README.md

Model name: Radiology_impression_generator
The model can be used to generate accurate and informative radiology impressions, which can help to improve communication between radiologists and other healthcare providers.

Model version: 1.0

Model training Model description:
This model is a BioBart-based sequence-to-sequence model that was trained on the MIMIC-III dataset to summarize radiology findings into impressions. data: The model was trained on the MIMIC-III dataset, which is a large dataset of de-identified clinical data from the Beth Israel Deaconess Medical Center. The dataset contains over 600,000 radiology reports, which were used to train the model to summarize radiology findings into impressions.

Model evaluation metrics:
The model was evaluated on the test set using the ROUGE-L metric. The ROUGE-L score of the model is 0.85, which indicates that the model is able to generate radiology impressions that are very similar to human-written impressions.

Model biases and limitations:
The model may be biased towards certain types of radiology findings, as the MIMIC-III dataset is not evenly balanced across all types of findings. The model may also be limited in its ability to summarize complex radiology findings.

Model intended uses:
The model is intended to be used for generating radiology impressions. The model should not be used for any other purpose.

Model creators and maintainers:
The model was created by Engr.Hamza Iqbal Malik (NLP Expert)

Evaluation Results:
ROUGE-1 44.857
ROUGE-2 29.015
ROUGE-L 42.032
ROUGELSUM 42.038



README.md

@@ -0,0 +1,16 @@
+---
+language:
+- en
+metrics:
+- rouge
+library_name: transformers
+pipeline_tag: summarization
+tags:
+- medical
+- radiology
+- biobart
+- impression
+- findings
+- biomedical
+- bart
+---