### Install Squidiff

Source: https://github.com/siyuh/squidiff/blob/main/README.md

Install the Squidiff package using pip.

```bash
pip install Squidiff
```

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

### Train Squidiff Model

Source: https://github.com/siyuh/squidiff/blob/main/README.md

Train the Squidiff model with specified parameters. Use --use_drug_structure True to incorporate drug structure information.

```bash
python train_squidiff.py --logger_path LOGGER_FIRE_NAME --data_path YOUR_ADATASET.h5ad --resume_checkpoint ptNAME --gene_size 500 --output_dim 500
```

```bash
python train_squidiff.py --logger_path logger_files/logger_sciplex_random_split_0 --data_path datasets/sci_plex_train_random_split_0.h5ad --resume_checkpoint sciplex_results_random_split_0 --use_drug_structure True --gene_size 200 --output_dim 200 --control_data_path datasets/sci_plex_train_random_split_0_control.h5ad
```

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

### Sample Squidiff Predictions

Source: https://github.com/siyuh/squidiff/blob/main/README.md

Generate predictions using a trained Squidiff model. Requires PyTorch and Scanpy. Ensure the model path and dataset are correctly specified.

```python
sampler = sample_squidiff.sampler(
    model_path = 'simu_results/model.pt',
    gene_size = 100,
    output_dim = 100,
    use_drug_structure = False
)

test_adata_scrna = sc.read_h5ad('datasets/sc_simu_test.h5ad')
z_sem_scrna = sampler.model.encoder(torch.tensor(test_adata_scrna.X).to('cuda'))

scrnas_pred = sampler.pred(z_sem_scrna, gene_size = test_adata_scrna.shape[1])

```

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