nielsr HF staff commited on
Commit
cc34856
1 Parent(s): c072a43

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +83 -0
README.md ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - vision
5
+ - image-classification
6
+ datasets:
7
+ - imagenet-21k
8
+ - imagenet-1k
9
+ widget:
10
+ - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
11
+ example_title: Tiger
12
+ - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg
13
+ example_title: Teapot
14
+ - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg
15
+ example_title: Palace
16
+ ---
17
+
18
+ # ConvNeXT (xlarge-sized model)
19
+
20
+ ConvNeXT model trained on ImageNet-1k at resolution 224x224. It was introduced in the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Liu et al. and first released in [this repository](https://github.com/facebookresearch/ConvNeXt).
21
+
22
+ Disclaimer: The team releasing ConvNeXT did not write a model card for this model so this model card has been written by the Hugging Face team.
23
+
24
+ ## Model description
25
+
26
+ ConvNeXT is a pure convolutional model (ConvNet), inspired by the design of Vision Transformers, that claims to outperform them. The authors started from a ResNet and "modernized" its design by taking the Swin Transformer as inspiration.
27
+
28
+ ![model image](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/convnext_architecture.png)
29
+
30
+ ## Intended uses & limitations
31
+
32
+ You can use the raw model for image classification. See the [model hub](https://huggingface.co/models?search=convnext) to look for
33
+ fine-tuned versions on a task that interests you.
34
+
35
+ ### How to use
36
+
37
+ Here is how to use this model to classify an image of the COCO 2017 dataset into one of the 1,000 ImageNet classes:
38
+
39
+ ```python
40
+ from transformers import ConvNextFeatureExtractor, ConvNextForImageClassification
41
+ import torch
42
+ from datasets import load_dataset
43
+
44
+ dataset = load_dataset("huggingface/cats-image")
45
+ image = dataset["test"]["image"][0]
46
+
47
+ feature_extractor = ConvNextFeatureExtractor.from_pretrained("facebook/convnext-xlarge-224-22k-1k")
48
+ model = ConvNextForImageClassification.from_pretrained("facebook/convnext-xlarge-224-22k-1k")
49
+
50
+ inputs = feature_extractor(image, return_tensors="pt")
51
+
52
+ with torch.no_grad():
53
+ logits = model(**inputs).logits
54
+
55
+ # model predicts one of the 1000 ImageNet classes
56
+ predicted_label = logits.argmax(-1).item()
57
+ print(model.config.id2label[predicted_label]),
58
+ ```
59
+
60
+ For more code examples, we refer to the [documentation](https://huggingface.co/docs/transformers/master/en/model_doc/convnext).
61
+
62
+ ### BibTeX entry and citation info
63
+
64
+ ```bibtex
65
+ @article{DBLP:journals/corr/abs-2201-03545,
66
+ author = {Zhuang Liu and
67
+ Hanzi Mao and
68
+ Chao{-}Yuan Wu and
69
+ Christoph Feichtenhofer and
70
+ Trevor Darrell and
71
+ Saining Xie},
72
+ title = {A ConvNet for the 2020s},
73
+ journal = {CoRR},
74
+ volume = {abs/2201.03545},
75
+ year = {2022},
76
+ url = {https://arxiv.org/abs/2201.03545},
77
+ eprinttype = {arXiv},
78
+ eprint = {2201.03545},
79
+ timestamp = {Thu, 20 Jan 2022 14:21:35 +0100},
80
+ biburl = {https://dblp.org/rec/journals/corr/abs-2201-03545.bib},
81
+ bibsource = {dblp computer science bibliography, https://dblp.org}
82
+ }
83
+ ```