# progress

**Repository Path**: Gurid/progress

## Basic Information

- **Project Name**: progress
- **Description**: 项目进展
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-06-09
- **Last Updated**: 2025-06-09

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# 1 完成的工作

## 环境配置

- Python              3.10
- PyTorch             2.1.1
- RDKit                 2022.09.5
- scikit-learn        1.3.0

## 源码

[OliXio/CMSSP · HF Mirror](https://hf-mirror.com/OliXio/CMSSP)

[OliXio/CMSSP at main](https://hf-mirror.com/OliXio/CMSSP/tree/main)

## 模型调试

### 训练

- 指令

  ```python
  python train.py cfg.json
  ```

- 输出

  ![](./train-results.png)

### 预测

- 指令

  ```python
  python predict.py 'F:\Projects\Python\CMSSP\data\CASMI2016\CASMI2016_Cat2and3_Challenge_1-neg.json' 'F:\Projects\Python\CMSSP\data\train-001\model-neg-v1.0.pth'
  ```

- 输出

  ![](./predict-results.png)

# 2 存在的问题

## 预测过程中的数据存在不兼容现象

如上图所示，部分分子预测时报错。

```shell
O=C(OCc1ccccc1[N+]([O-])=O)c2c(c(C(=O)OC(C)C)c(n2)C)C 'NoneType' object has no attribute 'GetNumAtoms'
```

## 实际运行结果较论文而言较差

```shell
['Top1: 59, top3: 69, top10: 76, top50: 79 of 81']
```