zkcplk/federe_ogrenme_ile_malware_tespiti

How zkcplk/federe_ogrenme_ile_malware_tespiti works

This is a federated-learning malware detection and classification application: clients train local FNN or LSTM models on malware datasets, exchange encrypted weights with a coordinating server, and receive a FedAvg global model. The main workflow runs from client polling and local training through authenticated round-based transfer, server aggregation, evaluation, logging, and either another round or completion. The sampled source supports these runtime links; dataset contents and some configuration details are represented only by the tree.

Main entry points: server_functions.py FedAvg round · client_functions.py Round loop · server_main.py Server listener · server_main.json Server config

Components 19

Client runtime

Client launcher Fed_Client/client_main.py
Round loop Fed_Client/client_functions.py
Malware datasets Fed_Client/datasets
Client protocol Fed_Client/client_functions.py

Server runtime

Server listener Fed_Server/server_main.py
Client handler Fed_Server/server_functions.py
FedAvg round Fed_Server/server_functions.py
Model evaluation Fed_Server/server_functions.py
Metrics log Fed_Server/server_functions.py
Evaluation datasets Fed_Server/datasets

Model learning

Client architectures Fed_Client/models/model.py
Server architectures Fed_Server/models/model.py

State and security

Client config Fed_Client/client_main.json
Server config Fed_Server/server_main.json
Client crypto and state Fed_Client/other.py
Server crypto and state Fed_Server/other.py
API credentials Fed_Server/apikeys/user.json

Other components

Client operator
Client participant

Connections 28

Client operator → Client launcher: starts
Client launcher → Client config: loads and updates
Client launcher → Client crypto and state: clears model files
Client launcher → Round loop: starts rounds
Round loop → Malware datasets: reads training data
Round loop → Client config: reads settings
Round loop → Client architectures: builds model
Round loop → Client protocol: checks and transfers
Client protocol → Server listener: TLS commands and weights
Client protocol → Client crypto and state: decrypts global weights
Round loop → Client crypto and state: encrypts local weights
Server listener → Server config: loads and updates
Server listener → Server architectures: builds initial model
Server listener → Server crypto and state: saves encrypted model
Server listener → Client handler: dispatches connection
Client handler → API credentials: checks credentials
Client handler → Server config: checks round
Client handler → Server crypto and state: reads and writes weights
Server listener → FedAvg round: updates after request
FedAvg round → Server config: reads round and submissions
FedAvg round → Server crypto and state: decrypts and encrypts weights
FedAvg round → Server architectures: loads and averages weights
FedAvg round → Model evaluation: tests global model
Model evaluation → Evaluation datasets: reads evaluation data
Model evaluation → Server config: reads evaluation settings
FedAvg round → Metrics log: records metrics
FedAvg round → Server config: advances round
Client participant → Client protocol: participates

More Python diagrams: zangobot/maltorch · zeekeey-jpeg/leroy-hq · zeeoale/promptcreatornode · zenpaiang/wos-bot · zxins/employmentcms

Drawn by GitDiagram on September 27, 2026 from the repository's file tree, README and source files. Read as Markdown ·