ubuntu16.0+cuda8.0+cudnn6.0 training ModelNet40 about kpconv

[HuguesTHOMAS]

Hello @Ailce8862

What version of Tensoflow are you using?

The op tf.batch_gather is now deprecated. So you have two solutions:

1. Use an older version of tensorflow (1.12)
2. Replace tf.batch_gather with tf.gather and use the batch_dims argument.

Best,
Hugues

[Ailce8862]

my tensorflow version is 1.4 ubuntu16.0+cuda8.0+cudnn6.0 running on conda. If I want to choose the second method, how to modify the program. | | 梁攀 | | 邮箱：[email protected] | 签名由 网易邮箱大师 定制 On 10/01/2019 19:17, Hugues THOMAS wrote: Hello @Ailce8862 What version of Tensoflow are you using? The op tf.batch_gather is now deprecated. So you have two solutions: Use an older version of tensorflow (1.12) Replace tf.batch_gather with tf.gather and use the batch_dims argument. Best, Hugues — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[HuguesTHOMAS]

Hello @Ailce8862

Actually TF 1.4 is too old and does not include the function "batch_gather". You will not be able to run the code unless you upgrade your version.

Best,
Hugues

[Ailce8862]

thank you for your help | | 梁攀 | | 邮箱：[email protected] | 签名由 网易邮箱大师 定制 On 10/03/2019 22:37, Hugues THOMAS wrote: Hello @Ailce8862 Actually TF 1.4 is too old and does not include the function "batch_gather". You will not be able to run the code unless you upgrade your version. Best, Hugues — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.