- Keras implementation of Github google-research/maxvit. Paper PDF 2204.01697 MaxViT: Multi-Axis Vision Transformer.
- Model weights ported from official publication.
| Model | Params | FLOPs | Input | Top1 Acc | Download |
|---|---|---|---|---|---|
| MaxViT_Tiny | 31M | 5.6G | 224 | 83.62 | tiny_224_imagenet.h5 |
| MaxViT_Tiny | 31M | 17.7G | 384 | 85.24 | tiny_384_imagenet.h5 |
| MaxViT_Tiny | 31M | 33.7G | 512 | 85.72 | tiny_512_imagenet.h5 |
| MaxViT_Small | 69M | 11.7G | 224 | 84.45 | small_224_imagenet.h5 |
| MaxViT_Small | 69M | 36.1G | 384 | 85.74 | small_384_imagenet.h5 |
| MaxViT_Small | 69M | 67.6G | 512 | 86.19 | small_512_imagenet.h5 |
| MaxViT_Base | 119M | 24.2G | 224 | 84.95 | base_224_imagenet.h5 |
| - imagenet21k | 135M | 24.2G | 224 | base_224_imagenet21k.h5 | |
| MaxViT_Base | 119M | 74.2G | 384 | 86.34 | base_384_imagenet.h5 |
| - imagenet21k-ft1k | 119M | 74.2G | 384 | 88.24 | base_384_21k-ft1k.h5 |
| MaxViT_Base | 119M | 138.5G | 512 | 86.66 | base_512_imagenet.h5 |
| - imagenet21k-ft1k | 119M | 138.5G | 512 | 88.38 | base_512_21k-ft1k.h5 |
| MaxViT_Large | 212M | 43.9G | 224 | 85.17 | large_224_imagenet.h5 |
| - imagenet21k | 233M | 43.9G | 224 | large_224_imagenet21k.h5 | |
| MaxViT_Large | 212M | 133.1G | 384 | 86.40 | large_384_imagenet.h5 |
| - imagenet21k-ft1k | 212M | 133.1G | 384 | 88.32 | large_384_21k-ft1k.h5 |
| MaxViT_Large | 212M | 245.4G | 512 | 86.70 | large_512_imagenet.h5 |
| - imagenet21k-ft1k | 212M | 245.4G | 512 | 88.46 | large_512_21k-ft1k.h5 |
| MaxViT_XLarge, imagenet21k | 507M | 97.7G | 224 | xlarge_224_imagenet21k.h5 | |
| MaxViT_XLarge, imagenet21k-ft1k | 475M | 293.7G | 384 | 88.51 | xlarge_384_21k-ft1k.h5 |
| MaxViT_XLarge, imagenet21k-ft1k | 475M | 535.2G | 512 | 88.70 | xlarge_512_21k-ft1k.h5 |
from keras_cv_attention_models import maxvit
# Will download and load pretrained imagenet weights.
mm = maxvit.MaxViT_Tiny(pretrained="imagenet")
# Run prediction
import tensorflow as tf
from tensorflow import keras
from skimage.data import chelsea
imm = keras.applications.imagenet_utils.preprocess_input(chelsea(), mode='torch') # Chelsea the cat
pred = mm(tf.expand_dims(tf.image.resize(imm, mm.input_shape[1:3]), 0)).numpy()
print(keras.applications.imagenet_utils.decode_predictions(pred)[0])
# [('n02124075', 'Egyptian_cat', 0.700786), ('n02123159', 'tiger_cat', 0.04504126), ...]Change input resolution.
from keras_cv_attention_models import maxvit
mm = maxvit.MaxViT_Tiny(input_shape=(117, 393, 3), pretrained="imagenet")
# >>>> Load pretrained from: ~/.keras/models/maxvit_tiny_224_imagenet.h5
# ...
# >>>> Reload mismatched weights: 224 -> (117, 393)
# >>>> Reload layer: stack_1_block_1/block_window_mhsa/pos_emb
# ...
# >>>> Reload layer: stack_4_block_2/grid_window_mhsa/pos_emb
# Run prediction
from skimage.data import chelsea
preds = mm(mm.preprocess_input(chelsea()))
print(mm.decode_predictions(preds))
# [('n02124075', 'Egyptian_cat', 0.77995235), ('n02123045', 'tabby', 0.017138876), ...]Actual num_classes for "imagenet21k" models is 21843.
from keras_cv_attention_models import maxvit
# Will download and load pretrained imagenet weights.
mm = maxvit.MaxViT_Large(pretrained="imagenet21k", num_classes=21843)
# Run prediction
from skimage.data import chelsea
preds = mm(mm.preprocess_input(chelsea()))
print(mm.decode_predictions(preds))
# [('n02124075', 'Egyptian_cat', 0.059920408), ('n02120997', 'feline, felid', 0.04957321), ...]