FBGEMM icon indicating copy to clipboard operation
FBGEMM copied to clipboard
verified publisher icon pytorch

support jagged tensor with no inner dense dim

Open jspark1105 opened this issue 3 years ago • 2 comments

Summary: jagged_to_padded_dense supports jagged tensor with inner dense dim is 1 and folded like the following example, but producing the error x_offsets.size(), 1 != num_jagged_dim, 0 in backward.

x_values = [5, 3, 7, 1] x_offsets = [0, 1, 4] max_length = 3 -> x_padded = [[5, 0, 0], [3, 7, 1]]

Differential Revision: D39121072

jspark1105 avatar Aug 30 '22 02:08 jspark1105

Deploy Preview for pytorch-fbgemm-docs canceled.

Name Link
Latest commit a535ff5ce6d4f485de59af1b40177b91c8a53bb2
Latest deploy log https://app.netlify.com/sites/pytorch-fbgemm-docs/deploys/63b1d6be44c5d10008d5c495

netlify[bot] avatar Aug 30 '22 02:08 netlify[bot]

This pull request was exported from Phabricator. Differential Revision: D39121072

facebook-github-bot avatar Aug 30 '22 02:08 facebook-github-bot

This pull request was exported from Phabricator. Differential Revision: D39121072

facebook-github-bot avatar Jan 01 '23 18:01 facebook-github-bot

This pull request has been merged in pytorch/FBGEMM@32b6a1249026cbb86121897d25d3ec1b08def1fa.

facebook-github-bot avatar Jan 01 '23 22:01 facebook-github-bot