gräfics

graefics's activity

• All
• Models
• Datasets
• Spaces
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

upvoted a paper 13 days ago

Slim attention: cut your context memory in half without loss of accuracy -- K-cache is all you need for MHA

Paper • 2503.05840 • Published Mar 7 • 3

upvoted an article about 1 month ago

view article

Article

Topic 33: Slim Attention, KArAt, XAttention and Multi-Token Attention Explained – What’s Really Changing in Transformers?

By and 1 other •

Apr 4

• 14

upvoted a paper about 1 month ago

Flash normalization: fast RMSNorm for LLMs

Paper • 2407.09577 • Published Jul 12, 2024 • 2

liked a model about 2 months ago

rasbt/llama-3.2-from-scratch

Updated Apr 16 • 271

liked a model 7 months ago

HuggingFaceTB/SmolLM-135M

Text Generation • Updated Aug 1, 2024 • 133k • 203

New activity in HuggingFaceTB/SmolLM-135M 8 months ago

Are there two identical embedding tensors, even though embeddings are shared?

#15 opened 8 months ago by

graefics

authored a paper 10 months ago

Flash normalization: fast RMSNorm for LLMs

Paper • 2407.09577 • Published Jul 12, 2024 • 2