Distributed ndcg #3054
Distributed ndcg #3054
Conversation
| discounted_gains = torch.tensor( | ||
| [_tie_averaged_dcg(y_p, y_t, discount_cumsum, device) for y_p, y_t in zip(y_pred, y_true)], device=device | ||
| ) |
There was a problem hiding this comment.
Let's check here: https://github.com/catalyst-team/catalyst/blob/master/catalyst/metrics/_ndcg.py if there is another way to implement this in a vectorized way, https://github.com/pytorch/ignite/pull/2632/files#r930048810
There was a problem hiding this comment.
It is purely perf reasons. For example, computing s1 below will be faster than s2:
tensor = torch.rand(100)
s1 = tensor.sum()
s2 = 0
for v in tensor:
s2 += v
ignite/metrics/recsys/ndcg.py
Outdated
| return normalized_gain | ||
|
|
||
|
|
||
| class NDCG(Metric): |
There was a problem hiding this comment.
Also, we need to write a docstring like here:
ignite/ignite/metrics/accuracy.py
Line 94 in 34a707e
Please read this section of contributing guide: https://github.com/pytorch/ignite/blob/master/CONTRIBUTING.md#writing-documentation, especially about .. versionadded::
|
I ve deleted "test_output_cuda", added "n_epoch" and alligned code format for now. I will work on vectorizing and wrting docstring @vfdev-5 |
…into distributed-ndcg
|
@ili0820 any blockers that we can help to move forward with this PR ? |
|
@vfdev-5 sorry for no updates for a long time, I was still struggling with vectorizing for loop. 😢 |
|
@ili0820 no worries, let's update the docstring and merge this PR. Vectorization part could be done later in a follow-up PR. |
|
@vfdev-5 wow sounds great. does it matter if i copy and paste docstring from sklearn or other opensource? |
|
@ili0820 yes, you can inspire from scikit-learn to write ignite's docstring. Please check this contrib part: https://github.com/pytorch/ignite/blob/master/CONTRIBUTING.md#writing-documentation |
|
@vfdev-5 thanks I ll look into that. 😄 |
…into distributed-ndcg
|
@ili0820 can you please also add the metric to this list: https://github.com/pytorch/ignite/blob/master/docs/source/metrics.rst Also please fix formatting issues: https://github.com/pytorch/ignite/actions/runs/6251715047/job/16981806158?pr=3054#step:11:104 |
|
@vfdev-5 I accidently deleted the branch................................ 🤢 |
|
restored. sorry for inconvenience |
|
@ili0820 can you update the PR according to the review, please? |
Fixes #2632
This PR is based on #2632 (thanks a lot to kamalojasv181!)
Description:
I ve tried to fix distributed NDCG tests like the examples. I am not really sure whether this is the way you guys wanted, but still, I tried my best and let me know which parts should be fixed (It's my first time contributing, so be kind plz 😯 )
Check list: