sometimesanotion/Qwenvergence-14B-v6-Prose

TEXT GENERATIONConcurrency Cost:1Model Size:14.8BQuant:FP8Ctx Length:32kPublished:Dec 26, 2024License:apache-2.0Architecture:Transformer Open Weights Cold

sometimesanotion/Qwenvergence-14B-v6-Prose is a 14.8 billion parameter merged language model based on the Qwen2.5-14B architecture, created using the TIES merge method. This model integrates multiple specialized models, including arcee-ai/Virtuoso-Small and sometimesanotion/Lamarck-14B-v0.3, to enhance its general prose generation capabilities. With a 32768 token context length, it is designed for diverse text generation tasks, leveraging the strengths of its constituent models.

Loading preview...

Qwenvergence-14B-v6-Prose: A Merged Language Model

This model, developed by sometimesanotion, is a 14.8 billion parameter language model built upon the Qwen2.5-14B base architecture. It was created using the TIES (Trimmed-mean-based Ensemble of Subnetworks) merge method, a technique designed to combine the strengths of multiple pre-trained models into a single, more capable model.

Key Capabilities & Merge Details

  • Base Model: Qwen/Qwen2.5-14B serves as the foundational architecture.
  • Merge Method: Utilizes the TIES method, which intelligently combines parameters from various models.
  • Constituent Models: The merge incorporates several specialized models, including:
    • arcee-ai/Virtuoso-Small
    • sometimesanotion/Lamarck-14B-v0.3
    • EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2
    • allura-org/TQ2.5-14B-Sugarquill-v1
    • oxyapi/oxy-1-small
    • v000000/Qwen2.5-Lumen-14B
    • sthenno-com/miscii-14b-1225
    • underwoods/medius-erebus-magnum-14b
    • huihui-ai/Qwen2.5-14B-Instruct-abliterated-v2
  • Context Length: Supports a substantial context window of 32768 tokens.

Good For

  • General Prose Generation: The integration of diverse models suggests an enhanced capability for generating varied and coherent text.
  • Applications requiring extended context: Its large context window makes it suitable for tasks involving longer documents or conversations.
  • Exploration of merged model performance: Ideal for researchers and developers interested in the efficacy of the TIES merging technique for combining specialized LLMs.