【Infer Symbolic Shape No.77】log_softmax #68025
Closed
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
|
你的PR提交成功，感谢你对开源项目的贡献! |
33 tasks
third_party/flashattn
Outdated
There was a problem hiding this comment.
麻烦提交时删除这些修改或拉一下最新的develop分支
| @@ -85,6 +85,8 @@ OP_DECLARE_INFER_SYMBOLIC_SHAPE(L1Norm_) | |||
| OP_DECLARE_INFER_SYMBOLIC_SHAPE(LpPool2d) | |||
| OP_DECLARE_INFER_SYMBOLIC_SHAPE(Logcumsumexp) | |||
| OP_DECLARE_INFER_SYMBOLIC_SHAPE(Logsumexp) | |||
| OP_DECLARE_INFER_SYMBOLIC_SHAPE(Log_softmax) | |||
There was a problem hiding this comment.
删除这个声明，只需使用下面的驼峰命名方式。
另外，带inplace版本的op是在驼峰命名的最后添加下划线
| @@ -1585,6 +1585,51 @@ bool LogsumexpOpInferSymbolicShape( | |||
| return details::ReduceInferDim(op, infer_context, axis, keepdim, reduce_all); | |||
| } | |||
|
|
|||
| bool Log_softmaxOpInferSymbolicShape( | |||
Comment on lines
+1589
to
+1625
| pir::Operation *op, pir::InferSymbolicShapeContext *infer_context) { | ||
| const auto &x_shape_or_data = | ||
| infer_context->GetShapeOrDataForValue(op->operand_source(0)); | ||
| const std::vector<symbol::DimExpr> &x_shape = x_shape_or_data.shape(); | ||
|
|
||
| int axis = op->attribute<pir::Int32Attribute>("axis").data(); | ||
| size_t rank = x_shape.size(); | ||
|
|
||
| if (rank > 0) { | ||
| PADDLE_ENFORCE_GE(axis, | ||
| -rank, | ||
| common::errors::InvalidArgument( | ||
| "Attr(axis) value should be in range [-R, R-1], " | ||
| "R is the rank of Input(X).")); | ||
| PADDLE_ENFORCE_LT(axis, | ||
| rank, | ||
| common::errors::InvalidArgument( | ||
| "Attr(axis) value should be in range [-R, R-1], " | ||
| "R is the rank of Input(X).")); | ||
| } else if (rank == 0) { | ||
| PADDLE_ENFORCE_GE(axis, | ||
| -1, | ||
| common::errors::InvalidArgument( | ||
| "Attr(axis) value should be in range [-1, " | ||
| "0] when input is 0D Tensor ")); | ||
| PADDLE_ENFORCE_LE(axis, | ||
| 0, | ||
| common::errors::InvalidArgument( | ||
| "Attr(axis) value should be in range [-1, " | ||
| "0] when input is 0D Tensor ")); | ||
| } | ||
|
|
||
| infer_context->SetShapeOrDataForValue( | ||
| op->result(0), | ||
| symbol::ShapeOrDataDimExprs{symbol::TensorShapeOrDataDimExprs(x_shape)}); | ||
|
|
||
| return true; |
There was a problem hiding this comment.
麻烦将这部分逻辑包装成一个推导函数，cumprod 和 log_softmax 都直接调用这个推导函数，参考命名：UnchangedCheckAxisInferSymbolicShape ？
# for free
to join this conversation on GitHub.
Already have an account?
# 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 Category
CINN
PR Types
Improvements
Description
算子
log_softmax符号推导