Skip to content

[ML] Add aten::split and aten::stack for question-answering models #3012

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
edsavage merged 1 commit into from
Mar 26, 2026
+10 −7
Merged

[ML] Add aten::split and aten::stack for question-answering models #3012

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions bin/pytorch_inference/CSupportedOperations.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,8 +124,10 @@ const CSupportedOperations::TStringViewSet CSupportedOperations::ALLOWED_OPERATI
"aten::size"sv,
"aten::slice"sv,
"aten::softmax"sv,
"aten::split"sv,
"aten::sqrt"sv,
"aten::squeeze"sv,
"aten::stack"sv,
"aten::str"sv,
"aten::sub"sv,
"aten::sum"sv,
Expand Down
11 changes: 6 additions & 5 deletions bin/pytorch_inference/unittest/testfiles/reference_model_ops.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1011,9 +1011,9 @@
"model_id": "deepset/tinyroberta-squad2",
"quantized": false,
"ops": [
"aten::Int",
"aten::add",
"aten::add_",
"aten::contiguous",
"aten::cumsum",
"aten::detach",
"aten::dropout",
Expand All @@ -1027,11 +1027,11 @@
"aten::ne",
"aten::reshape",
"aten::scaled_dot_product_attention",
"aten::select",
"aten::size",
"aten::slice",
"aten::split",
"aten::squeeze",
"aten::sub",
"aten::tanh",
"aten::to",
"aten::transpose",
"aten::type_as",
Expand All @@ -1040,9 +1040,10 @@
"prim::Constant",
"prim::GetAttr",
"prim::ListConstruct",
"prim::NumToTensor",
"prim::ListUnpack",
"prim::TupleConstruct"
]
],
"auto_class": "AutoModelForQuestionAnswering"
},
"qa-bart-large-mnli": {
"model_id": "facebook/bart-large-mnli",
Expand Down
2 changes: 1 addition & 1 deletion dev-tools/extract_model_ops/reference_models.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,7 @@
"elastic-test-elser-v2-quantized": {"model_id": "elastic/test-elser-v2", "quantized": true},

"_comment:qa-models": "Models from the Appex QA pytorch_tests suite. BART models require auto_class and config_overrides to trace correctly.",
"qa-tinyroberta-squad2": "deepset/tinyroberta-squad2",
"qa-tinyroberta-squad2": {"model_id": "deepset/tinyroberta-squad2", "auto_class": "AutoModelForQuestionAnswering"},
"qa-squeezebert-mnli": "typeform/squeezebert-mnli",
"qa-bart-large-mnli": {"model_id": "facebook/bart-large-mnli", "auto_class": "AutoModelForSequenceClassification", "config_overrides": {"use_cache": false}},
"qa-distilbart-mnli": {"model_id": "valhalla/distilbart-mnli-12-6", "auto_class": "AutoModelForSequenceClassification", "config_overrides": {"use_cache": false}}
Expand Down
2 changes: 1 addition & 1 deletion dev-tools/extract_model_ops/validation_models.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@
"es-dpr-question-encoder": "facebook/dpr-question_encoder-single-nq-base",

"_comment:qa-models": "Models from the Appex QA pytorch_tests suite. BART models require auto_class and config_overrides to trace correctly.",
"qa-tinyroberta-squad2": "deepset/tinyroberta-squad2",
"qa-tinyroberta-squad2": {"model_id": "deepset/tinyroberta-squad2", "auto_class": "AutoModelForQuestionAnswering"},
"qa-squeezebert-mnli": "typeform/squeezebert-mnli",
"qa-bart-large-mnli": {"model_id": "facebook/bart-large-mnli", "auto_class": "AutoModelForSequenceClassification", "config_overrides": {"use_cache": false}},
"qa-distilbart-mnli": {"model_id": "valhalla/distilbart-mnli-12-6", "auto_class": "AutoModelForSequenceClassification", "config_overrides": {"use_cache": false}}
Expand Down
Loading