Fail to convert onnx to HEF using hailo command

General
1

The result of the conversion is as shown below, an error has occurred.

Q1. Is an onnx vgg16_Opset18.onnx converted to HAR successfully?
Q2. Are there any failed processes in converting the HAR to vgg16_Opset18_optimized.har?
Q3. How can I solve this? Or, first of all, is this conversion even possible?

TIA

(v3.30.0) koch@DESKTOP:/work$ hailo parser onnx --hw-arch hailo8l ../../vgg16_Opset18.onnx
[info] Current Time: 11:39:16, 08/13/25
[info] CPU: Architecture: x86_64, Model: 11th Gen Intel(R) Core(TM) i7-11700KF @ 3.60GHz, Number Of Cores: 16, Utilization: 0.0%
[info] Memory: Total: 31GB, Available: 30GB
[info] System info: OS: Linux, Kernel: 6.6.87.2-microsoft-standard-WSL2
[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 --hw-arch hailo8l ../../vgg16_Opset18.onnx`
[info] Translation started on ONNX model vgg16_Opset18
[info] Restored ONNX model vgg16_Opset18 (completion time: 00:00:14.60)
[info] Extracted ONNXRuntime meta-data for Hailo model (completion time: 00:00:22.23)
[info] Start nodes mapped from original model: 'x': 'vgg16_Opset18/input_layer1'.
[info] End nodes mapped from original model: '/classifier/classifier.6/Gemm'.
[info] Translation completed on ONNX model vgg16_Opset18 (completion time: 00:00:24.67)
[info] Saved HAR to: /work/vgg16_Opset18.har

(v3.30.0) koch@DESKTOP:/work$ hailo optimize --hw-arch hailo8l --use-random-calib-set vgg16_Opset18.har
[info] Current Time: 11:42:06, 08/13/25
[info] CPU: Architecture: x86_64, Model: 11th Gen Intel(R) Core(TM) i7-11700KF @ 3.60GHz, Number Of Cores: 16, Utilization: 0.1%
[info] Memory: Total: 31GB, Available: 30GB
[info] System info: OS: Linux, Kernel: 6.6.87.2-microsoft-standard-WSL2
[info] Hailo DFC Version: 3.30.0
[info] HailoRT Version: Not Installed
[info] PCIe: No Hailo PCIe device was found
[info] Running `hailo optimize --hw-arch hailo8l --use-random-calib-set vgg16_Opset18.har`
[info] Found model with 3 input channels, using real RGB images for calibration instead of sampling random data.
[info] Starting Model Optimization
[warning] Reducing optimization level to 0 (the accuracy won't be optimized and compression won't be used) because there's no available GPU
[warning] Running model optimization with zero level of optimization is not recommended for production use and might lead to suboptimal accuracy results
[info] Model received quantization params from the hn
[info] MatmulDecompose skipped
[info] Starting Mixed Precision
[info] Model Optimization Algorithm Mixed Precision is done (completion time is 00:00:00.11)
[info] LayerNorm Decomposition skipped
[info] Starting Statistics Collector
[info] Using dataset with 64 entries for calibration
Calibration: 100%|███...███| 64/64 [00:32<00:00,  1.99entries/s]
[info] Model Optimization Algorithm Statistics Collector is done (completion time is 00:00:33.77)
[info] Starting Fix zp_comp Encoding
[info] Model Optimization Algorithm Fix zp_comp Encoding is done (completion time is 00:00:00.00)
[info] Matmul Equalization skipped
[info] No shifts available for layer vgg16_Opset18/conv13/conv_op, using max shift instead. delta=0.5329
[info] No shifts available for layer vgg16_Opset18/conv13/conv_op, using max shift instead. delta=0.2664
[info] No shifts available for layer vgg16_Opset18/fc1/conv_op, using max shift instead. delta=1.4097
[info] No shifts available for layer vgg16_Opset18/fc1/conv_op, using max shift instead. delta=0.7049
[info] No shifts available for layer vgg16_Opset18/conv13/conv_op, using max shift instead. delta=0.2664
[info] No shifts available for layer vgg16_Opset18/fc2/conv_op, using max shift instead. delta=1.4663
[info] No shifts available for layer vgg16_Opset18/fc2/conv_op, using max shift instead. delta=0.7331
[info] No shifts available for layer vgg16_Opset18/fc1/conv_op, using max shift instead. delta=0.7049
[info] No shifts available for layer vgg16_Opset18/fc3/conv_op, using max shift instead. delta=1.3297
[info] No shifts available for layer vgg16_Opset18/fc3/conv_op, using max shift instead. delta=0.6649
[info] No shifts available for layer vgg16_Opset18/fc2/conv_op, using max shift instead. delta=0.7331
[info] Finetune encoding skipped
[info] Bias Correction skipped
[info] Adaround skipped
[info] Quantization-Aware Fine-Tuning skipped
[info] Layer Noise Analysis skipped
[info] The calibration set seems to not be normalized, because the values range is [(0.0, 1.0), (0.0, 1.0), (0.0, 1.0)].
Since the neural core works in 8-bit (between 0 to 255), a quantization will occur on the CPU of the runtime platform.
Add a normalization layer to the model to offload the normalization to the neural core.
Refer to the user guide Hailo Dataflow Compiler user guide / Model Optimization / Optimization Related Model Script Commands / model_modification_commands / normalization for details.
[info] Model Optimization is done
[info] Saved HAR to: /work/vgg16_Opset18_optimized.har

