- Rewrite Version of clovaai/deep-text-recognition-benchmark. Make attempts of deployment on edge devices.
- Based on Pytorch(training), MNN(Inference) and NCNN(Inference)
- Based on
Pytorch1.3, Python3.6 - Original repo please check clovaai/deep-text-recognition-benchmark.
- Default training the TPS-ResNet-BiLSTM-Attn model.
- Please check the "When you need to train on your own dataset or Non-Latin language datasets." in the README.md of the original repo.
- Once when you prepare the datasets well, there should be "
data.mdb" and "lock.mdb" under the output folder. - Assign the value of "train_data" and "valid_data" with the absolute addresses of the output folders.
- Modify the values in
train.py:
| Parameters | Position | Remarks |
|---|---|---|
train_data |
train.py, Line:223 |
lmdb trainset folder |
valid_data |
train.py, Line:224 |
lmdb valset folder |
batch_max_length |
train.py, Line:246 |
width of the resized image |
imgH |
train.py, Line:247 |
height of the resized image |
imgW |
train.py, Line:248 |
width of the resized image |
character |
train.py, Line:250 |
all your characters of your vocabulary you want to recognize |
-
IMPORTANT:If there're special characters in your vocabulary(e.g."-"), please modify the
dataset.pyLine:170. -
The output weights (end with
.pth) are saved in the./saved_modelsby default. -
Then run
python train.py.
- You can use netron to open your
*.pthto view the model architecture.
- Please check the README.md under the
./inference.
- Be free to open issues or pull requests.
Not Finished yet!!!