DreadPoor/Irix-12B-Model_Stock

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:12BQuant:FP8Ctx Length:32kPublished:Mar 25, 2025Architecture:Transformer0.0K Warm

DreadPoor/Irix-12B-Model_Stock is a 12 billion parameter language model created by DreadPoor, formed by merging multiple pre-trained models using the Model Stock method. This model leverages yamatazen/EtherealAurora-12B-v2 as its base, integrating contributions from several other 12B models. It is designed to combine the strengths of its constituent models, offering a versatile foundation for various natural language processing tasks.

Loading preview...

Model Overview

DreadPoor/Irix-12B-Model_Stock is a 12 billion parameter language model developed by DreadPoor. This model is a product of a merge operation, specifically utilizing the Model Stock merge method, which combines the weights of several pre-trained language models. The base model for this merge was yamatazen/EtherealAurora-12B-v2.

Merge Details

The creation of Irix-12B-Model_Stock involved integrating four distinct models:

This merging strategy aims to consolidate the diverse capabilities and knowledge embedded within each contributing model into a single, more robust entity. The configuration used for the merge specified int8_mask: true and dtype: bfloat16, indicating considerations for efficiency and precision during deployment.

Potential Use Cases

As a merged model, Irix-12B-Model_Stock is intended for a broad range of applications where a consolidated knowledge base from multiple sources is beneficial. Its 12 billion parameters suggest suitability for tasks requiring nuanced understanding and generation of text.

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