File size: 10,416 Bytes
a1d409e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
<!---
Copyright 2022 The HuggingFace Team. All rights reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->

# Image pretraining examples

This directory contains Python scripts that allow you to pre-train Transformer-based vision models (like [ViT](https://huggingface.co/docs/transformers/model_doc/vit), [Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)) on your own data, after which you can easily load the weights into a [`AutoModelForImageClassification`](https://huggingface.co/docs/transformers/model_doc/auto#transformers.AutoModelForImageClassification). It currently includes scripts for:
- [SimMIM](#simmim) (by Microsoft Research)
- [MAE](#mae) (by Facebook AI).

NOTE: If you encounter problems/have suggestions for improvement, open an issue on Github and tag @NielsRogge.


## SimMIM

The `run_mim.py` script can be used to pre-train any Transformer-based vision model in the library (concretly, any model supported by the `AutoModelForMaskedImageModeling` API) for masked image modeling as proposed in [SimMIM: A Simple Framework for Masked Image Modeling](https://arxiv.org/abs/2111.09886) using PyTorch.

<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/simmim_architecture.jpg"
alt="drawing" width="300"/> 

<small> SimMIM framework. Taken from the <a href="https://arxiv.org/abs/2111.09886">original paper</a>. </small>

The goal for the model is to predict raw pixel values for the masked patches, using just a linear layer as prediction head. The model is trained using a simple L1 loss.

### Using datasets from πŸ€— datasets

Here we show how to pre-train a `ViT` from scratch for masked image modeling on the [cifar10](https://huggingface.co/datasets/cifar10) dataset.

Alternatively, one can decide to further pre-train an already pre-trained (or fine-tuned) checkpoint from the [hub](https://huggingface.co/). This can be done by setting the `model_name_or_path` argument to "google/vit-base-patch16-224-in21k" for example (and not specifying the `model_type` argument).

```bash
!python run_mim.py \
    --model_type vit \
    --output_dir ./outputs/ \
    --overwrite_output_dir \
    --remove_unused_columns False \
    --label_names bool_masked_pos \
    --do_train \
    --do_eval \
    --learning_rate 2e-5 \
    --weight_decay 0.05 \
    --num_train_epochs 100 \
    --per_device_train_batch_size 8 \
    --per_device_eval_batch_size 8 \
    --logging_strategy steps \
    --logging_steps 10 \
    --evaluation_strategy epoch \
    --save_strategy epoch \
    --load_best_model_at_end True \
    --save_total_limit 3 \
    --seed 1337
```

Here, we train for 100 epochs with a learning rate of 2e-5. Note that the SimMIM authors used a more sophisticated learning rate schedule, see the [config files](https://github.com/microsoft/SimMIM/blob/main/configs/vit_base__800ep/simmim_pretrain__vit_base__img224__800ep.yaml) for more info. One can easily tweak the script to include this learning rate schedule (several learning rate schedulers are supported via the [training arguments](https://huggingface.co/docs/transformers/main_classes/trainer#transformers.TrainingArguments)).

We can also for instance replicate the pre-training of a Swin Transformer using the same architecture as used by the SimMIM authors. For this, we first create a custom configuration and save it locally:

```python
from transformers import SwinConfig

IMAGE_SIZE = 192
PATCH_SIZE = 4
EMBED_DIM = 128
DEPTHS = [2, 2, 18, 2]
NUM_HEADS = [4, 8, 16, 32]
WINDOW_SIZE = 6

config = SwinConfig(
    image_size=IMAGE_SIZE,
    patch_size=PATCH_SIZE,
    embed_dim=EMBED_DIM,
    depths=DEPTHS,
    num_heads=NUM_HEADS,
    window_size=WINDOW_SIZE,
)
config.save_pretrained("path_to_config")
```

Next, we can run the script by providing the path to this custom configuration (replace `path_to_config` below with your path):

```bash
!python run_mim.py \
    --config_name_or_path path_to_config \
    --model_type swin \
    --output_dir ./outputs/ \
    --overwrite_output_dir \
    --remove_unused_columns False \
    --label_names bool_masked_pos \
    --do_train \
    --do_eval \
    --learning_rate 2e-5 \
    --num_train_epochs 5 \
    --per_device_train_batch_size 8 \
    --per_device_eval_batch_size 8 \
    --logging_strategy steps \
    --logging_steps 10 \
    --evaluation_strategy epoch \
    --save_strategy epoch \
    --load_best_model_at_end True \
    --save_total_limit 3 \
    --seed 1337
```

This will train a Swin Transformer from scratch.

