README, utils to src folder

README.md

@@ -15,16 +15,28 @@ This repository hosts a Hugging Face Space that provides an examplary API for su
 
 In this example, we trained a GIN classifier on the Tox21 targets and saved the trained model in the `assets/` folder.
 
-**Important:** For leaderboard submission, your Space does not need to include training code. It only needs to implement inference in the `predict()` function inside `predict.py`. The `predict()` function must keep the provided skeleton: it should take a list of SMILES strings as input and return a nested prediction dictionary as output, with SMILES as keys and dictionaries containing targetname-prediction pairs as values. Therefore, any preprocessing of SMILES strings must be executed on-the-fly during inference.
+**Important:** For leaderboard submission, your Space needs to include training code. The file `train.py` should train the model using the config specified inside the `config/` folder and save the final model parameters into a file inside the `checkpoints/` folder. The model should be trained using the [Tox21_dataset](https://huggingface.co/datasets/tschouis/tox21) provided on Hugging Face. The datasets can be loaded like this:
+```python
+from datasets import load_dataset
+ds = load_dataset("tschouis/tox21", token=token)
+train_df = ds["train"].to_pandas()
+val_df = ds["validation"].to_pandas()
+```
+ Additionally, the Space needs to implement inference in the `predict()` function inside `predict.py`. The `predict()` function must keep the provided skeleton: it should take a list of SMILES strings as input and return a nested prediction dictionary as output, with SMILES as keys and dictionaries containing targetname-prediction pairs as values. Therefore, any preprocessing of SMILES strings must be executed on-the-fly during inference.
 
 # Repository Structure
 - `predict.py` - Defines the `predict()` function required by the leaderboard (entry point for inference).
 - `app.py` - FastAPI application wrapper (can be used as-is).
+- `train.py` - trains and saves a model using the config in the `config/` folder.
+- `config/` - the config file used by `train.py`. 
+- `logs/` - all the logs of `train.py`, the saved model, and predictions on the validation set.
+- `checkpoints/` - the saved model that is used in `predict.py` is here.
 
 - `src/` - Core model & preprocessing logic:
-    - `preprocess.py` - SMILES preprocessing pipeline
-    - `model.py` - GIN classifier
-    - `seed.py` - used to ensure reproducibility 
+    - `preprocess.py` - SMILES preprocessing pipeline and dataset creation
+    - `train_evaluate.py` - train and evaluate model, compute metrics
+    - `seed.py` - set seed for everything
+    - `model.py` - contains the model class
 
 # Quickstart with Spaces
 
@@ -34,10 +46,15 @@ You can easily adapt this project in your own Hugging Face account:
 
 - Click "Duplicate this Space" (top-right corner).
 
+- Create a `.env` according to `.example.env`.
+
 - Modify `src/` for your preprocessing pipeline and model class
 
 - Modify `predict()` inside `predict.py` to perform model inference while keeping the function skeleton unchanged to remain compatible with the leaderboard.
 
+- Modify `train.py` according to your model and preprocessing pipeline.
+
+- Modify the file inside `config/` to contain all hyperparameters that are set in `train.py`.
 That’s it, your model will be available as an API endpoint for the Tox21 Leaderboard.
 
 # Installation
@@ -77,6 +94,6 @@ The output will be a nested dictionary in the format:
 
 # Notes
 
-- Only adapting `predict.py` for your model inference is required for leaderboard submission.
+- Adapting `predict.py`, `train.py`, `config/`, and `checkpoints/` is required for leaderboard submission.
 
-- Preprocessing (here inside `src/preprocess.py`) must be applied at inference time, not just predicting.
+- Preprocessing (here inside `src/preprocess.py`) must be done inside `predict.py` not just `train.py`.

checkpoints/best_gin_model.pt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ae536ad4a4c51b097a6411bb4ad68ebcc49ab70d570d2f38a8917f42efc768f0
-size 631874

predict.py

@@ -2,9 +2,9 @@ from torch_geometric.data import Batch
 from torch_geometric.utils import from_rdmol
 import torch
 
-from utils.model import GIN
-from utils.preprocess import create_clean_mol_objects
-from utils.seed import set_seed
+from src.model import GIN
+from src.preprocess import create_clean_mol_objects
+from src.seed import set_seed
 
 def predict(smiles_list):
     """

train.py

@@ -6,10 +6,10 @@ import json
 import os
 from dotenv import load_dotenv
 
-from utils.model import GIN
-from utils.preprocess import get_graph_datasets
-from utils.train_evaluate import train_model, evaluate, compute_roc_auc_avg_and_per_class
-from utils.seed import set_seed
+from src.model import GIN
+from src.preprocess import get_graph_datasets
+from src.train_evaluate import train_model, evaluate, compute_roc_auc_avg_and_per_class
+from src.seed import set_seed
 
 
 def train(config):