Updated video classification ref example with new transforms #2935
Conversation
…e-ref-video-classif
…e-ref-video-classif
There was a problem hiding this comment.
Thanks a lot!
I would have done the function / class differently. It's better if we have the reference codes to show what we would like users to do with the code.
Indeed, if we start showing that users can mix classes / functions, we will have less control in the future about what types of things we would be able to change / refactor.
On a different note, @bjuncek can you send a follow-up PR improving the readme with more information about how to train the model?
|
|
||
| ### Multiple GPUs | ||
|
|
||
| Run the training on a single node with 8 GPUs: |
There was a problem hiding this comment.
I believe @bjuncek trained those models on 64 GPUs, it might be good to add a mention here in a follow-up PR
| @@ -0,0 +1,33 @@ | |||
| # Video Classification | |||
|
|
|||
| TODO: Add some info about the context, dataset we use etc | |||
There was a problem hiding this comment.
@bjuncek can you enhance this with information that you used to train on the datasets? We could potentially refer to submitit
Codecov Report
@@ Coverage Diff @@
## master #2935 +/- ##
==========================================
- Coverage 73.41% 71.23% -2.18%
==========================================
Files 99 99
Lines 8801 8765 -36
Branches 1389 1389
==========================================
- Hits 6461 6244 -217
- Misses 1915 2105 +190
+ Partials 425 416 -9
Continue to review full report at Codecov.
|
…#2935) * [WIP] Update ref example video classification * [WIP] Updated video classification ref example * Replaced mem format conversion functions by classes
…#2935) * [WIP] Update ref example video classification * [WIP] Updated video classification ref example * Replaced mem format conversion functions by classes
Description: