cph-regression 0.1.0

A generic, config-driven regression training package using PyTorch Lightning

CPH Regression

A generic, config-driven regression training package using PyTorch Lightning. Train regression models on any tabular dataset by simply providing a YAML configuration file.

Features

• Fully Config-Driven: All settings (features, hyperparameters, paths) controlled via YAML files
• Generic & Reusable: Use the same codebase for any regression task (gemstone prices, house prices, etc.)
• Auto-Dimension Detection: Automatically calculates input dimensions from feature lists
• Production-Ready: Exports models to ONNX format with preprocessors for easy deployment
• PyTorch Lightning: Built on PyTorch Lightning for scalable, professional ML training

Installation

pip install cph-regression

Quick Start

1. Create a Configuration File

Create a config.yaml file:

# Your Project Configuration
seed_everything: true

trainer:
  callbacks:
    - class_path: lightning.pytorch.callbacks.ModelCheckpoint
      init_args:
        filename: "{epoch}-{val_loss:.2f}.best"
        monitor: "val_loss"
        mode: "min"
        save_top_k: 1
    - class_path: cph_regression.regression.callbacks.ONNXExportCallback
      init_args:
        output_dir: "models"
        model_name: "my_model"
        input_dim: null  # Auto-detected

  logger:
    class_path: lightning.pytorch.loggers.TensorBoardLogger
    init_args:
      save_dir: "lightning_logs"
      name: "MyProjectTraining"

  max_epochs: 50
  accelerator: auto
  devices: auto
  precision: 16-mixed

model:
  class_path: cph_regression.regression.modelmodule.ModelModuleRGS
  init_args:
    lr: 0.0001
    model:
      class_path: cph_regression.regression.modelfactory.RegressionModel
      init_args:
        input_dim: 0  # Auto-set from datamodule
        hidden_layers: [128, 64, 32]
        dropout_rates: [0.15, 0.1, 0.05]
        activation: "relu"

optimizer: 
  class_path: torch.optim.Adam
  init_args:
    lr: 0.001

data:
  class_path: cph_regression.regression.datamodule.DataModuleRGS
  init_args:
    csv_path: "data/your_data.csv"
    batch_size: 256
    val_split: 0.2
    categorical_cols:
      - column1
      - column2
    numeric_cols:
      - column3
      - column4
    target_col: "target"
    save_preprocessor: true
    preprocessor_path: "models/preprocessor.joblib"

fit:
  ckpt_path: null

test:
  ckpt_path: best

2. Run Training

cph-regression --config config.yaml

This will:

• Train the model
• Run validation
• Export the model to ONNX format
• Save the preprocessor

3. Alternative Commands

Training only:

cph-regression fit --config config.yaml

Testing only:

cph-regression test --config config.yaml

Configuration Guide

Data Configuration

• csv_path: Path to your CSV file
• batch_size: Batch size for training (default: 256)
• val_split: Validation split ratio (0.0 to 1.0, default: 0.2)
• categorical_cols: List of categorical feature column names
• numeric_cols: List of numeric feature column names
• target_col: Name of the target column to predict
• preprocessor_path: Where to save/load the preprocessor

Model Configuration

• hidden_layers: List of hidden layer sizes, e.g., [128, 64, 32]
• dropout_rates: List of dropout rates matching hidden layers, e.g., [0.15, 0.1, 0.05]
• activation: Activation function ("relu", "tanh", "gelu", "sigmoid", "leaky_relu", "elu")
• input_dim: Automatically set from datamodule (set to 0 in config)

Trainer Configuration

• max_epochs: Number of training epochs
• precision: Training precision ("16-mixed", "32", "bf16-mixed")
• accelerator: Hardware accelerator ("auto", "gpu", "cpu")
• devices: Number of devices ("auto", 1, [0, 1])

Output Files

After training, you'll find:

1. Models Directory (models/):

   • your_model_name.onnx: ONNX model for inference
   • preprocessor.joblib: Fitted preprocessor for data transformation

2. Checkpoints (lightning_logs/YourProjectTraining/version_X/checkpoints/):

   • epoch-X-val_loss=Y.best.ckpt: Best model checkpoint
   • epoch-X.last.ckpt: Last epoch checkpoint

3. Training Logs (lightning_logs/):

   • TensorBoard logs for visualization

Model Inference

After training, use the exported ONNX model and preprocessor:

import joblib
import onnxruntime as ort
import numpy as np
import pandas as pd

# Load preprocessor
preprocessor = joblib.load("models/preprocessor.joblib")

# Load ONNX model
session = ort.InferenceSession("models/your_model_name.onnx")

# Prepare input data
input_data = pd.DataFrame({
    'categorical_col': ['value1'],
    'numeric_col': [123.45],
    # ... other features
})

# Transform data
feature_cols = ['categorical_col', 'numeric_col']  # Your feature columns
transformed = preprocessor.transform(input_data[feature_cols])

# Predict
input_name = session.get_inputs()[0].name
output = session.run(None, {input_name: transformed.astype(np.float32)})
prediction = output[0][0][0]

print(f"Prediction: {prediction}")

Viewing Training Progress

TensorBoard

tensorboard --logdir lightning_logs

Then open http://localhost:6006 in your browser.

Example Projects

Gemstone Price Prediction

See the GemstonePricePrediction directory for a complete example.

Requirements

• Python >= 3.8
• PyTorch >= 2.0.0
• PyTorch Lightning >= 2.1.0

See requirements.txt for the complete list of dependencies.

License

MIT License - see LICENSE file for details.

Repository

• GitHub: https://github.com/imchandra11/cph-regression
• Author: chandra

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

For issues or questions:

1. Check the configuration file syntax
2. Verify CSV file format and column names
3. Check TensorBoard logs for training insights
4. Open an issue on GitHub