KiyoEditz/EVA-abliterated-TIES-Qwen2.5-14B
KiyoEditz/EVA-abliterated-TIES-Qwen2.5-14B is a 14.8 billion parameter language model merged from Qwen2.5-14B and other Qwen2.5-14B instruction-tuned variants using the TIES method. This model leverages the Qwen2.5 architecture and a 32K context length, combining the strengths of its constituent models. It is designed for general-purpose language tasks, supporting a wide array of languages including English, Chinese, French, Spanish, and more.
Loading preview...
Overview
KiyoEditz/EVA-abliterated-TIES-Qwen2.5-14B is a 14.8 billion parameter language model created by merging several pre-trained Qwen2.5-14B variants. This model utilizes the TIES (Trimmed, Iterative, and Selective) merge method, which combines the weights of multiple models to create a new, potentially more capable model. The base model for this merge is Qwen/Qwen2.5-14B, known for its robust performance and 32K token context length.
Key Capabilities
- Multilingual Support: Inherits broad language capabilities from its base models, supporting languages such as English, Chinese, French, Spanish, Portuguese, German, Italian, Russian, Japanese, Korean, Vietnamese, Thai, and Arabic.
- Merged Intelligence: Combines the instruction-following and general knowledge from
huihui-ai/Qwen2.5-14B-Instruct-abliterated-v2andEVA-UNIT-01/EVA-Qwen2.5-14B-v0.2. - Qwen2.5 Architecture: Benefits from the advanced architecture of the Qwen2.5 series, offering strong performance in various NLP tasks.
Good For
- General-purpose text generation: Suitable for a wide range of applications requiring coherent and contextually relevant text.
- Instruction-following tasks: Enhanced by the instruction-tuned components, making it effective for tasks requiring specific output formats or responses.
- Multilingual applications: Its extensive language support makes it a versatile choice for global use cases.