(v3.30.0) koch@DESKTOP:/work$ hailo compiler --hw-arch hailo8l vgg16_Opset18_optimized.har
[info] Current Time: 11:46:47, 08/13/25
[info] CPU: Architecture: x86_64, Model: 11th Gen Intel(R) Core(TM) i7-11700KF @ 3.60GHz, Number Of Cores: 16, Utilization: 0.0%
[info] Memory: Total: 31GB, Available: 30GB
[info] System info: OS: Linux, Kernel: 6.6.87.2-microsoft-standard-WSL2
[info] Hailo DFC Version: 3.30.0
[info] HailoRT Version: Not Installed
[info] PCIe: No Hailo PCIe device was found
[info] Running `hailo compiler --hw-arch hailo8l vgg16_Opset18_optimized.har`
[info] Compiling network
[info] To achieve optimal performance, set the compiler_optimization_level to "max" by adding performance_param(compiler_optimization_level=max) to the model script. Note that this may increase compilation time.
[info] Loading network parameters
[info] Starting Hailo allocation and compilation flow
[info] Finding the best partition to contexts...
[...........................<==>.........] Elapsed: 00:42:51
[error] Mapping Failed (Timeout, allocation time: 1h 0m 19s)
Compiler could not find a valid partition to contexts. Most common error is: Automri finished with too many resources on context_3 with 119/1662 failures.
Mapping Failed (Timeout, allocation time: 1h 0m 19s)

[error] Failed to produce compiled graph
[error] BackendAllocatorException: Compilation failed: Compiler could not find a valid partition to contexts. Most common error is: Automri finished with too many resources on context_3 with 119/1662 failures.
Mapping Failed (Timeout, allocation time: 1h 0m 19s)
2

Hi @Koch ,

The error you got is due a compilation timeout. Please try the following model script commands.

1) Extend the timeout

The timeout is calculated dynamically based on model size, but, if the model is difficult to compile, it is possible to increase the timeout using the following model script command:

allocator_param(timeout=VALUE)

For example:
allocator_param(timeout=6h)

2) Performance mode
Alternatively, you can try to run the compilation in performance mode, using the following command:

performance_param(compiler_optimization_level=max)

This will instruct the compiler to perform an exhaustive search at different resources utilization levels, in order to find the best FPS, and in cases like this it may help you find a valid allocation. The compilation time will be higher (a few hours), so I recommend running this overnight.

To run compilation with a model script, your command must be modified like this:

hailo compiler --hw-arch hailo8l vgg16_Opset18_optimized.har --model-script YOUR_SCRIPT.alls
3

Thank you for your advice with sample code easy to understand.

I’ll try it.