Skip to content

Commit

Permalink
feat: Support fine-tuning of NLU tasks (#284)
Browse files Browse the repository at this point in the history 
Co-authored-by: lorcanzhang <[email protected]>
Co-authored-by: luchun <[email protected]>
Co-authored-by: csunny <[email protected]>
  • Loading branch information
@fangyinc @zhanghy-sketchzh @csunny
4 people authored Aug 27, 2024
1 parent 91fe094 commit 1c50727
Show file tree
Hide file tree
Showing 103 changed files with 48,808 additions and 5,008 deletions.
66 changes: 0 additions & 66 deletions .github/workflows/ci.yml
View file

This file was deleted.

31 changes: 17 additions & 14 deletions .gitignore
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ __pycache__/

# C extensions
*.so
.gitkeep

# Distribution / packaging
.Python
Expand All @@ -13,21 +14,23 @@ data/spider
data/eval
output_pred/
wandb/
dbgpt_hub/data/*
src/dbgpt-hub-sql/dbgpt_hub_sql/data/*
# But track the data/eval_data folder itself
!dbgpt_hub/data/eval_data/
!dbgpt_hub/data/dataset_info.json
!dbgpt_hub/data/example_text2sql.json

# Ignore everything under dbgpt_hub/ouput/ except the adapter directory
dbgpt_hub/output/adapter/*
!dbgpt_hub/output/adapter/.gitkeep
dbgpt_hub/output/logs/*
!dbgpt_hub/output/logs/.gitkeep
dbgpt_hub/output/pred/*
!dbgpt_hub/output/pred/.gitkeep


!src/dbgpt-hub-sql/dbgpt_hub_sql/data/eval_data/
!src/dbgpt-hub-sql/dbgpt_hub_sql/data/dataset_info.json
!src/dbgpt-hub-sql/dbgpt_hub_sql/data/example_text2sql.json

# Ignore everything under dbgpt_hub_sql/ouput/ except the adapter directory
src/dbgpt-hub-sql/dbgpt_hub_sql/output/adapter/*
!src/dbgpt-hub-sql/dbgpt_hub_sql/output/adapter/.gitkeep
src/dbgpt-hub-sql/dbgpt_hub_sql/output/logs/*
!src/dbgpt-hub-sql/dbgpt_hub_sql/output/logs/.gitkeep
src/dbgpt-hub-sql/dbgpt_hub_sql/output/pred/*
!src/dbgpt-hub-sql/dbgpt_hub_sql/output/pred/.gitkeep

# Ignore NLU output
src/dbgpt-hub-nlu/output
src/dbgpt-hub-nlu/data

#
build/
Expand Down
31 changes: 31 additions & 0 deletions Makefile
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,31 @@
.DEFAULT_GOAL := help

SHELL=/bin/bash
VENV = venv

# Detect the operating system and set the virtualenv bin directory
ifeq ($(OS),Windows_NT)
VENV_BIN=$(VENV)/Scripts
else
VENV_BIN=$(VENV)/bin
endif

setup: $(VENV)/bin/activate

$(VENV)/bin/activate: $(VENV)/.venv-timestamp

$(VENV)/.venv-timestamp: src/dbgpt-hub-nlu/setup.py requirements
# Create new virtual environment if setup.py has changed
python3 -m venv $(VENV)
$(VENV_BIN)/pip install --upgrade pip
$(VENV_BIN)/pip install -r requirements/lint-requirements.txt
touch $(VENV)/.venv-timestamp


.PHONY: fmt
fmt: setup ## Format Python code
# TODO: Use isort to sort Python imports.
# https://github.com/PyCQA/isort
$(VENV_BIN)/isort src/
# https://github.com/psf/black
$(VENV_BIN)/black --extend-exclude="examples/notebook" .
86 changes: 47 additions & 39 deletions README.md
View file
Original file line number Diff line number Diff line change
@@ -1,6 +1,5 @@
# DB-GPT-Hub: Text-to-SQL parsing with LLMs


<div align="center">
<p>
<a href="https://github.com/eosphoros-ai/DB-GPT">
Expand All @@ -25,9 +24,17 @@


[**简体中文**](README.zh.md) | [**Discord**](https://discord.gg/7uQnPuveTY) | [**Wechat**](https://github.com/eosphoros-ai/DB-GPT/blob/main/README.zh.md#%E8%81%94%E7%B3%BB%E6%88%91%E4%BB%AC) | [**Huggingface**](https://huggingface.co/eosphoros) | [**Community**](https://github.com/eosphoros-ai/community) | [**Paper**](https://arxiv.org/abs/2406.11434)


