test2speech sample run fail with GPU , but pass on CPU

[jerryzhou0624]

test2speech sample run fail with GPU , but pass on CPU

openvino.genai\samples\python\speech_generation>python text2speech.py speecht5_tts --device GPU "This example demonstrates how to use the openvino_genai.Text2SpeechPipeline in Python to convert input text into speech. You can specify a target voice using a speaker embedding vector that captures the desired voice characteristics. Additionally, you can choose the inference device (e.g., CPU, GPU) to control where the model runs."
Traceback (most recent call last):
File "E:\AI-Video-Playground\comprehensive_vlm_eval\openvino.genai\samples\python\speech_generation\text2speech.py", line 49, in
main()
File "E:\AI-Video-Playground\comprehensive_vlm_eval\openvino.genai\samples\python\speech_generation\text2speech.py", line 28, in main
pipe = openvino_genai.Text2SpeechPipeline(args.model_dir, args.device)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: Exception from src\inference\src\cpp\core.cpp:129:
Exception from src\inference\src\dev\plugin.cpp:58:
Check 'false' failed at src\plugins\intel_gpu\src\plugin\program_builder.cpp:163:
[GPU] ProgramBuilder build failed!
Check 'it != format_traits_map.end()' failed at src\plugins\intel_gpu\src\runtime\format.cpp:174:
[GPU] Format description is missing in fmt traits

openvino 2025.3.0.dev20250814 19785
openvino-genai 2025.3.0.0.dev20250814
openvino-telemetry 2025.2.0
openvino-tokenizers 2025.3.0.0.dev20250814

[Wovchena]

I transferred to OpenVINO so GPU team could take a look