# AWQ model LoRA fine-tuning

**URL:** <https://discuss.huggingface.co/t/awq-model-lora-fine-tuning/74989>\
**Category:** Beginners\
**Created:** [February 27, 2024, 1:20pm UTC](https://discuss.huggingface.co/t/awq-model-lora-fine-tuning/74989 "2024-02-27T13:20:08Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![kklivil](https://avatars.discourse-cdn.com/v4/letter/k/bcef8e/32.png) [@kklivil](https://discuss.huggingface.co/u/kklivil)\
**Post date:** [February 27, 2024, 1:20pm UTC](https://discuss.huggingface.co/t/awq-model-lora-fine-tuning/74989/1 "2024-02-27T13:20:08Z")

</div>

Hello All, Is it possible to use LoRa fine-tuning directly on a an existing AWQ (e.g. `casperhansen/mistral-7b-instruct-v0.1-awq` ) model?

---

<div class="post-metadata">

**Author:** ![sparsh35](https://avatars.discourse-cdn.com/v4/letter/s/4da419/32.png) [@sparsh35](https://discuss.huggingface.co/u/sparsh35)\
**Post date:** [April 8, 2024, 7:13pm UTC](https://discuss.huggingface.co/t/awq-model-lora-fine-tuning/74989/2 "2024-04-08T19:13:12Z")

</div>

No it is currently not possible , but i think it is possible for GPTQ.