ldwang/DeepScaleR-1.5B-Preview-Reproduce

DeepScaleR-1.5B-Preview-Reproduce

Overview

This model is a reproduction of the agentica-project/deepscaler project. We have reproduced the results in the repo on an 8x80G node, achieving an average score of 56.7.

Training

export VLLM_ATTENTION_BACKEND=XFORMERS

# Run 8K context length training, 580 steps
export MODEL_PATH="deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B"
nohup bash run_deepscaler_1.5b_8k.sh --model $MODEL_PATH > stage1.log 2>&1 &

# Run 16K context length training, 430 steps
export MODEL_PATH="./checkpoints/deepscaler/deepscaler-1.5b-8k/actor/global_step_580"
nohup bash run_deepscaler_1.5b_16k.sh --model $MODEL_PATH > stage2.log 2>&1 &

# Run 24K context length training, 430 steps
export MODEL_PATH="./checkpoints/deepscaler/deepscaler-1.5b-16k/actor/global_step_430"
nohup bash run_deepscaler_1.5b_24k.sh --model $MODEL_PATH > stage3.log 2>&1 &

Evaluation

Model	AIME 2024	MATH 500	AMC 2023	Minerva Math	OlympiadBench	Avg.
Qwen-2.5-7B-Instruct	13.3	79.8	50.6	34.6	40.7	43.8
rStar-Math-7B	26.7	78.4	47.5	-	47.1	-
Eurus-2-7B-PRIME	26.7	79.2	57.8	38.6	42.1	48.9
Qwen2.5-7B-SimpleRL	26.7	82.4	62.5	39.7	43.3	50.9
DeepSeek-R1-Distill-Qwen-1.5B	28.8	82.8	62.9	26.5	43.3	48.9
Still-1.5B	32.5	84.4	66.7	29.0	45.4	51.6
DeepScaleR-1.5B-Preview	43.1	87.8	73.6	30.2	50.0	57.0
DeepScaleR-1.5B-Preview-Reproduce	40.4	87.9	72.0	31.5	50.2	56.4
🎉 DeepScaleR-1.5B-Preview-Reproduce	42.3	88.0	73.2	30.3	49.7	56.7
O1-Preview	40.0	81.4	-	-	-	-

Citation

@misc{deepscaler2025,
  title={DeepScaleR: Surpassing O1-Preview with a 1.5B Model by Scaling RL},
  author={Michael Luo and Sijun Tan and Justin Wong and Xiaoxiang Shi and William Y. Tang and Manan Roongta and Colin Cai and Jeffrey Luo and Tianjun Zhang and Li Erran Li and Raluca Ada Popa and Ion Stoica},
  year={2025},
  howpublished={\url{https://pretty-radio-b75.notion.site/DeepScaleR-Surpassing-O1-Preview-with-a-1-5B-Model-by-Scaling-RL-19681902c1468005bed8ca303013a4e2}},
  note={Notion Blog}
  year={2025}
}