Samsung AI Lab (SAIL) Montreal

SamsungSAILMontreal's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Articles

bknyaz 

updated a collection 21 days ago

REAM

Collection

Compressed MoE models with a reduced number of experts. See additional models at https://huggingface.co/bknyaz. • 12 items • Updated 21 days ago • 6

bknyaz 

updated a collection 25 days ago

REAM

Collection

Compressed MoE models with a reduced number of experts. See additional models at https://huggingface.co/bknyaz. • 12 items • Updated 21 days ago • 6

bknyaz 

in SamsungSAILMontreal/Qwen3-30B-A3B-Instruct-2507-REAM about 1 month ago

keep doing it!

❤️ 1

2

#4 opened about 1 month ago by

pirola

bknyaz 

updated a collection about 1 month ago

REAM

Collection

Compressed MoE models with a reduced number of experts. See additional models at https://huggingface.co/bknyaz. • 12 items • Updated 21 days ago • 6

bknyaz 

updated 2 models about 1 month ago

SamsungSAILMontreal/Qwen3-Coder-Next-REAP

Text Generation • 60B • Updated Apr 9 • 76 • 3

SamsungSAILMontreal/GLM-4.5-Air-REAP

Text Generation • 82B • Updated Apr 9 • 37

bknyaz 

published a model about 1 month ago

SamsungSAILMontreal/GLM-4.5-Air-REAP

Text Generation • 82B • Updated Apr 9 • 37

bknyaz 

updated a model about 1 month ago

bknyaz/GLM-4.5-Air-REAM

Text Generation • 82B • Updated Apr 9 • 40

bknyaz 

published a model about 1 month ago

bknyaz/GLM-4.5-Air-REAM

Text Generation • 82B • Updated Apr 9 • 40

bknyaz 

submitted a paper to Daily Papers about 1 month ago

REAM: Merging Improves Pruning of Experts in LLMs

Paper • 2604.04356 • Published Apr 6 • 9

bknyaz 

updated 2 models about 1 month ago

SamsungSAILMontreal/Qwen3-30B-A3B-Instruct-2507-REAM

Text Generation • 23B • Updated Apr 8 • 156 • 7

SamsungSAILMontreal/Qwen3-30B-A3B-Instruct-2507-REAP

Text Generation • 23B • Updated Apr 8 • 94

bknyaz 

authored 6 papers about 1 month ago

(Almost) Free Modality Stitching of Foundation Models

Paper • 2507.10015 • Published Jul 14, 2025 • 2

M^4olGen: Multi-Agent, Multi-Stage Molecular Generation under Precise Multi-Property Constraints

Paper • 2601.10131 • Published Jan 15 • 17

Recursive Autoconvolution for Unsupervised Learning of Convolutional Neural Networks

Paper • 1606.00611 • Published Jun 2, 2016

Celo2: Towards Learned Optimization Free Lunch

Paper • 2602.19142 • Published Feb 22

Celo: Training Versatile Learned Optimizers on a Compute Diet

Paper • 2501.12670 • Published Jan 22, 2025

REAM: Merging Improves Pruning of Experts in LLMs

Paper • 2604.04356 • Published Apr 6 • 9

bknyaz 

in bknyaz/Qwen3-Next-80B-A3B-Instruct-REAM 3 months ago

Has REAM been checked for their resilience towards quantization?

4

#1 opened 3 months ago by

TomLucidor