[dygraph qat] Refine saving output scale to infer program #31784
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
|
Thanks for your contribution! |
gfwm2013
reviewed
Mar 23, 2021
|
|
||
| # Note that, the items have priority in corresponding_dict | ||
| corresponding_dict = { | ||
| 'conv2d_tranpose': [['conv2d_tranpose'], None], |
There was a problem hiding this comment.
conv2d_transpose 对应是两个Op：conv2d_transpose / depthwise_ conv2d_transpose
另外：应该是conv2d_transpose，而不是conv2d_tranpose
gfwm2013
reviewed
Mar 23, 2021
| for key, value in corresponding_dict.items(): | ||
| if key in scale_name: | ||
| return (op.type in value[0]) and \ | ||
| (len(value) == 1 or value[1] is None or value[1](op)) |
There was a problem hiding this comment.
为什么要添加判断value[1] is None or value[1](op)呢？
There was a problem hiding this comment.
如果有必要，可以新增lambda函数来判断静态图op是否满足特定条件，目前这里仅仅使用了op的名字，所以默认都是None
gfwm2013
reviewed
Mar 23, 2021
| 'BatchNorm': paddle.nn.BatchNorm, | ||
| 'BatchNorm2D': paddle.nn.BatchNorm2D, | ||
| 'GroupNorm': paddle.nn.GroupNorm, | ||
| 'InstanceNorm2D': paddle.nn.InstanceNorm2D, |
There was a problem hiding this comment.
静态图量化中并不支持Instance_norm等Op，那动态图中新增了对于这些Op的支持，静态图量化中也请一并加上
gfwm2013
reviewed
Mar 23, 2021
| if not self._is_scale_op_matched(scale_name, op, global_block): | ||
| op_idx += 1 | ||
| else: | ||
| weight_ops = ["conv2d", "depthwise_conv2d", "matmul"] |
There was a problem hiding this comment.
weight_ops中还需添加conv2d_transpose 和depthwise_conv2d_transpose
… refine_calc_output_scale2
juncaipeng
commented
Mar 23, 2021
| if not self._is_scale_op_matched(scale_name, op, global_block): | ||
| op_idx += 1 | ||
| else: | ||
| weight_ops = ["conv2d", "depthwise_conv2d", "matmul"] |
|
|
||
| # Note that, the items have priority in corresponding_dict | ||
| corresponding_dict = { | ||
| 'conv2d_tranpose': [['conv2d_tranpose'], None], |
| for key, value in corresponding_dict.items(): | ||
| if key in scale_name: | ||
| return (op.type in value[0]) and \ | ||
| (len(value) == 1 or value[1] is None or value[1](op)) |
There was a problem hiding this comment.
如果有必要，可以新增lambda函数来判断静态图op是否满足特定条件，目前这里仅仅使用了op的名字，所以默认都是None
| 'BatchNorm': paddle.nn.BatchNorm, | ||
| 'BatchNorm2D': paddle.nn.BatchNorm2D, | ||
| 'GroupNorm': paddle.nn.GroupNorm, | ||
| 'InstanceNorm2D': paddle.nn.InstanceNorm2D, |
juncaipeng
force-pushed
the
refine_calc_output_scale2
branch
from
2d9bced
to
85f18ef
Compare
gfwm2013
approved these changes
Mar 24, 2021
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.
PR types
Others
PR changes
Others
Describe
Refine saving output scale to infer program