Fill - Get op from aclOpMap failed due to hashMap_ is empty · Issue #I4DKLV · Ascend/modelzoo

一、问题现象（附报错日志上下文）：
xxxx

二、软件版本:
-- CANN 版本 (e.g., CANN 3.0.x，5.x.x): CANN 5.0.3.alpha003
--Tensorflow/Pytorch/MindSpore 版本: N/A
--Python 版本 (e.g., Python 3.7.5): 3.7
-- MindStudio版本 (e.g., MindStudio 2.0.0 (beta3)): N/A
--操作系统版本 (e.g., Ubuntu 18.04):  Ubuntu 18.04

三、测试步骤：

Fill算子在INT64数据类型下，在CANN 5.0.2.alpha005上可以成功运行，在CANN 5.0.3.alpha003失败。

四、日志信息:
xxxx
请根据自己的运行环境参考以下方式搜集日志信息，如果涉及到算子开发相关的问题，建议也提供UT/ST测试和单算子集成测试相关的日志。

日志提供方式:
将日志打包后作为附件上传。若日志大小超出附件限制，则可上传至外部网盘后提供链接。

获取方法请参考wiki：
https://gitee.com/ascend/modelzoo/wikis/%E5%A6%82%E4%BD%95%E8%8E%B7%E5%8F%96%E6%97%A5%E5%BF%97%E5%92%8C%E8%AE%A1%E7%AE%97%E5%9B%BE?sort_id=4097825

错误日志为：

[WARNING] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.238 [op_model_manager.cpp:297]8360 Get: Get op from aclOpMap failed due to hashMap_ is empty when seed = 18274212371189895759, op = OpType: Fill, InputDesc[0]: [TensorDesc] DataType = 9, Format = 0, StorageFormat = 0, Shape = [2], StorageShape = [2], shapeRange = [], memtype = 1, isConst = 0 InputDesc[1]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [1], StorageShape = [1], shapeRange = [], memtype = 0, isConst = 0 OutputDesc[0]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [123, 92], StorageShape = [123, 92], shapeRange = [], memtype = 0, isConst = 0

[ERROR] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.265 [op_model_manager.cpp:810]8360 MatchDynamicOpModel: [FINAL][FINAL][Match][OpModel]MatchOpModel fail from static map or dynamic map

