IGIE的GoogleNet和Inception_v3的int8模型编译时报错, centernet acc报错

### 该问题是怎么引起的？

反馈已在4.3.0修复，待更新

### 重现步骤

### 报错信息

```
data types int8 and float32 do not match in BroadcastRel
data types int8 and float32 do not match in BroadcastRel
Traceback (most recent call last):
  File "/workspace/deepsparkinference/models/cv/classification/googlenet/igie/../../igie_common/inference.py", line 186, in <module>
    main()
  File "/workspace/deepsparkinference/models/cv/classification/googlenet/igie/../../igie_common/inference.py", line 130, in main
    lib = tvm.runtime.load_module(args.engine)
  File "/usr/local/corex-4.2.0/lib64/python3/dist-packages/tvm/runtime/module.py", line 775, in load_module
    raise ValueError(f"cannot find file {path}")
ValueError: cannot find file googlenet_bs_32_int8.so

```

```
root@32eb7cec7ef9:~/deepsparkinference/models/cv/object_detection/centernet/igie# bash scripts/infer_centernet_fp16_accuracy.sh
/usr/local/corex-4.2.0/lib64/python3/dist-packages/tvm/relay/frontend/onnx.py:122: DeprecationWarning: `mapping.TENSOR_TYPE_TO_NP_TYPE` is now deprecated and will be removed in a future release.To silence this warning, please use `helper.tensor_dtype_to_np_dtype` instead.
  return str(TENSOR_TYPE_TO_NP_TYPE[elem_type])
WARNING:autotvm:One or more operators have not been tuned. Please tune your model for better performance. Use DEBUG logging level to see more details.
[02:59:27] /home/corex/sw_home/apps/igie/src/relay/transforms/convert_layout.cc:99: Warning: Desired layout(s) not specified for op: nn.max_pool2d
[02:59:34] /home/corex/sw_home/apps/igie/src/relay/transforms/convert_layout.cc:99: Warning: Desired layout(s) not specified for op: nn.max_pool2d
The Relay type checker is unable to show the following types match:
  Tensor[(32, 80, 175, 175), float32]
  Tensor[(32, 80, 168, 168), float32]
In particular:
  dimension 2 conflicts: 175 does not match 168.  dimension 3 conflicts: 175 does not match 168.
The Relay type checker is unable to show the following types match.
In particular `Tensor[(32, 80, 168, 168), float32]` does not match `Tensor[(32, 80, 175, 175), float32]`
note: run with `TVM_BACKTRACE=1` environment variable to display a backtrace.
Traceback (most recent call last):
  File "/root/deepsparkinference/models/cv/object_detection/centernet/igie/inference.py", line 151, in <module>
    main()
  File "/root/deepsparkinference/models/cv/object_detection/centernet/igie/inference.py", line 84, in main
    lib = tvm.runtime.load_module(args.engine)
  File "/usr/local/corex-4.2.0/lib64/python3/dist-packages/tvm/runtime/module.py", line 775, in load_module
    raise ValueError(f"cannot find file {path}")
ValueError: cannot find file centernet_bs_32_fp16.so
```

DeepSpark/DeepSparkInference

内容风险标识

评论 (0)

DeepSpark/DeepSparkInference .gitee-modal { width: 500px !important; }

内容风险标识