gachon-CCLab · z8086486 · May 10, 2026 · May 10, 2026
diff --git a/docs/FedOps-Aggregation-method/FedOps-Aggregation-method.md b/docs/FedOps-Aggregation-method/FedOps-Aggregation-method.md
@@ -1,7 +1,7 @@
 ---
 layout: default
 title: FedOps Aggregation method
-nav_order: 7
+nav_order: 8
 has_children: true
 permalink: docs/FedOps-Aggregation-method
 ---

diff --git a/docs/FedOps-Fitbit-Health-Pipeline/FedOps-Fitbit-Health-Pipeline.md b/docs/FedOps-Fitbit-Health-Pipeline/FedOps-Fitbit-Health-Pipeline.md
@@ -1,7 +1,7 @@
 ---
 layout: default
 title: FedOps Fitbit Health Pipeline
-nav_order: 10
+nav_order: 11
 has_children: true
 permalink: docs/FedOps-Fitbit-Health-Pipeline
 ---

diff --git a/docs/FedOps-Hyperparameter-Optimize/FedOps-Hyperparameter-Optimize.md b/docs/FedOps-Hyperparameter-Optimize/FedOps-Hyperparameter-Optimize.md
@@ -1,7 +1,7 @@
 ---
 layout: default
 title: FedOps Clustering Tuning
-nav_order: 8
+nav_order: 9
 has_children: true
 permalink: docs/FedOps-Clustering-Tuning
 ---

