Cannot compile ONNX OCR model to HEF format

General
1

Hi, I’m trying to compile OCR model from ONNX format that I found here: GitHub - ankandrew/fast-plate-ocr: Lightweight & fast OCR models for license plate text recognition.

When I’m trying to run the command:

hailo parser onnx license-plate-ocr.onnx --hw-arch hailo8l --input-format input=BHWC --tensor-shapes input=[1,70,140,1]

or use the script like:

from hailo_sdk_client import ClientRunner
from hailo_sdk_client.exposed_definitions import Dims

onnx_model_name = 'license-plate-ocr'
onnx_path = './license-plate-ocr.onnx'
har_path = './license-plate-ocr.har'

chosen_hw_arch = 'hailo8l'

runner = ClientRunner(hw_arch=chosen_hw_arch)
hn, npz = runner.translate_onnx_model(
    onnx_path,
    onnx_model_name,
    net_input_shapes={'input': [1, 70, 140, 1]},
    net_input_format={'input': [Dims.BATCH,Dims.HEIGHT,Dims.WIDTH,Dims.CHANNELS]}
    )

runner.save_har(har_path)
print(f"Model successfully translated and saved as {har_path}")

I see the error:

[info] Current Time: 20:29:46, 04/12/25
[info] CPU: Architecture: x86_64, Model: DO-Premium-Intel, Number Of Cores: 2, Utilization: 0.0%
[info] Memory: Total: 3GB, Available: 3GB
[info] System info: OS: Linux, Kernel: 5.4.0-205-generic
[info] Hailo DFC Version: 3.30.0
[info] HailoRT Version: Not Installed
[info] PCIe: No Hailo PCIe device was found
[info] Running `hailo parser onnx license-plate-ocr.onnx --hw-arch hailo8l --input-format input=BHWC --tensor-shapes input=[1,70,140,1]`
[info] Translation started on ONNX model license-plate-ocr
[info] Restored ONNX model license-plate-ocr (completion time: 00:00:00.06)
[info] Extracted ONNXRuntime meta-data for Hailo model (completion time: 00:00:00.26)
[info] Simplified ONNX model for a parsing retry attempt (completion time: 00:00:02.95)
Traceback (most recent call last):
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 239, in translate_onnx_model
    parsing_results = self._parse_onnx_model_to_hn(
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 320, in _parse_onnx_model_to_hn
    return self.parse_model_to_hn(
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/sdk_backend/parser/parser.py", line 371, in parse_model_to_hn
    fuser = HailoNNFuser(converter.convert_model(), net_name, converter.end_node_names)
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/model_translator/translator.py", line 83, in convert_model
    self._create_layers()
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/model_translator/edge_nn_translator.py", line 39, in _create_layers
    self._add_direct_layers()
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/model_translator/edge_nn_translator.py", line 121, in _add_direct_layers
    self._layer_callback_from_vertex(vertex)
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/model_translator/onnx_translator/onnx_translator.py", line 407, in _layer_callback_from_vertex
    consumed_vertices = self._create_pooling_layer(vertex)
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/model_translator/onnx_translator/onnx_translator.py", line 970, in _create_pooling_layer
    vertex.validate_reduce_mean_as_pooling_layer()
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/model_translator/onnx_translator/onnx_graph.py", line 5144, in validate_reduce_mean_as_pooling_layer
    dims = [pred.output_format[axis] for axis in axes]
  File "/root/venv_hailo/lib/python3.10/site-packages/hailo_sdk_client/model_translator/onnx_translator/onnx_graph.py", line 5144, in <listcomp>
    dims = [pred.output_format[axis] for axis in axes]
TypeError: 'NoneType' object is not subscriptable

Do you know what I am doing wrong? Or is it a problem with the model and it’s not compatible with Hailo?
Thanks!