-
I have tried ABCNet use 2 pre-trained model `ctw1500_attn_R_50.pth` and `tt_e2e_attn_R_50.pth`, but I found there is some diffs between the size of these 2 `pth` file.
`ctw1500_attn_R_50.pth` is 282M…
-
py test.py --trained_model=craft_mlt_25k.pth --refiner_model=craft_refiner_CTW1500.pth --test_folder=test --refine
Loading weights from checkpoint (craft_mlt_25k.pth)
Loading weights of refiner fro…
-
Hi, good day,
I have achieved your promising results in the ablation study for CTW1500 (baseline+GCN **r**:81.45, **p**:83.75, **f**:82.58 ) by following the instructions in the paper. (pretrained…
-
Hello, I pull the lastest mirrors by docker. But after run the command python setup.py build develop, a problem occured as below:
![1](https://user-images.githubusercontent.com/37755235/98454049-1f…
-
I am training ABCNet on 1500 images having single word in an image. Can you tell me for how many steps I could train model. Is there any specific criteria for defining total number of steps?
-
感谢开源这么优秀的工作。关于自己实践过程中遇到的一些小问题，寻求帮助。
1）直接用repo里面提供的model_ctw.pth进行测试，得到HMean 0.78左右，低太多。（model_ctw.pth-1.9G为何比model_ic15.pth-336M大了那么多）
2）用4卡titanxp 分布式训练，在ctw1500数据集上训练得到 loss nan。（ic15数据集上收敛到hm…
-
I used this script to run the demo, but this error occurred:
script:
`python demo/demo.py \
--config-file configs/BAText/CTW1500/attn_R_50.yaml \
--input datasets/CTW1500/ctwtest_text_imag…
-
How to use "Bezier_curve_generation.py" and "generate_bezier_json.py" to generage abcnet train-set?
-
I'm running demo.py like this:
`python demo/demo.py \
--config-file configs/BAText/TotalText/attn_R_50.yaml \
--input datasets/ctw1500/test/text_image/ \
--opts MODEL.WEIGHTS tt_e2e_attn_R_50.pt…
-
When I try to train drrg in dataset totaltext, There are some problems while they didn't happen in icdar2015 and ctw1500. I checked the totaltext dataset and discovered that it seems labled irrefularl…