DLHUB Model Training
Load dataset
Once you complete the data preparation, click the ... icon and browse to the training data file/folder and import to DLHUB
- DLHUB will display a small data sample to visualize how it looks
- Alphabet categories will be auto-converted to numerical value for training purposes.
If your dataset is images, you can:
- Select Apply Transfer Learning; the input shape array will be automatically populated
- Select Image Augmentation to create more samples if you have a small dataset.
Once the DATA IS GOOD notification appears, click Next to continue.
Configure neural network
To construct the neural network, you can use our neural network template by clicking the Load Template button or start from scratch using Neural Network Modules.
To start constructing a new model, click New.
Add modules
- Select from the left-side panel the module/function of interest.
- Once the module is selected, its parameters will appear in the configuration panel.
- You can modify these parameters to match your application.
- Press
ADDto add the module to your neural network. The added module will appear in the neural network listed at the bottom. - Repeat until you have fully developed your Neural Network.
- You can use the editing tools to shift, duplicate and delete the selected modules in your model.
Save model
After you have finished the configuration, ensure to SAVE the model for reuse next time or share it with others.
Verify model
After saving your model, click Verify Model to start neural network verification
If it's green (the neural network is valid), click NEXT to continue
If it's red (the neural network is invalid), you can refer to the status info at the bottom of the screen for feedback and make adjustment.
Train neural network
Once the Deep Neural Network is configured, it is ready to be trained to learn features from the dataset.
Training Parameter
The training parameters are required for training algorithms to work and perform well. Specific parameter settings are only available for certain Training Algorithms.
Stopping Criteria
The stopping criteria contain parameters that control when to end/stop the training process.
In short, the Neural Network training is stopped when:
- Training goal is reached.
- Max Fail Count is reached.
You can manually stop the training if needed. Stopping criteria is just settings to stop training automatically.
Start Training
Training will begin immediately once you click Start Training
You may monitor the progress and the Loss and Evaluation values from the graph.
During the training:
- The training dataset is used to adjust the Neural Network weights to optimize the cost/loss function during the training process. The value of this Loss function is called a performance index or performance of a training dataset.
- The validation dataset (extracted from the training dataset) prevents over-fitting/over-trained issues.
Once you finish training, click Next to continue
Evaluate neural network
After a Neural Network has been trained, it will be evaluated to ensure it can cope with a completely new dataset.
Prepare a new dataset with the same format as the training dataset and load it into DLHUB for evaluation.
Detail about the dataset will be shown in the table for review, and a summary will be displayed for the user.
Once the dataset is loaded, click Evaluate to start the evaluation. DLHUB will calculate the model accuracy based on the loaded dataset
You can retrain the model or click Next to continue to the next step
Test model (optional)
Users can still manually test prediction results before exporting.
- Browse and load your Test dataset. The dataset will be listed in the
Test Filestable - Select each file to see prediction result, displaying on the right-side panel, with the raw calculated score for each class (output)
Export neural network for deployment
Click Export to export the model as .zip file that can be loaded directly into C++, C#, Python for deployment