upgrade for bscvrq to support non-quant #113
Open
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
xymyeah
reviewed
Aug 21, 2024
| // add for bscvrq | ||
| // embedx_concate_filter:true && quant_ratio_valid=false && need_filter=true && embed_threshold_filter=false | ||
| template <typename T> | ||
| struct sum_pooling_concate<T, false, false, true> { |
There was a problem hiding this comment.
need_filter=true or false?
need_filter和embed_threshold_filter是否反了？
| // embedx_concate_filter:true && quant_ratio_valid=false && need_filter=false | ||
| template <typename T, bool use_cvm, bool clk_filter, bool need_filter, | ||
| bool embed_threshold_filter, typename T2> | ||
| struct do_sum_pooling_and_cvm<T, use_cvm, clk_filter, need_filter, false, embed_threshold_filter, true, T2> { |
There was a problem hiding this comment.
缺少了do_sum_pooling_and_cvm_with_large_dim的实现
xymyeah
reviewed
Aug 21, 2024
| // add for bscvrq | ||
| // embedx_concate_filter:true && quant_ratio_valid=false && need_filter=false && embed_threshold_filter=false | ||
| template <typename T> | ||
| struct sum_pooling_concate<T, false, true, false, false> { |
There was a problem hiding this comment.
应该按第一版的实现就行，第一版里的注释和模版参数的值没有对应上
xymyeah
reviewed
Aug 21, 2024
| }; | ||
|
|
||
| // add for bscvrq | ||
| // embedx_concate_filter:true && quant_ratio_valid=false && need_filter=true && embed_threshold_filter=false |
There was a problem hiding this comment.
这个跟上面的可以合并成一个，模版参数写成<false, need_filter, false>
xymyeah
reviewed
Aug 21, 2024
| // add for bscvrq | ||
| // embedx_concate_filter:true && quant_ratio_valid=false && need_filter=true && embed_threshold_filter=true | ||
| template <typename T> | ||
| struct sum_pooling_concate<T, false, true, true> { |
There was a problem hiding this comment.
同理，这个跟现有的也可以合并，<quant_ratio_valid, true, true>
jack603047588
pushed a commit
that referenced
this pull request
Oct 11, 2024
* fused_seqpool_cvm_with_conv support filter by threshold * add fill zero in fused_seqpool_cvm * add fused seq tensor && support transpose batch fc weight * fix fused query seq tensor compare case --------- Co-authored-by: mojingcj <ChengJing_dhu@163.com> Co-authored-by: jiaoxuewu <jiaoxuewu@163.com> Co-authored-by: yuandong1998 <1377526365@qq.com> Co-authored-by: shangzhongbin <shangzhongbin@baidu.com>
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
PR changes
paddle/fluid/operators/fused/fused_seqpool_cvm_kernel.kps
Describe
支持非量化的分支，用于给bscvrq使用