| license: other | |
| [airoboros-33b-gpt4-1.2](https://huggingface.co/jondurbin/airoboros-33b-gpt4-1.2) merged with bhenrym14's [airoboros-33b-gpt4-1.4.1-lxctx-PI-16384-LoRA](https://huggingface.co/bhenrym14/airoboros-33b-gpt4-1.4.1-lxctx-PI-16384-LoRA), full model (FP16) | |
| More info about the LoRA [Here](https://huggingface.co/bhenrym14/airoboros-33b-gpt4-1.4.1-lxctx-PI-16384-LoRA). This is an alternative to SuperHOT 8k LoRA trained with LoRA_rank 64 and context extended to 16K, with airoboros 1.4.1 dataset. |