-
是不是只可以再resnet上进行训练,不能在自己已经训练好的模型上,再进行训练
-
又遇到两个问题。麻烦作者有空看一下哈。
第一个问题:python2 eval_metric.py (用python2运行)
报错:('checkpoint_path:', 'Logs/train/')
[u'Logs/train/model.ckpt-149999']
==========run progrss 0 / 1============
Traceback (most rec…
-
RuntimeError: Cannot compile pse: H:\tensorflow_PSENet-master\pse
-
I find that the pseNet may segment a long text into several segments.
could you advise that how to adjust the pse parameters to alleviate the problem.
Thanks a lot.
-
when I run eval_metric.py,I get this wrong
File "/home/PSENet-Tensorflow-master/eval_metric.py", line 12, in
from model import inference
File "/home/PSENet-Tensorflow-master/model/infere…
-
老哥,不管是用gpu还是cpu,运行predict.py每次都是吐核,很无语,你遇到过吗?
-
What network framework are you using? tensorflow or pytorch? I failed to convert with tensorflow framework.using liuheng92/tensorflow_PSENet code.
-
-
model_path = os.path.join(FLAGS.checkpoint_path, os.path.basename(ckpt_state.model_checkpoint_path))
AttributeError: 'NoneType' object has no attribute 'model_checkpoint_path'
does it because no …
-
作者你好,我在用CTW数据训练的时候,会报这个错误:" index 8 is out of bounds for axis 0 with size 8 "。
这个错误发生在dataset/preprocess.py的150行左右的“if not label[i]:”处。应该是box的shape[0]和label的个数对不上。请问这个该怎么修改?