wphuirtp/paper_helper

TEXT GENERATIONConcurrency Cost:1Model Size:14.8BQuant:FP8Ctx Length:32kPublished:Feb 5, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

wphuirtp/paper_helper is a 14.8 billion parameter Qwen2-based language model, fine-tuned by wphuirtp from unsloth/DeepSeek-R1-Distill-Qwen-14B-unsloth-bnb-4bit. It was trained using Unsloth and Huggingface's TRL library, achieving 2x faster training. The model's training data includes content from books on harmonic maps and geometric analysis, alongside the unsloth/OpenMathReasoning-mini dataset, suggesting a specialization in mathematical reasoning and complex analytical topics. It features a substantial context length of 131072 tokens, making it suitable for processing extensive documents.

Loading preview...

Overview

wphuirtp/paper_helper is a 14.8 billion parameter language model, fine-tuned by wphuirtp. It is based on the Qwen2 architecture, specifically leveraging unsloth/DeepSeek-R1-Distill-Qwen-14B-unsloth-bnb-4bit as its base model. This model was developed with the assistance of Unsloth and Huggingface's TRL library, which enabled a 2x faster training process.

Key Capabilities

  • Specialized Training Data: Fine-tuned on a unique dataset comprising content from books on harmonic maps and geometric analysis, combined with the unsloth/OpenMathReasoning-mini dataset (approximately 6000 + 5000 QA pairs).
  • Enhanced Training Efficiency: Utilizes Unsloth for significantly faster training, optimizing resource usage.
  • Large Context Window: Features a substantial context length of 131072 tokens, allowing for the processing and understanding of very long texts.

Good For

  • Mathematical and Geometric Analysis: Its specialized training data makes it particularly well-suited for tasks involving complex mathematical reasoning, especially in areas like harmonic maps and geometric analysis.
  • Academic Research Assistance: Ideal for researchers and students working with extensive academic papers or books in its domain of expertise, leveraging its large context window.
  • Question Answering on Technical Subjects: Capable of answering questions based on the specific technical content it was trained on, particularly in mathematical reasoning.

Finetunes
2 models