liuhengli

Results 7 comments of liuhengli

tf.nn.sparse_softmax_cross_entropy_with_logits(pred_labels, true_labels)

Thank you very much

tf.nn.sparse_softmax_cross_entropy_with_logits(pred_labels, true_labels)

I find your trainer.py file when FLAGS.mode == "image", it has error occurred （boxes_, confidences_ = ssd.single_image(sample)）. your single_image function finally get re but here is boxes_ and confidences_.

请问这个repo怎么测试demo?

@gittigxuy @BobLiu20 請問一下是否在視頻上測試過， 我發現您放出的模型誤檢要比原模型高

感谢分享，请问网盘给出的model和deploy不能加载成功是什么原因？

@zyc4me I also have this error, you have solved.

感谢分享，请问网盘给出的model和deploy不能加载成功是什么原因？

@lyxok1 OK, Thanks. I think this line needs to be deleted in 'video_detection_demo.py' file，‘ transformer.set_channel_swap('data', (2,1,0)) # the reference model has channels in BGR order instead of RGB’. opencv read...

Have you used coco 2017 keypoint data?

@RodrigoGantier why download the modified json files, but some picture can't found

We couldn't connect to 'https://huggingface.co' to load this file

@ljf12345678910 [bert-base-uncased](https://huggingface.co/bert-base-uncased)你直接把这个下载下来就可以了