FreedomIntelligence/AceGPT-7B

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kPublished:Sep 14, 2023License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

AceGPT-7B is a 7 billion parameter generative text model developed by researchers from CUHKSZ, SRIBD, and KAUST. Based on the Llama2 architecture, this model is specifically fine-tuned for the Arabic language domain. It demonstrates strong performance in Arabic benchmarks, particularly in dialogue applications, and is part of a collection that includes optimized chat versions. AceGPT-7B is designed for Arabic natural language generation and understanding tasks.

Loading preview...

AceGPT-7B: A Llama2-Based Model for Arabic Language Tasks

AceGPT-7B is a 7 billion parameter model from the AceGPT family, developed by researchers at the Chinese University of Hong Kong, Shenzhen (CUHKSZ), the Shenzhen Research Institute of Big Data (SRIBD), and King Abdullah University of Science and Technology (KAUST). This model is a fully fine-tuned generative text model built upon the Llama2 architecture, with a primary focus on the Arabic language domain.

Key Capabilities and Features

  • Arabic Language Specialization: AceGPT-7B is specifically optimized for generating and understanding text in Arabic.
  • Dialogue Optimization: The AceGPT family includes -chat versions, which are specifically designed and optimized for dialogue applications, demonstrating superior performance among open-source Arabic dialogue models.
  • Benchmark Performance: In evaluations on Arabic MMLU and EXAMs, AceGPT-7B-base shows competitive performance against other Llama2 variants and Bloomz, with the larger AceGPT-13B-base achieving the second-best scores across multiple categories, only surpassed by ChatGPT.
  • Text-to-Text: The model accepts text as input and generates text as output.

Ideal Use Cases

  • Arabic Content Generation: Creating various forms of text content in Arabic.
  • Arabic Chatbots and Dialogue Systems: Particularly the -chat variants are well-suited for building conversational AI in Arabic.
  • Arabic Language Understanding: Applications requiring comprehension and analysis of Arabic text.
  • Research and Development: As a strong baseline for further fine-tuning or research in Arabic NLP.

Finetunes
3 models