rivaldotimoty/skin_cancer_classifier

How rivaldotimoty/skin_cancer_classifier works

This repository contains two implemented products: a financial-document RAG analyzer and a nine-class skin-lesion classifier. The analyzer turns uploaded reports into indexed, retrievable context for summaries and financial questions. The classifier supports both model training/evaluation and image inference through a Streamlit app or CLI, with Grad-CAM explanations and test-time augmentation. The graph follows documented and indexed references; several classifier orchestration details and the financial frontend are unsampled, so uncertain links are omitted.

Main entry points: pipeline.py RAG Pipeline · optimize_pipeline.py Training Pipeline · main.py Pipeline Entry · inference.py Image Inference

Components 26

Financial Ingestion

Analyzer Entry LLM/financial_analyzer/main.py
Financial UI LLM/financial_analyzer/frontend
Financial Report LLM/financial_analyzer/data/raw/BUMI - Laporan Keuangan Q1 31 Mar 2025.pdf
Document Loader LLM/financial_analyzer/src/loader.py
Text Chunking LLM/financial_analyzer/src/chunking.py

Financial Retrieval

RAG Pipeline LLM/financial_analyzer/src/pipeline.py
Retriever LLM/financial_analyzer/src/retriever.py
Embedding LLM/financial_analyzer/src/embedding.py
FAISS Index LLM/financial_analyzer/vector_store/faiss_index
Ollama
LLM Client LLM/financial_analyzer/src/llm.py

Classifier Training

Configuration skin_cancer_classifier/config.py
Dataset Prep skin_cancer_classifier/dataset.py
Data Analysis skin_cancer_classifier/eda.py
Training Loop skin_cancer_classifier/train.py
Training Pipeline skin_cancer_classifier/optimize_pipeline.py
Evaluation skin_cancer_classifier/evaluate.py
Pipeline Entry skin_cancer_classifier/main.py
Model Outputs skin_cancer_classifier/outputs

Classifier Inference

Streamlit App skin_cancer_classifier/app.py
Image Inference skin_cancer_classifier/inference.py
Classifier Model skin_cancer_classifier/model.py
Grad-CAM skin_cancer_classifier/gradcam.py

Other components

Analyst
Image User
CLI User

Connections 33

Analyst → Financial UI: uses
Financial UI → Analyzer Entry: submits requests
Analyzer Entry → RAG Pipeline: invokes LLM/financial_analyzer/main.py
Analyzer Entry → LLM Client: uses LLM/financial_analyzer/main.py
RAG Pipeline → Document Loader: loads documents LLM/financial_analyzer/src/pipeline.py
RAG Pipeline → Text Chunking: chunks documents LLM/financial_analyzer/src/pipeline.py
RAG Pipeline → Retriever: builds index LLM/financial_analyzer/src/pipeline.py
RAG Pipeline → Retriever: retrieves context LLM/financial_analyzer/src/pipeline.py
RAG Pipeline → LLM Client: generates answers LLM/financial_analyzer/src/pipeline.py
Retriever → Embedding: embeds queries LLM/financial_analyzer/src/retriever.py
RAG Pipeline → FAISS Index: loads and saves index LLM/financial_analyzer/src/pipeline.py
LLM Client → Ollama: calls API LLM/financial_analyzer/src/llm.py
Image User → Streamlit App: uploads image
CLI User → Image Inference: runs inference skin_cancer_classifier/inference.py
Streamlit App → Image Inference: requests prediction skin_cancer_classifier/app.py
Image Inference → Classifier Model: runs model skin_cancer_classifier/inference.py
Image Inference → Grad-CAM: generates explanation skin_cancer_classifier/inference.py
Image Inference → Configuration: reads settings skin_cancer_classifier/inference.py
Classifier Model → Configuration: reads model settings skin_cancer_classifier/model.py
Dataset Prep → Configuration: reads dataset settings skin_cancer_classifier/dataset.py
Data Analysis → Configuration: reads paths and classes skin_cancer_classifier/eda.py
Training Loop → Configuration: reads training settings skin_cancer_classifier/train.py
Pipeline Entry → Dataset Prep: prepares splits skin_cancer_classifier/main.py
Pipeline Entry → Data Analysis: runs analysis skin_cancer_classifier/main.py
Pipeline Entry → Training Loop: trains model skin_cancer_classifier/main.py
Pipeline Entry → Evaluation: evaluates predictions skin_cancer_classifier/main.py
Pipeline Entry → Grad-CAM: visualizes samples skin_cancer_classifier/main.py
Training Pipeline → Dataset Prep: creates dataloaders skin_cancer_classifier/optimize_pipeline.py
Training Pipeline → Classifier Model: creates model skin_cancer_classifier/optimize_pipeline.py
Training Pipeline → Training Loop: trains model skin_cancer_classifier/optimize_pipeline.py
Training Pipeline → Evaluation: evaluates results skin_cancer_classifier/optimize_pipeline.py
Training Pipeline → Grad-CAM: visualizes results skin_cancer_classifier/optimize_pipeline.py
Training Pipeline → Model Outputs: writes reports and models

More Python diagrams: digitalservicebund/steuerlotse · dohaison/racnn · dteam-top/tsw-cli · dvlshah/tokenx · dwavesystems/penaltymodel

Drawn by GitDiagram on September 30, 2026 from the repository's file tree, README and source files. Read as Markdown ·