OLMo-3-7B-Think-LOPD

This model is fine-tuned from OLMo-3-7B-Think using Latent On-Policy Self-Distillation (LOPD) on the code generation domain.

For more details, see our GitHub repository and paper.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("liunanfu1992/OLMo-3-7B-Think-LOPD", torch_dtype="bfloat16", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("liunanfu1992/OLMo-3-7B-Think-LOPD")

Downloads last month: 35

Safetensors

Model size

7B params

Tensor type

BF16

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for liunanfu1992/OLMo-3-7B-Think-LOPD

Quantizations

1 model

Paper for liunanfu1992/OLMo-3-7B-Think-LOPD

Latent On-Policy Self-Distillation

Paper • 2608.13040 • Published 3 days ago