add steps in readme.md
Browse files
Graph/GraphMAE_MQ9/README.md
CHANGED
|
@@ -2,12 +2,15 @@
|
|
| 2 |
## Overview
|
| 3 |
We run the Graph Mask AutoEncoder on QM9 Dataset for pretraining. We use the atom position of each atom and the embedding of their element type as the input feature (dim=7) and predict the input feature by using the GraphSage with 4-dim hidden representation.
|
| 4 |
## How to run
|
| 5 |
-
|
|
|
|
|
|
|
|
|
|
| 6 |
|
| 7 |
```bash
|
| 8 |
python prepare_QM9_dataset.py --label_keys "mu" "gap"
|
| 9 |
```
|
| 10 |
-
- step2. Train the Graph Mask AutoEncoder on the preprocessed dataset
|
| 11 |
```bash
|
| 12 |
python run.py [--dataset_path] [--batch_size] [--epochs] [--device] [--save_dir]
|
| 13 |
```
|
|
|
|
| 2 |
## Overview
|
| 3 |
We run the Graph Mask AutoEncoder on QM9 Dataset for pretraining. We use the atom position of each atom and the embedding of their element type as the input feature (dim=7) and predict the input feature by using the GraphSage with 4-dim hidden representation.
|
| 4 |
## How to run
|
| 5 |
+
### If you do not want to re-train the model again
|
| 6 |
+
- **Unzip the model.zip** to get the model weight & embedded graph in each epoch
|
| 7 |
+
### If you want to try out the training process
|
| 8 |
+
- step1. **Preprocess the dataset** (we have provided the preprocessed as well)
|
| 9 |
|
| 10 |
```bash
|
| 11 |
python prepare_QM9_dataset.py --label_keys "mu" "gap"
|
| 12 |
```
|
| 13 |
+
- step2. **Train the Graph Mask AutoEncoder on the preprocessed dataset**
|
| 14 |
```bash
|
| 15 |
python run.py [--dataset_path] [--batch_size] [--epochs] [--device] [--save_dir]
|
| 16 |
```
|