metadata
language: id
tags:
- mistral
- merged
- lora
- linear-merge
- vision-language
base_model: mistralai/Ministral-3-3B-Base-2512
PAD Model — Linear Adapter Merge
Merged model from 3 LoRA adapters on mistralai/Ministral-3-3B-Base-2512 using linear delta merging (Task Arithmetic).
Merge Details
| Adapter | Weight |
|---|---|
| PAD-1 checkpoint-672 | 0.3 |
| PAD-2 checkpoint-3382 | 0.2 |
| VLM-IGRS | 0.5 |
Merge Formula
final = base + (adapter1 - base)*0.3 + (adapter2 - base)*0.2 + (adapter3 - base)*0.5
Chat Template
Borrowed from unsloth/Ministral-3-3B-Instruct-2512.
Usage
from transformers import AutoModelForImageTextToText, AutoProcessor
model = AutoModelForImageTextToText.from_pretrained("nuresens/pad_model_linier")
processor = AutoProcessor.from_pretrained("nuresens/pad_model_linier")