Describe the bug
|
std::shared_ptr<Tensor> input_ids = tokenizer.EncodeBatch(payload.prompts); |
doesn't check
payload.prompt. It should check that
payload.prompt or
payload.prompts is set and use the correct one.
To Reproduce
Run with a single prompt set.
Expected behavior
No divide by zero error.
Describe the bug
onnxruntime-genai/src/models/whisper_processor.cpp
Line 41 in 053d7d7
payload.prompt. It should check thatpayload.promptorpayload.promptsis set and use the correct one.To Reproduce
Run with a single prompt set.
Expected behavior
No divide by zero error.