[LPT] MoveFakeQuantize #6723
Merged
[LPT] MoveFakeQuantize #6723
Conversation
This file contains 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
openvino-pushbot
added
ExternalPR
|
Looks like you have a lot of redundant changes in this PR. |
ndemashov
force-pushed
the
nd/lpt/move_fake_quantize
branch
from
8654a22 to
dbb72f8
Compare
ndemashov
force-pushed
the
nd/lpt/move_fake_quantize
branch
from
35128ff to
bdfed62
Compare
eshoguli
reviewed
Aug 5, 2021
inference-engine/tests/ngraph_helpers/lpt_ngraph_functions/src/move_fake_quantize_function.cpp
Outdated
Show resolved
Hide resolved
eshoguli
reviewed
Aug 5, 2021
...ctional/inference_engine/lp_transformations/move_fake_quantize_for_concat_transformation.cpp
Outdated
Show resolved
Hide resolved
eshoguli
reviewed
Aug 5, 2021
inference-engine/src/low_precision_transformations/src/move_fake_quantize.cpp
Outdated
Show resolved
Hide resolved
ndemashov
force-pushed
the
nd/lpt/move_fake_quantize
branch
from
a81accf to
50d20ef
Compare
eshoguli
approved these changes
Sep 12, 2021
|
|
||
| #include "shared_test_classes/base/low_precision_transformations/layer_transformation.hpp" | ||
| #include "lpt_ngraph_functions/common/fake_quantize_on_data.hpp" | ||
| #include "lpt_ngraph_functions/common/fake_quantize_on_weights.hpp" |
inference-engine/tests/ngraph_helpers/lpt_ngraph_functions/src/move_fake_quantize_function.cpp
Outdated
Show resolved
Hide resolved
Comment on lines
+46
to
+75
| if (!fqOnData1.empty()) { | ||
| if (operation == "relu") { | ||
| auto relu1 = std::make_shared<ngraph::opset1::Relu>(input1->output(0)); | ||
| parent1 = makeFakeQuantize(relu1, inputPrecision, fqOnData1); | ||
| } else { | ||
| parent1 = makeFakeQuantize(input1, inputPrecision, fqOnData1); | ||
| } | ||
| parent1->set_friendly_name("concat_fq1"); | ||
| if (!convert1.empty()) { | ||
| parent1 = std::make_shared<opset1::Convert>(parent1, convert1.outPrecision); | ||
| } | ||
| if (!dequantization1.empty()) { | ||
| parent1 = makeDequantization(parent1, dequantization1); | ||
| } | ||
| } | ||
| if (!fqOnData2.empty()) { | ||
| if (operation == "relu") { | ||
| auto relu2 = std::make_shared<ngraph::opset1::Relu>(input2->output(0)); | ||
| parent2 = makeFakeQuantize(relu2, inputPrecision, fqOnData2); | ||
| } else { | ||
| parent2 = makeFakeQuantize(input1, inputPrecision, fqOnData2); | ||
| } | ||
| parent2->set_friendly_name("concat_fq2"); | ||
| if (!convert2.empty()) { | ||
| parent1 = std::make_shared<opset1::Convert>(parent2, convert2.outPrecision); | ||
| } | ||
| if (!dequantization1.empty()) { | ||
| parent2 = makeDequantization(parent2, dequantization2); | ||
| } | ||
| } |
Comment on lines
+85
to
+95
| if (!fqOnData3.empty()) { | ||
| std::shared_ptr<Node> fq; | ||
| if (operation == "relu") { | ||
| auto relu = std::make_shared<ngraph::opset1::Relu>(concat->output(0)); | ||
| fq = makeFakeQuantize(relu, inputPrecision, fqOnData3); | ||
| } else { | ||
| fq = makeFakeQuantize(concat, inputPrecision, fqOnData3); | ||
| } | ||
| fq->set_friendly_name("fakeQuantizeAfter"); | ||
| parent = fq; | ||
| } |
There was a problem hiding this comment.
minor:
if (operation == "relu") {
...
}
if (!fqOnData1.empty()) {
...
}
inference-engine/src/low_precision_transformations/src/move_fake_quantize.cpp
Outdated
Show resolved
Hide resolved
inference-engine/src/low_precision_transformations/src/move_fake_quantize.cpp
Outdated
Show resolved
Hide resolved
akuporos
pushed a commit
to akuporos/openvino
that referenced
this pull request
Sep 29, 2021
* add move_fake_quantize_for_concat_transformation, mfk and mfk_function * fix relu_transformation.cpp * backup * add change * add cpu test * [LPT] MoveFakeQuantizeTransformation: fixes * get InferenceEngine::NotImplemented * fix ieFuncTests * try without new cpu_test * fix cpuFuncTests and ieFuncTests * fix tests * fix lin * add cpu test * fix link and matcher in move_fake_quantize.cpp * update matcher * add gpu test * naming fix * move_fake_quantize.cpp add set_fr_name for new_concat * naming new fq fix * fix NetworkHelper::copyInfo naming * concat.cpp naming fix * gpu tests fix * rm network_helper changes * rm extra output * resolve conversations * resolve other conversations * add multi inputs for concat * fix lin * fix move_fake_qunatize naming * rm maxpool from mfk_function * mkldnn update * fix style * rm extra change * fix concat matcher * rm mkldnn_plugin changes * fix conversations * fix interval * fix and add isQuantizedStatic, add attribute and negative tests * add negative plugin tests * fix style: Co-authored-by: Edward Shogulin <[email protected]>
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.
Details:
Tickets: