ViperEkura
/
AstrAI
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

docs: 修改 README

This commit is contained in:
ViperEkura 2026-03-04 20:51:09 +08:00
parent c7d0448822
commit 5a8c442315
1 changed files with 35 additions and 82 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

117
README.md
View File

@ -12,39 +12,39 @@
<h2 id="english">English Version</h2> <h2 id="english">English Version</h2>
This is a Chinese-English bilingual Transformer model supporting both languages. It contains model configurations and training workflows, completing training by loading parameters defined in `param_path/config.json`. The training script `train.py` parses command-line arguments, including dataset root directory, number of training epochs, batch size, checkpoint interval, and checkpoint directory. A training and inference framework for autoregressive Transformer language models.
**Model Download Options (Choose One):** **Model Download Options (choose one):**
1. Visit [HuggingFace](https://huggingface.co/ViperEk/KHAOSZ) to access **Files and versions** 1. Visit [HuggingFace](https://huggingface.co/ViperEk/KHAOSZ) and check **Files and versions**
2. Run `scripts/download.py` to download parameters 2. Run `scripts/download.py` to download model parameters
**Demo Video:** [bilibili](https://www.bilibili.com/video/BV1z5RPYHEkd) **Demo Video:** [bilibili](https://www.bilibili.com/video/BV1z5RPYHEkd)
Training dataset sources are listed in the **Model Card** section of the HuggingFace download link. For training data sources, please refer to the **Model Card** section on the HuggingFace download page.
**License:** Code follows Apache-2.0 protocol. Please credit the source code when used. **License:** The code follows the GPL-3.0 license. Please provide attribution when using it.
- **📊 Device Selection:** Uses CUDA for training by default
- **🌐 Performance Optimization:** Enable `dtype=torch.bfloat16` to accelerate training and reduce memory usage. Ensure your hardware supports this feature
- **🤖 Language Support:** The model supports training in Chinese and English. Since the BBPE tokenizer hasn't been trained on multilingual text, OOV (Out-of-Vocabulary) issues are minimal for Chinese and English, but may exist for other languages
- **📊 Device Selection:** Code defaults to CUDA training
- **🌐 Performance Optimization:** `dtype=torch.bfloat16` is enabled to accelerate training and reduce memory usage. Ensure hardware supports this feature.
- **🤖 Language Support:** Model supports Chinese and English training. The BBPE tokenizer was trained without multilingual text, so OOV (out-of-vocabulary) issues are minimized for these languages but may exist for others.
### 📌 Training Guide ### 📌 Training Guide
To train this Transformer model, follow these steps: To train this Transformer model, follow these steps:
**(1). Prepare Dataset:** **(1). Prepare the Dataset:**
Place datasets in the designated root directory. Files should be text documents in Chinese, English, or mixed. Format should align with model input requirements - preferably pre-tokenized token_ids stored as `torch.Tensor` (using `torch.Tensor` saves memory compared to Python lists, which default to 64-bit precision). Place the dataset in the specified root directory. This system uses the BBPE tokenizer for tokenization and requires training with pre-tokenized segments (stored as *.h5 format files).
**(2). Install Dependencies:** **(2). Install Dependencies:**
```bash ```bash
pip install -r requirements.txt pip install -e .
pip install .
``` ```
**(3). Run Training Script:** **(3). Run the Training Script:**
```bash ```bash
python train.py \ python train.py \
@ -55,29 +55,29 @@ python train.py \
--batch_size=8 \ --batch_size=8 \
--max_lr=2e-4 \ --max_lr=2e-4 \
--checkpoint_interval=10000 \ --checkpoint_interval=10000 \
--checkpoint_dir=checkpoints --checkpoint_dir=checkpoints
``` ```
**Parameters Explanation:** **Parameter Explanation:**
- `--train_type`: Training type (seq, sft, dpo) - `--train_type`: Training type (seq, sft, dpo)
- `--data_root_path`: Root directory of the dataset - `--data_root_path`: Dataset root directory
- `--param_path`: Path to the model training parameters - `--param_path`: Path to model training parameters
- `--n_epoch`: Total number of training epochs - `--n_epoch`: Total number of training epochs
- `--batch_size`: Batch size - `--batch_size`: Batch size
- `--accumulation_steps`: Number of batches per training step - `--accumulation_steps`: Number of batches per training step
- `--warmup_steps`: Number of warmup steps - `--warmup_steps`: Warmup steps
