A commercial-friendly small language model by NVIDIA optimized for roleplay, RAG QA, and function calling.

Tools

14.9K Pulls Updated 2 weeks ago

2 weeks ago

ed76ab18784f · 2.7GB

model
nemotron
·
4.19B
·
Q4_K_M
template
{{- if (or .Tools .System) }}<extra_id_0>System {{ if .System }}{{ .System }} {{ end }} {{- if .Tools }} {{- range .Tools }}<tool> {{ . }} </tool>{{ end }} {{ end }} {{- end }} {{- range $i, $m := .Messages }} {{- $last := eq (len (slice $.Messages $i)) 1 -}} {{- if eq .Role "user" }}<extra_id_1>User {{ .Content }} {{- if $last }} <extra_id_1>Assistant {{- end }} {{ else if eq .Role "tool" }}<extra_id_1>Tool {{ .Content }} {{- if $last }} <extra_id_1>Assistant {{- end }} {{ else if eq .Role "assistant" }}<extra_id_1>Assistant {{- if .ToolCalls }} {{ range .ToolCalls }}<toolcall> {"name": "{{ .Function.Name }}", "arguments": {{ .Function.Arguments }}} </toolcall> {{ end }} {{ else }} {{ .Content }} {{- if not $last }} {{ end }} {{- end }} {{- end }} {{- end }}
license
NVIDIA AI Foundation Models Community License Agreement IMPORTANT NOTICE – PLEASE READ AND AGREE BEFORE USING THE NVIDIA AI FOUNDATION MODELS. This license agreement (“Agreement”) is a legal agreement between you, whether an individual or entity (“you”) and NVIDIA Corporation (“NVIDIA”) and governs the use of the NVIDIA AI Foundation Models, consisting of large language models including software, checkpoints, learnt weights, algorithms, parameters, configuration files, datasets and documentation, if and when available (collectively, “NVIDIA AI Foundation Models”). This Agreement can be accepted only by an adult of legal age of majority in the country in which the NVIDIA AI Foundation Models are used. If you don’t have the required age or authority to accept this Agreement, or if you don’t accept all the terms and conditions of this Agreement, do not use the NVIDIA AI Foundation Models. 1. License Grant. Subject to the terms of this Agreement, NVIDIA grants you a non-exclusive, revocable, nontransferable, non-sublicensable (except as expressly granted in this Agreement), license to: 1.1 (a) install and use copies of the NVIDIA AI Foundation Models, (b) create Derivative Models(as defined below), and (c) distribute the NVIDIA AI Foundation Models or Derivative Models in each case with a copy of this Agreement, all the foregoing for non-production uses, including test and evaluation uses; and 1.2 (a) install and use copies of the NVIDIA AI Foundation Models, (b) create Derivative Models (as defined below), and (c) distribute the NVIDIA AI Foundation Models or Derivative Models in each case with a copy of this Agreement, all the foregoing in production uses with a paid subscription to NVIDIA NeMo Framework available through NVIDIA AI Enterprise. Contact NVIDIA sales for NVIDIA AI Enterprise (see https://www.nvidia.com/enus/data-center/products/ai-enterprise/). “Derivative Models” mean all derivatives, adaptations, extensions or enhancements to the NVIDIA AI Foundation Models, works based on the NVIDIA AI Foundation Models, or any other model which is created or initialized by transfer of patterns of the weights, parameters, activations or output of the NVIDIA AI Foundation Models, to the other model, in order to cause the other model to perform similarly to the NVIDIA AI Foundation Models, including - but not limited to - distillation methods entailing the use of intermediate data representations or methods based on the generation of synthetic data by the NVIDIA AI Foundation Models for training the other model. 2. Use and Distribution. The following terms apply to the use or distribution of the NVIDIA AI Foundation Models and Derivative Models you develop as authorized in Section 1, and you are responsible for the consequences of non-conformance with these terms: 2.1 The NVIDIA AI Foundation Models and Derivative Models must only be used or distributed for purposes that are: permitted by this Agreement, consistent with the rights of others, and do not violate any applicable laws or regulations. 2.2 You may not use the NVIDIA AI Foundation Models, Derivative Models or any output or results of them to develop or improve any other large language model (excluding NVIDIA AI Foundation Models or Derivative Models). 2.3 You may not change or remove copyright or other proprietary notices in the NVIDIA AI Foundation Models or misrepresent the authorship of the NVIDIA AI Foundation Models, and you must cause any modified files to carry prominent notices stating that you changed the files such that modifications are not misrepresented as an original NVIDIA AI Foundation Models. Except as necessary to give attribution to works as described in this section, you are not granted any trademark license under this Agreement. 2.4 You may not use the NVIDIA AI Foundation Models in any manner that would cause them to become subject to an open source software license; subject to the terms in the “Components Under Other Licenses” section below. 