TAIDE Model

Llama3-TAIDE models License Agreement
Preambles.
National Institutes of Applied Research (hereinafter referred to as "NIAR") has developed and implemented Llama3-TAIDE models. When You obtain or start using Llama3-TAIDE models and its materials (hereinafter collectively referred to as the "TAIDE L models"), it means that You have read, understood, and agreed to accept all the contents of this Agreement. If You do not agree with the contents of this Agreement, or if Your country or region excludes all or part of the contents of this Agreement, You should immediately stop using TAIDE L models.
1. Definitions.
1.1 License: means NIAR agrees that You may use, reproduce, modify, and distribute TAIDE L models, subject to the terms of this Agreement.
1.2 Licensor: means the legal entity that owns all rights and intellectual property rights to TAIDE L models. The licensor under this Agreement is NIAR or the individuals or other legal entities authorized by NIAR to sublicense to Third Parties.
1.3 You (or "Your"): means the authorized person under this Agreement, namely individuals or other legal entities who may exercise the rights granted in this Agreement and/or use or access TAIDE L models for any purpose, in any field.
1.4 Third Party: means any individual or legal entity other than the "Licensor" or "You."
1.5 TAIDE L models: mean artificial intelligence-based language and dialogue model, software, and algorithm, including machine learning model code, parameters of the trained model, driver code, etc. You may adjust or modify TAIDE L models by using the content or data that You own or that You are authorized to use, in order to develop systems or models that meet Your requirements.
1.6 Derived Model: means a model built by transferring the parameters or output modes of TAIDE L models, a model trained by using synthetic data generated by using TAIDE L models, or a model built by modifying or altering TAIDE L models and having similar performance to TAIDE L models.
1.7 Personal Data: means data that can directly or indirectly identify a specific natural person. The processing of personal data includes but is not limited to recording, inputting, storing, editing, correcting, duplicating, retrieving, deleting, outputting, connecting, or internally transferring.
Load MoreGemma -TAIDE Models License Agreement
Preambles.
The National Institutes of Applied Research, Taiwan ("NIAR") has developed and deployed the “Gemma – TAIDE Model” (hereinafter referred to as the “TAIDE G Model”). By obtaining , accessing or using the TAIDE G Model its associated materials, you hereby acknowledge and agree that you have read, understood, and accepted all of the terms and conditions set forth in this License Agreement , and you further agree to be legally bound by them. If you do not accept or be bound by this License Agreement, in whole or in part, or if the laws or regulations of your jurisdiction prohibit or restrict your acceptance of any part of this License Agreement, you must immediately cease all access to and use of the TAIDE G Model.
1. Definitions.
1.1 License Grant: Refers to NIAR’s consent, under the conditions set forth in this License Agreement, allowing you to use, reproduce, modify, distribute, and otherwise exploit the TAIDE G Model.
1.2 Licensor: Refers to the legal entity holding ownership and intellectual property rights in the TAIDE G Model. Under this License Agreement, the Licensor is the National Institutes of Applied Research (NIAR), or any individual or legal entity authorized by NIAR to grant sublicenses to third parties.
1.3 You (or “Your”): Refers to the licensee under this Agreement—any individual or legal entity exercising the rights granted herein and using or accessing the TAIDE G Model for any purpose or in any domain.
1.4 Third Party: Refers to any individual or legal entity other than the Licensor and You.
1.5 TAIDE G Model: Refers to an artificial intelligence-based foundational model for language and dialogue, including machine learning model code, parameters of the trained model, driver code, and other associated software and algorithms. It also includes content or data that you own or are authorized to use, which may be adapted or modified using these materials to develop systems or models suited to your specific needs.
1.6 Derivative Model: Refers to any model constructed by transferring the parameters or output patterns of the TAIDE G Model, any model trained using synthetic data generated by the TAIDE G Model, or any model developed by modifying or altering the TAIDE G Model to achieve similar functionality.
1.7 Personal Data: Refers to any information that can identify a specific natural person, directly or indirectly. The processing of personal data includes, but is not limited to, recording, input, storage, editing, correction, duplication, retrieval, deletion, export, linking, or internal transmission.
Load Morenotes:
TAIDE-LX-7B、TAIDE-LX-7B-Chat、TAIDE-LX-7B-Chat-4bit、Llama3-TAIDE-LX-8B-Chat-Alpha1、Llama3-TAIDE-LX-8B-Chat-Alpha1-4bit、Llama-3.1-TAIDE-LX-8B-Chat
Gemma-TAIDE models License Agreement applies to the released model:
Gemma-3-TAIDE-12b-Chat。
·TAIDE Collaboration Model
TAIDE is actively seeking partners to jointly promote the development of Taiwan's Large Language Models (LLMs).
We are earnestly inviting industry professionals, academicians, and anyone with relevant technical expertise to participate in this milestone collaboration!
TAIDE aims to ground itself in Taiwanese culture, incorporating unique linguistic elements, values, and customs. This approach enables the generative AI engine to better understand and respond to local user needs, creating a trustworthy foundational model for generative AI engines. These models can be applied across various domains to meet the diverse needs of users.
The basic collaboration plan offered by TAIDE includes providing partners with a large language model as a powerful tool for applications in various fields, such as natural language processing, machine translation, intelligent customer service, and other creative possibilities.
TAIDE represents a new opportunity to advance technology and expand its applications, bringing tangible value and outcomes to partners and society at large. We look forward to responses from interested parties worldwide, eager to explore the limitless potential of LLMs.
If you or your organization are interested in this collaboration, please contact us for more detailed information and discuss potential partnerships.
TAIDE Project Team: TAIDE@narlabs.org.tw
- What Does TAIDE Offer?
- Large Language Model: Provision of a 7-billion-parameter language model to assist partners in achieving their application goals.
- Technical Support and Training: Offering technical interactions, knowledge sharing, relevant technical support and training to ensure smooth integration and usage by partners.
- Confidentiality and Security Management: TAIDE will sign clear cooperation agreements with partners to protect their rights and intellectual property, covering data usage rights, intellectual property ownership, and confidentiality.
- What Do Partners Need to Prepare?
- Dataset Requirements: Partners need to provide their datasets for training and fine-tuning the model, ensuring appropriate data support for practical applications.
- Relevant Information Capabilities: Partners should possess relevant information capabilities and resources, including data processing, model training expertise and computational resources to ensure the quality and security of datasets.
- Application Proposal: Partners are required to provide a proposal compatible with the TAIDE model, aligning with the objectives of promoting generative AI technology development and application.
The following types of applications will be prioritized by TAIDE in its call for projects, but of course, TAIDE always welcomes more creative contributions:
- Automated Knowledge Base Management
- Software News Translation Tools
- Multilingual Support Tools
- AI Customer Service
- API Service
- Web Service