-
Date: 19/03
Task:
- [x] Prepare data: Crop image into batches of text-lines given ground truth coordinates @patrickphatnguyen
- [x] Train CRNN baseline model: Train a CRNN model with prepared data…
-
您好，我想要使用动态图来实现CRNN+CTC，目前参考https://github.com/PaddlePaddle/models/tree/a92e212932b764e500a833527e0fb772ac9a491a/dygraph/ocr_recognition 这个项目实现了CRNN部分，但是该部分的输出是一个3-D的定长tensor，我看了fluid.layers.warpctc op…
wwjjy updated
4 years ago
-
更新了一下项目结构，在PyTorch1.2.0,cuda10.0下用中文数据集训练了一个Epoch，准确率达到了94%，没有出现NaN的情况。
-
Dear Erdene-Ochir Tuguldur,
You are doing great job! I mean also your TTS project.
The codes are so plain, training is so quick, but at the same time the solutions are powerful and effective.
Ma…
-
在训练的时候这几个参数['reduce_sum_0.tmp_0', 'reduce_sum_2.tmp_0', 'elementwise_sub_0']对应的值为 [nan, 4.791813, 5]，其中第一个参数reduce_sum_0.tmp_0的值为nan这是为什么？之前的训练都是正常的，在executor执行run后输出的结果中出现Nan，求解答
-
First, thanks for your great work :) ! You've done a good job!
Here's my question, I've retrained the model with the option as:
"--select_data MJ-ST --batch_ratio 0.5-0.5 --Transformation None --F…
-
Hello,
i am trying to implement this with keras,
in your code (Model.py) you have mentioned the layer filters, but i am not sure about the rnn and ctc layers.
please have a look at this
![mo…
-
Traceback (most recent call last):
File "/home/ici/Downloads/crnn_chinese_characters_rec-master/crnn_main.py", line 195, in
training(crnn,train_loader,criterion,optimizer)
File "/home/ici/…
-
模型测试加调整一周左右了，做的是日文半角的模型和纯数字英文的模型。测试了很久，各种调试参数，但是loss一直稳定在6~7左右，accuracy也未超过0.5，希望前辈们分享一下各自的经验，有哪些需要注意的细节。本人确实在模型设计上没什么基础，目前只能在参数上动动手脚。谢谢
![2019-12-23_104106](https://user-images.githubusercontent.com…
-
I understand that removing/ deactivating the Theano backend is not an option right now.
I don't have enough insight into these parts to know whether moving the corresponding files into a separate fo…