You Zhiyuan
You Zhiyuan
Dear authors, Thanks very much for your amazing work! I have some questions about the batch size selection. The batch size is defaulted as 4 per GPU. This is a...
Hi, In train.py, there is a localization option shown like, `parser.add_argument('--localization', default='False', choices=('True', 'False'), help='If True train on (64,64) cropped patches')` However, `if 'False'` actually means True because `'False'` is...
Hi~ Thanks for your excellent work. But I think what you do in Fig. 4 for DETR is ***fully wrong***. In Fig. 4, you conduct dynamic experiments with DETR, and...
Hi～ Thanks for your nice work. I have run your training code of auto-encoder. I checked the code relating metrics, and I found that **_the metrics are calculated with the...
Hi～ Thanks for your great work! I have a small question about the LPIPS evaluation. Actually, you do the LPIPS evaluation with `evaluate_lpips()` in: https://github.com/phizaz/diffae/blob/master/experiment.py#L606. The function `evaluate_lpips()` is defined...
fix dataset bug: bedroom dataset is mistaken as horse dataset
Hi~ Thanks for your excellent work! It certainly helps me a lot. However, ***I could not find the real-world rainy dataset you collected*** (consisted of 50 rainy images used for...
In `models/p2pnet.py`: Line 278: `losses['loss_point'] = loss_bbox.sum() / num_points` Line 335: `weight_dict = {'loss_ce': 1, 'loss_points': args.point_loss_coef}` one is "***loss_points***" and another is "***loss_point***" (an "s" is different) ! This...
博主，你好！ 非常感谢你的开源代码！对于我理解DDPG有非常大的帮助！ 我使用你的开源代码在新的environment中进行DDPG学习的时候，出现了输出的动作总是动作边界值的情况。经过查找资料，我认为batch normalization可以解决这个问题。 于是，我在不同的层之间，使用tf.contrib.layers.batch_norm函数进行batch normalization。但是改动之后，我收到了很长的一堆报错。 ``` Traceback (most recent call last): File "/home/youzhiyuan/anaconda3/envs/pool/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1356, in _do_call return fn(*args) File "/home/youzhiyuan/anaconda3/envs/pool/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1341, in _run_fn options, feed_dict, fetch_list, target_list,...
Hello, Thanks for your great work. I have found that the demo does not support in-painting tasks. So I wonder whether this repo contains the in-painting scripts ? Thanks again.