Small fixes related to BF16 fusion_gru and fusion_lstm #33295
Conversation
|
Thanks for your contribution! |
| @@ -48,7 +48,8 @@ static int BuildFusion(Graph* graph, const std::string& name_scope, | |||
|
|
|||
| // Create New OpDesc | |||
| auto gru_creater = [&](Node* gru, Node* x, Node* weight_x, Node* weight_h, | |||
| Node* bias, Node* hidden, Node* fc_bias) { | |||
| Node* bias, Node* hidden, Node* fc_bias, | |||
| const bool& use_mkldnn) { | |||
There was a problem hiding this comment.
Maybe passing use_mkldnn as value would be better? Same with lstm_creater
There was a problem hiding this comment.
You are right, for such small types, it is better to pass by value.
|
@wozna Hi is this PR ready to review? If yes please ask members to review? because last time we informed Baidu that at least 2 LGTM the PR will be merged. Because I m not sure if this is ready so if it's ready you ask other members to review ok? Thanks |
|
LGTM. Thank you very much the fix! And thank you for calling and explaining it, the reviewing process get much faster. @juncaipeng Could you please review and merge this PR. If you like, we can call and speed up the reviewing process. |
| @@ -48,7 +48,8 @@ static int BuildFusion(Graph* graph, const std::string& name_scope, | |||
|
|
|||
| // Create New OpDesc | |||
| auto gru_creater = [&](Node* gru, Node* x, Node* weight_x, Node* weight_h, | |||
There was a problem hiding this comment.
I found the name "gru_creater" a bit confusing. Maybe you could use "gru_creator" ?
There was a problem hiding this comment.
Oh yes, it could be a spelling mistake.
PR types
Bug fixes
PR changes
OPs
Describe
In this PR there are small fixes related to fusion_gru and fusion_lstm:
use_mkldnnattribute infc_gru_fuse_passandfc_lstm_fuse_passand tests for that. Thanks to this, I could removemkldnn_placement_passin test paddle/fluid/inference/tests/api/analyzer_lexical_analysis_gru_tester.cc that runs twice.fusion_lstmto the list of bf16 operatorsmkldnn_data_typeattribute tofusion_lstm, which is necessary to run the model on bfloat16