|
--- |
|
license: apache-2.0 |
|
language: |
|
- th |
|
- en |
|
metrics: |
|
- accuracy |
|
datasets: |
|
- AIAT/The_Scamper-train |
|
pipeline_tag: table-question-answering |
|
|
|
|
|
--- |
|
# Model Card for Model ID |
|
|
|
<!-- Provide a quick summary of what the model is/does. --> |
|
|
|
|
|
## Model Details |
|
|
|
### Model Description |
|
|
|
<!-- Provide a longer summary of what this model is. --> |
|
|
|
|
|
|
|
- **Developed by:** The Scamper |
|
- **Model type:** Transformer |
|
- **Language(s) (NLP):** Thai, English |
|
- **License:** apache-2.0 |
|
- **Finetuned from model:** OpenThaiGPT-1.0.0 70B (https://huggingface.co/openthaigpt/openthaigpt-1.0.0-70b-chat) |
|
|
|
|
|
## Uses |
|
|
|
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> |
|
|
|
The Tubular Question Answering Large Language Model is based on OpenThaiGPT and fine-tuned for converting natural language questions into SQL queries. It learns to map the nuances of Thai language to SQL structures, enabling efficient retrieval of information from databases. |
|
|
|
model2_path ="AIAT/The_Scamper-opt70bqt" |
|
tokenizer = AutoTokenizer.from_pretrained(model2_path, padding_side="right",use_fast=False) |
|
model = AutoModelForCausalLM.from_pretrained(model2_path, |
|
device_map="auto") |
|
|
|
|
|
### Recommendations |
|
|
|
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> |
|
|
|
|
|
|
|
## How to Get Started with the Model |
|
|
|
Use the code below to get started with the model. |
|
|
|
[More Information Needed] |
|
|
|
## Training Details |
|
|
|
### Training Data |
|
|
|
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> |
|
|
|
[More Information Needed] |
|
|
|
### Training Procedure |
|
|
|
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> |
|
|
|
The methodology for fine-tuning involves a dataset with two columns: "question" and "SQL syntax". Here's a brief outline of the process: |
|
|
|
1. **Data Collection**: Gather a dataset containing pairs of questions and their corresponding SQL queries. Ensure the questions cover various topics and query types, while the SQL queries represent the desired actions on a database. |
|
|
|
2. **Pre-processing**: Clean and preprocess the data to remove noise, standardize formatting, and handle any inconsistencies. Tokenize the text and encode it into a format suitable for training. |
|
|
|
3. **Model Architecture**: Utilize OpenThaiGPT 1.0.0 70B as the base model. |
|
|
|
4. **Fine-tuning Setup**: Divide the dataset into training (90%) and test sets (10%). We define the training procedure, including hyperparameters such as learning rate, batch size, and number of training epochs. |
|
|
|
5. **Fine-tuning Process**: Train the model on the question-SQL pairs using the defined setup. During training, the model learns to predict the SQL query corresponding to a given question by minimizing a suitable loss function. |
|
|
|
6. **Testing**: Evaluate the final model on a held-out test set to assess its generalization performance on unseen data. |
|
|
|
7. **Deployment**: Deploy the fine-tuned model for text-to-SQL tasks in real-world applications, where it can generate SQL queries from natural language questions effectively and efficiently. |
|
|
|
By following this methodology, the model can be fine-tuned to accurately convert natural language questions into SQL syntax, enabling seamless interaction with structured databases. |
|
|
|
|
|
|
