Conversation
facebook-github-bot
added
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D49262422 |
…nguish (facebookincubator#929) Summary: For vdd, it seems that the jagged tensor batch dim is identical to dense tensor batch dim, which caused issue in bmm kernel, that it cannot handle batch size as large as 2^16. This fix adds a flag `deduce_jagged_tensor_with_graph_analysis` so that when it is turnt on, we depend on graph analysis, i.e. `try_getting_jagged_tensor_map`, to deduce batch dim for jagged tensor. This can be more reliable than deducing based on value. Differential Revision: D49262422
ac3913e to
6d87d33
Compare
|
This pull request was exported from Phabricator. Differential Revision: D49262422 |
|
Hi @tissue3! Thank you for your pull request. We require contributors to sign our Contributor License Agreement, and yours needs attention. You currently have a record in our system, but the CLA is no longer valid, and will need to be resubmitted. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks! |
This branch has not been deployed
2 participants
Summary:
For vdd, it seems that the jagged tensor batch dim is identical to dense tensor batch dim, which caused issue in bmm kernel, that it cannot handle batch size as large as 2^16.
This fix adds a flag
deduce_jagged_tensor_with_graph_analysisso that when it is turnt on, we depend on graph analysis, i.e.try_getting_jagged_tensor_map, to deduce batch dim for jagged tensor. This can be more reliable than deducing based on value.Differential Revision: D49262422