diff --git a/docs/FedOps-VLM/FedOps-VLM.md b/docs/FedOps-VLM/FedOps-VLM.md
@@ -0,0 +1,8 @@
+---
+layout: default
+title: FedOps VLM
+nav_order: 7
+has_children: true
+permalink: docs/FedOpsVLM
+---
+# FedOps VLM
diff --git a/docs/FedOps-VLM/How-to-use-FedOps-VLM-Framework.md b/docs/FedOps-VLM/How-to-use-FedOps-VLM-Framework.md
@@ -0,0 +1,147 @@
+---
+layout: default
+title: How to use FedOps VLM Framework?
+parent: FedOps VLM
+nav_order: 2
+---
+# 📝 FedOps VLM Guide
+
+<br />
+
+## 1. Download the Codebase Setup
+
+{: .highlight }
+💡 Go to https://gachon-cclab.github.io/fedopsMultimodal/ 
+Select Your model, Dataset, FL Configs, aggregation strategy and  Click Download Codebase.
+
+select as below: 
+
+- Model: ***OneVision 0.5B***
+
+- Dataset: ***VQA-RAD***
+
+- Aggregation Strategy: ***FedMAP***
+
+<br />
+
+![image.png](../../img/FedOps-VLM/image1.png)
+
+ and then click:
+
+![image.png](../../img/FedOps-VLM/image2.png)
+
+now you will automatically download a zip folder with whole codebase files.
+Unzip it.
+
+## **2. Now, the codebase include following files.**
+
+`conf/config.yaml` - This is where configurations reside (including the task ID)
+
+`client_main.py` - This is the main file the client side logics reside
+
+`client_manager_main.py`  - This is the connection check file of client side
+
+`data_preparation.py` - Data loading, preprocessing logics reside here
+
+`models.py`, `generate_paramshape.py` - file, which will generate model’s parameter shape
+
+`parameter_names.json` - this is needed for the client side to distinguish parameter types
+
+`parameter_shapes.json` - Generated parameter shape of the model
+
+`setup.sh` - The file, which sets up the environment including dependencies of client side
+
+
+## **3. Open the folder in your IDE (VSCode)**
+
+Open your bash terminal,
+
+then run
+
+```
+conda create -n *fedops_multimodal_env python=3.12.7
+
+conda activate fedops_multimodal_env
+
+chmod +x setup.sh
+
+bash setup.sh
+```
+
+![image.png](../../img/FedOps-VLM/image3.png)
+
+and then run `python client_main.py`
+
+![image.png](../../img/FedOps-VLM/image4.png)
+
+Now open  a new terminal in the same folder path, and run 
+
+`python client_manager_main.py`
+
+
+## **4. Server Side Code Management**
+
+Visit: https://ccl.gachon.ac.kr/fedops/task
+
+signin and create a task named task id: ***”fedopstune2”***. and do the followings
+
+![image.png](../../img/FedOps-VLM/image5.png)
+
+Leave the rest as it is.
+
+Go to Server Management
+Press Below shown green button to create a Server pod in Fedops K8 Environment
+
+![image.png](../../img/FedOps-VLM/image6.png)
+
+Scale the Resources(memory): please do enter Scale Resources button after entering 10Gi as memory.
+
+![image.png](../../img/FedOps-VLM/image7.png)
+
+Now check the server status through below picture. keep refreshing the status and wait until you see similar to this.
+
+![image.png](../../img/FedOps-VLM/image8.png)
+
+Wait for around 6-7 mins until you see yellow colored log in Server log section below in same server management tab (shown in picture). Keep refreshing the logs to see real time log.
+
+![image.png](../../img/FedOps-VLM/image9.png)
+
+**Then, start editing server side files**
+
+{: .highlight }
+💡 You can check server side files structure by typing file browser path as below and clicking browse button
+
+![image.png](../../img/FedOps-VLM/image10.png)
+
+{: .highlight }
+💡 Paste the **file names with path** in **yellow** colored text **space** → **Copy & paste** the file content **from downloaded folder** content **into code space** → **Press save file** button. Likewise this, **do for each and every file path** mentioned below.
+
+filenames with path as follows:
+
+`/app/code/conf/config.yaml`
+
