Skip to content

Commit

Permalink
- fix
Browse files Browse the repository at this point in the history
  • Loading branch information
@jczaja
jczaja committed Oct 28, 2021
1 parent 6f8b5ea commit 951afab
Showing 1 changed file with 2 additions and 1 deletion.
3 changes: 2 additions & 1 deletion paddle/fluid/operators/mkldnn/conv_mkldnn_op.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -647,7 +647,8 @@ class ConvMKLDNNHandlerT
MKLDNNMemoryFormat::x);

// Get Bias scales for int8
auto p_tupple = platform::is_int8<T>() ? (get_int8_bias_scales(ctx, dev_ctx, key)) : std::make_shared<std::tuple<float, std::vector<float>>>(std::make_tuple(0.0f, {1.0f}));
// std::make_shared<std::tuple<float, std::vector<float>>>(std::make_tuple(static_cast<float>(mask_reorder), std::vector<float>(count)));
auto p_tupple = platform::is_int8<T>() ? (get_int8_bias_scales(ctx, dev_ctx, key)) : std::make_shared<std::tuple<float, std::vector<float>>>(std::make_tuple(0.0f, std::vector<float>(1,1.0f)));

return this->AcquireMemoryWithReorder(
dev_ctx, user_bias_md, this->fwd_pd_->bias_desc(),
Expand Down

1 comment on commit 951afab

@paddle-bot-old
Copy link

@paddle-bot-old paddle-bot-old bot commented on 951afab Oct 28, 2021 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🕵️ CI failures summary

🔍 PR: #36595 Commit ID: 951afab contains failed CI.

🔹 Failed: PR-CI-APPROVAL

approve_failed
2021-10-28 23:21:32 长度：5 [application/octet-stream]

2021-10-28 23:21:32 正在保存至: “bk.txt”

2021-10-28 23:21:32      0K                                                       100% 2.42M=0s

2021-10-28 23:21:32 2021-10-28 23:21:32 (2.42 MB/s) - 已保存 “bk.txt” [5/5])

2021-10-28 23:21:40 ****************

2021-10-28 23:21:40 0. The error message you wrote in PADDLE_ENFORCE{**} or PADDLE_THROW does not meet our error message writing specification. Possible errors include 1. the error message is empty / 2. the error message is too short / 3. the error type is not specified. Please read the specification [ /~https://github.com/PaddlePaddle/Paddle/wiki/Paddle-Error-Message-Writing-Specification ], then refine the error message. If it is a mismatch, please request chenwhql (Recommend), luotao1 or lanxianghit review and approve.

2021-10-28 23:21:40 The PADDLE_ENFORCE{**} or PADDLE_THROW entries that do not meet the specification are as follows:

2021-10-28 23:21:40 PADDLE_ENFORCE_NE(bias->format(), MKLDNNMemoryFormat::undef, + "Got wrong format for Bias tensor."));

2021-10-28 23:21:40 PADDLE_ENFORCE_NE(bias->format(), MKLDNNMemoryFormat::undef, + "Got wrong format for Bias tensor."));

2021-10-28 23:21:40 PADDLE_ENFORCE_EQ(ct.Analyze(3), true, + PADDLE_ENFORCE_EQ(ct.Analyze(6), true, +++ b/paddle/fluid/platform/mkldnn_reuse.h + bool is_persistent = false, + const std::shared_ptrmkldnn::memory& target_memory_p) { + auto reorder_p = + std::make_sharedmkldnn::reorder(*user_memory_p, *target_memory_p);

2021-10-28 23:21:40 There are 1 approved errors.

2021-10-28 23:21:40 ****************

2021-10-28 23:21:40 + EXCODE=6

2021-10-28 23:21:40 + echo 'EXCODE: 6'

2021-10-28 23:21:40 EXCODE: 6

2021-10-28 23:21:40 + echo 'ipipe_log_param_EXCODE: 6'

2021-10-28 23:21:40 ipipe_log_param_EXCODE: 6

2021-10-28 23:21:40 + exit 6

🔹 Failed: PR-CI-Static-Check

code_style_failed
2021-10-28 23:37:43     with_avx: ON

2021-10-28 23:37:43     with_gpu: ON

2021-10-28 23:37:43     with_mkl: ON

2021-10-28 23:37:43     with_mkldnn: ON

2021-10-28 23:37:43     with_python: ON'

2021-10-28 23:37:43 + set +x

2021-10-28 23:37:43 ========================================

2021-10-28 23:37:43 summary problems:

2021-10-28 23:37:43 There is 1 error: Code format error.

2021-10-28 23:37:43 ========================================

2021-10-28 23:37:43 Code format error Please fix it according to the diff information:

2021-10-28 23:37:43 code format error

2021-10-28 23:37:43 diff --git a/paddle/fluid/operators/mkldnn/conv_mkldnn_op.cc b/paddle/fluid/operators/mkldnn/conv_mkldnn_op.cc

2021-10-28 23:37:43 index a88af38..eb678a2 100644

2021-10-28 23:37:43 --- a/paddle/fluid/operators/mkldnn/conv_mkldnn_op.cc

2021-10-28 23:37:43 +++ b/paddle/fluid/operators/mkldnn/conv_mkldnn_op.cc

2021-10-28 23:37:43 @@ -418,9 +418,9 @@ class ConvMKLDNNHandlerT

2021-10-28 23:37:43      // Scales for int8 bias are to be cached to avoid

2021-10-28 23:37:43      // computing them each iteration

2021-10-28 23:37:43      auto bias_scale_tuple =

2021-10-28 23:37:43 -        std::static_pointer_cast<std::tuple<float, std::vector > >(dev_ctx.GetBlob(key_bs));

Please sign in to comment.