Model Card for manjaveca/emilyrae140922

This repository contains a PEFT adapter that fine-tunes the
NousResearch/Hermes-3-Llama-3.2-3B
large language model to emulate the “Emily Rae” persona.

Model Details

Model Description

A lightweight adapter that, when applied to the base Llama-3.2-3B model,
steers its responses to match the style, tone, and content strategy
defined for the “Emily Rae” conversational agent.

• Developed by: Manja Veca
• Model type: Causal language model (PEFT-adapter)
• Language(s): English
• License: Apache-2.0
• Finetuned from: NousResearch/Hermes-3-Llama-3.2-3B

Uses

Direct Use

Load the adapter on top of the base model to get “Emily Rae” behavior:

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

tokenizer = AutoTokenizer.from_pretrained("manjaveca/emilyrae140922")
base = AutoModelForCausalLM.from_pretrained(
    "NousResearch/Hermes-3-Llama-3.2-3B",
    torch_dtype=torch.float16,
    device_map="auto"
)
model = PeftModel.from_pretrained(base, "manjaveca/emilyrae140922", torch_dtype=torch.float16)
2