+`/app/code/parameter_shapes.json` -Just copy paste this content ,so it will automatically gets created
+
+`/app/code/parameter_names.json` -Just copy paste this content ,so it will automatically gets created
+
+`/app/code/server_main.py`
+
+
+![image.png](../../img/FedOps-VLM/image11.png)
+
+Install below dependency as below by pressing execute:
+
+![image.png](../../img/FedOps-VLM/image12.png)
+
+## **5. Start the FL Server**
+
+**Start the actual server:**
+
+{: .highlight }
+💡 Click Start FL Server button to prepare the command to run the FL server. you can then see log below says FL server created.
+
+{: .highlight }
+💡 you must type **/app/data/logs/serverlog.txt** in File content field and press load button to see real time server side logs, and monitor server side FL global model training process.
+
+{: .highlight }
+💡 You can stop the server by “**stop FL server**” button ,if you want to stop the server in middle.
diff --git a/docs/FedOps-VLM/What-is-FedOps-VLM-Framework.md b/docs/FedOps-VLM/What-is-FedOps-VLM-Framework.md
@@ -0,0 +1,59 @@
+---
+layout: default
+title: What is FedOps VLM Framework?
+nav_order: 1
+parent: FedOps VLM
+---
+# ✅ What is FedOps VLM Framework??
+<br />
+
+**FedOps VLM Framework** is an open, cloud-native federated learning (FL) platform purpose-built for **Vision-Language Models (VLMs)**. It bridges the gap between large-scale multimodal AI and privacy-preserving distributed training — enabling researchers and practitioners to fine-tune powerful VLMs across decentralized edge devices without ever centralizing raw data.
+
+Built on top of the **FedOps** ecosystem (PyPI `fedops`), the framework combines:
+
+- **Flower** — for robust FL communication and client–server orchestration
+- **PEFT / LoRA / QLoRA** — for parameter-efficient fine-tuning of large VLMs on resource-constrained devices
+- **FedMAP (Parameter-Aware Federated Aggregation)** — a novel component-aware weighted aggregation theorem that achieves tighter convergence bounds than flat averaging when LoRA and projection parameters have different learning dynamics.
+
+The result is a full-stack FL solution that works from a **web portal download** all the way to a **Kubernetes-managed cloud FL server** — reducing the barrier to federated VLM research to just a few configuration steps.
+
+---
+
+## Core Concepts
+
+### Federated Learning for VLMs
+
+Traditional VLM fine-tuning requires pooling sensitive data into a central server. FedOps VLM Framework flips this: **the data never leaves the device**. Only lightweight parameter updates are sent to the server for aggregation.
+
+### Plugin-Based Architecture
+
+**Plugin-based** — models, datasets, and aggregation strategies are swappable. Pick your combination from the web portal, download, and train. No boilerplate wiring required.
+
+### Supported Models
+
+| Model | Size | Quantization |
+| --- | --- | --- |
+| OneVision | 0.5B | QLoRA (4-bit) |
+| PhiVA | 3.8B | QLoRA (4-bit) |
+
+> Custom models via `models.py`
+
+### Supported Datasets
+
+| Dataset | Domain | Size |
+| --- | --- | --- |
+| VQA-RAD | Medical radiology | 1793 train, 451 test |
+| VQAv2 | General vision | Large-scale |
+| PathVQA | Pathology | ~19.7K train |
+| SLAKE | Multi-organ medical | ~4.9K train, 1K test |
+| ChartQA | Chart reasoning | ~18K train |
+| DocVQA | Document understanding | ~10K train |
+
+> Custom datasets via `data_preparation.py`
+
+### Aggregation Strategies
+
+| Strategy | Description |
+| --- | --- |
+| FedAvg | Standard weighted average |
+| FedMAP | Component-aware QP — separate λ for LoRA vs. projection layers |
diff --git a/docs/FedOps-XAI/FedOps-XAI.md b/docs/FedOps-XAI/FedOps-XAI.md
@@ -1,7 +1,7 @@
 ---
 layout: default
 title: FedOps XAI
-nav_order: 9
+nav_order: 10
 has_children: true
 permalink: docs/FedOps-XAI
 ---

