Hi Deepti_Prasad
thank you for reviewing the code. I apologize for the inconvenience I am causing you.
Regarding the first correction, I checked the code and there is not a hint on the use of tf.greater_equal (please see below). Instead, by instance, for the use of [tf.math.argmax], it’s very clear.
Hello Mart,
you didn’t cause me any inconvenience, not all hints will be instructions where you are mentioning, I had shared a statement in my first comment
apply tensorflow function (with probability >= threshold) for this code which is part of the instruction given just before the code line you have written. review your grader cell again
This is what is written in the assignment:
My point is that I would not do this if I do not ask someone in the forum. I did not know about the existence of such tf function until I asked.
The mask should have the same dimension as box_class_scores, so you must have use tf. function here right? and to create a condition True with the probability i.r.t. threshold, you need to use tf function here.
I agree some of the code parts is not clearly or directly mentioned that is when videos, ungraded labs and these instructions would help you. the message you shared mentions to create a filtering mask based on “box_class_scores” by using “threshold”
if you see the next section, you will understand why you need to use tf function for the filtering mask
Use TensorFlow to apply the mask to box_class_scores , boxes and box_classes to filter out the boxes you don’t want. You should be left with just the subset of boxes you want to keep.
(I know you might tell this is for the next code line but if you check the third instruction which mentions
Create a mask by using a threshold. As a reminder: ([0.9, 0.3, 0.4, 0.5, 0.1] < 0.4) returns: [False, True, False, False, True]. The mask should be True for the boxes you want to keep.
So by using tf function for the filtering mask, you are only choosing boxes which holds true value for threshold value.
I hope now you understood!! 
Do not think I am annoyed. We are here to help learner’s doubt. Remember the more doubt one has, the more we tend to remember that point.
If you still have doubt, feel free to ask!!!
Keep Learning!!!
Regards
DP
Hi all,
I did the corrections that Deepti_Prasad kindly suggested but I cannot get my assignment graded. This time the grader lasts forever.
Specifically, these are my corrections (as before, all tests are passing):
1.
2.
3.
Cheers
what are you missing in compare to the image.
Regards
DP
@Mubsi I have instructed on the correction of his codes. Kindly look into this post as after all these corrections he still seems to fail the grader.
Regards
DP
Hello… thanks for the suggestions… I will check up the code disparities if any and resubmit. Will update everyone if i am able to go past it… Thanks a lot for your help
Hello,
you have similar errors related to the other learner, so you can refer the same suggestions, except one error has a typo error.
# maximum number of boxes set to max_boxes and a threshold of iou_threshold (≈1 line) this code lines has a typo error / kindly remove it
Let me know once you pass the grader. I have sent you file with corrections, so you need to review each line as I have mentioned the correct codes with ####CORRECT CODE IN YOUR ASSIGNMENT BOOK
Do not send assignment notebook for other assignments. try looking for issues in search section related to your issue. When you don’t get the issue resolved, create a post regarding that assignment with appropriate explanation.
A bigger part of coding or programming is ability to find your own error and ability to search for solution here in the community as well as on google.
If you still don’t get, then ask for help in discourse community.
Regards
DP
Can I know what do you mean by this statement??
Hi @Deepti_Prasad,
I have same problem with this exercise and I’m trying to understand your previous comments with fixes:
Why you mention numpy function here? Both box_confidence and box_class_probs are TF tensors, so shouldn’t be tf.math.multiply used here to calculate the element-wise product?
Second question is about tf.image.non_max_suppression. According to provided comment in the notebook max_boxes_tensor should be used as third parameter:
But from the comments in this thread it seems that max_boxes should be used instead. Why?
And by the way, are you sure the grader is working fine? I have also A2 exercise failing with same message…
we need to compute the box scores by doing element wise product AS DESCRIBED IN FIGURE 4 (PXC). SO REFER THAT FIGURE, YOU WILL UNDERSTAND WHICH NUMPY FUNCTION NEED TO BE RECALLED.
FOR YOUR SECOND QUESTION, IF YOU NOTICE THE MAX_BOX_TENSOR HAS BEEN ALREADY RECALLED BEFORE YOU WROTE THE CODE FOR THAT CELL, SO YOU ONLY NEED TO USE MAX_BOXES AND NOT MAX_BOXES_TENSOR AS THIRD PARAMETER. ALSO IF YOU SEE MAX_BOXES HAS BEEN GIVEN SOME VALUE WITH OTHER PARAMETERS, SO YOU NEED TO IMPLEMENT THEM WITH THE SAME NAME, RATHER THAN WITH TENSOR AS FOR THAT CODE LINE YOU USE TF.IMAGE.NON_MAX_SUPPRESSION()
READ THE INSTRUCTIONS, YOU WILL UNDERSTAND.
If your all codes are correct and still your grader shows error, then we can raise any issue. Sometimes while typing the codes we end up editing the metadata, so that is why we tell learner to get fresh copy and write their codes again if they have passed all the test.
I am not sure which exercise you stating about, you can send the notebook, let me have a look.
Regards
DP
Hello everyone,
after reviewing and correcting few of the learner’s notebook and still encountering the submission grader error. I eventually have got 2-3 more learner’s having similar issue for both the assignments related to week 3 of course 4 of DLS specialisation. So I have notified the learner technologist @Mubsi about the same via personal DM. He will look into the issue once from his side.
Until then kindly have patience for him to check what is going wrong.
Learner who want to try and for whom I have done code correction, please do the steps below and try once submitting assignment.
Try this as last option, if still stuck, then wait for Mubsi to have a look and address the issue.
Regards
DP
Please note that the grader does not need to see the output from running the notebook. That can actually cause problems in some cases particularly if the output is large. The only reason for the “Run All” is for you to check the results of the local test cases. Then the most reliable way to submit is:
The reason for the “Save” is that in some notebooks, the “Submit” operation does not do an “autosave” for you. It’s not consistent about that. In DLS Course 1, I think it does, but not in some of the assignments in C4. So it’s better to be safe and make sure you do the save manually.
Hi all,
I have the same issue with both of the assignments for week 3 of course 4.
I just want to say I am encountering the same error for both of my assignments at week 3.
Coursera seems to be having technical issues at this time.
Facing similar issue in both the assignments of Week 3. YOLO shows “Grader feedback not found”, and YOLO one shows " Platform Error: We’re experiencing some technical difficulties"
Nothing has changed since I posted on this thread two hours ago.
