# 03/06/25 Meeting Notes #24
# This Week's Progress 這週進度
## Datasets
- Change **"Bohemian Style"** to **"Bohemian and Hippie Style"** due to insufficient datasets for Bohemian. `(因為波西米亞風格的數據集不足)`
- Bohemian and Hippie styles share significant similarities, so merging them ensures a more sufficient dataset. `Bohemian 和 Hippie 風格高度相似，合併後能確保數據集更加充足`
- Current total:
| Item Type | Count |
| -------- | -------|
| Bohemian and Hippie | 1092 |
| Casual | 1586 |
| Semi Formal | 1703 |
| Formal | 1527 |
| Streetwear | 1561 |
| Sporty | 1598 |
## Classification Model 2
- Trained the model but is overfitting. `已訓練模型，但出現過擬合問題`
- Trained with:
Full dataset count:

820 files in bohemian_and_hippie 
1560 files in streetwear
1572 files in sporty
1527 files in formal
1586 files in casual
1703 files in semi-formal
total: 8768

Split dataset count:

train dataset: 
    574 files in bohemian_and_hippie
    1092 files in streetwear
    1100 files in sporty
    1068 files in formal
    1110 files in casual
    1192 files in semi-formal
total: 6136

val dataset: 
    123 files in bohemian_and_hippie
    234 files in streetwear
    235 files in sporty
    229 files in formal
    237 files in casual
    255 files in semi-formal
total: 1313

test dataset: 
    123 files in bohemian_and_hippie
    234 files in streetwear
    237 files in sporty
    230 files in formal
    239 files in casual
    256 files in semi-formal
total: 1319

- Currently tuning the model to prevent overfitting. `目前正在調整模型以防止過擬合`

## Generative Model
- The model can now differentiate between front and back images of tops (sweater, polo, T-shirt, jacket) based on the filename. `模型現在可以根據檔案名稱區分上衣（毛衣、Polo 衫、T 恤、夾克）的前後圖片。`
- It assigns different landmark points accordingly. `可對應不同的標記點`
- The model successfully outputs a JSON file with landmark points.`成功輸出包含標記點的 JSON 文件`
- Tested with the main generative code and it works perfectly. `已與主要生成模型測試，運行正常`

## Frontend
- Tried to manually do the frontend design, but some designs were too time-consuming. `嘗試手動設計前端，但部分設計過於耗時`
- Used DhWise to convert Figma design to Flutter; very messy but captured the overall look of the app. `使用 DhWise 將 Figma 設計轉換為 Flutter，整體外觀大致呈現`
- Problem: they blocked downloading and copying the code. Will need to pay for premium to get the code. (1 year 150usd ~= 5000ntd) `程式碼無法下載或複製,需要付費升級為高級會員才能獲取程式碼`
- Figma design still needs significant alterations for the details before development. `Figma 設計仍需大量細節修改，才能進行開發`
- Also currently facing problems with Android Studio. `目前也遇到 Android Studio 的問題`

## Backend
- In progress. Continuation from last week's progress. `進行中。延續上週的進度`

# To Do 需做
- A input and output for generative model.
- Rerun classification model 2 with the newest set of datasets. See the accuracy.
- Try to get boho hippie dataset count to 1.5k.
- figma template with code: take components that we would like to use then can add more details after
# Next Weeks's Meeting 下週會議

---
Previous:
Next:
Full Content List [here](https://hackmd.io/@emps-113up/full-list)