ynklab/Qwen2.5-7B-Sep_2c1t

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 7, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

ynklab/Qwen2.5-7B-Sep_2c1t is a 7.6 billion parameter language model developed by ynklab, fine-tuned from Qwen/Qwen2.5-7B-Instruct. It specializes in multilingual chunk-level machine translation, supporting 10 languages including English, Chinese, German, and French, with a 32768 token context length. This model is specifically designed for document-level translation by processing current source chunks with two preceding source-language context chunks.

Loading preview...

Overview

ynklab/Qwen2.5-7B-Sep_2c1t is a 7.6 billion parameter model, fine-tuned from Qwen/Qwen2.5-7B-Instruct, specifically for multilingual chunk-level machine translation. Developed by ynklab as part of the Doc2FRC project, this Sep_2c1t variant processes a current source chunk by leveraging two preceding source-language context chunks, enabling length-consistent document-level translation.

Key Capabilities

  • Document-Level Machine Translation: Optimized for translating documents by considering context from previous chunks, crucial for maintaining coherence.
  • Multilingual Support: Translates between English and 9 other languages, including Chinese, German, Spanish, French, Italian, Korean, Dutch, Portuguese, and Russian.
  • Contextual Translation: Utilizes a unique prompting format where two preceding source-language context chunks (Context1, Context2) are provided alongside the current source text for translation.
  • Fixed-Range Chunking: Trained on examples derived from the sardinelab/DocBlocks dataset using fixed-range chunks of 256–512 tokens.
  • High Context Length: Features a maximum sequence length of 32,768 tokens, allowing for substantial contextual input.

Good For

  • Developers and researchers working on document-level machine translation tasks.
  • Applications requiring context-aware translation across supported language pairs.
  • Integration into systems that can manage fixed-range chunking for translation workflows. For detailed inference scripts and document chunking procedures, refer to the Doc2FRC GitHub repository.

Finetunes
1 models