yamatazen/EsotericSage-12B

TEXT GENERATIONConcurrency Cost:1Model Size:12BQuant:FP8Ctx Length:32kPublished:May 23, 2025Architecture:Transformer0.0K Cold

EsotericSage-12B is a 12 billion parameter language model developed by yamatazen, created by merging yamatazen/LinearWriter-12B and yamatazen/ForgottenMaid-12B using the NearSwap method. This model is a product of combining pre-trained language models to leverage their respective strengths. Its primary characteristic is its origin as a merge, suggesting a focus on synthesizing capabilities from its constituent models.

Loading preview...

EsotericSage-12B Overview

EsotericSage-12B is a 12 billion parameter language model developed by yamatazen. It was created through a merging process using mergekit, specifically employing the NearSwap merge method.

Key Characteristics

  • Merge-based Architecture: This model is a composite, built by combining existing pre-trained language models.
  • Base Model: The merging process utilized yamatazen/LinearWriter-12B as its foundational base model.
  • Merged Component: yamatazen/ForgottenMaid-12B was integrated into the base model to form EsotericSage-12B.
  • Merge Method: The NearSwap technique was applied, with specific t parameters configured during the merge process.

Intended Use Cases

Given its construction from merged models, EsotericSage-12B is likely intended for applications that benefit from the combined strengths of its constituent models. Developers interested in exploring the outcomes of model merging, particularly with the NearSwap method, may find this model useful for experimentation and specific task fine-tuning where the characteristics of LinearWriter-12B and ForgottenMaid-12B are desirable.