onnxruntime error
: INVALID_ARGUMENT : Invalid Feed Input Name:token_type_ids
I am getting the above error of input as invalid argument, for this model. Please let me know how can I fix this!
Hi, can you paste some code to reproduce ? Thank you
from optimum.pipelines import pipeline
classifier = pipeline(task="zero-shot-classification", model="sileod/deberta-v3-base-tasksource-nli", accelerator="ort")
text = "Robot inspection"
labels = ["materials", "robots", "silicon"]
results = classifier(text.lower(),
candidate_labels=labels,
multi_label=False)
But if you change the model to "facebook/bart-large-mnli", the above code works fine. From my research, there is a difference between the onnx model input field and the input field created by the model.
Last version (optimum 1.13.1) says that zero shot NLI is not supported, so this seems to be a optimum problem
But isn't the other model that I provided, "facebook/bart-large-mnli" is also zero shot NLI?
Yes, but the underlying models are different (deberta vs bart), so they might have noticed that support was not good enough then marked it as not supported
Okay got it!
Please let me know if I can find any solution for that!
Thanks for your fast reply and the efficient models!