Skip to main content

Training

Can I initiate a training using a local environment?​

Yes, we support training through both Google Colab and local environments using Jupyter Notebook. To set up your local environment, refer to our Jupyter Notebook training guide.

In Step 3. of the Training Guide, I get the error that Model and Dataset parameters dont match?​

If you encounter the following error:

Assignment failed!
DataSet 'dataset ID' expected parameters :
classes : 5, shape: 224
'model name' parameters :
classes : 6, shape: 224
Please change model parameters to match expected dataset parameters.

This means the number of classes in your dataset does not match the number specified in your model. To fix this, update the number of output classes in your model to match the dataset. For guidance, refer to the Model Structure Guide.

How to navigate in the Anaconda Prompt?​

InputDescription
dirDisplays directory contents
cd + "folder"Navigate into a folder
cd..Move up one directory
conda activate + "Environment"Activates your Anaconda environment
jupyter-notebookOpens Jupyter Notebook in your environment

What are pre-trained weights and how can I use them in my model?​

Pre-trained weights refer to previously trained model weights. Your weights file should contain only the model's weights, not the entire model object.

For PyTorch models: Using PyTorch, the weights file for your model needs to have the .pth extension. To create your own pre-trained weights, use:

model = MyModel() # load model class here
torch_weight_file = ("<filename>_weights.pth")

torch.save(model.state_dict(), torch_weight_file)

For TensorFlow models: Using TensorFlow, the weights file for your model needs to have the .pkl extension. To create your own pre-trained weights, use:

w = model.get_weights() 

with open('<filename>_weights.pkl', 'wb') as output: 
    pickle.dump(w, output)

By following these examples, you'll be able to create a .pkl/.pth file that is compatible with our federated learning infrastructure.

Can I utilize pre-trained weights from models like ImageNet directly?​

Unfortunately, third-party libraries other than list provided in supported are not supported directly. However, if you wish to use a model architecture from any third party library this is not supported, you can reimplement it using PyTorch or TensorFlow. Ensure that your reimplemented model follows our model submission guidelines

What kind of training environment does tracebloc use for experiments?​

Tracebloc utilizes a multi-core CPU environment for each experiment. This means multiple CPU cores work together to efficiently run and manage training processes, resulting in faster and more optimized experiments. For some use cases we enable GPU environment also, please check collaboration for more information about the environment.

Can I replicate the Tracebloc environment on my local system?​

No, it is not currently feasible to replicate the unique configuration and infrastructure of the Tracebloc platform on a local system.

What types of model file extensions are supported?​

Tracebloc supports TensorFlow and PyTorch models with .py or .zip extensions uploaded via the tracebloc_package.

Can I use multiple subfolders in my model .zip file?​

No. When uploading a .zip model file, it should not contain subfolders. All model files must be in a single folder, with the main file following the required structure.

What operations are not allowed in the model file?​

The following operations are not permitted in the model file:

  • File operations (e.g., open, read, write)
  • Imports beyond TensorFlow, Keras, PyTorch, and TorchVision
  • Executable operations
  • Subdirectories
  • Subprocess operations
  • Terminal commands
On This Page