 6 - Validation Accuracy: 0.4750\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.901\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 7\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.990\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 7 - Validation Accuracy: 0.4750\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:44.995\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 8\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.085\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 8 - Validation Accuracy: 0.5000\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.090\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 9\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.179\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 9 - Validation Accuracy: 0.4750\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.184\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 10\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.271\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 10 - Validation Accuracy: 0.5500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.276\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 11\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.363\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 11 - Validation Accuracy: 0.5750\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.368\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 12\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.455\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 12 - Validation Accuracy: 0.5500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.460\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 13\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.549\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 13 - Validation Accuracy: 0.5250\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.554\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 14\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.642\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 14 - Validation Accuracy: 0.5250\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.647\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 15\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.736\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 15 - Validation Accuracy: 0.5250\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.741\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 16\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.829\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 16 - Validation Accuracy: 0.5500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.834\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 17\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.921\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 17 - Validation Accuracy: 0.5500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:45.926\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 18\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:46.013\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 18 - Validation Accuracy: 0.5500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:46.018\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 19\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:46.106\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 19 - Validation Accuracy: 0.5500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:46.111\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mmerge_predictions_across_gpus\u001b[0m:\u001b[36m147\u001b[0m - \u001b[1mMerging predictions for epoch 20\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:46.198\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mpostprocess_epoch\u001b[0m:\u001b[36m211\u001b[0m - \u001b[1mEpoch 20 - Validation Accuracy: 0.5500\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:46.204\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36msummarize_epochs\u001b[0m:\u001b[36m234\u001b[0m - \u001b[1mCompleted post-processing and summary of all epochs.\u001b[0m\n",
      "\u001b[32m2024-10-15 21:43:46.207\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mcombine_epoch_results\u001b[0m:\u001b[36m61\u001b[0m - \u001b[1mCombined results saved to /home/ubuntu/scgenai/examples/testrun/test1/combined_epoch_results.csv.\u001b[0m\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot saved to /home/ubuntu/scgenai/examples/testrun/test1/train_summary.pdf\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "config_file = '/home/ubuntu/scgenai/tutorials/yaml/config_biofounction_context_template_llama.yaml'\n",
    "run_training_from_config(config_file)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fe550d5-16c0-425a-bfd7-d257c1ecfe1d",
   "metadata": {},
   "source": [
    "## Post-Training Model and Results\n",
    "\n",
    "Upon completion of the training, the trained model will be saved in the `model_dir` as defined in the configuration file. Within this directory, two subfolders will be created: `best_model` and `last_model`. The `best_model` folder contains the model that achieved the highest training performance, while the `last_model` contains the model from the final epoch. Additionally, a `combined_epoch_results.csv` file will be generated in the `model_dir`, summarizing the training performance metrics across all epochs."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cc88e128-eaba-4557-8fca-eae0966e482d",
   "metadata": {},
   "source": [
    "### Training Summary\n",
    "\n",
    "In addition to the output summary files generated after training, the training summary can be directly accessed in Python by loading the saved configuration file.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "9cd73cce-7ddc-47ab-b59f-fc4ac955b57f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scGenAI.config import Config\n",
