ryan0114/clin-bhc-summ

How ryan0114/clin-bhc-summ works

This repository is a research pipeline for hospital-course summarization: prepare MIMIC-IV-BHC notes, fine-tune or prompt language models, generate summaries, and score outputs. Its main runtime paths are the PEFT training and inference scripts and the discrete-prompting runner, with shared argument, configuration, and data-processing code. The README also documents GPT and Llama notebook workflows. Those notebook internals and several source modules were not sampled, so their connections are shown only where the README documents them.

Main entry points: train_peft.py PEFT training · calc_metrics.py Metric calculation · process.py Data and prompts · run_peft.py PEFT inference

Components 19

Data preparation

MIMIC-IV-BHC source
Dataset preprocessing utils/mimic_iv_bhc_preprocessing.py
Token filtering utils/filter_data_by_token.ipynb

Shared runtime

Models and cases src/constants.py
Arguments and paths src/parser.py
Data and prompts src/process.py

Model adaptation

PEFT training src/train_peft.py
Causal LM training src/run_clm.py
Adapter merging src/merge_adapter_weights.py
QLoRA notebook training/llama2_peft.ipynb
Model training script training/train_peft.sh

Summarization and evaluation

PEFT inference src/run_peft.py
Discrete prompting src/run_discrete.py
GPT prompting inference/gpt_inference.ipynb
Llama inference inference/llama2_inference.ipynb
Summary outputs
Metric calculation src/calc_metrics.py
Metric results

Other components

Researcher

Connections 27

Researcher → Dataset preprocessing: prepares dataset README.md
MIMIC-IV-BHC source → Dataset preprocessing: reads notes utils/mimic_iv_bhc_preprocessing.py
Dataset preprocessing → Token filtering
Researcher → QLoRA notebook: fine-tunes models README.md
Researcher → Model training script: fine-tunes models README.md
Model training script → PEFT training: runs training
Researcher → Causal LM training: trains model
PEFT training → Arguments and paths: parses arguments src/train_peft.py
PEFT training → Models and cases: reads settings src/train_peft.py
PEFT training → Data and prompts: loads and batches data src/train_peft.py
PEFT training → Adapter merging
Researcher → Discrete prompting: runs prompting README.md
Researcher → PEFT inference: runs tuned models README.md
Researcher → GPT prompting: runs prompting README.md
Researcher → Llama inference: runs inference README.md
Discrete prompting → Arguments and paths: parses arguments src/run_discrete.py
Discrete prompting → Models and cases: reads model settings src/run_discrete.py
Discrete prompting → Data and prompts: loads and saves data src/run_discrete.py
PEFT inference → Arguments and paths: parses arguments src/run_peft.py
PEFT inference → Models and cases: loads model settings src/run_peft.py
PEFT inference → Data and prompts: loads and saves data src/run_peft.py
Data and prompts → Summary outputs: writes summaries src/process.py
Researcher → Metric calculation: scores outputs README.md
Metric calculation → Arguments and paths: parses arguments src/calc_metrics.py
Metric calculation → Data and prompts: loads summaries src/calc_metrics.py
Metric calculation → Summary outputs: reads predictions src/calc_metrics.py
Metric calculation → Metric results: writes metrics src/calc_metrics.py

Drawn by GitDiagram on November 15, 2026 from the repository's file tree, README and source files. Read as Markdown ·