diff --git a/docs/FlowerHub with GFedOps/FedOps-Launcher/How-to-use-FedOps-Launcher.md b/docs/FlowerHub with GFedOps/FedOps-Launcher/How-to-use-FedOps-Launcher.md
@@ -18,6 +18,8 @@ This guide provides step-by-step instructions for using FedOps Launcher, a FedOp
 
 ### 1. Install Docker 
 
+FedOps Launcher runs as a Docker container. Docker is required to provide an isolated execution environment for the launcher, manage the workspace volume, and run the FL client consistently across different operating systems.
+
 [Docker Download](https://www.docker.com/get-started/)
 
 ![image.png](../../../img/FedOps-Launcher/image(1).png)
@@ -36,7 +38,7 @@ This guide provides step-by-step instructions for using FedOps Launcher, a FedOp
 docker rm -f fedops-launcher 2>/dev/null; docker run -d --name fedops-launcher -p 5600:5600 -e WORKSPACE_DIR=/workspace -v "$HOME/fedops-workspace:/workspace" joseongjin311/fedops-launcher:v1.0.0 && echo "Now Running On: http://127.0.0.1:5600"
 ```
 
-In my case, I use VSCode in a Windows WSL environment.
+The examples in this guide are performed in VSCode within a Linux environment provided by Windows WSL.
 
 Just copy the command and run it in your CLI environment.
 
@@ -118,7 +120,7 @@ If you are logged in, you can choose a Task what you created on FedOps website.
 
 ![image.png](../../../img/FedOps-Launcher/image(12).png)
 
-**"ssj" is my Task name, so you should select your own Task**
+**In this example, "ssj" is the Task name. Select the Task created for your own environment.**
 
 Next, select the FL project to which you want to apply the Task.
 
@@ -134,6 +136,14 @@ All settings are complete. Just click the **Run** button.
 
 ![image.png](../../../img/FedOps-Launcher/image(14).png)
 
+### 7. Stop FedOps Launcher
+
+When you finish using FedOps Launcher, stop and remove the running container with the following command:
+
+```bash
+docker rm -f fedops-launcher
+```
+
 **Thank you.**
 
-**You have successfully started the FedOps Launcher.**
+**You have successfully learned how to use the FedOps Launcher.**
diff --git a/docs/FlowerHub with GFedOps/FedOps-Launcher/What-is-FedOps-Launcher.md b/docs/FlowerHub with GFedOps/FedOps-Launcher/What-is-FedOps-Launcher.md
@@ -21,4 +21,4 @@ Key points:
 - FedOps Launcher provides a Web UI for preparing and running GFedOps Apps.
 - Users can import a GFedOps App from FlowerHub and create a local client project.
 - It supports code editing, dependency installation, FedOps task selection, and client execution in one workspace.
-- Users can join a FedOps federated learning task as a real client more easily.
+- Users can join a FedOps federated learning task as a real client more easily.
diff --git a/docs/FlowerHub with GFedOps/FlowerHub-with-GFedOps.md b/docs/FlowerHub with GFedOps/FlowerHub-with-GFedOps.md
@@ -1,8 +1,8 @@
 ---
 layout: default
 title: FlowerHub with GFedOps 
-nav_order: 11
+nav_order: 12
 has_children: true
-permalink: docs/FedOps-Launcher
+permalink: docs/FlowerHub-with-GFedOps
 ---
 # FlowerHub with GFedOps
diff --git a/docs/img/FedOps-Launcher/image(10).png b/docs/img/FedOps-Launcher/image(10).png
diff --git a/docs/img/FedOps-Launcher/image(11).png b/docs/img/FedOps-Launcher/image(11).png
diff --git a/docs/img/FedOps-Launcher/image(13).png b/docs/img/FedOps-Launcher/image(13).png
diff --git a/docs/img/FedOps-Launcher/image(14).png b/docs/img/FedOps-Launcher/image(14).png
diff --git a/docs/img/FedOps-Launcher/image(9).png b/docs/img/FedOps-Launcher/image(9).png
diff --git a/docs/img/FedOps-VLM/image1.png b/docs/img/FedOps-VLM/image1.png
diff --git a/docs/img/FedOps-VLM/image10.png b/docs/img/FedOps-VLM/image10.png
diff --git a/docs/img/FedOps-VLM/image11.png b/docs/img/FedOps-VLM/image11.png
diff --git a/docs/img/FedOps-VLM/image12.png b/docs/img/FedOps-VLM/image12.png
diff --git a/docs/img/FedOps-VLM/image2.png b/docs/img/FedOps-VLM/image2.png
diff --git a/docs/img/FedOps-VLM/image3.png b/docs/img/FedOps-VLM/image3.png
diff --git a/docs/img/FedOps-VLM/image4.png b/docs/img/FedOps-VLM/image4.png
diff --git a/docs/img/FedOps-VLM/image5.png b/docs/img/FedOps-VLM/image5.png
diff --git a/docs/img/FedOps-VLM/image6.png b/docs/img/FedOps-VLM/image6.png
diff --git a/docs/img/FedOps-VLM/image7.png b/docs/img/FedOps-VLM/image7.png
diff --git a/docs/img/FedOps-VLM/image8.png b/docs/img/FedOps-VLM/image8.png
diff --git a/docs/img/FedOps-VLM/image9.png b/docs/img/FedOps-VLM/image9.png
diff --git a/docs/industry-pack.md b/docs/industry-pack.md
@@ -1,7 +1,7 @@
 ---
 layout: default
 title: FedOps Industry Pack
-nav_order: 12
+nav_order: 13
 ---
 
 # FedOps Industry Pack

diff --git a/docs/package.md b/docs/package.md
@@ -1,7 +1,7 @@
 ---
 layout: default
 title: FedOps Package
-nav_order: 13
+nav_order: 14
 # h2부터 h5까지 제목을 표시합니다
 toc_min_heading_level: 2
 toc_max_heading_level: 5