ConvNeXt revisits and modernizes convolutional neural network design by incorporating architectural insights from Vision Transformers, such as large kernels, simplified blocks, and improved normalization, while retaining convolutional efficiency.

Original paper: A ConvNet for the 2020s, Liu et al., 2022

ConvNeXt-T

This model uses the ConvNeXt-Tiny variant, a lightweight configuration that delivers strong accuracy with relatively low computational cost. It is well suited for high-resolution image classification and as a general-purpose backbone for detection and segmentation tasks where CNN efficiency is preferred.

Model Configuration:

• Reference implementation: ConvNeXt_T
• Original Weight: ConvNeXt_Tiny_Weights.IMAGENET1K_V1
• Resolution: 3x224x224
• Support Cooper version:
  • Cooper SDK: [2.5.2]
  • Cooper Foundry: [2.2]

Model	Device	Model Link
ConvNeXt-T	N1-655	Model_Link
ConvNeXt-T	CV72	Model_Link
ConvNeXt-T	CV75	Model_Link