[Bug]: Supported precisions set is not available for Convert operation.

[starlitsky2010]

OpenVINO Version

Master

Operating System

Android System

Device used for inference

CPU

Framework

ONNX

Model used

mobelinet-v3-tf

Issue description

LD_LIBRARY_PATH=/data/local/tmp ./data/local/tmp/benchmark_app -d CPU -m /data/local/tmp/mobelinet-v3-tf/v3-small_224_1.0_float.xml -hint throughput

Step-by-step reproduction

According to https://github.com/openvinotoolkit/openvino/blob/master/docs/dev/build_android.md I built the android ONNX with ABI x86_64 and enable the benchmark_app compilation with latest version OpenVINO master baseline (commit ID 287ab98)

But when I try mobelinet-v3-tf example mentioned in the official document above, the error occurs.
Could OpenVINO guys give some tips about it?

Relevant log output

[Step 1/11] Parsing and validating input arguments
[ INFO ] Parsing input parameters
[Step 2/11] Loading OpenVINO Runtime
[ INFO ] OpenVINO:
[ INFO ] Build ................................. 2025.0.0-17426-287ab9883ac
[ INFO ]
[ INFO ] Device info:
[ INFO ] CPU
[ INFO ] Build ................................. 2025.0.0-17426-287ab9883ac
[ INFO ]
[ INFO ]
[Step 3/11] Setting device configuration
[Step 4/11] Reading model files
[ INFO ] Loading model files
[ INFO ] Read model took 33.87 ms
[ INFO ] Original model I/O parameters:
[ INFO ] Network inputs:
[ INFO ]     input:0 (node: input) : f32 / [...] / [1,224,224,3]
[ INFO ] Network outputs:
[ INFO ]     MobilenetV3/Predictions/Softmax:0 (node: MobilenetV3/Predictions/Softmax) : f32 / [...] / [1,1001]
[Step 5/11] Resizing model to match image sizes and given batch
[Step 6/11] Configuring input of the model
[ INFO ] Model batch size: 1
[ INFO ] Network inputs:
[ INFO ]     input:0 (node: input) : u8 / [N,H,W,C] / [1,224,224,3]
[ INFO ] Network outputs:
[ INFO ]     MobilenetV3/Predictions/Softmax:0 (node: MobilenetV3/Predictions/Softmax) : f32 / [...] / [1,1001]
[Step 7/11] Loading the model to the device
[ ERROR ] Exception from src/inference/src/cpp/core.cpp:107:
Exception from src/inference/src/dev/plugin.cpp:53:
Check 'jitter != jitters.end()' failed at src/common/snippets/src/lowered/target_machine.cpp:19:
Supported precisions set is not available for Convert operation.

### Issue submission checklist

- [X] I'm reporting an issue. It's not a question.
- [X] I checked the problem with the documentation, FAQ, open issues, Stack Overflow, etc., and have not found a solution.
- [X] There is reproducer code and related data files such as images, videos, models, etc.

[a-sidorova]

@starlitsky2010 Hi! Thank you for the reporting issue!

The error message shows that the exception has been thrown from the Graph Compiler "Snippets".
As temporary solution, I recommend to disable Snippets tokenization. To do that you need:

• Create the config file config.json with the content { "CPU" : {"SNIPPETS_MODE" : "DISABLE"} }
• Add the key -load_config config.json to your command line for benchmark_app:

LD_LIBRARY_PATH=/data/local/tmp ./data/local/tmp/benchmark_app -d CPU -m /data/local/tmp/mobelinet-v3-tf/v3-small_224_1.0_float.xml -hint throughput -load_config config.json

I believe it should help to temporary fix the problem.

@chenhu-wang May I ask you to take a look please at the exception in Snippets? Looks like Convert op on model input was not transformed to Snippets dialect: ConvertTruncation or ConvertSaturation. Thank you in advance!

[chenhu-wang]

@starlitsky2010 ,Could you please provide the CPU info you used and full command line to reproduce. I see you set layout but not full displayed in the pic? Thanks!

[chenhu-wang]

Hi @starlitsky2010, I can not reproduce locally. Anyway I analyzed the code and error message, identified possibility that report the error. Could you please try below PR if fix you issue. If not, could please provide detailed info: the host machine and OS you build binary, target machine you run app, the full command line?
#27948