Train on larger Llama-2 models, and comparison to Sqlcoder
#17
by
buddyroo30
- opened
You fine-tuned Llama-2-7B to generate this, but I'm just wondering if you have plans to fine tune some of the larger Llama-2 models similarly, i.e. 13B or 70B models (presumably would give better performance)? Also, how does your model compare to Sqlcoder which similarly does English to SQL generation: https://defog.ai/sqlcoder-demo/ ? Thanks.