kerasformers/swinv2_small_window16_256_ms_in1k

See our collection for all versions of Swin Transformer V2.

Run Swin Transformer V2 with Keras 3: JAX, PyTorch, or TensorFlow

kerasformers/swinv2_small_window16_256_ms_in1k

Paper: Swin Transformer V2: Scaling Up Capacity and Resolution (arXiv:2111.09883) · HF Papers

Swin V2 scales capacity and resolution with residual post-norm and scaled cosine attention. Classifier + hierarchical backbone.

For more details on the model, please go to the upstream model card.

Pure-Keras 3 conversion of timm/swinv2_small_window16_256.ms_in1k for kerasformers. One implementation runs unmodified on TensorFlow / Torch / JAX.

This is an image-classification / backbone checkpoint (SwinV2ImageClassify / SwinV2Model).

✨ Quick start

import os
os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

from PIL import Image
import numpy as np
from kerasformers.models.swinv2 import SwinV2ImageClassify, SwinV2Model

model = SwinV2ImageClassify.from_weights("kerasformers/swinv2_small_window16_256_ms_in1k")
backbone = SwinV2Model.from_weights(
    "kerasformers/swinv2_small_window16_256_ms_in1k", as_backbone=True
)

image = Image.open("your_image.jpg").convert("RGB")
image = image.resize((224, 224))
x = np.asarray(image, dtype="float32")[None]  # (1, H, W, 3)
print(model(x).shape)  # (1, num_classes)
feats = backbone(x)
print(len(feats), [tuple(f.shape) for f in feats])

Load any Swin Transformer V2 variant the same way with from_weights("kerasformers/<variant>"):

Variant	Hub
swinv2_base_window12_192_ms_in22k	kerasformers/swinv2_base_window12_192_ms_in22k
swinv2_base_window12to16_192to256_ms_in22k_ft_in1k	kerasformers/swinv2_base_window12to16_192to256_ms_in22k_ft_in1k
swinv2_base_window12to24_192to384_ms_in22k_ft_in1k	kerasformers/swinv2_base_window12to24_192to384_ms_in22k_ft_in1k
swinv2_base_window16_256_ms_in1k	kerasformers/swinv2_base_window16_256_ms_in1k
swinv2_base_window8_256_ms_in1k	kerasformers/swinv2_base_window8_256_ms_in1k
swinv2_large_window12_192_ms_in22k	kerasformers/swinv2_large_window12_192_ms_in22k
swinv2_large_window12to16_192to256_ms_in22k_ft_in1k	kerasformers/swinv2_large_window12to16_192to256_ms_in22k_ft_in1k
swinv2_large_window12to24_192to384_ms_in22k_ft_in1k	kerasformers/swinv2_large_window12to24_192to384_ms_in22k_ft_in1k
swinv2_small_window16_256_ms_in1k	kerasformers/swinv2_small_window16_256_ms_in1k
swinv2_small_window8_256_ms_in1k	kerasformers/swinv2_small_window8_256_ms_in1k
swinv2_tiny_window16_256_ms_in1k	kerasformers/swinv2_tiny_window16_256_ms_in1k
swinv2_tiny_window8_256_ms_in1k	kerasformers/swinv2_tiny_window8_256_ms_in1k

Tips

• Set KERAS_BACKEND before importing Keras / kerasformers.
• SwinV2ImageClassify returns class logits; SwinV2Model returns features (as_backbone=True for multi-scale stages).
• See docs and Loading Weights.
• Upstream / timm checkpoints: SwinV2ImageClassify.from_weights("hf:timm/swinv2_small_window16_256.ms_in1k").

Special Thanks

A huge thank you to the Swin Transformer V2 authors and the timm / Hub communities for creating and releasing these models.

License: see YAML license (usually matches the upstream checkpoint).