Hugging Face's logo
kerasformers 's Collections
GPT and GPT-2
Janus-Pro
DeepSeek-VL
DeBERTa (v1 / v2 / v3)
XLM-RoBERTa
BERT
RoBERTa
MobileViT-V2
MobileViT
MiT (SegFormer encoder)
Xception
ViT
VGG
EfficientNetV2
EfficientNet-Lite
EfficientNet
Swin Transformer V2
Swin Transformer
SE-ResNet and SE-ResNeXt
ResNeXt
ResNetV2 (BiT)
ResNet
ResMLP
Res2Net
PoolFormer
PiT
Next-ViT
MobileNetV3
MobileNetV2
MLP-Mixer
MaxViT
Inception-V4
Inception-V3
Inception-ResNet-V2
InceptionNeXt
FlexiViT
EfficientFormer
DenseNet
DeiT and DeiT3
ConvNeXt-V2
ConvNeXt
ConvMixer
CaiT
Gemma 2
Granite Speech Plus
Granite Speech
Moonshine
Speech2Text
Whisper
LocateAnything
SigLIP2
SigLIP
MetaCLIP 2
CLIP
Depth Anything V1 and V2
DINO V1,V2,V3
MobileViT V1 and V2
Mask2Former
MaskFormer
EoMT
DeeplabV3
OneFormer
SAM V1,V2,V3
SegFormer
OWL-ViT
OWLv2
RT-DETR V1 and V2
D-FINE
RF-DETR
DETR
Qwen2
Gemma

XLM-RoBERTa

Pure Keras 3, backend-agnostic (JAX / TensorFlow / PyTorch) ports of the multilingual XLM-RoBERTa encoders and masked-LM heads.