[参考译文] EDGE-AI-STUDIO：需要有关使用 edgeai-modelmaker 生成模型的说明

admin

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1494273/edge-ai-studio-need-clarification-on-making-model-using-edgeai-modelmaker

器件型号：EDGE-AI-STUDIO

工具与软件：

尊敬的专家：

我正在尝试使用 https://github.com/TexasInstruments/edgeai-modelmaker 创建自定义模型。我不能运行 run_convert_dataset.sh 似乎错误在一开始似乎它找不到一个图像,所以我从 json 文件中删除了该图像和注释。我仅取消了该脚本中的以下代码的注释。

python3 ./scripts/convert_dataset.py --source_format=labelstudio_classification --source_anno=./data/sample/sample.json --source_data=./data/sample/new --dest_anno=./data/sample/labels.json

但在运行时、似乎出现了如下错误。

argv: ['./scripts/convert_dataset.py', '--source_format=labelstudio_classification', '--source_anno=./data/sample/sample.json', '--source_data=./data/sample/new', '--dest_anno=./data/sample/labels.json']
Traceback (most recent call last):
  File "/home/user/modelmaker/edgeai-modelmaker/./scripts/convert_dataset.py", line 414, in <module>
    main(args)
  File "/home/user/modelmaker/edgeai-modelmaker/./scripts/convert_dataset.py", line 378, in main
    convert_labelstudio_classification(args)
  File "/home/user/modelmaker/edgeai-modelmaker/./scripts/convert_dataset.py", line 248, in convert_labelstudio_classification
    dataset_json_min = json.load(afp)
  File "/home/user/.pyenv/versions/3.10.16/lib/python3.10/json/__init__.py", line 293, in load
    return loads(fp.read(),
  File "/home/user/.pyenv/versions/3.10.16/lib/python3.10/json/__init__.py", line 335, in loads
    raise JSONDecodeError("Unexpected UTF-8 BOM (decode using utf-8-sig)",
json.decoder.JSONDecodeError: Unexpected UTF-8 BOM (decode using utf-8-sig): line 1 column 1 (char 0)

如何解决此问题？

步骤我做:
已从 GitHub 页面下载并安装所有必需内容。
使用 label studio 对一些图像进行注释并将其导出为 JSON-MIN 格式。
使用的图像和 json 文件将复制到 modelmaker 目录中的路径/data/sample 中。
然后尝试运行 run_convert_dataset.sh。

请帮我解决这个问题。

此致、
Sajan

6 个月前

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

您好！
我能否在从模型编写器导出的已训练模型中使用 json 文件。
此致、
Sajan

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

尊敬的 Sajan：

您可以将从 Model Composer 导出的数据集用作 Model Maker 的数据集输入。

我将把这个问题交给支持 Model Maker 的团队、以解决您遇到的错误。

Martin

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

Martin、您好！

您能否尽快连接支持团队。

此致、
Sajan

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

https://github.com/TexasInstruments/edgeai-modelmaker 已标记为存档、不再进行维护。它将立即被删除。

请从 https://github.com/TexasInstruments/edgeai-tensorlab 获取 edgeai-modelmaker

数据集的说明如下： https://github.com/TexasInstruments/edgeai-tensorlab/tree/main/edgeai-modelmaker#step-4-prepare-your-own-dataset-with-your-own-images-and-annotations

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

Manu、您好！
感谢您的答复。前几天我看到了,需要标签工作室进行注释。
但在这个存储库中、我找不到该步骤。

此致、
Sajan

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

还有一点、我在模型编写器中标注了一个大图像集。我可以将经过训练的模型导出给模型制作者吗

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

可以-您可以下载在模型编写器中注释的数据集、并在 modelmaker 中使用它。

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

[报价 userid="46292" url="~/support/processors-group/processors/f/processors-forum/1494273/edge-ai-studio-need-clarification-on-making-model-using-edgeai-modelmaker/5743833 #5743833"] 来自 https://github.com/TexasInstruments/edgeai-tensorlab的 edgeai-modelmaker

运行 ./ setup_cpu.sh 时出现错误

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
openxlab 0.1.2 requires setuptools~=60.2.0, but you have setuptools 73.0.0 which is incompatible.

