### Clone Repository and Create Conda Environment for FALCON

Source: https://github.com/asalmehradfar/falcon/blob/main/README.md

This sequence of commands clones the FALCON repository from GitHub, navigates into the project directory, and then creates and activates a new Conda environment using the `falcon.yml` file to manage project dependencies. This setup ensures all required libraries and dependencies are correctly installed for the project.

```bash
git clone https://github.com/AsalMehradfar/FALCON.git
cd FALCON
conda env create -f falcon.yml
conda activate falcon
```

--------------------------------

### Run Data Preprocessing Scripts for FALCON

Source: https://github.com/asalmehradfar/falcon/blob/main/README.md

These Python scripts are used for initial data preparation within the FALCON pipeline. `process_netlists.py` converts raw netlists into graph representations, and `generate_splits.py` creates index splits for datasets, both necessary if `graphs.json` or index splits are not already available.

```bash
python scripts/process_netlists.py
```

```bash
python scripts/generate_splits.py
```

--------------------------------

### Update Existing Conda Environment for FALCON

Source: https://github.com/asalmehradfar/falcon/blob/main/README.md

This command updates an existing Conda environment for FALCON, pruning any unused packages. It should be run if changes are made to the `falcon.yml` file or new packages are added to ensure the environment remains consistent with the project's dependency specifications.

```bash
conda env update -f falcon.yml --prune
```

--------------------------------

### Train Machine Learning Models for FALCON Pipeline

Source: https://github.com/asalmehradfar/falcon/blob/main/README.md

These Python scripts are used to train the core machine learning models within the FALCON pipeline. `train_mlp.py` trains the MLP for topology classification, `train_gnn.py` trains the GNN for performance prediction, and `finetune_gnn.py` optionally fine-tunes the GNN on unseen topologies to improve generalization.

```bash
python scripts/train_mlp.py
```

```bash
python scripts/train_gnn.py
```

```bash
python scripts/finetune_gnn.py
```

--------------------------------

### Evaluate Performance of FALCON Pipeline Stages

Source: https://github.com/asalmehradfar/falcon/blob/main/README.md

These Python scripts evaluate the performance of each stage in the FALCON pipeline. `mlp_eval.py` assesses topology selection accuracy, `gnn_forward_eval.py` evaluates forward performance prediction, and `gnn_backward_eval.py` checks inverse parameter inference with layout-aware optimization.

```bash
python evaluation/mlp_eval.py
```

```bash
python evaluation/gnn_forward_eval.py
```

```bash
python evaluation/gnn_backward_eval.py
```

--------------------------------

### Generate Model Input Data and Scalers for FALCON

Source: https://github.com/asalmehradfar/falcon/blob/main/README.md

These Python scripts prepare the necessary input data for the MLP and GNN models within the FALCON framework. `save_mlp_data.py` generates inputs for the Multi-Layer Perceptron, and `save_gnn_data.py` generates inputs for the Graph Neural Network, saving required scalers for both.

```bash
python scripts/save_mlp_data.py
```

```bash
python scripts/save_gnn_data.py
```

--------------------------------

### Cite FALCON Research Paper

Source: https://github.com/asalmehradfar/falcon/blob/main/README.md

This BibTeX entry provides the citation details for the FALCON research paper. It includes authors, title, year, and URL, suitable for inclusion in academic publications when referencing the FALCON framework.

```BibTeX
@article{Mehradfar2025FALCON,
      title={FALCON: An ML Framework for Fully Automated Layout-Constrained Analog Circuit Design}, 
      author={Asal Mehradfar and Xuzhe Zhao and Yilun Huang and Emir Ceyani and Yankai Yang and Shihao Han and Hamidreza Aghasi and Salman Avestimehr},
      year={2025},
      url={https://arxiv.org/abs/2505.21923}
}
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.