clips / fewshot-biomedical-names Goto Github PK

I was trying to run the code with given below arguments on colab---->
!python main.py --fasttext pub2fast.bin --outfile out1.txt --gpu 1
but the error is coming given below----->
Initializing encoder...
Traceback (most recent call last):
File "main.py", line 41, in
gpu_index=args.gpu,
File "/content/drive/MyDrive/fewshot-biomedical-names/encoder.py", line 22, in init
super().init(**kwargs)
File "/content/drive/MyDrive/fewshot-biomedical-names/encoder_base.py", line 138, in init
super().init(**kwargs)
TypeError: init() got an unexpected keyword argument 'triplet_margin'

while trying to fix this ,i commented out the line 35 in main.py but then error was it was not able to create object for fast text

Can you please provide instruction how to train the model with SNOMED training data and how to do continual learning i.e. how to train from SNOMED-CD to ICD-10 and vice versa.
and how to use pretrained name representation that got from BioBERT model and BNE Model as it is mentioned in the paper and in repository by only using fast text ,training has been done.

Using ICD-10
I used FastText embedding and 15 shots and I didn't got exactly as same result as you mentioned in the paper as you can see in the screenshot

5 shots results for UMNSRS and EHR-RelB are better as you can in following screenshot

And you Didn't mentioned in the paper how to use BNE and BioBert pretrainded Embedding and how to do continual
learning from SNOMED-CT to ICD-10
I want to know the instructions to see the results that is mentioned in your pape

Hello, Pieter

Can you please provide me access to the SNOMED training data set used in this model, as I was not able to find it anywhere, If available online, please provide the path.

you can also mail me on - [email protected]

Thanks,