### Using your own data

To use your own dataset, the training script expects the following directory structure:

```bash
root/dog/xxx.png
root/dog/xxy.png
root/dog/[...]/xxz.png

root/cat/123.png
root/cat/nsdf3.png
root/cat/[...]/asd932_.png
```

Note that you can put images in dummy subfolders, whose names will be ignored by default (as labels aren't required). You can also just place all images into a single dummy subfolder. Once you've prepared your dataset, you can run the script like this:

```bash
python run_mim.py \
    --model_type vit \
    --dataset_name nateraw/image-folder \
    --train_dir <path-to-train-root> \
    --output_dir ./outputs/ \
    --remove_unused_columns False \
    --label_names bool_masked_pos \
    --do_train \
    --do_eval
```

## MAE

The `run_mae.py` script can be used to pre-train a Vision Transformer as a masked autoencoder (MAE), as proposed in [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377). The script can be used to train a `ViTMAEForPreTraining` model in the Transformers library, using PyTorch. After self-supervised pre-training, one can load the weights of the encoder directly into a `ViTForImageClassification`. The MAE method allows for learning high-capacity models that generalize well: e.g., a vanilla ViT-Huge model achieves the best accuracy (87.8%) among methods that use only ImageNet-1K data.

The goal for the model is to predict raw pixel values for the masked patches. As the model internally masks patches and learns to reconstruct them, there's no need for any labels. The model uses the mean squared error (MSE) between the reconstructed and original images in the pixel space.

### Using datasets from πŸ€— `datasets`

One can use the following command to pre-train a `ViTMAEForPreTraining` model from scratch on the [cifar10](https://huggingface.co/datasets/cifar10) dataset:

```bash
python run_mae.py \
    --dataset_name cifar10 \
    --output_dir ./vit-mae-demo \
    --remove_unused_columns False \
    --label_names pixel_values \
    --mask_ratio 0.75 \
    --norm_pix_loss \
    --do_train \
    --do_eval \
    --base_learning_rate 1.5e-4 \
    --lr_scheduler_type cosine \
    --weight_decay 0.05 \
    --num_train_epochs 800 \
    --warmup_ratio 0.05 \
    --per_device_train_batch_size 8 \
    --per_device_eval_batch_size 8 \
    --logging_strategy steps \
    --logging_steps 10 \
    --evaluation_strategy epoch \
    --save_strategy epoch \
    --load_best_model_at_end True \
    --save_total_limit 3 \
    --seed 1337
```

Here we set:
- `mask_ratio` to 0.75 (to mask 75% of the patches for each image)
- `norm_pix_loss` to use normalized pixel values as target (the authors reported better representations with this enabled) 
- `base_learning_rate` to 1.5e-4. Note that the effective learning rate is computed by the [linear schedule](https://arxiv.org/abs/1706.02677): `lr` = `blr` * total training batch size / 256. The total training batch size is computed as `training_args.train_batch_size` * `training_args.gradient_accumulation_steps` * `training_args.world_size`.

This replicates the same hyperparameters as used in the original implementation, as shown in the table below.

<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/mae_pretraining_setting.png"
alt="drawing" width="300"/> 

<small> Original hyperparameters. Taken from the <a href="https://arxiv.org/abs/2111.06377">original paper</a>. </small>

Alternatively, one can decide to further pre-train an already pre-trained (or fine-tuned) checkpoint from the [hub](https://huggingface.co/). This can be done by setting the `model_name_or_path` argument to "facebook/vit-mae-base" for example.


### Using your own data

To use your own dataset, the training script expects the following directory structure:

```bash
root/dog/xxx.png
root/dog/xxy.png
root/dog/[...]/xxz.png

root/cat/123.png
root/cat/nsdf3.png
root/cat/[...]/asd932_.png
```

Note that you can put images in dummy subfolders, whose names will be ignored by default (as labels aren't required). You can also just place all images into a single dummy subfolder. Once you've prepared your dataset, you can run the script like this:

```bash
python run_mae.py \
    --model_type vit_mae \
    --dataset_name nateraw/image-folder \
    --train_dir <path-to-train-root> \
    --output_dir ./outputs/ \
    --remove_unused_columns False \
    --label_names pixel_values \
    --do_train \
    --do_eval
```

#### πŸ’‘ The above will split the train dir into training and evaluation sets
  - To control the split amount, use the `--train_val_split` flag.
  - To provide your own validation split in its own directory, you can pass the `--validation_dir <path-to-val-root>` flag.


## Sharing your model on πŸ€— Hub

0. If you haven't already, [sign up](https://huggingface.co/join) for a πŸ€— account

1. Make sure you have `git-lfs` installed and git set up.

```bash
$ apt install git-lfs
$ git config --global user.email "you@example.com"
$ git config --global user.name "Your Name"
```

2. Log in with your HuggingFace account credentials using `huggingface-cli`

```bash
$ huggingface-cli login
# ...follow the prompts
```

3. When running the script, pass the following arguments:

```bash
python run_xxx.py \
    --push_to_hub \
    --push_to_hub_model_id <name-of-your-model> \
    ...
```