Hang Dong

exactly same error here Environment: tensorflow(1.6.0), Python 3.6.2.

This is because of the fact that softmax is shift-invariant by a constant offset in the input. > softmax is invariant under translation by the same value in each coordinate...

I had the same question as the above. To concatenate the hidden states of last 4 layers, the right way in python for ```-pooling_layer``` is ``` args = get_args_parser().parse_args(['-model_dir', model_path,...

> you need > if **name** == '**main**': if \_\_name\_\_ == '\_\_main\_\_':

I have almost edited it to be runnable (see my fork and changes of [```models.py```](https://github.com/acadTags/caml-mimic/blob/master/learn/models.py)). But the results are not as good as the one on the paper. Also, the...

The study used both DIAGNOSES_ICD and PROCEDURES_ICD. Maybe 37.22 and 96.72 are from the procedure codes?

If you check the generated `notes_labeled.csv`, you will find that there is an empty row between every two records. It is the empty rows that cause the `row[1]` to have...

Check [```evaluate_model.sh```](https://github.com/jamesmullenbach/caml-mimic/blob/master/predictions/CAML_mimic3_full/evaluate_model.sh) for exact reproduction of results of CAML. The model parameters have been published, see [```model.pth```](https://github.com/jamesmullenbach/caml-mimic/blob/master/predictions/CAML_mimic3_full/model.pth) There is also one script in the DR-CAML folder. Best wishes, Acadtags

> Only active concepts will be translated to OWL classes. I expect you will find that there are 332,210 concepts with an active value of 1 in the sct2_Concept_Full_US1000124_20170301.txt file....

> The remaining discrepancy is probably due to attributes. In SNOMED CT attributes (relationship types) are represented as concepts in the concept file, however in OWL these do not become...