Skip to content

Conversation

@Tongkaio
Copy link
Contributor

@Tongkaio Tongkaio commented Mar 27, 2024

修改:

  • scale_kernel.cc:ScaleKernel 的参数列表中,bias 的类型已由 float 改为 Scalar,参考:#62598
  • test_expand_v2_op_npu.py:expand 已支持 uint8 等类型,同步修改单测,参考:#62849
  • test_transpose_op_npu.py:transpose 已经支持 stride,默认调用 Paddle 主框架中的 transpose,只有 npu 的 transpose 支持非标准格式,故设置 FLAGS_use_stride_kernel 为 False 来调用 npu 上的 transpose。

相关单测已测试通过;

  • test_elementwise_add_op_npu
  • test_matmulv2_op_npu
  • test_npu_identity_op
  • test_multinomial_op_npu
  • test_momentum_op_npu
  • test_rms_norm_npu
  • test_rope_npu
  • test_transpose_op_npu
  • test_unbind_op_npu
  • test_expand_v2_op_npu
  • test_flashattention_npu
  • test_adadelta_op_npu
  • test_LeNet_MNIST

@paddle-bot
Copy link

paddle-bot bot commented Mar 27, 2024

Thanks for your contribution!

Copy link
Contributor

@qili93 qili93 left a comment

Choose a reason for hiding this comment

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

LGTM

@qili93 qili93 merged commit 0067f85 into PaddlePaddle:develop Mar 28, 2024
@Tongkaio Tongkaio deleted the fix_npu_ut branch March 29, 2024 08:51
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants