Model card for convnext base.clip laion2b augreg ft in12k in1k 384 A ConvNeXt image classification model. CLIP image tower weights pretrained in OpenCLIP on LAION and fine tuned on ImageNet 12k followed by ImageNet 1k in timm bby Ross Wightman. Please see related OpenCLIP model cards for more details on pretrain: https://huggingface.co/laion/CLIP convnext xxlarge laion2B s34B b82K augreg soup https://huggingface.co/laion/CLIP convnext large d.laion2B s26B b102K augreg https://huggingface.co/laion/CLIP convnext base w laion2B s13B b82K augreg https://huggingface.co/laion/CLIP convnext base w 320 laion aesthetic s13B b82K Model Details Model Type: Image classification / feature backbone Model Stats: Params (M): 88.6 GMACs: 45.2 Activations (M): 84.5 Image size: 384 x 384 Papers: LAION 5B: An open large scale dataset for training next generation image text models: https://arxiv.org/abs/2210.08402 A ConvNet for the 2020s: https://arxiv.org/abs/2201.03545 Learning Transferable Visual Models From Natural Language Supervision: https://arxiv.org/abs/2103.00020 Original: https://github.com/mlfoundations/open clip Pretrain Dataset: LAION 2B Dataset: ImageNet 1k Model Usage Image Classificati…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy