Han-Bit Kang

hbkang's activity

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

upvoted a paper 7 days ago

The Other Half of the Memory Wall: Serving 35B MoEs from SSD with Trained Routing Prediction

Paper • 2609.18063 • Published 8 days ago • 18

updated a collection 7 days ago

cool-papers

Collection

122 items • Updated 7 days ago • 11

upvoted a paper 7 days ago

VC-Attention: Value Smoothing and Softmax Casting for Low-bit Attention

Paper • 2609.15810 • Published 10 days ago • 50

updated a collection 11 days ago

cool-models

Collection

8 items • Updated 11 days ago

upvoted a paper 11 days ago

DFlash: Block Diffusion for Flash Speculative Decoding

Paper • 2602.06036 • Published Feb 5 • 99

updated a collection 12 days ago

interesting architecture

Collection

50 items • Updated 12 days ago • 5

upvoted a paper 12 days ago

Graph Machine: Towards Better Pretraining via Edges

Paper • 2609.02881 • Published 22 days ago • 7

updated a collection 14 days ago

cool-papers

Collection

122 items • Updated 7 days ago • 11

upvoted a paper 14 days ago

OracleZoom: On-Policy Self-Distillation Inspired Reference-Constrained Recursive Image Super Resolution

Paper • 2609.06490 • Published 18 days ago • 9

upvoted a paper 15 days ago

Unlocking Lossless Speedups in LLMs via Discrete Diffusion

Paper • 2609.04010 • Published 21 days ago • 113

updated 3 collections 15 days ago

cool-papers

Collection

122 items • Updated 7 days ago • 11

interesting architecture

Collection

50 items • Updated 12 days ago • 5

cool-models

Collection

8 items • Updated 11 days ago

upvoted a paper 15 days ago

Marigold V2: Revisiting Diffusion Transformers for Monocular Depth Estimation

Paper • 2609.08084 • Published 16 days ago • 70

updated a collection 15 days ago

interesting architecture

Collection

50 items • Updated 12 days ago • 5

upvoted a paper 15 days ago

ENEAS: Embedding-guided Neural Ensemble for Adaptive Segmentation

Paper • 2609.03756 • Published 21 days ago • 25

upvoted a paper 17 days ago

Puro-2B: Poor Lab's Qwen2-1.5B Trained on RTX 5090 within $5090

Paper • 2608.27370 • Published 28 days ago • 40

updated a collection 21 days ago

cool-papers

Collection

122 items • Updated 7 days ago • 11

upvoted a paper 21 days ago

A Glance Is All You Need: Single-Pass Fine-Grained Image Captioning with SimLoss

Paper • 2609.00591 • Published 23 days ago • 18

updated a collection 23 days ago

cool-models

Collection

8 items • Updated 11 days ago