Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V7-70B

Hugging Face
TEXT GENERATIONPricing:Input $2.88 / Output $2.88Concurrent Unit Cost:4Model Size:70BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Dec 31, 2024License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Warm

Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V7-70B is a 70 billion parameter Japanese-enhanced language model developed by Mr. Yunsung Ji (Saxo) at Linkbricks. It was fine-tuned using SFT, DPO, and MERGE techniques on a 40 million Japanese news and wiki corpus. This model excels in cross-lingual tasks across Japanese, Korean, Chinese, and English, and is specifically trained to handle complex logical and mathematical problems, making it suitable for high-dimensional analysis of customer reviews, coding, writing, and logical reasoning.

Loading preview...

Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V7-70B Overview

Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V7-70B is a 70 billion parameter language model developed by Mr. Yunsung Ji (Saxo), a data scientist at Linkbricks. This model is an enhanced version, fine-tuned from the Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V6-70B base model using SFT, DPO, and MERGE training on 8 H100-80G GPUs. A key focus of its training involved a 40 million Japanese news and wiki corpus, alongside cross-training data for Japanese, Korean, Chinese, and English, and specialized data for mathematics and logical reasoning.

Key Capabilities

  • Multilingual Proficiency: Enhanced for cross-lingual tasks across Japanese, Korean, Chinese, and English.
  • Advanced Reasoning: Trained to address complex logical and mathematical problems.
  • High-Dimensional Analysis: Optimized for in-depth analysis of customer reviews and social media posts.
  • Code and Writing Assistance: Improved capabilities in coding and general writing tasks.
  • Function Calling: Supports function calling and tool calling, enabling integration with external systems.
  • Extended Context Window: Features a 128k-Context Window for processing longer inputs.

Good for

  • Applications requiring robust Japanese language understanding and generation.
  • Cross-lingual communication and analysis involving Japanese, Korean, Chinese, and English.
  • Tasks demanding strong logical reasoning and mathematical problem-solving.
  • Analyzing large volumes of customer feedback or social media data.
  • Developing intelligent agents that require function calling capabilities.

Finetunes
1 models