name | about | labels |
---|---|---|
Bug Report | Use this template for reporting a bug | kind/bug |
DQN 训练时算子报错(Select-op0 op dtype is not same, type1:int32, type2:int64)
Ascend
/GPU
/CPU
) / 硬件环境:Please delete the backend not involved / 请删除不涉及的后端:
/device ascend
Software Environment / 软件环境 (Mandatory / 必填):
-- MindSpore version (e.g., 1.7.0.Bxxx) :
-- Python version (e.g., Python 3.7.5) :
-- OS platform and distribution (e.g., Linux Ubuntu 16.04):
-- GCC/Compiler version (if compiled from source):
HiAI/Milan_C17/20240414
master_20240426230938_a8763
Excute Mode / 执行模式 (Mandatory / 必填)(PyNative
/Graph
):
Please delete the mode not involved / 请删除不涉及的模式:
/mode graph
test_ms_dqn_train_infer_0001
执行通过,性能及精度正常
[ERROR] OP(156528,python):2024-04-27-03:56:55.325.464 [/home/jenkins/agent-working-dir/workspace/Compile_Ascend_ARM_CentOS/mindspore/mindspore/ccsrc/plugin/device/ascend/kernel/aicpu/aicpu_ops/customize/op_proto/utils/util.cc:106][OP_PROTO][CheckTwoInputDtypeSame][156528] OpName:[Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0] "The Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0 op dtype is not same, type1:int32, type2:int64"
[ERROR] OP(156528,python):2024-04-27-03:56:55.325.495 [/home/jenkins/agent-working-dir/workspace/Compile_Ascend_ARM_CentOS/mindspore/mindspore/ccsrc/plugin/device/ascend/kernel/aicpu/aicpu_ops/customize/op_proto/selection_ops_proto.cc:147][OP_PROTO][SelectVerify][156528] OpName:[Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0] "call function CheckTwoInputDtypeSame failed, data type of input[x1] is not same as input[x2]"
[ERROR] GE(156528,python):2024-04-27-03:56:55.325.647 [infershape_pass.cc:133]156528 InferShapeAndType: ErrorNo: 4294967295(failed) [COMP][PRE_OPT][Call][Verify] Verifying Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0 failed.
[ERROR] GE(156528,python):2024-04-27-03:56:55.325.909 [infershape_pass.cc:123]156528 Infer: ErrorNo: 1343242270(Prepare Graph infershape failed) [COMP][PRE_OPT][Call][InferShapeAndType] for node:Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0(Select) failed
[ERROR] GE(156528,python):2024-04-27-03:56:55.325.924 [infer_base_pass.cc:112]156528 InferAndUpdate: ErrorNo: 1343242270(Prepare Graph infershape failed) [COMP][PRE_OPT][Call][Infer] failed for node Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0, ret: 1343242270
[ERROR] GE(156528,python):2024-04-27-03:56:55.325.935 [infer_base_pass.cc:69]156528 Run: ErrorNo: 1343242270(Prepare Graph infershape failed) [COMP][PRE_OPT][Call][InferAndUpdate] for node Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0 failed! ret: 1343242270
[ERROR] GE(156528,python):2024-04-27-03:56:55.325.963 [base_pass.cc:574]156528 RunPassesOnNode: ErrorNo: 1343225860(Internal errors) [COMP][PRE_OPT][Process][Pass] InferShapePass on node Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0 failed, result 4294967295, the passes will be terminated immediately.
[ERROR] GE(156528,python):2024-04-27-03:56:55.325.976 [base_pass.cc:522]156528 RunPassesNodeOnce: ErrorNo: 4294967295(failed) [COMP][PRE_OPT][Process][Passes] on node Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0 type Select failed, error code:4294967295
[ERROR] GE(156528,python):2024-04-27-03:56:55.325.986 [base_pass.cc:478]156528 RunPassesGraphRepass: ErrorNo: 4294967295(failed) [COMP][PRE_OPT][Process][Passes] on node Default/msrl-MSRL/actors-DQNActor/collect_policy-EpsilonGreedyPolicy/Select-op0 type Select failed, error code:4294967295
当前戴仁杰定界中
定界完成再转给对应责任人
Please assign maintainer to check this issue.
请为此issue分配处理人。
@wanlinhui_A
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。
感谢您的提问,您可以评论//mindspore-assistant更快获取帮助:
问题原因是算子yaml设置了输入类型一致性的校验,已删除该校验,后端算子能处理不同类型输入的场景
回归通过
HiAI/Milan_C17/20240414
mindspore-2.3.0rc2+20240504-cp37-cp37m-linux_aarch64.whl
登录 后才可以发表评论