- `--max_lr`: Maximum learning rate (using warmup + cosine decay) - `--max_lr`: Maximum learning rate (using warmup + cosine decay)
- `--checkpoint_interval`: Checkpoint saving interval - `--checkpoint_interval`: Checkpoint saving interval
- `--checkpoint_dir`: Directory to save checkpoints - `--checkpoint_dir`: Checkpoint saving directory
- `--resume_dir`: Resume training from the specified path - `--resume_dir`: Resume training from specified path
Training logs will be saved in `train_log.txt`. Checkpoints will be saved in the specified directory for resuming training or evaluation.
### 👉 Usage Guide ### 👉 Usage Guide
**(1). Chatting with the Model:** **(1). Chat with the Model:**
Open `chat.py` or use streaming/non-streaming interfaces: Open `chat.py` or use the streaming/non-streaming interfaces:
**Streaming Output:** **Streaming Output:**
```python ```python
@ -129,7 +129,7 @@ while True:
print(response) print(response)
``` ```
**(2) Retrieval-Augmented Generation (RAG):** **(2). Retrieval-Augmented Generation (RAG):**
```python ```python
import torch import torch
@ -148,29 +148,8 @@ retrieved_content = model.retrieve_generate(
print(retrieved_content) print(retrieved_content)
``` ```
### 📌 Model Specifications
This model is based on a 24-layer Transformer with parameters defined in `config.json`, totaling approximately 1.0 billion (1.0B) parameters.
**Key Design Choices:**
- Weight tying between embedding and final linear layers (standard for small models to save parameters)
- Embedding layer optimization: Without weight tying, a 10,000-word vocabulary would consume ~102M parameters (0.1B)
**Limitations:**
- May struggle with complex language phenomena due to smaller parameter size
- Prone to overfitting on specialized datasets
- Limited multilingual capabilities
**Advantages:**
- Runs efficiently on lower-spec hardware
- Shorter training time compared to larger models
**Training Pipeline:**
The model has completed pre-training + SFT (Supervised Fine-Tuning) + DPO (Direct Preference Optimization) workflows. All corresponding training code is included in the repository.
<h2 id="chinese">中文版本</h2> <h2 id="chinese">中文版本</h2>
这是一个支持中英文双语的 Transformer 模型，能够处理两种语言。模型包含配置文件和训练流程，通过加载 `param_path/config.json` 中定义的参数完成训练。训练脚本 `train.py` 支持命令行参数解析，包括数据集根目录、训练轮数（epochs）、批量大小（batch size）、检查点保存间隔、检查点目录等。 这是一个支持基于自回归模式的 Transfomer 语言模型训练以及推理框架
**模型下载选项（任选其一）：** **模型下载选项（任选其一）：**
@ -181,30 +160,28 @@ The model has completed pre-training + SFT (Supervised Fine-Tuning) + DPO (Direc
训练数据来源请参见 HuggingFace 下载页面中的 **Model Card** 部分。 训练数据来源请参见 HuggingFace 下载页面中的 **Model Card** 部分。
**许可证：** 代码遵循 Apache-2.0 协议，使用时请注明出处。 **许可证：** 代码遵循 GPL-3.0 协议，使用时请注明出处。
- **📊 设备选择：** 默认使用 CUDA 进行训练 - **📊 设备选择：** 默认使用 CUDA 进行训练
- **🌐 性能优化：** 启用 `dtype=torch.bfloat16` 以加速训练并减少内存占用，请确保硬件支持该特性 - **🌐 性能优化：** 启用 `dtype=torch.bfloat16` 以加速训练并减少内存占用，请确保硬件支持该特性
- **🤖 语言支持：** 模型支持中文和英文训练。由于 BBPE 分词器未使用多语言文本训练，因此中英文的 OOV（未登录词）问题较少，其他语言可能存在 OOV 问题 - **🤖 语言支持：** 模型支持中文和英文训练。由于 BBPE 分词器未使用多语言文本训练，因此中英文的 OOV（未登录词）问题较少，其他语言可能存在 OOV 问题
### 📌 训练指南 ### 📌 训练指南
要训练该 Transformer 模型，请按照以下步骤操作： 要训练该 Transformer 模型，请按照以下步骤操作：
#### **(1). 准备数据集：** **(1). 准备数据集：**
将数据集放置在指定的根目录下。文件应为包含中文、英文或混合文本的文本文档。格式应符合模型输入要求——建议使用预分词后的 `token_ids` 并以 `torch.Tensor` 格式保存（使用 `torch.Tensor` 相比 Python 列表更节省内存，列表默认为 64 位精度）。 将数据集放置在指定的根目录下， 本系统采用 BBPE 分词器进行分词，并且要求使用已经经过分词的 token 分段训练（分段存储为 *.h5 格式）
#### **(2). 安装依赖：** **(2). 安装依赖：**
```bash ```bash
pip install -r requirements.txt pip install -e .
pip install .
``` ```
#### **(3). 运行训练脚本：** **(3). 运行训练脚本：**
```bash ```bash
python train.py \ python train.py \
@ -231,13 +208,11 @@ python train.py \
- `--checkpoint_dir`: 检查点保存目录 - `--checkpoint_dir`: 检查点保存目录
- `--resume_dir`: 从指定路径恢复训练 - `--resume_dir`: 从指定路径恢复训练
训练日志将保存在 `train_log.txt` 中。检查点将保存在指定目录，用于恢复训练或评估。
### 👉 使用指南 ### 👉 使用指南
#### **(1). 与模型对话：** **(1). 与模型对话：**
打开 `chat.py` 或使用流式/非流式接口： 打开 `chat.py` 或使用流式/非流式接口：
@ -291,7 +266,7 @@ while True:
print(response) print(response)
``` ```
#### **(2). 基于检索的生成（RAG）：** **(2). 基于检索的生成（RAG）：**
```python ```python
import torch import torch
@ -308,26 +283,4 @@ retrieved_content = model.retrieve_generate(
top_p=0.95 top_p=0.95
) )
print(retrieved_content) print(retrieved_content)
``` ```
### 📌 模型规格说明（重复部分）
该模型基于一个 24 层的 Transformer 架构，参数配置定义在 `config.json` 中，总参数量约为 10 亿（1.0B）。
**关键设计选择：**
- 在嵌入层（embedding）与最终线性层之间进行权重绑定（weight tying），这是小型模型中常见的节省参数量的做法
- 嵌入层优化：若不进行权重绑定，一个包含 10,000 个词的词汇表将消耗约 1.02 亿（0.1B）参数
**局限性：**
- 由于参数规模较小，可能在处理复杂语言现象时表现受限
- 在特定领域的数据集上容易出现过拟合
- 多语言能力有限
**优势：**
- 可在低配置硬件上高效运行
- 相较于大型模型，训练时间更短
**训练流程：**
该模型已完成预训练（pre-training）+ 监督微调（SFT, Supervised Fine-Tuning）+ 直接偏好优化（DPO, Direct Preference Optimization）的全流程。所有相关的训练代码均已包含在代码库中。