Training Octopi Models¶
This guide covers everything you need to know about training 3D U-Net models with Octopi. Here, we are fixing the model architecture. For those interested in writing their own training functions, refer to the API.
Single Model Training¶
For specific use cases or when you have a known good architecture, you can train a single model directly. In this case, the command only allows for training U-Net models. To play around with more unique model configurations, or to try importing new model designs refer to the API or octopi model-explore.
Basic Training Command¶
octopi train \
--config config.json \
--voxel-size 10 --tomo-alg wbp --Nclass 8 \
--tomo-batch-size 50 --num-epochs 100 --val-interval 10 \
--target-info targets,octopi,1
Training Parameters¶
| Parameter | Description | Example Value |
|---|---|---|
--config | Path to copick configuration file | config.json |
--voxel-size | Tomogram voxel size (Angstroms) | 10 |
--tomo-alg | Tomographic reconstruction algorithm | wbp (weighted back projection) |
--Nclass | Number of output classes for segmentation | 8 |
--tomo-batch-size | Batch size for tomographic processing | 50 |
--num-epochs | Total number of training epochs | 100 |
--val-interval | Validation frequency (every N epochs) | 10 |
--target-info | Target information in format: name,framework,version | targets,octopi,1 |
Remember that Nclass is number of classes + 1 for background.
Fine Tuning Models¶
If we have base weights that we would like to fine-tune for new datasets, we can still use the train command. Instead of specifying the model architecture, we can simple point to the configuration file and weights to load the existing model to fine tune.
octopi train \
--config config.json \
--voxel-size 10 --tomo-alg wbp \
--model-config results/model_config.yaml \
--model-weights results/best_model_weights.pth
Fine-Tuning Parameters¶
The parameters for fine-tuning is equivalent to the base training, with the added option to provide the pre-trained model configuration and its weights.
| Parameter | Description | Example Value |
|---|---|---|
--model-config | Path to configuration file that the previous training run generated. | results/model_config.yaml |
--model-weights | Path to trained model weights. | results/best_model_weights.pth |
Training Outputs¶
Training Output During training, you'll see:
- Progress indicators: Real-time loss and accuracy metrics
- Validation results: Periodic evaluation on validation set
- Model checkpoints: Saved to results/ directory by default
- Training logs: Detailed logs for monitoring and debugging