I'm quite confused on how to implement this. Once converted the ner model to onnx I want to deploy it to a triton server. The issue is that I would like to get the full inference on the triton server but it looks like I can just receive the logits and then locally compute the entities.
Is there a way to "send" the Overcharged TokenClassificationPipeline to the triton server so that the Triton Inference call directly returns the dictionary of entities?
I'm quite confused on how to implement this. Once converted the ner model to onnx I want to deploy it to a triton server. The issue is that I would like to get the full inference on the triton server but it looks like I can just receive the logits and then locally compute the entities. Is there a way to "send" the Overcharged TokenClassificationPipeline to the triton server so that the Triton Inference call directly returns the dictionary of entities?