Improving Hugging Face Training Efficiency Through Packing with Flash Attention
•
35
IBM
company
Verified
AI & ML interests
Enterprise AI and ML, Foundation Models, Responsible AI
Recent Activity
View all activity
Articles
Welcome to IBM’s multi-modal foundation model for materials, FM4M, designed to support and advance research in materials science and chemistry.
Wishes datasets created by Genie
Datasets from "A Novel Metric for Measuring the Robustness of Large Language Models in Non-adversarial Scenarios" (https://arxiv.org/abs/2408.01963)
A suite of open-source biomedical foundation models. https://research.ibm.com/projects/biomedical-foundation-models
Dense & MoE LLMs trained with power learning rate scheduler.
-
Unitxt: Flexible, Shareable and Reusable Data Preparation and Evaluation for Generative AI
Paper • 2401.14019 • Published • 24 -
The Belebele Benchmark: a Parallel Reading Comprehension Dataset in 122 Language Variants
Paper • 2308.16884 • Published • 10 -
Genie: Achieving Human Parity in Content-Grounded Datasets Generation
Paper • 2401.14367 • Published • 8 -
ibm-nasa-geospatial/Prithvi-WxC-1.0-2300M
Updated • 209 • 73
Welcome to IBM’s multi-modal foundation model for materials, FM4M, designed to support and advance research in materials science and chemistry.
A suite of open-source biomedical foundation models. https://research.ibm.com/projects/biomedical-foundation-models
Dense & MoE LLMs trained with power learning rate scheduler.
Wishes datasets created by Genie
-
Unitxt: Flexible, Shareable and Reusable Data Preparation and Evaluation for Generative AI
Paper • 2401.14019 • Published • 24 -
The Belebele Benchmark: a Parallel Reading Comprehension Dataset in 122 Language Variants
Paper • 2308.16884 • Published • 10 -
Genie: Achieving Human Parity in Content-Grounded Datasets Generation
Paper • 2401.14367 • Published • 8 -
ibm-nasa-geospatial/Prithvi-WxC-1.0-2300M
Updated • 209 • 73
Datasets from "A Novel Metric for Measuring the Robustness of Large Language Models in Non-adversarial Scenarios" (https://arxiv.org/abs/2408.01963)