2.5 You agree to defend, indemnify and hold harmless NVIDIA and its affiliates, and their respective employees, contractors, agents, officers and directors, from and against any and all claims, damages, obligations, losses, liabilities, costs or debt, fines, restitutions and expenses (including but not limited to attorney’s fees and costs incident to establishing the right of indemnification) arising out of: (i) products or services that have been developed or deployed with or use the NVIDIA AI Foundation Models or Derivative Models (including results or data generated from such use), or claims that they violate laws, or infringe, violate, or misappropriate any third party rights; or (ii) a violation of the terms and conditions of this Agreement. 3. Your Privacy: Collection and Use of Information. 3.1 Privacy Policy. Please review the NVIDIA Privacy Policy, located at https://www.nvidia.com/en-us/aboutnvidia/privacy-policy, which explains NVIDIA’s policy for collecting and using data, as well as visit the NVIDIA Privacy Center, located at https://www.nvidia.com/en-us/privacy-center, to manage your consent and privacy preferences. 3.2 Collection Purposes. NVIDIA may require certain personal information such as name, email address or entitlement information to provide you access to the NVIDIA AI Foundation Models. 3.3 Third Party Privacy Practices. The NVIDIA AI Foundation Models may contain links to third party websites and services. NVIDIA encourages you to review the privacy statements on those sites and services that you choose to visit to understand how they may collect, use and share your data. NVIDIA is not responsible for the privacy statements or practices of third party sites or services. 4. AI Ethics. Technology can have a profound impact on people and the world, and NVIDIA is committed to enabling trust and transparency in AI development. NVIDIA encourages you to adopt principles of AI ethics and trustworthiness to guide your business decisions by doing the following: 4.1 Ethical Use. Ensure the product or service you develop, use, offer as a service or distribute meets the ethical requirements of the relevant industry or use case and that you have taken reasonable measures to address bias and inform users of the limitations of the product or service. 4.2 Development Practices. Ensure that in developing your product or service, you adopt adequate measures to mitigate the risk of harming more vulnerable groups, especially those that have been historically disadvantaged or are at risk of exclusion. 4.3 Transparency. Communicate information to stakeholders about the capabilities and limitations of your product or service, including transparency about the fact that stakeholders are dealing with an AI system. 4.4 Distributor or Reseller Duty. If you are a distributor or reseller, ensure you do not sell to anyone who you believe will use your products or services for an improper purpose. 5. Components Under Other Licenses. The NVIDIA AI Foundation Models may include or be distributed with components provided with separate legal notices or terms that accompany the components, such as open source software licenses and other license. The components are subject to the applicable other licenses, including any proprietary notices, disclaimers, requirements and extended use rights; except that this Agreement will prevail regarding the use of third-party open source software, unless a third-party OSS License requires its license terms to prevail. “Open source software license” means any software, data or documentation subject to any license identified as an open source license by the Open Source Initiative (http://opensource.org), Free Software Foundation (http://www.fsf.org) or other similar open source organization or listed by the Software Package Data Exchange (SPDX) Workgroup under the Linux Foundation (http://www.spdx.org). 6. Termination. This Agreement will automatically terminate without notice from NVIDIA if you fail to comply with any of the terms in this Agreement or if you commence or participate in any legal proceeding against NVIDIA with respect to the NVIDIA AI Foundation Models, Derivative Models or any output or results of them. Additionally, NVIDIA may terminate this Agreement with prior written notice to you if, in NVIDIA’s sole discretion, the continued use of the NVIDIA AI Foundation Models is no longer commercially viable or creates liabilities for NVIDIA. Upon any termination, you must stop using and destroy all copies of the NVIDIA AI Foundation Models and Derivative Models. Upon written request, you will certify in writing that you have complied with your commitments under this section. All provisions will survive termination, except for the licenses granted to you. 7. Ownership and Attribution. 7.1 NVIDIA Ownership. The NVIDIA AI Foundation Models, including all intellectual property rights, is and will remain the sole and exclusive property of NVIDIA or its licensors. Except as expressly granted in this Agreement, (i) NVIDIA reserves all rights, interests and remedies in connection with the NVIDIA AI Foundation Models, and (ii) no other license or right is granted to you by implication, estoppel or otherwise. 7.2 Your Ownership. Subject to the rights of NVIDIA and its suppliers in the NVIDIA AI Foundation Models and to the extent permitted by applicable law, as between you and NVIDIA you hold all rights, title and interest in and to: (i) Derivative Models that you develop as permitted in this Agreement including their respective intellectual property rights, and (ii) outputs of the NVIDIA AI Foundation Models or Derivative Models provided that due to the nature of machine learning, your output may not be unique, could be generated by other licensees using the NVIDIA AI Foundation Models or derivatives thereof, and outputs created by other licensees are not yours. 