[BYOC-DNNL] add post_sum pattern #12151
Merged
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
crazydemo
force-pushed
the
upstream-sum_pattern
branch
from
e8d8a67
to
b9ed19d
Compare
crazydemo
force-pushed
the
upstream-sum_pattern
branch
3 times, most recently
from
fe68b77
to
63a44ba
Compare
tests/python/contrib/test_dnnl.py
Outdated
| @@ -192,6 +192,7 @@ def run_and_verify(mod, input, params, target, run_module, subgraph_num=None, te | |||
| if use_dnnl: | |||
| processed_mod = partition_for_dnnl(processed_mod, params, alter_layout) | |||
| check_dnnl_used(processed_mod) | |||
| print(processed_mod) | |||
crazydemo
force-pushed
the
upstream-sum_pattern
branch
from
63a44ba
to
d89db44
Compare
yangulei
reviewed
Jul 28, 2022
python/tvm/relay/op/contrib/dnnl.py
Outdated
Comment on lines
398
to
403
| dnnl_patterns.append( | ||
| ("dnnl.conv2d_bias_sum_relu", make_conv_bias_sum_relu_pattern("nn.conv2d"), make_predicate(add_checker)) | ||
| ), | ||
| dnnl_patterns.append( | ||
| ("dnnl.conv2d_bias_sum", make_conv_bias_sum_relu_pattern("nn.conv2d", False), make_predicate(add_checker)) | ||
| ), |
There was a problem hiding this comment.
Suggested change
| dnnl_patterns.append( | |
| ("dnnl.conv2d_bias_sum_relu", make_conv_bias_sum_relu_pattern("nn.conv2d"), make_predicate(add_checker)) | |
| ), | |
| dnnl_patterns.append( | |
| ("dnnl.conv2d_bias_sum", make_conv_bias_sum_relu_pattern("nn.conv2d", False), make_predicate(add_checker)) | |
| ), | |
| dnnl_patterns.append( | |
| ( | |
| "dnnl.conv2d_bias_sum_relu", | |
| make_conv_bias_sum_relu_pattern("nn.conv2d"), | |
| make_predicate(add_checker), | |
| ) | |
| ) | |
| dnnl_patterns.append( | |
| ( | |
| "dnnl.conv2d_bias_sum", | |
| make_conv_bias_sum_relu_pattern("nn.conv2d", False), | |
| make_predicate(add_checker), | |
| ) | |
| ) |
yangulei
reviewed
Jul 28, 2022
| @@ -361,6 +367,10 @@ class DNNLJSONRuntime : public JSONRuntimeBase { | |||
|
|
|||
| // TODO(@apeskov): Simulation of inplace primitive. just as PoC. | |||
| auto sum_in_tr = GetInputByName(nid, "sum_idx").TreatAs(dst_layout); | |||
| if (op_name.find("_sum") != std::string::npos) { | |||
| sum_in_tr = GetInput(nid, node.GetInputs().size()-1); | |||
There was a problem hiding this comment.
Suggested change
| sum_in_tr = GetInput(nid, node.GetInputs().size()-1); | |
| sum_in_tr = GetInput(nid, node.GetInputs().size() - 1); |
yangulei
reviewed
Jul 28, 2022
tests/python/contrib/test_dnnl.py
Outdated
| param_lst += ["data1"] | ||
| return relay.nn.relu(out), dic, param_lst | ||
|
|
||
| conv2d_bn_sum_relu, dic, param_lst = get_conv2d_bn_sum_relu(x_shape, k_shape, sum_shape=(1, 16, 6, 6), dtype=dtype) |
There was a problem hiding this comment.
Suggested change
| conv2d_bn_sum_relu, dic, param_lst = get_conv2d_bn_sum_relu(x_shape, k_shape, sum_shape=(1, 16, 6, 6), dtype=dtype) | |
| conv2d_bn_sum_relu, dic, param_lst = get_conv2d_bn_sum_relu( | |
| x_shape, k_shape, sum_shape=(1, 16, 6, 6), dtype=dtype | |
| ) |
yangulei
reviewed
Jul 28, 2022
tests/python/contrib/test_dnnl.py
Outdated
| config = conv2d_bn_sum_relu, dic, param_lst | ||
| run_and_verify_func(config, run_module=run_module, dtype=dtype) | ||
|
|
||
| conv2d_bn_sum_relu, dic, param_lst = get_conv2d_bn_sum_relu(x_shape, k_shape, sum_shape=(1, 16, 1, 1), dtype=dtype) |
There was a problem hiding this comment.
Suggested change
| conv2d_bn_sum_relu, dic, param_lst = get_conv2d_bn_sum_relu(x_shape, k_shape, sum_shape=(1, 16, 1, 1), dtype=dtype) | |
| conv2d_bn_sum_relu, dic, param_lst = get_conv2d_bn_sum_relu( | |
| x_shape, k_shape, sum_shape=(1, 16, 1, 1), dtype=dtype | |
| ) |
yangulei
reviewed
Jul 28, 2022
tests/python/contrib/test_dnnl.py
Outdated
Comment on lines
1774
to
1775
| # tvm.testing.main() | ||
| test_conv2d_bias_sum_relu(True) |
There was a problem hiding this comment.
Suggested change
| # tvm.testing.main() | |
| test_conv2d_bias_sum_relu(True) | |
| tvm.testing.main() | |
crazydemo
force-pushed
the
upstream-sum_pattern
branch
from
654bab6
to
b2dca37
Compare
crazydemo
force-pushed
the
upstream-sum_pattern
branch
from
b2dca37
to
001898f
Compare
|
@masahi Could you please review this PR? This PR adds |
masahi
approved these changes
Aug 1, 2022
xinetzone
pushed a commit
to daobook/tvm
that referenced
this pull request
Nov 25, 2022
* add post_sum pattern * add checkers for sum pattern * fix lint * fix error in test_pass_partition_graph * fix lint error
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.
This PR add
conv2d-add-sum-relupattern, and the corresponding test case is added.