Datasent, Inc.

company

Verified

http://www.datasent.com

AI & ML interests

None defined yet.

Articles

Compressing visual tokens in vision-language models: 3x more requests per GPU on Qwen2-VL

updated a Space 2 months ago

Polygen RAG -- Encode Once, Query Many

A visual corpus encoded once, queried many times

updated a Space 4 months ago

Datasent's Polygen Tokenization Demo

Tokenize any file into a compact Polygen token

published an article 5 months ago

Article

Compressing visual tokens in vision-language models: 3x more requests per GPU on Qwen2-VL

datas3nt

•

Jun 17

• 1

updated 2 Spaces 5 months ago

Polygen for Qwen2-VL

Compress visual tokens. Same answer, proven per image.

Datasent

published 3 Spaces 5 months ago

Datasent

Datasent's Polygen Tokenization Demo

Tokenize any file into a compact Polygen token

Polygen RAG -- Encode Once, Query Many

A visual corpus encoded once, queried many times

updated 2 models 5 months ago

datas3nt/qwen2vl-polygen-lora-r16-1000

Image-Text-to-Text • Updated Jun 1 • 9

datas3nt/qwen2vl-polygen-7b-lora-r16-1000

Image-Text-to-Text • Updated Jun 1 • 15

published a model 5 months ago

datas3nt/qwen2vl-polygen-7b-lora-r16-1000

Image-Text-to-Text • Updated Jun 1 • 15

published a Space 5 months ago

Polygen for Qwen2-VL

Compress visual tokens. Same answer, proven per image.

published a model 5 months ago

datas3nt/qwen2vl-polygen-lora-r16-1000

Image-Text-to-Text • Updated Jun 1 • 9