8. Feedback. You may, but are not obligated to, provide suggestions, requests, fixes, modifications, enhancements, or other feedback regarding or in connection with NVIDIA AI Foundation Models (“Feedback”). Feedback, even if designated as confidential by you, will not create any confidentiality obligation for NVIDIA or its affiliates. If you provide Feedback, you hereby grant NVIDIA, its affiliates and its designees a non-exclusive, perpetual, irrevocable, sublicensable, worldwide, royalty-free, fully paid-up and transferable license, under your intellectual property rights, to publicly perform, publicly display, reproduce, use, make, have made, sell, offer for sale, distribute (through multiple tiers of distribution), import, create derivative works of and otherwise commercialize and exploit the Feedback at NVIDIA’s discretion. 9. Disclaimer of Warranties. THE NVIDIA AI FOUNDATION MODELS, OUTPUT OR RESULTS ARE PROVIDED “AS IS”. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, NVIDIA DISCLAIMS ALL WARRANTIES AND REPRESENTATIONS OF ANY KIND, WHETHER EXPRESS, IMPLIED OR STATUTORY, RELATING TO OR ARISING UNDER THIS AGREEMENT, INCLUDING, WITHOUT LIMITATION, THE WARRANTIES OF TITLE, NONINFRINGEMENT, MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, USAGE OF TRADE AND COURSE OF DEALING. WITHOUT LIMITING THE FOREGOING, NVIDIA DOES NOT WARRANT THAT THE NVIDIA AI FOUNDATION MODELS, OUTPUT OR RESULTS WILL MEET YOUR REQUIREMENTS OR THAT ANY DEFECTS OR ERRORS WILL BE CORRECTED. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE NVIDIA AI FOUNDATION MODELS, OUTPUTS OR RESULTS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE NVIDIA AI FOUNDATION MODELS, OUTPUTS OR RESULTS. 10. Limitations of Liability. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, IN NO EVENT WILL NVIDIA BE LIABLE FOR ANY (I) INDIRECT, PUNITIVE, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES, OR (II) DAMAGES FOR THE (A) COST OF PROCURING SUBSTITUTE GOODS OR (B) LOSS OF PROFITS, REVENUES, USE, DATA OR GOODWILL ARISING OUT OF OR RELATED TO THIS AGREEMENT, WHETHER BASED ON BREACH OF CONTRACT, TORT (INCLUDING NEGLIGENCE), STRICT LIABILITY, OR OTHERWISE, AND EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES AND EVEN IF A PARTY'S REMEDIES FAIL THEIR ESSENTIAL PURPOSE. ADDITIONALLY, TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, NVIDIA’S TOTAL CUMULATIVE AGGREGATE LIABILITY FOR ANY AND ALL LIABILITIES, OBLIGATIONS OR CLAIMS ARISING OUT OF OR RELATED TO THIS AGREEMENT WILL NOT EXCEED FIVE U.S. DOLLARS (US$5). 11. Governing Law and Jurisdiction. This Agreement will be governed in all respects by the laws of the United States and the laws of the State of Delaware, without regard to conflict of laws principles or the United Nations Convention on Contracts for the International Sale of Goods. The state and federal courts residing in Santa Clara County, California will have exclusive jurisdiction over any dispute or claim arising out of or related to this Agreement, and the parties irrevocably consent to personal jurisdiction and venue in those courts; except that either party may apply for injunctive remedies or an equivalent type of urgent legal relief in any jurisdiction. 12. Trade and Compliance. You agree to comply with all applicable export, import, trade and economic sanctions laws and regulations, including U.S. Export Administration Regulations and Office of Foreign Assets Control regulations. You confirm that you will not export or reexport any products or technology, directly or indirectly, without first obtaining any required license or other approval from appropriate authorities, (i) to any countries that are subject to any U.S. or local export restrictions (currently including, but not necessarily limited to, Cuba, Iran, North Korea, Syria, the Region of Crimea, Donetsk People’s Republic Region and Luhansk People’s Republic Region); (ii) to any end user who you know or have reason to know will utilize them in the design, development or production of nuclear, chemical or biological weapons, missiles, rocket systems, unmanned air vehicles, or any weapons of mass destruction; (iii) to any end-user who has been prohibited from participating in the U.S. or local export transactions by any governing authority; or (iv) to any known military or military-intelligence end-user or for any known military or military-intelligence end-use in accordance with U.S. trade compliance laws and regulations. Use of the NVIDIA AI Foundation Models provided under this Agreement will be consistent with NVIDIA’s Human Rights Policy.

Readme

Nemotron-Mini-4B-Instruct is a model for generating responses for roleplaying, retrieval augmented generation, and function calling. It is a small language model (SLM) optimized through distillation, pruning and quantization for speed and on-device deployment.

This instruct model is optimized for roleplay, RAG QA, and function calling in English. It supports a context length of 4,096 tokens. This model is ready for commercial use.

References

Blog

HuggingFace