    "from scGenAI.data.postprocess import TrainSummary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9bc20420-8143-4ff7-b474-d4a3f6936991",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Epoch</th>\n",
       "      <th>Validation Accuracy</th>\n",
       "      <th>Train Loss</th>\n",
       "      <th>Train Accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.450</td>\n",
       "      <td>1.048929</td>\n",
       "      <td>0.586957</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>0.425</td>\n",
       "      <td>0.769282</td>\n",
       "      <td>0.739130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>0.500</td>\n",
       "      <td>0.704915</td>\n",
       "      <td>0.717391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>0.475</td>\n",
       "      <td>0.643454</td>\n",
       "      <td>0.739130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>0.450</td>\n",
       "      <td>0.531550</td>\n",
       "      <td>0.804348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>0.475</td>\n",
       "      <td>0.349914</td>\n",
       "      <td>0.891304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>0.475</td>\n",
       "      <td>0.179244</td>\n",
       "      <td>0.978261</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>0.500</td>\n",
       "      <td>0.091448</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>0.475</td>\n",
       "      <td>0.047549</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>0.550</td>\n",
       "      <td>0.032353</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>11</td>\n",
       "      <td>0.575</td>\n",
       "      <td>0.020925</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>12</td>\n",
       "      <td>0.550</td>\n",
       "      <td>0.017287</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>13</td>\n",
       "      <td>0.525</td>\n",
       "      <td>0.012231</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>14</td>\n",
       "      <td>0.525</td>\n",
       "      <td>0.009602</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>15</td>\n",
       "      <td>0.525</td>\n",
       "      <td>0.008264</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>16</td>\n",
       "      <td>0.550</td>\n",
       "      <td>0.007202</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>17</td>\n",
       "      <td>0.550</td>\n",
       "      <td>0.006351</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>18</td>\n",
       "      <td>0.550</td>\n",
       "      <td>0.005655</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>19</td>\n",
       "      <td>0.550</td>\n",
       "      <td>0.005074</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>20</td>\n",
       "      <td>0.550</td>\n",
       "      <td>0.004585</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Epoch  Validation Accuracy  Train Loss  Train Accuracy\n",
       "0       1                0.450    1.048929        0.586957\n",
       "1       2                0.425    0.769282        0.739130\n",
       "2       3                0.500    0.704915        0.717391\n",
       "3       4                0.475    0.643454        0.739130\n",
       "4       5                0.450    0.531550        0.804348\n",
       "5       6                0.475    0.349914        0.891304\n",
       "6       7                0.475    0.179244        0.978261\n",
       "7       8                0.500    0.091448        1.000000\n",
       "8       9                0.475    0.047549        1.000000\n",
       "9      10                0.550    0.032353        1.000000\n",
       "10     11                0.575    0.020925        1.000000\n",
       "11     12                0.550    0.017287        1.000000\n",
       "12     13                0.525    0.012231        1.000000\n",
       "13     14                0.525    0.009602        1.000000\n",
       "14     15                0.525    0.008264        1.000000\n",
       "15     16                0.550    0.007202        1.000000\n",
       "16     17                0.550    0.006351        1.000000\n",
       "17     18                0.550    0.005655        1.000000\n",
       "18     19                0.550    0.005074        1.000000\n",
       "19     20                0.550    0.004585        1.000000"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "config = Config(config_file)\n",
    "summary = TrainSummary(config.model_dir)\n",
    "dfsummary = summary.table()\n",
    "dfsummary"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a0ba9299-9ce7-4625-8f21-c49a6e3fd75a",
   "metadata": {},
   "source": [
    "It can also be plotted using the `summary.plot`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "24962a8b-d8b7-4b38-b982-10633416b86e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot saved to /home/ubuntu/scgenai/examples/testrun/test1/train_summary.pdf\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "summary.plot(saveplot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "346b4f66-b17c-462e-947f-0a3d08972cb4",
   "metadata": {},
   "source": [
    "## Prediction Using the Trained Model\n",
    "\n",
    "Similar to the training step, predictions can be run using the `run_prediction_from_config` function in scGenAI with a simple configuration file. Unlike the training configuration, the prediction configuration file only requires the following parameters to be defined: `mode`, `predict_file`, `batch_size`, `model_dir`, and `outputfile`. All other parameters will be adopted from the trained model.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "10cd73c8-8ea4-44cf-a2c5-6b2f4aecd53b",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scGenAI.prediction.predict import run_prediction_from_config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "6d8cab06-6cec-49b1-8fa8-53aa426c38df",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model_dir: /home/ubuntu/scgenai/examples/testrun/test1/best_model/\n",
