S1K sc LoRA adapter
LoRA adapter for Qwen/Qwen3-1.7B, exported from training checkpoint step 2.
- Variant:
sc - Source run:
s1k30-data42-a40-lora16-seed42-4168506-sc - LoRA rank: 16
- LoRA alpha: 32
This repository contains adapter weights. Load them with the base model:
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
base_model = "Qwen/Qwen3-1.7B"
tokenizer = AutoTokenizer.from_pretrained(base_model)
model = AutoModelForCausalLM.from_pretrained(
base_model, torch_dtype="auto", device_map="auto"
)
model = PeftModel.from_pretrained(model, "talzoomanzoo/s1k-sc")
model.eval()
- Downloads last month
- 9