---
license: cc-by-4.0
language:
- en
pipeline_tag: text-generation
---
quant of [defog's](https://huggingface.co/defog) [sqlcoder-34b-alpha](https://huggingface.co/defog/sqlcoder-34b-alpha)

https://huggingface.co/datasets/knowrohit07/know_sql used as calibration dataset.

since the original model didn't provide a `tokenizer.model` i used the one from codellama 34b. i hope that is the correct way to do it. it seems to work fine at least.