metadata
library_name: peft
model_name: gemma-roleplay-v2-lora
tags:
- base_model:adapter:google/gemma-4-E4B-it
- lora
- sft
- transformers
- trl
- roleplay
license: gemma
base_model: google/gemma-4-E4B-it
pipeline_tag: text-generation
Gemma Roleplay v2 LoRA adapter
This is the PEFT adapter for Gemma Roleplay v2, trained from google/gemma-4-E4B-it with QLoRA SFT. It is intended for fictional consenting-adult roleplay and creative chat. See the parent model card for usage, limitations, and the live hosted inference endpoint.
Quick start
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
base = "google/gemma-4-E4B-it"
tokenizer = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, torch_dtype="auto", device_map="auto")
model = PeftModel.from_pretrained(model, "text-generator/llmtrain", subfolder="adapter")
inputs = tokenizer.apply_chat_template(
[{"role": "user", "content": "Write a short scene in a haunted hotel."}],
add_generation_prompt=True, return_tensors="pt",
).to(model.device)
output = model.generate(inputs, max_new_tokens=128, do_sample=True, temperature=0.85)
print(tokenizer.decode(output[0, inputs.shape[-1]:], skip_special_tokens=True))
Training procedure
This model was trained with SFT.
Framework versions
- PEFT 0.18.0
- TRL: 1.8.0
- Transformers: 5.5.0
- Pytorch: 2.9.1
- Datasets: 4.0.0
- Tokenizers: 0.22.2
Citations
Cite TRL as:
@software{vonwerra2020trl,
title = {{TRL: Transformers Reinforcement Learning}},
author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
license = {Apache-2.0},
url = {https://github.com/huggingface/trl},
year = {2020}
}