You can deploy this model and start earning money today!

kmseong/
llama3_2_3b-instruct-safe-lora-math_safety_later20_lr2e-4

Name: llama3_2_3b-instruct-safe-lora-math_safety_later20_lr2e-4
Availability: InStock

Views

No views yet

llama3_2_3b-safe-lora-math_safety_later20_lr2e-4

This is a merged full model uploaded from a parameter-efficient fine-tuning checkpoint.

Model Description

Base Model: kmseong/llama3_2_3b-instruct-SSFT-lr5e-5
Fine-tuning Method: Safe LoRA (MATH)
Upload Source: /home/yonsei_jong/SafeLoRA/safe_lora_models/llama2-7b-safe-lora-math-20260429-005256
Upload Date: 2026-04-29 01:13:51

Notes

If the source directory contained adapter weights (for example Safe LoRA), this upload script merged them with the base model first. So this repository contains directly loadable full model weights.

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "kmseong/llama3_2_3b-safe-lora-math_safety_later20_lr2e-4"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Generate text
prompt = "How can I help you today?"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_length=100)
print(tokenizer.decode(outputs[0]))

Safety Note

This model was produced with Safe LoRA (MATH) and uploaded as a merged full model.

License

This model is licensed under the Apache 2.0 License. See the base model (kmseong/llama3_2_3b-instruct-SSFT-lr5e-5) for more details.

References

Base model: https://huggingface.co/kmseong/llama3_2_3b-instruct-SSFT-lr5e-5
Method: Safe LoRA (MATH)

kmseong/llama3_2_3b-instruct-safe-lora-math_safety_later20_lr2e-4