[**Text2SQL**](README.md) | [**Text2NLU**](src/dbgpt-hub-nlu/README.zh.md)
</div>

## 🔥🔥🔥 News
- Support [Text2NLU](src/dbgpt-hub-nlu/README.zh.md) fine-tuning to improve semantic understanding accuracy.

## Baseline

Text2SQL eval execution accuracy (ex) metric, and we will move this to `src/dbgpt_hub_sql`
- update time: 2023/12/08
- metric: execution accuracy (ex)
- more details refer to [docs/eval-llm-result.md](https://github.com/eosphoros-ai/DB-GPT-Hub/blob/main/docs/eval_llm_result.md)
Expand Down Expand Up @@ -381,8 +388,9 @@ git clone https://github.com/eosphoros-ai/DB-GPT-Hub.git
cd DB-GPT-Hub
conda create -n dbgpt_hub python=3.10
conda activate dbgpt_hub
pip install poetry
poetry install
cd src/dbgpt_hub_sql
pip install -e .
```
### 3.2 Quick Start

Expand All @@ -392,13 +400,13 @@ Firstly, install `dbgpt-hub` with the following command

Then, set up the arguments and run the whole process.
```python
from dbgpt_hub.data_process import preprocess_sft_data
from dbgpt_hub.train import start_sft
from dbgpt_hub.predict import start_predict
from dbgpt_hub.eval import start_evaluate
from dbgpt_hub_sql.data_process import preprocess_sft_data
from dbgpt_hub_sql.train import start_sft
from dbgpt_hub_sql.predict import start_predict
from dbgpt_hub_sql.eval import start_evaluate

# Config the input datasets
data_folder = "dbgpt_hub/data"
data_folder = "dbgpt_hub_sql/data"
data_info = [
{
"data_source": "spider",
Expand All @@ -424,7 +432,7 @@ train_args = {
"template": "llama2",
"lora_rank": 64,
"lora_alpha": 32,
"output_dir": "dbgpt_hub/output/adapter/CodeLlama-13b-sql-lora",
"output_dir": "dbgpt_hub_sql/output/adapter/CodeLlama-13b-sql-lora",
"overwrite_cache": True,
"overwrite_output_dir": True,
"per_device_train_batch_size": 1,
Expand All @@ -443,20 +451,20 @@ predict_args = {
"model_name_or_path": "codellama/CodeLlama-13b-Instruct-hf",
"template": "llama2",
"finetuning_type": "lora",
"checkpoint_dir": "dbgpt_hub/output/adapter/CodeLlama-13b-sql-lora",
"predict_file_path": "dbgpt_hub/data/eval_data/dev_sql.json",
"predict_out_dir": "dbgpt_hub/output/",
"checkpoint_dir": "dbgpt_hub_sql/output/adapter/CodeLlama-13b-sql-lora",
"predict_file_path": "dbgpt_hub_sql/data/eval_data/dev_sql.json",
"predict_out_dir": "dbgpt_hub_sql/output/",
"predicted_out_filename": "pred_sql.sql",
}

# Config evaluation parameters
evaluate_args = {
"input": "./dbgpt_hub/output/pred/pred_sql_dev_skeleton.sql",
"gold": "./dbgpt_hub/data/eval_data/gold.txt",
"gold_natsql": "./dbgpt_hub/data/eval_data/gold_natsql2sql.txt",
"db": "./dbgpt_hub/data/spider/database",
"table": "./dbgpt_hub/data/eval_data/tables.json",
"table_natsql": "./dbgpt_hub/data/eval_data/tables_for_natsql2sql.json",
"input": "./dbgpt_hub_sql/output/pred/pred_sql_dev_skeleton.sql",
"gold": "./dbgpt_hub_sql/data/eval_data/gold.txt",
"gold_natsql": "./dbgpt_hub_sql/data/eval_data/gold_natsql2sql.txt",
"db": "./dbgpt_hub_sql/data/spider/database",
"table": "./dbgpt_hub_sql/data/eval_data/tables.json",
"table_natsql": "./dbgpt_hub_sql/data/eval_data/tables_for_natsql2sql.json",
"etype": "exec",
"plug_value": True,
"keep_distict": False,
Expand All @@ -479,15 +487,15 @@ start_evaluate(evaluate_args)

DB-GPT-Hub uses the information matching generation method for data preparation, i.e. the SQL + Repository generation method that combines table information. This method combines data table information to better understand the structure and relationships of the data table, and is suitable for generating SQL statements that meet the requirements.

