plenari

Frobenius 范数计算的相加的顺序1,2是不是应该为2,1？计算中发现有错误，定位到这里了。 还有BIlstm里若pooling为all，outp为什么需要交换一下0,1轴呢？ 你在SelfAttentiveEncoder标记的输出大小也与交换后的不符合啊size = outp.size() # [bsz, len, nhid*2]？ 请问其中由什么道理吗？

distributed training

4

Does keras support distributed training? Can I use tensorflow's distributed training tools?

correct_yolo_boxes

5

Detetion.custom.__init__.CustomDetectionUtils.correct_yolo_boxes @staticmethod def correct_yolo_boxes(boxes, image_h, image_w, net_h, net_w): new_w, new_h = net_w, net_h for i in range(len(boxes)): x_offset, x_scale = (net_w - new_w) / 2. / net_w, float(new_w) / net_w...

@ operator

2

hello,how does the operator @ work?

FileWriter Warning

6

Finished trainig!Saving model to disk.保存模型的时候 attempting to use a closed FileWriter .The operation will be a noop unless the FileWriter is explicitly reopened.

Hi @grahamrow： Thank you very much for creating such a wonderful work. However, I found that muview2 could not produce a synchronous coordinate system . Excuse me, can you create...