ERROR: Cannot install -r requirements_freeze.txt (line 55), -r requirements_freeze.txt (line 80) and onnx==1.14.0 because these package versions have conflicting dependencies.

ERROR: ResolutionImpossible: for help visit pip.pypa.io/.../

谢谢、现在它处于 EPOCH 15时代、培训仍在继续、setup_cpu.sh 出现版本冲突错误。我能否将在./data/projects 中创建的模型使用到 AM62A opt/model_zoo/中所需的任何其他培训流程的电子邮件。
我没有明确指出 SDK 版本是10.1。此模型与 SDK 10.1兼容。或任何我看起来不是 SDK 版本的地方

此致、
Sajan

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

您好！

建立关系。编译有一些错误。

(py310) user@M122DTRV:~/modelmaker/edgeai-tensorlab/edgeai-modelmaker$ ./run_modelmaker.sh AM62A config_classification.yaml 
Number of AVX cores detected in PC: 12
AVX compilation speedup in PC     : 1
Target device                     : AM62A
PYTHONPATH                        : .:
TIDL_TOOLS_PATH                   : ../edgeai-benchmark/tools/tidl_tools_package/AM62A/tidl_tools
LD_LIBRARY_PATH                   : ../edgeai-benchmark/tools/tidl_tools_package/AM62A/tidl_tools:
argv: ['./scripts/run_modelmaker.py', 'config_classification.yaml', '--target_device', 'AM62A']
---------------------------------------------------------------------
INFO: ModelMaker - task_type:classification model_name:regnet_x_800mf dataset_name:DMS run_name:20250401-123741/regnet_x_800mf
- Model: regnet_x_800mf
- TargetDevices & Estimated Inference Times (ms): {'TDA4VM': 2.95, 'AM62A': 5.95, 'AM67A': '5.95 (with 1/2 device capability)', 'AM68A': 2.92, 'AM69A': '2.85 (with 1/4th device capability)'}
- This model can be compiled for the above device(s).
---------------------------------------------------------------------
downloading from download.pytorch.org/.../regnet_x_800mf-ad17e45c.pth to ./data/downloads/pretrained/torch/hub/checkpoints/regnet_x_800mf-ad17e45c.pth
100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 29302593/29302593 [00:01<00:00, 15028495.50B/s]
assuming the given download_url is a valid path: /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/datasets/DMS
INFO: ModelMaker - dataset split sizes {'train': 8799, 'val': 2253}
INFO: ModelMaker - max_num_files is set to: 10000
INFO: ModelMaker - dataset split sizes are limited to: {'train': 8000, 'val': 2000}
INFO: ModelMaker - dataset loading OK
INFO: ModelMaker - run params is at: /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250401-123741/regnet_x_800mf/run.yaml
INFO: ModelMaker - running training - for detailed info see the log file: /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250401-123741/regnet_x_800mf/training/run.log
TASKS TOTAL=1, NUM_RUNNING=1:   0%|                                                        | 0/1 [2:15:45<?, ?it/s, postfix={'RUNNING': ['20250401-123741/regnet_x_800mf:training'], 'COMPLETED': []}]TASKS TOTAL=1, NUM_RUNNING=1:   0%|                                                        | 0/1 [5:16:51<?, ?it/s, postfix={'RUNNING': ['20250401-123741/regnet_x_800mf:training'], 'COMPLETED': []}]
TASKS TOTAL=1, NUM_RUNNING=0: 100%|██████████████████████████████████████████████████████████████████████| 1/1 [8:45:37<00:00, 31537.33s/it, postfix={'RUNNING': [], 'COMPLETED': ['regnet_x_800mf']}]
Trained model is at: /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250401-123741/regnet_x_800mf/training

SUCCESS: ModelMaker - Training completed.
INFO: ModelMaker - running compilation - for detailed info see the log file: /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250401-123741/regnet_x_800mf/compilation/AM62A/work/cl-6170/run.log

