File size: 2,138 Bytes
9f2b5ae |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 |
---
base_model:
- 152334H/miqu-1-70b-sf
- NeverSleep/MiquMaid-v1-70B
- Sao10K/WinterGoddess-1.4x-70B-L2
library_name: transformers
tags:
- mergekit
- merge
---
# aranea-ancilla-116b-v1.0
**aka MiquMaid-v1-70B + interleaved WinterGoddess-1.4x-70B-L2**
A [mergekit](https://github.com/arcee-ai/mergekit) frankenmerge based on [NeverSleep/MiquMaid-v1-70B](https://huggingface.co/NeverSleep/MiquMaid-v1-70B) with interleaved layers of [Sao10K/WinterGoddess-1.4x-70B-L2](https://huggingface.co/Sao10K/WinterGoddess-1.4x-70B-L2).
This was the top performing model from a series of merge experiments to create a highly coherant creative writing model.
Tests consisted of a series of private benchmarks and manual comparisons. A number of different base models, interleave models and layer offsets were compared.
- Usable context ~32768
- Recommended context ~16384
Non frankenstein miqu-1 finetunes generally outperform their frankenstein counterparts at very long contexts due to coherency loss.
As a rough suggestion I might suggest swapping out to either [NeverSleep/MiquMaid-v1-70B](https://huggingface.co/NeverSleep/MiquMaid-v1-70B) or [152334H/miqu-1-70b-sf](https://huggingface.co/152334H/miqu-1-70b-sf) after 16k context.
Layers: 136
### License
No license. Component models based on the [Mistral AI Miqu-1](https://huggingface.co/miqudev/miqu-1-70b/tree/main) llama2 finetune that was released without license.
### Interesting observations from benchmarking
- 10 layer interleave stride with a 20 layer interleave width consistently outperformed alternatives combinations.
- Offsetting the interleaved model's first set of layers generally improved coherency. [14-30] reliably beat the [10-30] mergekit slice configuration for various combinations of models.
- Quality of resulting merges can vary wildly. Whilst a merge of two strong models tends to produce a strong frankenstein model, this rule does not always hold true.
### Quantizations
Exllamav2 quants will be available when bandwidth permits.
|