DeusImperator/Lyra4-Gutenberg-12B_exl2_8bpw_max

Lyra4-Gutenberg-12B - EXL2 8bpw max

This is a 8bpw EXL2 quant of nbeerbower/Lyra4-Gutenberg-12B

This quant was made using exllamav2-0.2.1 with default dataset. I used a slightly modified quantization script to force use of highest bpw methods for all layers in the model (which is usually "1:8b_128g s4") to ensure max quality.

I also added a small fix in config file to set max default context at 128k as original Mistral-Nemo should have.

I tested this quant shortly in some random RPs (including ones over 8k context) and it seems to work fine.

Prompt Templates

Uses ChatML or modified mistral format like mentioned in original Lyra v4. I tested it with ChatML.

Original readme below

---

Lyra4-Gutenberg-12B

Sao10K/MN-12B-Lyra-v4 finetuned on jondurbin/gutenberg-dpo-v0.1.

Method

ORPO Finetuned using an RTX 3090 + 4060 Ti for 3 epochs.

Fine-tune Llama 3 with ORPO