nbeerbower/EVA-abliterated-TIES-Qwen2.5-1.5B
nbeerbower/EVA-abliterated-TIES-Qwen2.5-1.5B is a 1.5 billion parameter language model based on the Qwen2.5 architecture, created by nbeerbower using the TIES merge method. This model combines huihui-ai/Qwen2.5-1.5B-Instruct-abliterated and EVA-UNIT-01/EVA-Qwen2.5-1.5B-v0.0, leveraging the Qwen/Qwen2.5-1.5B as its base. It supports a 32768 token context length and is designed for multilingual applications, including English, Chinese, French, Spanish, and other languages.
Loading preview...
Model Overview
nbeerbower/EVA-abliterated-TIES-Qwen2.5-1.5B is a 1.5 billion parameter language model built upon the Qwen2.5 architecture. This model was developed by nbeerbower through a merge operation using the TIES (Trimmed-mean of Ensembles of Subnetworks) method, which combines the strengths of multiple pre-trained models.
Key Capabilities
- Merged Architecture: Created by merging
huihui-ai/Qwen2.5-1.5B-Instruct-abliteratedandEVA-UNIT-01/EVA-Qwen2.5-1.5B-v0.0, usingQwen/Qwen2.5-1.5Bas the foundational base model. - Multilingual Support: Inherits broad language capabilities from its base models, supporting languages such as Chinese (zho), English (eng), French (fra), Spanish (spa), Portuguese (por), German (deu), Italian (ita), Russian (rus), Japanese (jpn), Korean (kor), Vietnamese (vie), Thai (tha), and Arabic (ara).
- Context Length: Features a substantial context window of 32768 tokens, enabling processing of longer inputs and generating more coherent, extended responses.
Good For
- Applications requiring a compact yet capable multilingual model.
- Scenarios where combining specific instruction-tuned and general-purpose models is beneficial.
- Tasks that can leverage a 1.5B parameter model with a large context window for efficiency and performance.