EVA-Tissint-v1.1-14B

This is a merge of pre-trained language models created using mergekit.

New merge of EVA v0.2, now using Tissint v1.1. Hopefully the new version of Tissint and slight change of merge settings make an improvement over the first iteration.

I recommend the samplers provided on Tissint's model card.

If you'd like to use XTC, I recommend a threshold of 0.2. Lower thresholds seem to adversely affect the coherency.

Quantisations

Static: https://huggingface.co./mradermacher/EVA-Tissint-v1.1-14B-GGUF

Imatrix: https://huggingface.co./mradermacher/EVA-Tissint-v1.1-14B-i1-GGUF

Merge Method

This model was merged using the della_linear merge method using EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2 as a base.

Models Merged

The following models were included in the merge:

• Ttimofeyka/Tissint-14B-v1.1-128k-RP

Configuration

The following YAML configuration was used to produce this model:

models:
    - model: Ttimofeyka/Tissint-14B-v1.1-128k-RP
      parameters:
        density: 0.45
        weight: 0.3
    - model: EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2
      parameters:
        density: 0.55
        weight: 0.7

merge_method: della_linear
base_model: EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2
parameters:
    epsilon: 0.05
    lambda: 1
dtype: bfloat16