---
license: other
language:
- en
---

[EXL2](https://github.com/turboderp/exllamav2/tree/master#exllamav2) Quantization of [Undi95's's MM-ReMM-L2-20B](https://huggingface.co/Undi95/MM-ReMM-L2-20B).


## Model details

Quantized at 3.18bpw with hb 6, This one can actually go full 4K context on 16GB VRAM, will redo the other 20b models later.

Perplexity:

Base = 6.9504

3.18 h6 = 7.0138

Dataset = [wikitext](https://huggingface.co/datasets/wikitext/resolve/refs%2Fconvert%2Fparquet/wikitext-2-v1/test/0000.parquet)

## Prompt Format

```
Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:

```