How to Retrain Classification model?

General
1

Hi

I wonder if there is a guide that summarizes how to retrain the classification model

In the link (hailo_model_zoo/docs/public_models/HAILO8/HAILO8_classification.rst at master · hailo-ai/hailo_model_zoo · GitHub) I found, I found only the tflte model file that was created after it had already been quantized

As far as I know, it’s impossible to retrain these models

Is it possible to export the classification model made through Yolo8n-cls.pt like object detection to onnx and then compile it?

And I also want to know the classification model input image size limit (I understand that object detection is only estimated at 640.)

2

We provide retraining Docker for a few selected models only.

GitHub - Hailo Model Zoo - Retrain on custom dataset

For all other models you will need to consult the original source. The links are available for each model in the Model Zoo. See page you linked above.

If you need some additional help, ChatGPT can give you some really good pointers in many cases.

3

Thank you for your reply

I want to know selected classification model re-train method

Is there no way to re-train the Hailo-supported Classification model?

4

As I said before we only provide retraining for a selected few models.

Did you try ChatGPT? I tested it with ResNet v1 18. I got a complete description of all steps necessary.

5

Thank you for your help.

Using the paid version of GPT, I successfully trained a Vision Transformer model (vit_base_patch16_224) for 2 classes and exported it to ONNX format (opset version 14), referring to GPT and the official documentation throughout the process.

  • YAML used: vit_base_bn.yaml
  • ALLS file used: vit_base_bn.alls
  • Command used:
hailomz compile --ckpt /hailo_compile/BK/ViT/vit_custom_model.onnx \
                --calib-path /hailo_compile/BK/CAS-ViT/classification/data/PetImages_ImageNet/val \
                --yaml /hailo_compile/BK/ViT/vit_base_bn.yaml \
                --classes 2

An error occurred when I ran the command above.

(venv) root@ca0b87193902:/hailo_compile# hailomz compile --ckpt /hailo_compile/BK/ViT/vit_custom_model.onnx --calib-path /hailo_compile/BK/CAS-ViT/classification/data/PetImages_ImageNet/val --yaml /hailo_compile/BK/ViT/vit_base_bn.yaml --classes 2

<Hailo Model Zoo INFO> Start run for network vit_base_bn ...
<Hailo Model Zoo INFO> Initializing the hailo8 runner...
[info] Translation started on ONNX model vit_base_bn
[info] Restored ONNX model vit_base_bn (completion time: 00:00:01.56)
[info] Extracted ONNXRuntime meta-data for Hailo model (completion time: 00:00:12.02)
[info] Simplified ONNX model for a parsing retry attempt (completion time: 00:00:26.92)
Traceback (most recent call last):
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 220, in translate_onnx_model
    parsing_results = self._parse_onnx_model_to_hn(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 300, in _parse_onnx_model_to_hn
    return self.parse_model_to_hn(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 340, in parse_model_to_hn
    converter = ONNXConverter(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/onnx_translator/onnx_translator.py", line 170, in __init__
    super().__init__(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/edge_nn_translator.py", line 27, in __init__
    super().__init__(graph, start_node_names, end_node_names)
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/translator.py", line 51, in __init__
    self._calculate_valid_subgraph_scope()
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/translator.py", line 388, in _calculate_valid_subgraph_scope
    current_vertex.in_valid_subgraph = True
AttributeError: 'NoneType' object has no attribute 'in_valid_subgraph'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/hailo_test/venv/bin/hailomz", line 33, in <module>
    sys.exit(load_entry_point('hailo-model-zoo', 'console_scripts', 'hailomz')())
  File "/hailo_test/hailo_model_zoo/hailo_model_zoo/main.py", line 122, in main
    run(args)
  File "/hailo_test/hailo_model_zoo/hailo_model_zoo/main.py", line 111, in run
    return handlers[args.command](args)
  File "/hailo_test/hailo_model_zoo/hailo_model_zoo/main_driver.py", line 219, in compile
    _ensure_optimized(runner, logger, args, network_info)
  File "/hailo_test/hailo_model_zoo/hailo_model_zoo/main_driver.py", line 57, in _ensure_optimized
    _ensure_parsed(runner, logger, network_info, args)
  File "/hailo_test/hailo_model_zoo/hailo_model_zoo/main_driver.py", line 83, in _ensure_parsed
    parse_model(runner, network_info, ckpt_path=args.ckpt_path, results_dir=args.results_dir, logger=logger)
  File "/hailo_test/hailo_model_zoo/hailo_model_zoo/core/main_utils.py", line 95, in parse_model
    model_name = translate_model(runner, network_info, ckpt_path, tensor_shapes=start_node_shapes)
  File "/hailo_test/hailo_model_zoo/hailo_model_zoo/utils/parse_utils.py", line 28, in translate_model
    runner.translate_onnx_model(ckpt_path, model_name,
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_common/states/states.py", line 16, in wrapped_func
    return func(self, *args, **kwargs)
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/runner/client_runner.py", line 1158, in translate_onnx_model
    parser.translate_onnx_model(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 260, in translate_onnx_model
    parsing_results = self._parse_onnx_model_to_hn(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 300, in _parse_onnx_model_to_hn
    return self.parse_model_to_hn(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 340, in parse_model_to_hn
    converter = ONNXConverter(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/onnx_translator/onnx_translator.py", line 170, in __init__
    super().__init__(
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/edge_nn_translator.py", line 27, in __init__
    super().__init__(graph, start_node_names, end_node_names)
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/translator.py", line 51, in __init__
    self._calculate_valid_subgraph_scope()
  File "/hailo_test/venv/lib/python3.10/site-packages/hailo_sdk_client/model_translator/translator.py", line 388, in _calculate_valid_subgraph_scope
    current_vertex.in_valid_subgraph = True
AttributeError: 'NoneType' object has no attribute 'in_valid_subgraph'

I did not modify the .alls file. In the .yaml file, the only changes I made were to update the network_path and alls_script fields under paths to absolute paths.

How should I proceed with the compilation?
(The transformed onnx model had no problem with image inference)