README.md

# t5_wikisql_SQL2en
---
language: en
datasets:
- wikisql
---

[Googles T5](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) fine-tuned on [WikiSQL](https://github.com/salesforce/WikiSQL) for **English** to **SQL** **translation**.


This is a [t5-small] (https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) fine-tuned version on the [wikisql dataset] (https://huggingface.co/datasets/wikisql).

The model can be loaded like so:

```python
from transformers import AutoTokenizer, AutoModelWithLMHead
tokenizer = AutoTokenizer.from_pretrained("dbernsohn/t5_wikisql_SQL2en")
model = AutoModelWithLMHead.from_pretrained("dbernsohn/t5_wikisql_SQL2en")
```

You can then use this model to translate SQL queries into plain english.

```python
query = "SELECT COUNT Params from model where location=HF-Hub"
input_text = f"translate English to SQL: {query} </s>"
features = tokenizer([input_text], return_tensors='pt')

output = model.generate(input_ids=features['input_ids'].cuda(), 
                        attention_mask=features['attention_mask'].cuda())

tokenizer.decode(output[0])
```

The whole training process and hyperparameters are in my [GitHub repo] (https://github.com/DorBernsohn)