Model Upload/Download

Overview

Upload Guidelines

• Open HubAI and navigate to the Models section by clicking on the dotted icon in the left sidebar.

• Click the Add Model button in the top right corner of the page.

• Select the upload model type:
  • YOLO, if uploading PyTorch (.pt) weights of a YOLO model. Please, first check that your model is one of the supported PyTorch YOLO models.
  • Custom, if uploading an ONNX (.onnx), OpenVINO IR (.xml and .bin), or TensorFlow Lite (.tflite) or any RVC compiled format model executable(s).

• If the YOLO option was selected, the model will be converted to the appropriate RVC compiled format during the conversion. A pop-up window is opened where you can choose the RVC Platform the model should be converted for.

• Next, a pop-up window is opened. Fill-in the empty fields with model descriptors. Depending on the selected model type and platform, you will be asked to fill-in some of the following parameters:
  • Model Name - Name used to reference the model through HubAI.
  • Short Description - Short description of the model shown under the model tumbnail.
  • Description - Comprehensive description of the model (Model Card for public models).
  • Model Image (OPTIONAL) - Upload image that illustrates model performance.
  • Tasks - List machine learning tasks that the model tackles (e.g. classification, object detection, etc.).
  • License - Choose the license that governs the use/distribution of the model (e.g. MIT).
  • Public - Choose whether the model should be public (visible to other HubAI users). Default is private (team-owned).
  • YOLO-specific parameters:
    • Model Instance Name - Name of the converted model (e.g. <model>:<platform>).
    • Version - YOLO version of the uploaded model (e.g. YOLOv5).
    • Class Names - Add class names one-by-one by clicking the Add button.
    • Shape - Input shape of the model.
    • Quantization Data (OPTIONAL) - Models converted for RVC3 and RVC4 platforms can be quantized to reduce their computational and memory costs. This process requires example input data to be passed to the model. Choose the most appropriate dataset from the provided options (the one most similar to the model’s training data) or leave the empty if you do not want to quantize the model.
    • POT Target Device - Target device for POT (RVC3 only). e
• Once done, upload the Model File(s) (click to upload or drag and drop to the Model File section) and click on the Add button (or Export button for YOLO) to upload the model to HubAI.

• HubAI allows upload of multiple versions of the model (e.g. if models are exported with different input shapes, trained on a different datasets, etc.). A new model version can be added by clicking the Add Version button in the bottom right corner of the page. It's suggested to follow a uniform version naming for public models. The version name should not include the model name and consist only of relevant descriptors used to differentiate between variants (e.g. name of the training dataset, input shape etc.). Descriptors shoulf be separated by a single space. Make sure that the input shape (input_height x input_width) always comes as the last descriptor.

• Once done uploading all model versions, you can reference to the specific version using the reference obtained by clicking the Copy:

Download Guidelines

• Open HubAI and navigate to the Models section by clicking on the dotted icon in the left sidebar.

• Click the icon of the desired public or private (team-owned) model, scroll down to Model Versions, and click on the desired model version.

Base Model

• To download the un-converted model, click on the Base Model button in the top right corner of the page.

• This opens up a pop-up window. Click on the model version name to start the download.

Converted Model

• To download a converted model, click on the name of the converted instance.

• This opens up a pop-up window. Click on the name of the .tar file to start the download.