Viredery/tf-eager-fasterrcnn
Faster R-CNN R-101-FPN model was implemented with TensorFlow2.0 eager execution.
Jupyter NotebookMIT
Issues
- 1
- 2
Train error
#20 opened by AlessandroZoia-dev - 3
Can't load weights when training the model
#27 opened by khatchad - 0
Training return loss "nan"
#26 opened by ZeroSum0x00 - 1
- 0
请问作者有没有遇到在训练COCO数据集时出现loss=nan的问题
#24 opened by 1522315 - 1
关于loss计算的疑问
#21 opened by Julius-ZCJ - 1
How to train own data set?
#22 opened by 10207011 - 0
- 0
作者您好,请问这套代码是怎么计算Map的?
#19 opened by jxuepeng - 3
请问作者,这个代码能使用resnet50的网络结构吗?不使用与训练模型训练
#18 opened by jxuepeng - 1
- 0
- 1
InvalidArgumentError: indices[204] = [204, -1] does not index into shape [256,81,4] [Op:TensorScatterUpdate]
#15 opened by arunraja-hub - 2
- 1
Training losses all nan values
#14 opened by LMD93 - 1
when I run section #overfit a sample# in train_model.ipynb, it appears. How can I solve it?
#13 opened by ily666666 - 2
rpn层输出的proposals中带有nan值问题
#11 opened by Julius-ZCJ - 8
what is img_metas?
#6 opened by QiuSYang - 1
Question about pretrained model
#8 opened by Noba1anc3 - 3
Problem in train_model.ipynb
#10 opened by AndrewNest - 6
Problems in inspect_model.ipynb
#7 opened by kuailedagongzai - 6
- 2
Cannot download weights from Baidu
#5 opened by MichelHalmes - 5
- 9
- 5
different padded size on training and test
#2 opened by dragen1860 - 1
TypeError: simple_test_bboxes() missing 1 requiresd positional argument `proposals`
#1 opened by dragen1860