ConvNeXt Base CLIP LAION-2B

JSON →
timm vision
image

ConvNeXt base model trained with CLIP on LAION-2B, augmented and regularized, then fine-tuned on ImageNet-12k and ImageNet-1k.

vision