日志上下文如下：

```bash
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.104 [hash_utils.cpp:45]8360 GetTensorDescHash: GetTensorDescHash out, seed = 17484447184844283459
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.107 [hash_utils.cpp:61]8360 GetAclOpHash: After Output GetTensorDescHash, seed = 17484447184844283459
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.110 [hash_utils.cpp:69]8360 GetAclOpHash: After get attr combine hash, seed = 1683616286680390730
[WARNING] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.131 [op_model_manager.cpp:297]8360 Get: Get op from aclOpMap failed due to hashMap_ is empty when seed = 1683616286680390730, op = OpType: Fill, InputDesc[0]: [TensorDesc] DataType = 9, Format = 0, StorageFormat = 0, Shape = [2], StorageShape = [2], shapeRange = [], memtype = 1, isConst = 1 , isConst = true, Const Len = 4 ,Const data = 123,0,92,0, InputDesc[1]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [1], StorageShape = [1], shapeRange = [], memtype = 0, isConst = 0 OutputDesc[0]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [123, 92], StorageShape = [123, 92], shapeRange = [], memtype = 0, isConst = 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.136 [op_model_manager.cpp:430]8360 operator(): do nothing when call reset of shared_ptr
[INFO] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.140 [attr_utils.cpp:718]8360 SaveConstToAttr: begin to insert constDataBuf in aclopAttr
[INFO] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.143 [attr_utils.cpp:39]8360 ClearConstBuf: clear constBuf
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.162 [hash_utils.cpp:54]8360 GetAclOpHash: GetAclOpHash start, aclOp is OpType: Fill, InputDesc[0]: [TensorDesc] DataType = 9, Format = 0, StorageFormat = 0, Shape = [2], StorageShape = [2], shapeRange = [], memtype = 1, isConst = 0 InputDesc[1]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [1], StorageShape = [1], shapeRange = [], memtype = 0, isConst = 0 OutputDesc[0]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [123, 92], StorageShape = [123, 92], shapeRange = [], memtype = 0, isConst = 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.166 [hash_utils.cpp:18]8360 GetTensorDescHash: GetTensorDescHash in, num = 2, seed = 11128535453279901141
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.169 [hash_utils.cpp:31]8360 GetTensorDescHash: i = 0, dims size is 1
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.172 [hash_utils.cpp:35]8360 GetTensorDescHash: i = 0, shapeRange size is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.175 [hash_utils.cpp:40]8360 GetTensorDescHash: i = 0, isConst is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.178 [hash_utils.cpp:42]8360 GetTensorDescHash: i = 0, memtype is 1
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.181 [hash_utils.cpp:31]8360 GetTensorDescHash: i = 1, dims size is 1
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.183 [hash_utils.cpp:35]8360 GetTensorDescHash: i = 1, shapeRange size is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.186 [hash_utils.cpp:40]8360 GetTensorDescHash: i = 1, isConst is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.189 [hash_utils.cpp:42]8360 GetTensorDescHash: i = 1, memtype is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.192 [hash_utils.cpp:45]8360 GetTensorDescHash: GetTensorDescHash out, seed = 8399953208672443207
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.195 [hash_utils.cpp:58]8360 GetAclOpHash: After Input GetTensorDescHash, seed = 8399953208672443207
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.198 [hash_utils.cpp:18]8360 GetTensorDescHash: GetTensorDescHash in, num = 1, seed = 8399953208672443207
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.201 [hash_utils.cpp:31]8360 GetTensorDescHash: i = 0, dims size is 2
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.204 [hash_utils.cpp:35]8360 GetTensorDescHash: i = 0, shapeRange size is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.207 [hash_utils.cpp:40]8360 GetTensorDescHash: i = 0, isConst is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.210 [hash_utils.cpp:42]8360 GetTensorDescHash: i = 0, memtype is 0
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.213 [hash_utils.cpp:45]8360 GetTensorDescHash: GetTensorDescHash out, seed = 18377841822433365061
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.215 [hash_utils.cpp:61]8360 GetAclOpHash: After Output GetTensorDescHash, seed = 18377841822433365061
[DEBUG] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.218 [hash_utils.cpp:69]8360 GetAclOpHash: After get attr combine hash, seed = 18274212371189895759
[WARNING] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.238 [op_model_manager.cpp:297]8360 Get: Get op from aclOpMap failed due to hashMap_ is empty when seed = 18274212371189895759, op = OpType: Fill, InputDesc[0]: [TensorDesc] DataType = 9, Format = 0, StorageFormat = 0, Shape = [2], StorageShape = [2], shapeRange = [], memtype = 1, isConst = 0 InputDesc[1]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [1], StorageShape = [1], shapeRange = [], memtype = 0, isConst = 0 OutputDesc[0]: [TensorDesc] DataType = 0, Format = 0, StorageFormat = 0, Shape = [123, 92], StorageShape = [123, 92], shapeRange = [], memtype = 0, isConst = 0
[INFO] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.243 [op_model_manager.cpp:736]8360 MatchStaticOpModel: Match static opModels fail, begin to match model from dynamic opModels. opType = Fill
[INFO] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.247 [op_model_manager.cpp:755]8360 MatchDynamicOpModel: aclOp.numInputs is 2, aclOp.numOutputs is 1
[INFO] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.250 [op_model_manager.cpp:263]8360 GetTensorShapeStatus: GetTensorShapeStatus opType is Fill, size of shapeStatus is 0
[ERROR] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.793.265 [op_model_manager.cpp:810]8360 MatchDynamicOpModel: [FINAL][FINAL][Match][OpModel]MatchOpModel fail from static map or dynamic map
[INFO] ASCENDCL(8360,python3.7):2021-10-12-17:02:35.821.090 [device.cpp:63]8360 aclrtSetDevice: start to execute aclrtSetDevice, deviceId = 0.
[INFO] RUNTIME(8360,python3.7):2021-10-12-17:02:35.821.111 [npu_driver.cc:218] 8360 GetDeviceCount: device count=4
```

Ascend/modelzoo
暂停

内容风险标识

Fill - Get op from aclOpMap failed due to hashMap_ is empty

评论 (9)

Ascend/modelzoo暂停 .gitee-modal { width: 500px !important; }

内容风险标识

Fill - Get op from aclOpMap failed due to hashMap_ is empty

评论 (9)

搜索帮助

Ascend/modelzoo
暂停