Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V3-70B

TEXT GENERATIONPricing:Input $3.5 / Cached $0.7 / Output $8.3Concurrent Unit Cost:4Model Size:70BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Dec 10, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V3-70B is a 70 billion parameter Japanese-enhanced language model developed by Yunsung Ji (Saxo) at Linkbricks. It was fine-tuned using SFT, DPO, and MERGE methods on 50 million Japanese news and Wiki corpora, alongside cross-lingual data for Japanese, Korean, Chinese, and English, and data for mathematics and logical reasoning. This model excels at high-dimensional analysis of customer reviews and social posts, coding, writing, mathematics, and complex logical problem-solving, featuring a 128k context window and Function Calling support.

Loading preview...

Overview

Saxo/Linkbricks-Horizon-AI-Japanese-Pro-V3-70B is a 70 billion parameter, Japanese-enhanced language model developed by Yunsung Ji (Saxo), a data scientist at Linkbricks. This model was fine-tuned from the Saxo/Linkbricks-Horizon-AI-Japanese-Base-70B using SFT, DPO, and MERGE techniques on 8 H100-80G GPUs. Its training incorporated 50 million Japanese news and Wiki corpora, along with cross-lingual data for Japanese, Korean, Chinese, and English, and specialized data for mathematics and logical reasoning, enabling it to handle complex logical problems and cross-language tasks.

Key Capabilities

  • Enhanced Japanese Processing: Optimized for Japanese language tasks, leveraging extensive Japanese news and Wiki corpora.
  • Multilingual Cross-Learning: Trained with Japanese, Korean, Chinese, and English cross-learning data for improved performance across these languages.
  • Advanced Analytical Skills: Excels in high-dimensional analysis of customer reviews and social posts.
  • Coding and Writing: Strong capabilities in code generation and various writing tasks.
  • Mathematical and Logical Reasoning: Specifically trained to address complex mathematical and logical problems.
  • Extended Context Window: Supports a 128k context window for processing longer inputs.
  • Function Calling: Includes support for Function Calling and Tool Calling.

Good for

  • Applications requiring deep analysis of Japanese text, such as customer feedback or social media monitoring.
  • Multilingual projects involving Japanese, Korean, Chinese, and English.
  • Tasks demanding strong logical reasoning and mathematical problem-solving.
  • Development of intelligent agents that require Function Calling capabilities.

Finetunes
1 models