blaze999

/

Medical-NER

like 164

@DeepMount00 , you can find it on config.json file

i want to use this model but don't know how to get complete code from this.. will you please let me know about this??

@Thelion what do you mean with complete code?

@Thelion , what code are you referring to ?

can you provide how a dataset should look like, so it can be used to train this model?

Thank you for the work you did on this model. I would like to know 1) what does 'I' and 'B' refer to in medical entities? What is the significance?; and 2) are these medical entities mapped to any existing medical coding system such as LOINC or SNOMED CT?

@yburan

• B- (Beginning): Marks the beginning of a named entity. The tag is followed by a label that indicates the type of the entity (e.g., PER for person, LOC for location).
  Example: "Barack" in "Barack Obama" would be tagged as "B-PER".

• I- (Inside): Indicates that a token is inside a named entity and is a continuation of an entity started by a "B-" tag. The tag is followed by the same label as the preceding "B-" or "I-" tag to show it is part of the same entity.
  Example: "Obama" in "Barack Obama" would be tagged as "I-PER".

can I know how did you train this model?

In the description of the model its said that it could detect 41 entities...can you tell me what are those entities?

@NlpSpeech is here:
https://huggingface.co/Clinical-AI-Apollo/Medical-NER/blob/main/config.json

Here you can see the tag

But these are 82 labels ..i want exact 41 entities that are mentioned in model that it can detect, when trained on PubMed dataset.

These are the labels, and there are 41 of them. Please refer to the documentation to understand how the NER model works.

These are the labels, and there are 41 of them. Please refer to the documentation to understand how the NER model works.

Thanks i got it ..

Can i know where can i find the dataset to train this?

no the dataset is private