INFO:20250401-212324: number of configs - 1
TASKS TOTAL=1, NUM_RUNNING=1:   0%|                                                                                   | 0/1 [00:02<?, ?it/s, postfix={'RUNNING': ['cl-6170:import'], 'COMPLETED': []}]
ERROR:20250401-212327: model_id:cl-6170 run_import:True run_inference:False - [ONNXRuntimeError] : 1 : FAIL : Non-zero status code returned while running TIDL_0 node. Name:'TIDLExecutionProvider_TIDL_0_0' Status Message: TIDL Compute Import Failed.
Traceback (most recent call last):
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/pipeline_runner.py", line 273, in _run_pipeline
    result = cls._run_pipeline_impl(settings, pipeline_config, description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/pipeline_runner.py", line 308, in _run_pipeline_impl
    result = accuracy_pipeline(description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/accuracy_pipeline.py", line 79, in __call__
    param_result = self._run(description=description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/accuracy_pipeline.py", line 104, in _run
    self._import_model(description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/accuracy_pipeline.py", line 168, in _import_model
    output, info_dict = session.run_import(input_data, info_dict)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/sessions/onnxrt_session.py", line 62, in run_import
    output = ONNXRuntimeWrapper.run_import(self, input_data)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/core/onnxrt_runtime.py", line 58, in run_import
    output = self._run(input_data, output_keys)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/core/onnxrt_runtime.py", line 91, in _run
    outputs = self.interpreter.run(output_keys, input_data)
  File "/home/user/.pyenv/versions/py310/lib/python3.10/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 217, in run
    return self._sess.run(output_names, input_feed, run_options)
onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Non-zero status code returned while running TIDL_0 node. Name:'TIDLExecutionProvider_TIDL_0_0' Status Message: TIDL Compute Import Failed.
WARNING: terminating the process - cl-6170:import - Status Message: TIDL Compute Import Failed
TASKS TOTAL=1, NUM_RUNNING=1:   0%|                                                                                    | 0/1 [00:04<?, ?it/s, postfix={'RUNNING': ['cl-6170:infer'], 'COMPLETED': []}]
ERROR:20250401-212331: Error occurred: cl-6170:infer - Error Code: -11 at /home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/utils/parallel_runner.py
WARNING: terminating the process - cl-6170:infer - Status Message: TIDL Compute Import Failed
TASKS TOTAL=1, NUM_RUNNING=0: 100%|██████████████████████████████████████████████████████████████████████████████████| 1/1 [00:07<00:00,  2.00s/it, postfix={'RUNNING': [], 'COMPLETED': ['cl-6170']}]
WARNING: Benchmark - completed: 0/1
TASKS TOTAL=1, NUM_RUNNING=0: 100%|██████████████████████████████████████████████████████████████████████████████████| 1/1 [00:07<00:00,  7.28s/it, postfix={'RUNNING': [], 'COMPLETED': ['cl-6170']}]
INFO: packaging artifacts to /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250401-123741/regnet_x_800mf/compilation/AM62A/pkg please wait...
WARNING:20250401-212331: could not package - /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250401-123741/regnet_x_800mf/compilation/AM62A/work/cl-6170
Traceback (most recent call last):
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/./scripts/run_modelmaker.py", line 153, in <module>
    main(config)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/./scripts/run_modelmaker.py", line 88, in main
    model_runner.run()
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/edgeai_modelmaker/ai_modules/vision/runner.py", line 193, in run
    self.model_compilation.run()
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/edgeai_modelmaker/ai_modules/vision/compilation/edgeai_benchmark.py", line 163, in run
    edgeai_benchmark.interfaces.package_artifacts(self.settings, self.work_dir, out_dir=self.package_dir, custom_model=True)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/interfaces/run_package.py", line 271, in package_artifacts
    with open(os.path.join(out_dir,'artifacts.yaml'), 'w') as fp:
FileNotFoundError: [Errno 2] No such file or directory: '/home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250401-123741/regnet_x_800mf/compilation/AM62A/pkg/artifacts.yaml'

来源于 run.log

2025-04-01 21:23:27.065009786 [E:onnxruntime:, sequential_executor.cc:514 ExecuteKernel] Non-zero status code returned while running TIDL_0 node. Name:'TIDLExecutionProvider_TIDL_0_0' Status Message: TIDL Compute Import Failed.

