| license: apache-2.0 | |
| library_name: transformers | |
| base_model: | |
| - arcee-ai/Trinity-Mini | |
| tags: | |
| - trinity | |
| - moe | |
| - merged-lora | |
| # Trinity Mini Quirky | |
| This repository contains `arcee-ai/Trinity-Mini` with the `lora-1-quirky` adapter merged into the base weights. | |
| ## Merge Details | |
| - Base model: `arcee-ai/Trinity-Mini` | |
| - Adapter source: `lora-1-quirky/Quirky-cr8a7qzr4ay8yhtvjup2ijl1.zip` | |
| - Merge script: `merge_lora.py` | |
| - Confirmed applied tensors: `11,686 / 11,686` | |
| - LoRA rank: `16` | |
| - LoRA alpha: `64` | |
| - Merge scale: `4.0` | |
| ## Usage | |
| Load this model the same way you would load `arcee-ai/Trinity-Mini`, including `trust_remote_code=True` when required by your environment. | |