---
base_model: Qwen/Qwen3-8B-Base
library_name: transformers
license: apache-2.0
tags:
- llama-factory
- full
- generated_from_trainer
- math-reasoning
pipeline_tag: text-generation
model-index:
- name: DiffGen-8B
  results: []
---

Paper: [ScaleDiff: Scaling Difficult Problems for Advanced Mathematical Reasoning](https://arxiv.org/abs/2509.21070)

Code: https://github.com/QizhiPei/ScaleDiff

# DiffGen-8B

This model is a fine-tuned version of [Qwen/Qwen3-8B-Base](https://huggingface.co/Qwen/Qwen3-8B-Base).

## Model description

DiffGen-8B is a specialized difficult problem generator developed as part of the ScaleDiff pipeline, an approach designed to scale the creation of challenging mathematical problems for advanced mathematical reasoning. The model is trained on a filtered dataset of difficult problems, enabling it to efficiently produce a vast number of new, complex mathematical problems. This process eliminates the need for complex, per-instance prompting and its associated high API costs, addressing the scarcity of high-quality, difficult training data for Large Reasoning Models (LRMs).

## Intended uses & limitations

**Intended Uses**: DiffGen-8B is primarily intended for generating large-scale datasets of challenging mathematical problems. These generated problems are then used to augment training data for Large Reasoning Models (LRMs), thereby enhancing their mathematical reasoning capabilities. It serves as a crucial component in pipelines focused on improving LRM performance on difficult benchmarks by providing a continuous supply of intricate reasoning challenges.

**Limitations**: While DiffGen-8B excels at generating difficult problems, its primary scope is mathematical problem generation. The quality and relevance of the generated problems are further ensured through subsequent solution distillation and filtering steps within the broader ScaleDiff pipeline. Its performance may not be optimized for other general text generation tasks.

## Training and evaluation data

DiffGen-8B is a fine-tuned version of [Qwen/Qwen3-8B-Base](https://huggingface.co/Qwen/Qwen3-8B-Base). It was trained on a subset of difficult problems selected from the [AM-Qwen3-Distilled](https://huggingface.co/datasets/a-m-team/AM-Qwen3-Distilled) dataset. This selection was performed efficiently using [AdaptThink](https://huggingface.co/THU-KEG/AdaptThink-7B-delta0.05), an adaptive thinking model that perceives problem difficulty with only a single forward pass, eliminating the need for solutions during selection. The problems generated by DiffGen-8B contribute to the creation of the [ScaleDiff-Math](https://huggingface.co/datasets/QizhiPei/ScaleDiff-Math) dataset.

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- total_train_batch_size: 128
- total_eval_batch_size: 64
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1.0

### Training results



### Framework versions

- Transformers 4.52.0.dev0
- Pytorch 2.6.0+cu124
- Datasets 2.21.0
- Tokenizers 0.21.1