      "gmtfile: /home/ubuntu/scgenai/examples/testrun/test1/best_model/biofounction_context\n",
      "predict_file: /home/ubuntu/scgenai/examples/data/example_val.h5ad\n",
      "log_dir: /home/ubuntu/scgenai/examples/logs\n",
      "output_dir: /home/ubuntu/scgenai/examples/output\n",
      "seed: 1314521\n",
      "mode: Predict\n",
      "min_cells: 2\n",
      "multiomics: No\n",
      "keepIntermediateFiles: No\n",
      "savelog: No\n",
      "savesetting: Yes\n",
      "model_backbone_name: llama\n",
      "model_backbone_size: small\n",
      "context_method: biofounction\n",
      "num_bins: 10\n",
      "optimizer: AdamW\n",
      "weight_decay: 0.01\n",
      "depth: 2\n",
      "max_length: 1024\n",
      "batch_size: 1\n",
      "learning_rate: 1e-05\n",
      "num_epochs: 30\n",
      "world_size: 1\n",
      "evaluate_during_training: False\n",
      "hidden_size: 768\n",
      "num_layers: 12\n",
      "num_heads: 12\n",
      "intermediate_size: 1024\n",
      "stride: 512\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-10-15 21:45:34.194 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 3, Total GPUs used: 4\n",
      "2024-10-15 21:45:34.199 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 0, Total GPUs used: 4\n",
      "2024-10-15 21:45:34.206 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 2, Total GPUs used: 4\n",
      "2024-10-15 21:45:34.210 | INFO     | scGenAI.utils.distributed:setup_distributed:26 - Distributed setup complete for GPU 1, Total GPUs used: 4\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processed 40/40 cells\r"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "GPU 2 Predicting: 100%|██████████| 46/46 [00:01<00:00, 27.32it/s]\n",
      "GPU 1 Predicting: 100%|██████████| 46/46 [00:01<00:00, 27.30it/s]\n",
      "GPU 0 Predicting: 100%|██████████| 46/46 [00:01<00:00, 27.28it/s]\n",
      "GPU 3 Predicting: 100%|██████████| 46/46 [00:01<00:00, 27.25it/s]\n",
      "2024-10-15 21:45:42.891 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "2024-10-15 21:45:42.891 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "2024-10-15 21:45:42.891 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "2024-10-15 21:45:42.891 | INFO     | scGenAI.utils.distributed:cleanup:33 - Cleaned up DDP environment\n",
      "\u001b[32m2024-10-15 21:45:44.084\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mscGenAI.data.postprocess\u001b[0m:\u001b[36mcombine_prediction_results\u001b[0m:\u001b[36m103\u001b[0m - \u001b[1mCompleted post-processing for prediction\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "pred_config_file = '/home/ubuntu/scgenai/tutorials/yaml/config_prediction_template.yaml'\n",
    "run_prediction_from_config(pred_config_file)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7fa3e6fc-55c8-4ef4-b64e-f6347b84fbb4",
   "metadata": {},
   "source": [
    "The prediction output is a CSV file, as defined in the configuration file. It contains the original metadata extracted from the input prediction file (`obs` slot) along with three additional columns: `context_id`, `PredictedFeature`, and `prediction_score`. \n",
    "\n",
    "- `context_id` represents the context used to determine the prediction for the corresponding cell. \n",
    "- `PredictedFeature` is the final predicted feature for the cell using the trained model. \n",
    "- `prediction_score` indicates the confidence level of the prediction, with a maximum value of 1.\n",
    "\n",
    "For example in this tutorials:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b4c89ab7-232d-41a1-8cb4-f59851b5fe6d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>index</th>\n",
       "      <th>orig.ident</th>\n",
       "      <th>nCount_RNA</th>\n",
       "      <th>nFeature_RNA</th>\n",
       "      <th>RNA_snn_res.0.8</th>\n",
       "      <th>seurat_clusters</th>\n",
       "      <th>sample</th>\n",
       "      <th>sampleId</th>\n",
       "      <th>projectId</th>\n",
       "      <th>sampleType</th>\n",
       "      <th>...</th>\n",
       "      <th>subjectId</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>processStatus</th>\n",
       "      <th>predicted_cell_type</th>\n",
       "      <th>ct</th>\n",
       "      <th>cell_barcode</th>\n",
       "      <th>context_id</th>\n",
       "      <th>PredictedFeature</th>\n",
       "      <th>prediction_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TGCCCTAGTCTTGATG-1--GSM5515744</td>\n",
       "      <td>SeuratProject</td>\n",
       "      <td>2306</td>\n",
       "      <td>871</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSE181989</td>\n",
       "      <td>Normal</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>QC pass</td>\n",
       "      <td>9:granulocyte</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>TGCCCTAGTCTTGATG-1--GSM5515744</td>\n",
       "      <td>1_TGCCCTAGTCTTGATG-1--GSM5515744_window_DENDRI...</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>0.610014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TGTTCCGGTCACACGC-1--GSM5515744</td>\n",
       "      <td>SeuratProject</td>\n",
       "      <td>1637</td>\n",
       "      <td>665</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSE181989</td>\n",
       "      <td>Normal</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>QC pass</td>\n",
