| BART Large CNN model trained for converting NLP queries to SQL queries. | |
| The model was trained on the Spider dataset Link: https://yale-lily.github.io/spider | |
| The model was trained using Google colab. | |
| Hyperparameters: | |
| "num epochs" = 3 | |
| "learning rate" = 1e-5 | |
| "batch size" = 8 | |
| "weight decay" = 0.01 | |
| "max input length" = 256 | |
| "max target length" = 256 | |
| "model name" : "facebook/bart-large-cnn" | |
| use code : | |
| from typing import List | |
| from transformers import AutoTokenizer, AutoModelForSeq2SeqLM | |
| tokenizer = AutoTokenizer.from_pretrained("upadhyay/sql") | |
| model = AutoModelForSeq2SeqLM.from_pretrained("upadhyay/sql") | |
| def prepare_input(question: str, table: List[str]): | |
| table_prefix = "table:" | |
| question_prefix = "question:" | |
| join_table = ",".join(table) | |
| inputs = f"{question_prefix} {question} {table_prefix} {join_table}" | |
| input_ids = tokenizer(inputs, max_length=700, return_tensors="pt").input_ids | |
| return input_ids | |
| def inference(question: str, table: List[str]) -> str: | |
| input_data = prepare_input(question=question, table=table) | |
| input_data = input_data.to(model.device) | |
| outputs = model.generate(inputs=input_data, num_beams=10, top_k=10, max_length=700) | |
| result = tokenizer.decode(token_ids=outputs[0], skip_special_tokens=True) | |
| return result | |
| print(inference(question="what is salary?", table=["id", "name", "age"])) | |