还有一个问题、我可以通过修改 CONFIG_CLASSIFIC.YAML 或任何文件来运行编译、跳过成功完成的训练吗。

此致、
Sajan

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

您好！

我将 Onxruntime 升级到1.21.0。现在 TIDLCompilationProvider 出现错误。在搜索时、似乎需要两次导出到 edgeai-tivox-modules、但 edgeai-tensorlab 中没有目录。详细日志如下所示。

Terminal Print

SUCCESS: ModelMaker - Training completed.
INFO: ModelMaker - running compilation - for detailed info see the log file: /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250402-172810/regnet_x_800mf/compilation/AM62A/work/cl-6170/run.log

INFO:20250402-173443: number of configs - 1
TASKS TOTAL=1, NUM_RUNNING=1:   0%|                                                                                    | 0/1 [00:21<?, ?it/s, postfix={'RUNNING': ['cl-6170:infer'], 'COMPLETED': []}]
ERROR:20250402-173507: model_id:cl-6170 run_import:False run_inference:True - unsupported operand type(s) for *: 'NoneType' and 'float'
Traceback (most recent call last):
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/pipeline_runner.py", line 273, in _run_pipeline
    result = cls._run_pipeline_impl(settings, pipeline_config, description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/pipeline_runner.py", line 308, in _run_pipeline_impl
    result = accuracy_pipeline(description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/accuracy_pipeline.py", line 79, in __call__
    param_result = self._run(description=description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/accuracy_pipeline.py", line 128, in _run
    output_list = self._infer_frames(description)
  File "/home/user/modelmaker/edgeai-tensorlab/edgeai-benchmark/edgeai_benchmark/pipelines/accuracy_pipeline.py", line 250, in _infer_frames
    self.infer_stats_dict.update({'perfsim_time_ms': stats_dict['perfsim_time'] * constants.MILLI_CONST})
TypeError: unsupported operand type(s) for *: 'NoneType' and 'float'
TASKS TOTAL=1, NUM_RUNNING=0: 100%|██████████████████████████████████████████████████████████████████████████████████| 1/1 [00:23<00:00,  2.04s/it, postfix={'RUNNING': [], 'COMPLETED': ['cl-6170']}]
WARNING: Benchmark - completed: 0/1
TASKS TOTAL=1, NUM_RUNNING=0: 100%|██████████████████████████████████████████████████████████████████████████████████| 1/1 [00:24<00:00, 24.14s/it, postfix={'RUNNING': [], 'COMPLETED': ['cl-6170']}]
INFO: packaging artifacts to /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250402-172810/regnet_x_800mf/compilation/AM62A/pkg please wait...
SUCCESS:20250402-173508: finished packaging - /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250402-172810/regnet_x_800mf/compilation/AM62A/work/cl-6170
Compiled model is at: /home/user/modelmaker/edgeai-tensorlab/edgeai-modelmaker/data/projects/DMS/run/20250402-172810/regnet_x_800mf/compilation/AM62A/pkg/20250402-172810_regnet_x_800mf_onnxrt_AM62A.tar.gz

WARNING: ModelMaker - Compilation completed with errors.

-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Error from the log file created is below


/home/user/.pyenv/versions/py310/lib/python3.10/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:118: UserWarning: Specified provider 'TIDLCompilationProvider' is not in available provider names.Available providers: 'AzureExecutionProvider, CPUExecutionProvider'
  warnings.warn(
*************** EP Error ***************
EP Error Unknown Provider Type: TIDLCompilationProvider when using ['TIDLCompilationProvider', 'CPUExecutionProvider']
Falling back to ['CPUExecutionProvider'] and retrying.

帮助我解决问题。

此致、
Sajan

0 admin 6 个月前

TI__Guru**** 2393725 points

请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。

>> EP 错误未知提供程序类型：使用['TIDLComplationProvider'、'CPUExecutionProvider']时的 TIDLCompilationProvider
>>返回至['CPUExecutionProvider']并重试。

当 TIDL_TOOLS_PATH 环境变量未正确设置或 TIDL_TOOLS 未正确下载时、会发生这种情况

处理器（参考译文帖）

处理器（参考译文帖）(Read Only)

[参考译文] EDGE-AI-STUDIO：需要有关使用 edgeai-modelmaker 生成模型的说明