Novelist1.0-27b-Adapter
LoRA for Qwen/Qwen3.8-27B: the same Novelist1.0 writer as the merged 16-bit repo — literary roleplay and fiction in English and Turkish.
r = 64, α = 64, dropout 0. Targets: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj.
Load this adapter on the base model, or skip PEFT and use the merged weights. Prompting, sampling, and what the model is for: see the merged card.
- Downloads last month
- 32
Model tree for Dxniz/Novelist1.0-27b-Adapter
Base model
Qwen/Qwen3.8-27B
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.8-27B") model = PeftModel.from_pretrained(base_model, "Dxniz/Novelist1.0-27b-Adapter")