Download the [Spider dataset]((https://drive.google.com/uc?export=download&id=1TqleXec_OykOYFREKKtschzY29dUcVAQ)) from the Spider dataset link. By default, after downloading and extracting the data, place it in the dbgpt_hub/data directory, i.e., the path should be `dbgpt_hub/data/spider`.
Download the [Spider dataset]((https://drive.google.com/uc?export=download&id=1TqleXec_OykOYFREKKtschzY29dUcVAQ)) from the Spider dataset link. By default, after downloading and extracting the data, place it in the dbgpt_hub_sql/data directory, i.e., the path should be `dbgpt_hub_sql/data/spider`.

For the data preprocessing part, simply **run the following script** :
```bash
## generate train and dev(eval) data
poetry run sh dbgpt_hub/scripts/gen_train_eval_data.sh
sh dbgpt_hub_sql/scripts/gen_train_eval_data.sh
```

In the directory `dbgpt_hub/data/`, you will find the newly generated training file example_text2sql_train.json and testing file example_text2sql_dev.json, containing 8659 and 1034 entries respectively. For the data used in subsequent fine-tuning, set the parameter `file_name` value to the file name of the training set in dbgpt_hub/data/dataset_info.json, such as example_text2sql_train.json
In the directory `dbgpt_hub_sql/data/`, you will find the newly generated training file example_text2sql_train.json and testing file example_text2sql_dev.json, containing 8659 and 1034 entries respectively. For the data used in subsequent fine-tuning, set the parameter `file_name` value to the file name of the training set in dbgpt_hub_sql/data/dataset_info.json, such as example_text2sql_train.json


The data in the generated JSON looks something like this:
Expand All @@ -500,43 +508,43 @@ The data in the generated JSON looks something like this:
"history": []
},
```
The data processing code of `chase`, `cosql` and `sparc` has been embedded in the data processing code of the project. After downloading the data set according to the above link, you only need to add ` in `dbgpt_hub/configs/config.py` Just loosen the corresponding code comment in SQL_DATA_INFO`.
The data processing code of `chase`, `cosql` and `sparc` has been embedded in the data processing code of the project. After downloading the data set according to the above link, you only need to add ` in `dbgpt_hub_sql/configs/config.py` Just loosen the corresponding code comment in SQL_DATA_INFO`.

### 3.4. Model fine-tuning

The model fine-tuning supports both LoRA and QLoRA methods. We can run the following command to fine-tune the model. By default, with the parameter --quantization_bit, it uses the QLoRA fine-tuning method. To switch to LoRAs, simply remove the related parameter from the script.
Run the command:

```bash
poetry run sh dbgpt_hub/scripts/train_sft.sh
sh dbgpt_hub_sql/scripts/train_sft.sh
```

After fine-tuning, the model weights will be saved by default in the adapter folder, specifically in the dbgpt_hub/output/adapter directory.
After fine-tuning, the model weights will be saved by default in the adapter folder, specifically in the dbgpt_hub_sql/output/adapter directory.

If you're using **multi-GPU training and want to utilize deepseed**, you should modify the default content in train_sft.sh. The change is:

