SamsungSAILMontreal/nino
Code for "Accelerating Training with Neuron Interaction and Nowcasting Networks" [ICLR 2025]
This project offers a method to speed up the training of large AI models, particularly for tasks involving language and vision. It takes past states of a model's parameters and uses them to predict future optimal parameter settings. This allows machine learning engineers and researchers to train complex models more efficiently, reducing the time and computational resources needed.
Use this if you are a machine learning engineer or researcher looking to significantly reduce the training time for large language models (LLMs) or vision models, especially when using optimizers like Adam.
Not ideal if you are working with very small, simple models or if your primary bottleneck is not training time but rather data processing or model architecture design.
Stars
28
Forks
10
Language
Python
License
MIT
Category
Last pushed
Feb 20, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/SamsungSAILMontreal/nino"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related models
graphdeeplearning/graphtransformer
Graph Transformer Architecture. Source code for "A Generalization of Transformer Networks to...
vijaydwivedi75/gnn-lspe
Source code for GNN-LSPE (Graph Neural Networks with Learnable Structural and Positional...
snap-stanford/relgt
Relational Graph Transformer
omron-sinicx/crystalframer
The official code respository for "Rethinking the role of frames for SE(3)-invariant crystal...
SamsungSAILMontreal/ghn3
Code for "Can We Scale Transformers to Predict Parameters of Diverse ImageNet Models?" [ICML 2023]