# C5W4 Programming Ass UNQ\_C3 **URL:** \ **Category:** Sequence Models\ **Tags:** coursera-platform\ **Created:** [August 6, 2021, 1:16am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458 "2021-08-06T01:16:39Z")\ **Posts on this page:** 9\ **Page:** 1
**Author:** ![walrus](https://avatars.discourse-cdn.com/v4/letter/w/51bf81/32.png) [@walrus](https://community.deeplearning.ai/u/walrus)\ **Post date:** [August 6, 2021, 1:16am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/1 "2021-08-06T01:16:40Z")
I’m stuck. Can’t find the problem. I’m applying the mask and -1e9 when computing the scaled\_attention\_logits. I don’t understand why it is += instead of multiplication. Keep getting the following error. I’m using tf.nn.softmax to compute attention\_weights on the scaled\_attention\_logits, axis = 1. * * * AssertionError Traceback (most recent call last) in 1 # UNIT TEST ----\> 2 scaled\_dot\_product\_attention\_test(scaled\_dot\_product\_attention) ~/work/W4A1/public\_tests.py in scaled\_dot\_product\_attention\_test(target) 73 assert np.allclose(weights, [[0.30719590187072754, 0.5064803957939148, 0.0, 0.18632373213768005], 74 [0.3836517333984375, 0.3836517333984375, 0.0, 0.2326965481042862], —\> 75 [0.3836517333984375, 0.3836517333984375, 0.0, 0.2326965481042862]]), “Wrong masked weights” 76 assert np.allclose(attention, [[0.6928040981292725, 0.18632373213768005], 77 [0.6163482666015625, 0.2326965481042862], AssertionError: Wrong masked weights ---
**Author:** ![TMosh](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/tmosh/32/9321_2.png) [@TMosh](https://community.deeplearning.ai/u/TMosh)\ **Post date:** [August 6, 2021, 1:52am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/2 "2021-08-06T01:52:09Z")
Did you read this part of the instructions carefully? ![image](https://global.discourse-cdn.com/dlai/original/2X/3/3ea8e9249d50e7cadbd38aa9c44ed6e183c41643.png) ---
**Author:** ![walrus](https://avatars.discourse-cdn.com/v4/letter/w/51bf81/32.png) [@walrus](https://community.deeplearning.ai/u/walrus)\ **Post date:** [August 6, 2021, 2:41am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/3 "2021-08-06T02:41:06Z")
I read, multiple them, so I have the following (1.-mask) \* -1e9 ---
**Author:** ![TMosh](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/tmosh/32/9321_2.png) [@TMosh](https://community.deeplearning.ai/u/TMosh)\ **Post date:** [August 6, 2021, 2:51am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/4 "2021-08-06T02:51:02Z")
> [@walrus](#): > > I’m using tf.nn.softmax to compute attention\_weights Try using tf.keras.activations.softmax(…) instead. You don’t need an “axis” argument. ---
**Author:** ![walrus](https://avatars.discourse-cdn.com/v4/letter/w/51bf81/32.png) [@walrus](https://community.deeplearning.ai/u/walrus)\ **Post date:** [August 6, 2021, 3:50am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/5 "2021-08-06T03:50:54Z")
wow! so I don’t need an axis, that solved it, thank you! ---
**Author:** ![walrus](https://avatars.discourse-cdn.com/v4/letter/w/51bf81/32.png) [@walrus](https://community.deeplearning.ai/u/walrus)\ **Post date:** [August 6, 2021, 3:52am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/6 "2021-08-06T03:52:13Z")
Why is axis a problem here? ---
**Author:** ![TMosh](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/tmosh/32/9321_2.png) [@TMosh](https://community.deeplearning.ai/u/TMosh)\ **Post date:** [August 6, 2021, 6:27am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/7 "2021-08-06T06:27:43Z")
If you’re going to specify an axis, you need to be sure it’s the correct one. In this case, the default axis is the correct one. ---
**Author:** ![wayne9598](https://avatars.discourse-cdn.com/v4/letter/w/c0e974/32.png) [@wayne9598](https://community.deeplearning.ai/u/wayne9598)\ **Post date:** [August 6, 2021, 3:08pm UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/8 "2021-08-06T15:08:13Z")
what does this: Multiply (1. - mask) by -1e9 before applying the softmax. even mean?? ---
**Author:** ![TMosh](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/tmosh/32/9321_2.png) [@TMosh](https://community.deeplearning.ai/u/TMosh)\ **Post date:** [February 24, 2024, 7:42am UTC](https://community.deeplearning.ai/t/c5w4-programming-ass-unq-c3/25458/9 "2024-02-24T07:42:15Z")
Closed on February 24, 2024, 7:42am UTC