MiniMoog/Mergerix-7b-v0.2

TEXT GENERATIONConcurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Apr 2, 2024License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

Mergerix-7b-v0.2 by MiniMoog is a 7 billion parameter language model created by merging alpindale/Mistral-7B-v0.2-hf, liminerity/M7-7b, and rwitz/experiment26-truthy-iter-0 using the model_stock merge method. This model leverages the Mistral-7B-v0.2-hf as its base, offering a context length of 8192 tokens. It is designed for general text generation tasks, benefiting from the combined strengths of its constituent models.

Loading preview...

Mergerix-7b-v0.2 Overview

Mergerix-7b-v0.2 is a 7 billion parameter language model developed by MiniMoog. This model is a product of merging three distinct models: alpindale/Mistral-7B-v0.2-hf, liminerity/M7-7b, and rwitz/experiment26-truthy-iter-0. The merge was performed using the model_stock method within LazyMergekit, with alpindale/Mistral-7B-v0.2-hf serving as the base architecture.

Key Characteristics

  • Architecture: Based on the Mistral-7B-v0.2-hf model family.
  • Parameter Count: 7 billion parameters.
  • Context Length: Supports an 8192-token context window.
  • Merge Method: Utilizes the model_stock merging technique to combine the strengths of its constituent models.

Usage and Application

This model is suitable for a variety of text generation tasks, leveraging the combined capabilities inherited from its merged components. Developers can integrate Mergerix-7b-v0.2 into their applications using the Hugging Face transformers library, with support for bfloat16 data types for efficient inference. The provided usage example demonstrates how to set up a text generation pipeline and apply chat templates for conversational AI scenarios.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p