Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2407.14561 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916668901425152 |
|---|---|
| author | Fiotto-Kaufman, Jaden Loftus, Alexander R. Todd, Eric Brinkmann, Jannik Pal, Koyena Troitskii, Dmitrii Ripa, Michael Belfki, Adam Rager, Can Juang, Caden Mueller, Aaron Marks, Samuel Sharma, Arnab Sen Lucchetti, Francesca Prakash, Nikhil Brodley, Carla Guha, Arjun Bell, Jonathan Wallace, Byron C. Bau, David |
| author_facet | Fiotto-Kaufman, Jaden Loftus, Alexander R. Todd, Eric Brinkmann, Jannik Pal, Koyena Troitskii, Dmitrii Ripa, Michael Belfki, Adam Rager, Can Juang, Caden Mueller, Aaron Marks, Samuel Sharma, Arnab Sen Lucchetti, Francesca Prakash, Nikhil Brodley, Carla Guha, Arjun Bell, Jonathan Wallace, Byron C. Bau, David |
| contents | We introduce NNsight and NDIF, technologies that work in tandem to enable scientific study of the representations and computations learned by very large neural networks. NNsight is an open-source system that extends PyTorch to introduce deferred remote execution. The National Deep Inference Fabric (NDIF) is a scalable inference service that executes NNsight requests, allowing users to share GPU resources and pretrained models. These technologies are enabled by the Intervention Graph, an architecture developed to decouple experimental design from model runtime. Together, this framework provides transparent and efficient access to the internals of deep neural networks such as very large language models (LLMs) without imposing the cost or complexity of hosting customized models individually. We conduct a quantitative survey of the machine learning literature that reveals a growing gap in the study of the internals of large-scale AI. We demonstrate the design and use of our framework to address this gap by enabling a range of research methods on huge models. Finally, we conduct benchmarks to compare performance with previous approaches.
Code, documentation, and tutorials are available at https://nnsight.net/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_14561 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internals Fiotto-Kaufman, Jaden Loftus, Alexander R. Todd, Eric Brinkmann, Jannik Pal, Koyena Troitskii, Dmitrii Ripa, Michael Belfki, Adam Rager, Can Juang, Caden Mueller, Aaron Marks, Samuel Sharma, Arnab Sen Lucchetti, Francesca Prakash, Nikhil Brodley, Carla Guha, Arjun Bell, Jonathan Wallace, Byron C. Bau, David Machine Learning Artificial Intelligence We introduce NNsight and NDIF, technologies that work in tandem to enable scientific study of the representations and computations learned by very large neural networks. NNsight is an open-source system that extends PyTorch to introduce deferred remote execution. The National Deep Inference Fabric (NDIF) is a scalable inference service that executes NNsight requests, allowing users to share GPU resources and pretrained models. These technologies are enabled by the Intervention Graph, an architecture developed to decouple experimental design from model runtime. Together, this framework provides transparent and efficient access to the internals of deep neural networks such as very large language models (LLMs) without imposing the cost or complexity of hosting customized models individually. We conduct a quantitative survey of the machine learning literature that reveals a growing gap in the study of the internals of large-scale AI. We demonstrate the design and use of our framework to address this gap by enabling a range of research methods on huge models. Finally, we conduct benchmarks to compare performance with previous approaches. Code, documentation, and tutorials are available at https://nnsight.net/. |
| title | NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internals |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2407.14561 |