metadata
license: apache-2.0
For convenience, I'm keeping smaller LoRA adapters I've extracted from various models for Qwen2.5 14B in this single repository:
- CultriX/Qwen2.5-14B-Wernicke - Base Qwen2.5-14B-Instruct, rank 128.
- huihui-ai/Qwen2.5-Coder-14B-Instruct-abliterated - Base Qwen2.5-Coder-14B-Instruct, rank 32.
- tanliboy/lambda-qwen2.5-14b-dpo-test - Base Qwen2.5-14B-Instruct, rank 128.
- tanliboy/lambda-qwen2.5-14b-dpo-test - Base Qwen2.5-14B-Instruct, rank 32.
This repo's file structor follows thes naming scheme: author/model name/base model author/base model name-rank##