caphe/affine-5HdchXepaTzbM97fuYqXXjWS955xk1dkJT2ZjasDgwjfvuvd

TEXT GENERATIONConcurrent Unit Cost:3Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jul 13, 2026Architecture:Transformer Featherless Exclusive Cold

caphe/affine-5HdchXepaTzbM97fuYqXXjWS955xk1dkJT2ZjasDgwjfvuvd is a 35.1 billion parameter language model created by caphe, merged using the Model Breadcrumbs with TIES method. It combines components from vera6/affine-5GmvsbydEDvdhHy6w8JHMCorYyJPiAnvzHWaa8Aj6iTdzzns, OronoCris/affine-83-5ETkv1LSG1sMVBma1oPQUUKPkzmLDHQN7zwf8sWqCQEps3yR, and DuWenJing/affine-model-5CrVf11voFhxLs2nM7LBVUKPfD4gNwYugag4Dmwe8JM423RC. This model is designed for general language understanding and generation tasks, leveraging its merged architecture for potentially enhanced performance across various applications.

Loading preview...

Model Overview

caphe/affine-5HdchXepaTzbM97fuYqXXjWS955xk1dkJT2ZjasDgwjfvuvd is a 35.1 billion parameter language model developed by caphe. It was created using the mergekit tool, specifically employing the Model Breadcrumbs with TIES merge method.

Merge Details

This model is a composite of several pre-trained language models, with vera6/affine-5GmvsbydEDvdhHy6w8JHMCorYyJPiAnvzHWaa8Aj6iTdzzns serving as the base model. The merge incorporated two additional models:

  • DuWenJing/affine-model-5CrVf11voFhxLs2nM7LBVUKPfD4gNwYugag4Dmwe8JM423RC: This component was given a higher weight (0.7) and a density of 0.95, indicating a significant contribution to the final model's characteristics.
  • OronoCris/affine-83-5ETkv1LSG1sMVBma1oPQUUKPkzmLDHQN7zwf8sWqCQEps3yR: This model had a lower weight (0.3) and a density of 0.8, suggesting a more subtle influence.

Configuration

The merge process utilized a bfloat16 data type and specific parameters for density (0.92) and gamma (0.02) within the Breadcrumbs with TIES method. This configuration aims to optimize the integration of the constituent models.

Potential Use Cases

Given its merged nature and 35.1B parameters, this model is suitable for a range of natural language processing tasks, including but not limited to:

  • Text generation
  • Summarization
  • Question answering
  • General conversational AI

Finetunes
1 models