rwightman HF staff commited on
Commit
c7eccd6
1 Parent(s): 3634810
Files changed (4) hide show
  1. README.md +130 -0
  2. config.json +35 -0
  3. model.safetensors +3 -0
  4. pytorch_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,130 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - image-classification
4
+ - timm
5
+ library_name: timm
6
+ license: apache-2.0
7
+ datasets:
8
+ - imagenet-1k
9
+ ---
10
+ # Model card for pvt_v2_b2_li
11
+
12
+ A PVT-v2 (Pyramid Vision Transformer) image classification model. Trained on ImageNet-1k by paper authors.
13
+
14
+ ## Model Details
15
+ - **Model Type:** Image classification / feature backbone
16
+ - **Model Stats:**
17
+ - Params (M): 22.6
18
+ - GMACs: 3.9
19
+ - Activations (M): 27.6
20
+ - Image size: 224 x 224
21
+ - **Papers:**
22
+ - PVT v2: Improved Baselines with Pyramid Vision Transformer: https://arxiv.org/abs/2106.13797
23
+ - **Dataset:** ImageNet-1k
24
+ - **Original:** https://github.com/whai362/PVT
25
+
26
+ ## Model Usage
27
+ ### Image Classification
28
+ ```python
29
+ from urllib.request import urlopen
30
+ from PIL import Image
31
+ import timm
32
+
33
+ img = Image.open(urlopen(
34
+ 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
35
+ ))
36
+
37
+ model = timm.create_model('pvt_v2_b2_li', pretrained=True)
38
+ model = model.eval()
39
+
40
+ # get model specific transforms (normalization, resize)
41
+ data_config = timm.data.resolve_model_data_config(model)
42
+ transforms = timm.data.create_transform(**data_config, is_training=False)
43
+
44
+ output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1
45
+
46
+ top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5)
47
+ ```
48
+
49
+ ### Feature Map Extraction
50
+ ```python
51
+ from urllib.request import urlopen
52
+ from PIL import Image
53
+ import timm
54
+
55
+ img = Image.open(urlopen(
56
+ 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
57
+ ))
58
+
59
+ model = timm.create_model(
60
+ 'pvt_v2_b2_li',
61
+ pretrained=True,
62
+ features_only=True,
63
+ )
64
+ model = model.eval()
65
+
66
+ # get model specific transforms (normalization, resize)
67
+ data_config = timm.data.resolve_model_data_config(model)
68
+ transforms = timm.data.create_transform(**data_config, is_training=False)
69
+
70
+ output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1
71
+
72
+ for o in output:
73
+ # print shape of each feature map in output
74
+ # e.g.:
75
+ # torch.Size([1, 64, 56, 56])
76
+ # torch.Size([1, 128, 28, 28])
77
+ # torch.Size([1, 320, 14, 14])
78
+ # torch.Size([1, 512, 7, 7])
79
+
80
+ print(o.shape)
81
+ ```
82
+
83
+ ### Image Embeddings
84
+ ```python
85
+ from urllib.request import urlopen
86
+ from PIL import Image
87
+ import timm
88
+
89
+ img = Image.open(urlopen(
90
+ 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
91
+ ))
92
+
93
+ model = timm.create_model(
94
+ 'pvt_v2_b2_li',
95
+ pretrained=True,
96
+ num_classes=0, # remove classifier nn.Linear
97
+ )
98
+ model = model.eval()
99
+
100
+ # get model specific transforms (normalization, resize)
101
+ data_config = timm.data.resolve_model_data_config(model)
102
+ transforms = timm.data.create_transform(**data_config, is_training=False)
103
+
104
+ output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor
105
+
106
+ # or equivalently (without needing to set num_classes=0)
107
+
108
+ output = model.forward_features(transforms(img).unsqueeze(0))
109
+ # output is unpooled, a (1, 512, 7, 7) shaped tensor
110
+
111
+ output = model.forward_head(output, pre_logits=True)
112
+ # output is a (1, num_features) shaped tensor
113
+ ```
114
+
115
+ ## Model Comparison
116
+ Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results).
117
+
118
+ ## Citation
119
+ ```bibtex
120
+ @article{wang2021pvtv2,
121
+ title={Pvtv2: Improved baselines with pyramid vision transformer},
122
+ author={Wang, Wenhai and Xie, Enze and Li, Xiang and Fan, Deng-Ping and Song, Kaitao and Liang, Ding and Lu, Tong and Luo, Ping and Shao, Ling},
123
+ journal={Computational Visual Media},
124
+ volume={8},
125
+ number={3},
126
+ pages={1--10},
127
+ year={2022},
128
+ publisher={Springer}
129
+ }
130
+ ```
config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "pvt_v2_b2_li",
3
+ "num_classes": 1000,
4
+ "num_features": 512,
5
+ "global_pool": "avg",
6
+ "pretrained_cfg": {
7
+ "custom_load": false,
8
+ "input_size": [
9
+ 3,
10
+ 224,
11
+ 224
12
+ ],
13
+ "fixed_input_size": false,
14
+ "interpolation": "bicubic",
15
+ "crop_pct": 0.9,
16
+ "crop_mode": "center",
17
+ "mean": [
18
+ 0.485,
19
+ 0.456,
20
+ 0.406
21
+ ],
22
+ "std": [
23
+ 0.229,
24
+ 0.224,
25
+ 0.225
26
+ ],
27
+ "num_classes": 1000,
28
+ "pool_size": [
29
+ 7,
30
+ 7
31
+ ],
32
+ "first_conv": "patch_embed.proj",
33
+ "classifier": "head"
34
+ }
35
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7ad7e1de410c024ac81e631a4fe6a570125877f434e1da037f23064d3711e85e
3
+ size 90248458
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:43d943de3200a1ab70957f4294170aff6257e3a66abd418262534e17d2293729
3
+ size 90340533