-
Notifications
You must be signed in to change notification settings - Fork 62
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Erorr in reg_loss #14
Comments
Are you doing this without modifying anything or are you trying to implement your own code? |
without modifying anything |
Probably tf version mismatch, try tf 2.0 not others. |
my tf version is 2.3.1 is that mismatch and i am working with cpu not gpu?? |
Yes this doesn’t work in anything besides tf 2.0
Get Outlook for iOS<https://aka.ms/o0ukef>
…________________________________
From: Saleh Shmali <notifications@github.com>
Sent: Monday, December 14, 2020 4:37:56 AM
To: FurkanOM/tf-faster-rcnn <tf-faster-rcnn@noreply.github.com>
Cc: Oygenblik, David <davido@gatech.edu>; Comment <comment@noreply.github.com>
Subject: Re: [FurkanOM/tf-faster-rcnn] Erorr in reg_loss (#14)
my tf version is 2.3.1 is that mismatch and i am working with cpu not gpu??
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub<#14 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ANM2UXYW3ZTHEN3VHGQB75DSUXMHJANCNFSM4UZLRQAA>.
|
Also you’re gonna have a hell of a long time training it on cpu, on my powerful gpu training on my custom data set takes roughly a day and a half
Get Outlook for iOS<https://aka.ms/o0ukef>
…________________________________
From: Oygenblik, David <davido@gatech.edu>
Sent: Monday, December 14, 2020 5:19:46 AM
To: FurkanOM/tf-faster-rcnn <reply@reply.github.com>; FurkanOM/tf-faster-rcnn <tf-faster-rcnn@noreply.github.com>
Cc: Comment <comment@noreply.github.com>
Subject: Re: [FurkanOM/tf-faster-rcnn] Erorr in reg_loss (#14)
Yes this doesn’t work in anything besides tf 2.0
Get Outlook for iOS<https://aka.ms/o0ukef>
________________________________
From: Saleh Shmali <notifications@github.com>
Sent: Monday, December 14, 2020 4:37:56 AM
To: FurkanOM/tf-faster-rcnn <tf-faster-rcnn@noreply.github.com>
Cc: Oygenblik, David <davido@gatech.edu>; Comment <comment@noreply.github.com>
Subject: Re: [FurkanOM/tf-faster-rcnn] Erorr in reg_loss (#14)
my tf version is 2.3.1 is that mismatch and i am working with cpu not gpu??
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub<#14 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ANM2UXYW3ZTHEN3VHGQB75DSUXMHJANCNFSM4UZLRQAA>.
|
i tried on tf2.0 it was aproblems on tensorflow datasets like this : |
what version of tfds is suitable? |
Hi, |
The batch_size problem with tf 2.4 is the result of a change in the huber loss function: tf2.0:tf.losses.Huber(reduction=tf.losses.Reduction.NONE)(tf.ones(shape=(2,16500,4)),tf.ones(shape=(2,16500,4))).shape tf 2.4:tf.losses.Huber(reduction=tf.losses.Reduction.NONE)(tf.ones(shape=(2,16500,4)),tf.ones(shape=(2,16500,4))).shape Fixing the bug for tf 2.4 is easy: Remove the additional reduce_sum in the reg_loss in line 218. |
Anyone know with Tensorflow 2.8.0, seems the error now change to
|
Hi @ruman1609, |
Hi @colindecourt I already do like you do, and like @wr0112358 did. Still the error occurred so right now I using 2.1.0 version. The error log was just like I sent before is using TF 2.8.0. That's why I downgrade my TF to 2.1.0 |
1- AssertionError: Could not compute output Tensor("rpn_reg_loss/truediv:0", shape=(), dtype=float32) in faster rcnn
2-Incompatible shapes: [8,9216] vs. [8]
[[node gradient_tape/reg_loss/mul/BroadcastGradientArgs (defined at e:/faster_rcnn/rpn_trainer.py:60) ]] [Op:__inference_train_function_12539] in rpn alone
The text was updated successfully, but these errors were encountered: