ajtaltarabukin2022/merge_v10_27_73_9

TEXT GENERATIONConcurrency Cost:2Model Size:32BQuant:FP8Ctx Length:32kPublished:Apr 22, 2026Architecture:Transformer Cold

The ajtaltarabukin2022/merge_v10_27_73_9 is a 32 billion parameter language model created by ajtaltarabukin2022, formed by merging pre-trained models using the DARE TIES method. This model integrates components from a base model and an additional source, aiming to combine their respective strengths. With a 32768 token context length, it is designed for general language understanding and generation tasks, leveraging its merged architecture for potentially enhanced performance.

Loading preview...

Model Overview

The ajtaltarabukin2022/merge_v10_27_73_9 is a 32 billion parameter language model developed by ajtaltarabukin2022. It was constructed using the DARE TIES (Disentangled Representation Editing for Text-to-Image Synthesis) merge method, a technique designed to combine the capabilities of multiple pre-trained language models. The merging process utilized MergeKit, a tool for creating new models from existing ones.

Merge Details

This model is a merge of two distinct sources:

  • A base model identified as /root/finetuneqwen/prexpertMJDD.
  • An additional model, /root/finetuneqwen/distillgptNsH2.

The DARE TIES method was applied with specific weighting parameters, assigning a 0.55 weight to the base model and a 0.45 weight to the additional model across layers 0 to 64. The configuration also specified a bfloat16 data type and a density of 0.9 for the merged parameters. This approach aims to synthesize the knowledge and capabilities present in the constituent models into a single, more robust model.

Key Characteristics

  • Architecture: Merged model based on pre-trained language models.
  • Parameter Count: 32 billion parameters.
  • Context Length: Supports a substantial context window of 32768 tokens.
  • Merge Method: Utilizes the DARE TIES technique for combining model weights.

Potential Use Cases

Given its merged nature and significant parameter count, this model is suitable for a broad range of natural language processing tasks, including:

  • General text generation and completion.
  • Understanding and responding to complex prompts.
  • Applications requiring a large context window for processing longer texts.