Search Results for author:
Found 11 papers, 7 papers with code
u-$μ$P: The Unit-Scaled Maximal Update Parametrization
The Maximal Update Parametrization ($\mu$P) aims to make the optimal hyperparameters (HPs) of a model independent of its size, allowing them to be swept using a cheap proxy model rather than the full-size target model.
T-FREE: Subword Tokenizer-Free Generative LLMs via Sparse Representations for Memory-Efficient Embeddings
Tokenizers are crucial for encoding information in Large Language Models, but their development has recently stagnated, and they contain inherent weaknesses.
Efficient Parallelization Layouts for Large-Scale Distributed Model Training
In this work, we conduct a comprehensive ablation study of possible training configurations for large language models.
Tokenizer Choice For LLM Training: Negligible or Crucial?
The recent success of Large Language Models (LLMs) has been predominantly driven by curating the training dataset composition, scaling of model architectures and dataset sizes and advancements in pretraining objectives, leaving tokenizer influence as a blind spot.
MultiFusion: Fusing Pre-Trained Models for Multi-Lingual, Multi-Modal Image Generation
The recent popularity of text-to-image diffusion models (DM) can largely be attributed to the intuitive interface they provide to users.
AtMan: Understanding Transformer Predictions Through Memory Efficient Attention Manipulation
Generative transformer models have become increasingly complex, with large numbers of parameters and the ability to process multiple input modalities.
M-VADER: A Model for Diffusion with Multimodal Context
We introduce M-VADER: a diffusion model (DM) for image generation where the output can be specified using arbitrary combinations of images and text.
GPT-NeoX-20B: An Open-Source Autoregressive Language Model
We introduce GPT-NeoX-20B, a 20 billion parameter autoregressive language model trained on the Pile, whose weights will be made freely and openly available to the public through a permissive license.
Ranked #40 on
Multi-task Language Understanding
on MML
MAGMA -- Multimodal Augmentation of Generative Models through Adapter-based Finetuning
Large-scale pretraining is fast becoming the norm in Vision-Language (VL) modeling.
Domain-Level Explainability -- A Challenge for Creating Trust in Superhuman AI Strategies
For strategic problems, intelligent systems based on Deep Reinforcement Learning (DRL) have demonstrated an impressive ability to learn advanced solutions that can go far beyond human capabilities, especially when dealing with complex scenarios.
Deep Reinforcement Learning
Explainable Artificial Intelligence (XAI)
