# 01/23/25 Meeting Notes #19
# This Week's Progress 這週進度
## Datasets
### Classification Model #1
| Item Type | Count |
| --- | --- |
| 01Dress | 3761 |
| 02Shirt | 4017 |
| 03Tshirt | 4377 |
| 04Hoodie | 3343 |
| 05Sweater | 3052 |
| 06PoloShirt | 3032 |
| 07Pants | 4899 |
| 08Shorts | 3849 |
| 09Skirt | 3022 |
| 10Flats | 3871 |
| 11Heels | 3010 |
| 12Shoes | 3165 |
| 13Sneakers | 3816 |
- Will have slight changes due to misclassification of some images.
` 由于某些图像的错误分类，将会有轻微的变化。`
- Will continue model2 dataset after sorting out model1 dataset.
` 整理完 model1 数据集后，将继续 model2 数据集。`

## Classification Model 1
### Hierarchy Structure `层次结构`
### Comparison between with and without using hierarchy structure `使用和不使用层次结构的比较`

### Loss and accuracy graph without hierarchy structure `无层次结构的损失和准确率图`

| Model | Confusion Matrix | Loss & Accuracy Graph |
|:---------------:|:-----------------------------------------------------------------------:|:------------------------------------------------------------------------:|
| big_category.pt |  |  |
| top.pt |  |  |
| bottom.pt |  |  |
| footwear.pt |  |  |

| | without hierarchy | big_category | top | bottom | footwear |
|:-------------------:|:-----------------:|:------------:|:------:|:------:|:--------:|
| Training accuracy | 88.02% | 97.21% | 88.08% | 93.74% | 91.24% |
| Validation accuracy | 88.65% | 97.51% | 87.27% | 93.88% | 91.91% |
| Testing accuracy | 88.19% | 97.18% | 86.30% | 93.83% | 91.07% |

### Training, Validating, and Training dataset:
| | Train (70%) | Val (15%) | Test (15%) |
|:----------- |:-----------:|:---------:|:----------:|
| 01Dress | 2632 | 564 | 565 |
| 02Shirt | 2811 | 602 | 604 |
| 03Tshirt | 3063 | 656 | 658 |
| 04Hoodie | 2340 | 501 | 502 |
| 05Sweater | 2136 | 457 | 459 |
| 06PoloShirt | 2122 | 454 | 456 |
| 07Pants | 3429 | 734 | 736 |
| 08Shorts | 2694 | 577 | 578 |
| 09Skirt | 2115 | 453 | 454 |
| 10Flats | 2709 | 580 | 582 |
| 11Heels | 2107 | 451 | 452 |
| 12Shoes | 2215 | 474 | 476 |
| 13Sneakers | 2671 | 572 | 573 |
| total | 3304 | 7075 | 7095 |

## Classification Model 2
|Confusion Matrix|Loss & Acc Graph|
|---|---|
|||

## NSTC Proposal
1. Question regarding number paging.
(四)頁碼編寫:請對摘要及目錄部分用羅馬字 I 、II、 III......標在每頁下方中央;報告內容至附錄部分請以阿拉伯數字 1.2.3.......順序標在每頁下方中央。
2. Second update on proposal due?

# To Do 需做
- Modify proposal
- Submit proposal 5feb, meeting with prof 6feb
- Change location of misplaced datasets in model1
- VISUELLE dataset divide top

# Next Weeks's Meeting 下週會議
- Finish sorting dataset for model1
- Continue dataset for model2
---
Previous:
Next:
Full Content List [here](https://hackmd.io/@emps-113up/full-list)