[ONNX] Add eliminate_nop_cast pass #3376
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
andrey-churkin
commented
Mar 25, 2025
| @@ -348,6 +349,7 @@ def create_nncf_graph(onnx_model: onnx.ModelProto) -> NNCFGraph: | |||
| """ | |||
| onnx_model = GraphConverter._replace_empty_node_name(onnx_model) | |||
| onnx_model = onnx.shape_inference.infer_shapes(onnx_model) | |||
| onnx_model = onnxoptimizer.optimize(onnx_model, ["eliminate_nop_cast"]) | |||
There was a problem hiding this comment.
There are other passes that can be useful as well.
"eliminate_nop_dropout",
"eliminate_nop_flatten",
"eliminate_if_with_const_cond",
"eliminate_nop_monotone_argmax",
"eliminate_nop_pad",
"eliminate_nop_concat",
"eliminate_nop_split",
"eliminate_nop_expand",
"eliminate_shape_gather",
"eliminate_slice_after_shape",
"eliminate_nop_transpose",
"eliminate_nop_reshape",
"eliminate_nop_with_unit",
"eliminate_common_subexpression",
"eliminate_deadend",
"eliminate_identity",
"eliminate_shape_op",
"eliminate_unused_initializer",
"eliminate_duplicate_initializer",
andrey-churkin
commented
Mar 25, 2025
Comment on lines
351
to
352
| onnx_model = onnx.shape_inference.infer_shapes(onnx_model) | ||
| onnx_model = onnxoptimizer.optimize(onnx_model, ["eliminate_nop_cast"]) |
There was a problem hiding this comment.
The pass should be applied after onnx.shape_inference.infer_shapes() call. Otherwise, not all no-op Cast nodes will be found.
andrey-churkin
commented
Mar 25, 2025
| @@ -348,6 +349,7 @@ def create_nncf_graph(onnx_model: onnx.ModelProto) -> NNCFGraph: | |||
| """ | |||
| onnx_model = GraphConverter._replace_empty_node_name(onnx_model) | |||
| onnx_model = onnx.shape_inference.infer_shapes(onnx_model) | |||
| onnx_model = onnxoptimizer.optimize(onnx_model, ["eliminate_nop_cast"]) | |||
There was a problem hiding this comment.
The following code snippet can be used to apply that pass to the initial ONNX model.
import onnx
import onnxoptimizer
model = onnx.load(MODEL_PATH)
model = onnx.shape_inference.infer_shapes(model)
model = onnxoptimizer.optimize(model, ["eliminate_nop_cast"])
|
@kshpv Please review |
andrey-churkin
force-pushed
the
ac/eliminate_nop_cast
branch
from
4650dc4 to
fa04951
Compare
alexsu52
reviewed
Apr 21, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Changes
Apply the
eliminate_nop_castoptimization pass fromonnxoptimizerbefore building the NNCF graph. This pass aims to eliminate no-op cast nodes in the graph. A no-op cast is a cast that doesn't change the value, such as casting a tensor to its own type.Reason for changes
Statistics cannot be collected after the no-op Cast nodes because such nodes are removed from the ONNX inference graph during the session.
Related tickets
Ref: 164211