> [ABINet: Read Like Humans: Autonomous, Bidirectional and Iterative Language Modeling for Scene Text Recognition](https://openaccess.thecvf.com/content/CVPR2021/papers/Fang_Read_Like_Humans_Autonomous_Bidirectional_and_Iterative_Language_Modeling_for_CVPR_2021_paper.pdf)
> Shancheng Fang and Hongtao Xie and Yuxin Wang and Zhendong Mao and Yongdong Zhang
@@ -66,6 +66,7 @@ Supported text recognition algorithms (Click the link to get the tutorial):
...
@@ -66,6 +66,7 @@ Supported text recognition algorithms (Click the link to get the tutorial):
-[x] [SEED](./algorithm_rec_seed_en.md)
-[x] [SEED](./algorithm_rec_seed_en.md)
-[x] [SVTR](./algorithm_rec_svtr_en.md)
-[x] [SVTR](./algorithm_rec_svtr_en.md)
-[x] [ViTSTR](./algorithm_rec_vitstr_en.md)
-[x] [ViTSTR](./algorithm_rec_vitstr_en.md)
-[x] [ABINet](./algorithm_rec_abinet_en.md)
Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation result of these above text recognition (using MJSynth and SynthText for training, evaluate on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE) is as follow:
Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation result of these above text recognition (using MJSynth and SynthText for training, evaluate on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE) is as follow:
...
@@ -85,6 +86,7 @@ Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation r
...
@@ -85,6 +86,7 @@ Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation r
-[3. Model Training / Evaluation / Prediction](#3)
-[3.1 Training](#3-1)
-[3.2 Evaluation](#3-2)
-[3.3 Prediction](#3-3)
-[4. Inference and Deployment](#4)
-[4.1 Python Inference](#4-1)
-[4.2 C++ Inference](#4-2)
-[4.3 Serving](#4-3)
-[4.4 More](#4-4)
-[5. FAQ](#5)
<aname="1"></a>
## 1. Introduction
Paper:
> [ABINet: Read Like Humans: Autonomous, Bidirectional and Iterative Language Modeling for Scene Text Recognition](https://openaccess.thecvf.com/content/CVPR2021/papers/Fang_Read_Like_Humans_Autonomous_Bidirectional_and_Iterative_Language_Modeling_for_CVPR_2021_paper.pdf)
> Shancheng Fang and Hongtao Xie and Yuxin Wang and Zhendong Mao and Yongdong Zhang
> CVPR, 2021
Using MJSynth and SynthText two text recognition datasets for training, and evaluating on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE datasets, the algorithm reproduction effect is as follows:
Please refer to ["Environment Preparation"](./environment_en.md) to configure the PaddleOCR environment, and refer to ["Project Clone"](./clone_en.md) to clone the project code.
<aname="3"></a>
## 3. Model Training / Evaluation / Prediction
Please refer to [Text Recognition Tutorial](./recognition_en.md). PaddleOCR modularizes the code, and training different recognition models only requires **changing the configuration file**.
Training:
Specifically, after the data preparation is completed, the training can be started. The training command is as follows:
```
#Single GPU training (long training period, not recommended)
First, the model saved during the ABINet text recognition training process is converted into an inference model. ( [Model download link]())), you can use the following command to convert:
- If you are training the model on your own dataset and have modified the dictionary file, please pay attention to modify the `character_dict_path` in the configuration file to the modified dictionary file.
- If you modified the input size during training, please modify the `infer_shape` corresponding to ABINet in the `tools/export_model.py` file.
After the conversion is successful, there are three files in the directory:
```
/inference/rec_r45_abinet/
├── inference.pdiparams
├── inference.pdiparams.info
└── inference.pdmodel
```
For ABINet text recognition model inference, the following commands can be executed:
After executing the command, the prediction result (recognized text and score) of the image above is printed to the screen, an example is as follows:
After executing the command, the prediction result (recognized text and score) of the image above is printed to the screen, an example is as follows:
The result is as follows:
The result is as follows:
```shell
```shell
Predicts of ./doc/imgs_words_en/word_10.png:('pain', 0.9265879392623901)
Predicts of ./doc/imgs_words_en/word_10.png:('pain', 0.9465042352676392)
```
```
<aname="4-2"></a>
<aname="4-2"></a>
...
@@ -121,12 +122,146 @@ Not supported
...
@@ -121,12 +122,146 @@ Not supported
1. In the `NRTR` paper, Beam search is used to decode characters, but the speed is slow. Beam search is not used by default here, and greedy search is used to decode characters.
1. In the `NRTR` paper, Beam search is used to decode characters, but the speed is slow. Beam search is not used by default here, and greedy search is used to decode characters.
<aname="6"></a>
## 6. Release Note
1. The release/2.6 version updates the NRTR code structure. The new version of NRTR can load the model parameters of the old version (release/2.5 and before), and you may use the following code to convert the old version model parameters to the new version model parameters:
```python
params=paddle.load('path/'+'.pdparams')# the old version parameters
state_dict=model.state_dict()# the new version model parameters
