README.md

---
base_model: grimjim/rogue-enchantress-32k-7B
library_name: transformers
quanted_by: grimjim
license: cc-by-nc-4.0
pipeline_tag: text-generation
---
# rogue-enchantress-32k-7B-GGUF

This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).

An ambition of this merge was to augment text generation with the potential creative richness of the WizardLM-2 7B and Zephyr-7B-Beta models, the reasoning of the Starling-LM-7B-beta model, and extended context length of Mistral v0.2.

The resulting model is very attentive to character card descriptions and capable of applying reasoning. This model is in the smarter side, following context and formatting. The model is creative and "wants" to write, incorporating details cooperatively and driving plot, with occasional runaway narration if it finds that the prompt leans that way.

Tested with ChatML Instruct prompts, temperature 1.0, and minP 0.02.

- Full weights: [grimjim/rogue-enchantress-32k-7B](https://huggingface.co/grimjim/rogue-enchantress-32k-7B)
- GGUF quants: [grimjim/rogue-enchantress-32k-7B-GGUF](https://huggingface.co/grimjim/rogue-enchantress-32k-7B-GGUF)

## Merge Details
### Merge Method

This model was merged using the SLERP merge method.

### Models Merged

The following models were included in the merge:
* [grimjim/zephyr-wizard-kuno-royale-BF16-merge-7B](https://huggingface.co/grimjim/zephyr-wizard-kuno-royale-BF16-merge-7B)
* [grimjim/cuckoo-starling-7B](https://huggingface.co/grimjim/cuckoo-starling-7B)

### Configuration

The following YAML configuration was used to produce this model:

```yaml
slices:
  - sources:
    - model: grimjim/zephyr-wizard-kuno-royale-BF16-merge-7B
      layer_range: [0,32]
    - model: grimjim/cuckoo-starling-7B
      layer_range: [0,32]
merge_method: slerp
base_model: grimjim/zephyr-wizard-kuno-royale-BF16-merge-7B
parameters:
  t:
    - value: 0.5
dtype: bfloat16

```