```
CUDA_VISIBLE_DEVICES=0 python dbgpt_hub/train/sft_train.py \
CUDA_VISIBLE_DEVICES=0 python dbgpt_hub_sql/train/sft_train.py \
--quantization_bit 4 \
...
```
change to ：
```
deepspeed --num_gpus 2 dbgpt_hub/train/sft_train.py \
--deepspeed dbgpt_hub/configs/ds_config.json \
deepspeed --num_gpus 2 dbgpt_hub_sql/train/sft_train.py \
--deepspeed dbgpt_hub_sql/configs/ds_config.json \
--quantization_bit 4 \
...
```

if you need order card id
```
deepspeed --include localhost:0,1 dbgpt_hub/train/sft_train.py \
--deepspeed dbgpt_hub/configs/ds_config.json \
deepspeed --include localhost:0,1 dbgpt_hub_sql/train/sft_train.py \
--deepspeed dbgpt_hub_sql/configs/ds_config.json \
--quantization_bit 4 \
...
```

The other parts that are omitted (…) can be kept consistent. If you want to change the default deepseed configuration, go into the `dbgpt_hub/configs` directory and make changes to ds_config.json as needed,the default is stage2.
The other parts that are omitted (…) can be kept consistent. If you want to change the default deepseed configuration, go into the `dbgpt_hub_sql/configs` directory and make changes to ds_config.json as needed,the default is stage2.

In the script, during fine-tuning, different models correspond to key parameters lora_target and template, as shown in the following table:

Expand All @@ -563,10 +571,10 @@ In the script, during fine-tuning, different models correspond to key parameters

> quantization_bit: Indicates whether quantization is applied, with valid values being [4 or 8].
> model_name_or_path: The path of the LLM (Large Language Model).
> dataset: Specifies the name of the training dataset configuration, corresponding to the outer key value in dbgpt_hub/data/dataset_info.json, such as example_text2sql.
> dataset: Specifies the name of the training dataset configuration, corresponding to the outer key value in dbgpt_hub_sql/data/dataset_info.json, such as example_text2sql.
> max_source_length: The length of the text input into the model. If computing resources allow, it can be set as large as possible, like 1024 or 2048.
> max_target_length: The length of the SQL content output by the model; 512 is generally sufficient.
> output_dir: The output path of the Peft module during SFT (Supervised Fine-Tuning), set by default to `dbgpt_hub/output/adapter/` .
> output_dir: The output path of the Peft module during SFT (Supervised Fine-Tuning), set by default to `dbgpt_hub_sql/output/adapter/` .
> per_device_train_batch_size: The size of the batch. If computing resources allow, it can be set larger; the default is 1.
> gradient_accumulation_steps: The number of steps for accumulating gradients before an update.
> save_steps: The number of steps at which model checkpoints are saved; it can be set to 100 by default.
Expand All @@ -575,10 +583,10 @@ In the script, during fine-tuning, different models correspond to key parameters

### 3.5. Model Predict

Under the project directory ./dbgpt_hub/output/pred/, this folder is the default output location for model predictions(if not exist, just mkdir).
Under the project directory ./dbgpt_hub_sql/output/pred/, this folder is the default output location for model predictions(if not exist, just mkdir).

```bash
poetry run sh ./dbgpt_hub/scripts/predict_sft.sh
sh ./dbgpt_hub_sql/scripts/predict_sft.sh
```

In the script, by default with the parameter `--quantization_bit`, it predicts using QLoRA. Removing it switches to the LoRA prediction method.
Expand All @@ -593,7 +601,7 @@ You can find the second corresponding model weights from Huggingface [hg-eospho
If you need to merge the weights of the trained base model and the fine-tuned Peft module to export a complete model, execute the following model export script:

```bash
poetry run sh ./dbgpt_hub/scripts/export_merge.sh
sh ./dbgpt_hub_sql/scripts/export_merge.sh
```

Be sure to replace the parameter path values in the script with the paths corresponding to your project.
Expand All @@ -602,7 +610,7 @@ Be sure to replace the parameter path values in the script with the paths corres
To evaluate model performance on the dataset, default is spider dev dataset.
Run the following command:
```bash
poetry run python dbgpt_hub/eval/evaluation.py --plug_value --input Your_model_pred_file
python dbgpt_hub_sql/eval/evaluation.py --plug_value --input Your_model_pred_file
```
You can find the results of our latest review and part of experiment results [here](docs/eval_llm_result.md)
**Note**: The database pointed to by the default code is a 95M database downloaded from [Spider official website] (https://yale-lily.github.io/spider). If you need to use Spider database (size 1.27G) in [test-suite](https://github.com/taoyds/test-suite-sql-eval), please download the database in the link to the custom directory first, and run the above evaluation command which add parameters and values ​​like `--db Your_download_db_path`.
Expand Down Expand Up @@ -644,13 +652,13 @@ We warmly invite more individuals to join us and actively engage in various aspe

Before submitting your code, please ensure that it is formatted according to the black style by using the following command:
```
poetry run black dbgpt_hub
black dbgpt_hub
```

If you have more time to execute more detailed type checking and style checking of your code, please use the following command:
```
poetry run pyright dbgpt_hub
poetry run pylint dbgpt_hub
pyright dbgpt_hub
pylint dbgpt_hub
```

If you have any questions or need further assistance, don't hesitate to reach out. We appreciate your involvement!
Expand Down
Loading

0 comments on commit 1c50727

Please sign in to comment.