       "      <td>13:cDC2</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>TGTTCCGGTCACACGC-1--GSM5515744</td>\n",
       "      <td>3_TGTTCCGGTCACACGC-1--GSM5515744_window_MONOCY...</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>0.504304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TGTCCCAGTTCAGACT-1--GSM5515744</td>\n",
       "      <td>SeuratProject</td>\n",
       "      <td>1935</td>\n",
       "      <td>805</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSE181989</td>\n",
       "      <td>Normal</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>QC pass</td>\n",
       "      <td>13:cDC2</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>TGTCCCAGTTCAGACT-1--GSM5515744</td>\n",
       "      <td>0_TGTCCCAGTTCAGACT-1--GSM5515744_window_CD34_T...</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>0.533765</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AACACGTGTGCACGAA-1--GSM5515744</td>\n",
       "      <td>SeuratProject</td>\n",
       "      <td>2053</td>\n",
       "      <td>780</td>\n",
       "      <td>25</td>\n",
       "      <td>25</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSE181989</td>\n",
       "      <td>Normal</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>QC pass</td>\n",
       "      <td>13:cDC2</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>AACACGTGTGCACGAA-1--GSM5515744</td>\n",
       "      <td>3_AACACGTGTGCACGAA-1--GSM5515744_window_0</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>0.435774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ATCATGGCAAAGAATC-1--GSM5515744</td>\n",
       "      <td>SeuratProject</td>\n",
       "      <td>1983</td>\n",
       "      <td>788</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSM5515744</td>\n",
       "      <td>GSE181989</td>\n",
       "      <td>Normal</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>QC pass</td>\n",
       "      <td>4:hematopoietic stem cell</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>ATCATGGCAAAGAATC-1--GSM5515744</td>\n",
       "      <td>2_ATCATGGCAAAGAATC-1--GSM5515744_window_ERYTHR...</td>\n",
       "      <td>Granulocyte-monocyte progenitor (GMP)</td>\n",
       "      <td>0.411171</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                            index     orig.ident  nCount_RNA  nFeature_RNA   \n",
       "0  TGCCCTAGTCTTGATG-1--GSM5515744  SeuratProject        2306           871  \\\n",
       "1  TGTTCCGGTCACACGC-1--GSM5515744  SeuratProject        1637           665   \n",
       "2  TGTCCCAGTTCAGACT-1--GSM5515744  SeuratProject        1935           805   \n",
       "3  AACACGTGTGCACGAA-1--GSM5515744  SeuratProject        2053           780   \n",
       "4  ATCATGGCAAAGAATC-1--GSM5515744  SeuratProject        1983           788   \n",
       "\n",
       "   RNA_snn_res.0.8  seurat_clusters      sample    sampleId  projectId   \n",
       "0               11               11  GSM5515744  GSM5515744  GSE181989  \\\n",
       "1                6                6  GSM5515744  GSM5515744  GSE181989   \n",
       "2                6                6  GSM5515744  GSM5515744  GSE181989   \n",
       "3               25               25  GSM5515744  GSM5515744  GSE181989   \n",
       "4                6                6  GSM5515744  GSM5515744  GSE181989   \n",
       "\n",
       "  sampleType  ... subjectId age gender  processStatus   \n",
       "0     Normal  ...       NaN NaN    NaN        QC pass  \\\n",
       "1     Normal  ...       NaN NaN    NaN        QC pass   \n",
       "2     Normal  ...       NaN NaN    NaN        QC pass   \n",
       "3     Normal  ...       NaN NaN    NaN        QC pass   \n",
       "4     Normal  ...       NaN NaN    NaN        QC pass   \n",
       "\n",
       "         predicted_cell_type                                     ct   \n",
       "0              9:granulocyte  Granulocyte-monocyte progenitor (GMP)  \\\n",
       "1                    13:cDC2  Granulocyte-monocyte progenitor (GMP)   \n",
       "2                    13:cDC2  Granulocyte-monocyte progenitor (GMP)   \n",
       "3                    13:cDC2  Granulocyte-monocyte progenitor (GMP)   \n",
       "4  4:hematopoietic stem cell  Granulocyte-monocyte progenitor (GMP)   \n",
       "\n",
       "                     cell_barcode   \n",
       "0  TGCCCTAGTCTTGATG-1--GSM5515744  \\\n",
       "1  TGTTCCGGTCACACGC-1--GSM5515744   \n",
       "2  TGTCCCAGTTCAGACT-1--GSM5515744   \n",
       "3  AACACGTGTGCACGAA-1--GSM5515744   \n",
       "4  ATCATGGCAAAGAATC-1--GSM5515744   \n",
       "\n",
       "                                          context_id   \n",
       "0  1_TGCCCTAGTCTTGATG-1--GSM5515744_window_DENDRI...  \\\n",
       "1  3_TGTTCCGGTCACACGC-1--GSM5515744_window_MONOCY...   \n",
       "2  0_TGTCCCAGTTCAGACT-1--GSM5515744_window_CD34_T...   \n",
       "3          3_AACACGTGTGCACGAA-1--GSM5515744_window_0   \n",
       "4  2_ATCATGGCAAAGAATC-1--GSM5515744_window_ERYTHR...   \n",
       "\n",
       "                        PredictedFeature prediction_score  \n",
       "0  Granulocyte-monocyte progenitor (GMP)         0.610014  \n",
       "1  Granulocyte-monocyte progenitor (GMP)         0.504304  \n",
       "2  Granulocyte-monocyte progenitor (GMP)         0.533765  \n",
       "3  Granulocyte-monocyte progenitor (GMP)         0.435774  \n",
       "4  Granulocyte-monocyte progenitor (GMP)         0.411171  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "preconfig = Config(pred_config_file)\n",
    "dfprediction = pd.read_csv(preconfig.outputfile)\n",
    "dfprediction.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0cd6dab7-bb05-4204-bc26-b730ad4461a4",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}