Ascend
/GPU
/CPU
): Ascend+pynativeUncomment only one
/device <>
line, hit enter to put that in a new line, and remove leading whitespaces from that line:/device ascend
def test_asin_input_1x12x8x4x16x10x24_fp64():
def test_asin_input_20x7x13x9x88_fp64():
def test_asin_input_20x7x88_fp64():
fact = AsinFactory(input_shape=(20, 7, 88), dtype=np.float64)
fact.forward_cmp()
> fact.grad_cmp()
ascend+pynative , dtype是fp64时, 报错RuntimeError,具体信息如下:
ascend图模式下是pass的, cpu上也是pass的
def test_asin_input_20x7x88_fp64():
fact = AsinFactory(input_shape=(20, 7, 88), dtype=np.float64)
fact.forward_cmp()
> fact.grad_cmp()
test_asin.py:123:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
../share/ops/primitive/asin_ops.py:72: in grad_cmp
input_grad_mindspore = self.grad_mindspore_impl()
../share/ops/primitive/asin_ops.py:50: in grad_mindspore_impl
input_grad = grad_net(input_ms, out_grad)
/root/archiconda3/envs/caory_3.7/lib/python3.7/site-packages/mindspore/nn/cell.py:483: in __call__
raise err
/root/archiconda3/envs/caory_3.7/lib/python3.7/site-packages/mindspore/nn/cell.py:480: in __call__
output = self.run_construct(cast_inputs, kwargs)
/root/archiconda3/envs/caory_3.7/lib/python3.7/site-packages/mindspore/nn/cell.py:363: in run_construct
output = self.construct(*cast_inputs, **kwargs)
../share/grad.py:27: in construct
return self.grad(self.network)(*inputs)
/root/archiconda3/envs/caory_3.7/lib/python3.7/site-packages/mindspore/ops/composite/base.py:395: in after_grad
return grad_(fn)(*args, **kwargs)
/root/archiconda3/envs/caory_3.7/lib/python3.7/site-packages/mindspore/common/api.py:78: in wrapper
results = fn(*arg, **kwargs)
/root/archiconda3/envs/caory_3.7/lib/python3.7/site-packages/mindspore/ops/composite/base.py:384: in after_grad
out = _pynative_executor(fn, *args, **kwargs)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <mindspore.common.api._PynativeExecutor object at 0xffff9ae19450>, obj = Asin<>
args = (Tensor(shape=[20, 7, 88], dtype=Float64, value=
[[[ 2.64771472e-01, 1.84741329e-01, 3.00367079e-01 ... 1.70422975e...5e-01],
[-3.12819027e-01, 4.64458469e-01, -6.54240419e-01 ... 1.23265315e+00, 7.13418732e-01, 7.93989728e-01]]]))
kwargs = {}
def __call__(self, obj, *args, **kwargs):
args = args + tuple(kwargs.values())
> return self._executor(obj, args)
E RuntimeError: mindspore/ccsrc/backend/session/ascend_session.cc:1734 SyncStream] Sync stream error!
[ERROR] DEVICE(69583,fffe30ff91e0,python3.7):2021-11-17-14:49:02.062.226 [mindspore/ccsrc/runtime/device/ascend/ascend_kernel_runtime.cc:612] TaskFailCallback] Execute TaskFailCallback failed. task_fail_info or current_graph_ is nullptr
[ERROR] DEVICE(69583,fffe30ff91e0,python3.7):2021-11-17-14:49:02.063.968 [mindspore/ccsrc/runtime/device/ascend/ascend_kernel_runtime.cc:612] TaskFailCallback] Execute TaskFailCallback failed. task_fail_info or current_graph_ is nullptr
[ERROR] DEVICE(69583,fffe30ff91e0,python3.7):2021-11-17-14:49:02.064.057 [mindspore/ccsrc/runtime/device/ascend/ascend_kernel_runtime.cc:958] SyncStream] Call runtime rtStreamSynchronize error.
[CRITICAL] SESSION(69583,fffe30ff91e0,python3.7):2021-11-17-14:49:02.064.072 [mindspore/ccsrc/backend/session/ascend_session.cc:1734] SyncStream] Sync stream error!
[ERROR] SESSION(69583,fffe30ff91e0,python3.7):2021-11-17-14:49:02.064.265 [mindspore/ccsrc/backend/session/ascend_session.cc:1763] ReportErrorMessage] Ascend error occurred, error message:
E39999: Inner Error!
Aicpu kernel execute failed, device_id=3, stream_id=3, task_id=5, fault so_name=libaicpu_kernels.so, fault kernel_name=RunCpuKernel, fault op_name=, extend_info=[FUNC:GetError][FILE:stream.cc][LINE:712]
Stream synchronize failed, stream = 0xfffe240a6640[FUNC:StreamSynchronize][FILE:logger.cc][LINE:285]
rtStreamSynchronize execute failed, reason=[aicpu exception][FUNC:ReportFuncErrorReason][FILE:error_message_manage.cc][LINE:41]
执行pass无异常
Please add labels (comp or sig), also you can visit https://gitee.com/mindspore/community/blob/master/sigs/dx/docs/labels.md to find more.
为了让代码尽快被审核,请您为Pull Request打上 组件(comp)或兴趣组(sig) 标签,打上标签的PR可以直接推送给责任人进行审核。
更多的标签可以查看https://gitee.com/mindspore/community/blob/master/sigs/dx/docs/labels.md
以组件相关代码提交为例,如果你提交的是data组件代码,你可以这样评论:
//comp/data
当然你也可以邀请data SIG组来审核代码,可以这样写:
//sig/data
另外你还可以给这个PR标记类型,例如是bugfix或者是特性需求:
//kind/bug or //kind/feature
恭喜你,你已经学会了使用命令来打标签,接下来就在下面的评论里打上标签吧!
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。
问题回归不通过
问题原因:
AsinFactory方式下,so文件名识别错误
解决方案:
写死so文件名
回归通过
登录 后才可以发表评论