Views
No views yet
LoRA adapter for Qwen/Qwen2.5-32B-Instruct — the builder personality from the Buleyean affective-RL series. Powered by Aether / Edgework.ai.
| Property | Value |
|---|---|
| Type | LoRA adapter (PEFT) |
| Base model | Qwen/Qwen2.5-32B-Instruct |
| Personality | builder |
| Adapter | lora/adapter_model.safetensors |
| License | apache-2.0 |
1from peft import PeftModel
2from transformers import AutoModelForCausalLM, AutoTokenizer
3
4base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-32B-Instruct")
5model = PeftModel.from_pretrained(base, "forkjoin-ai/buleyean-qwen2.5-32b-builder", subfolder="lora")
6tok = AutoTokenizer.from_pretrained("forkjoin-ai/buleyean-qwen2.5-32b-builder", subfolder="lora")