ArkidMitra commited on
Commit
b28b6f8
·
verified ·
1 Parent(s): e8f9ded

org card: lead with a short welcome and a start-here line, as Marqo does

Browse files
Files changed (1) hide show
  1. README.md +11 -14
README.md CHANGED
@@ -10,23 +10,20 @@ short_description: SOTA fashion retrieval, measured and open.
10
  ---
11
 
12
  <section>
13
- <h1>Hopit AI — fashion retrieval, measured properly</h1>
 
 
 
14
  <p>
15
- We build the <b>MODA</b> family of fashion retrieval models and benchmark them
16
- the way we wish everyone did: <b>full corpus, one shared harness, every
17
- competitor under identical protocol — including the cells we lose.</b>
18
- </p>
19
- <p>
20
- <a href="https://hopit-ai.github.io/Moda/">Full benchmark page</a>
21
- ·
22
- <a href="https://huggingface.co/spaces/HopitAI/moda-fashion-search">Interactive demo</a>
23
- ·
24
- <a href="https://github.com/hopit-ai/Moda">Code &amp; methodology</a>
25
- ·
26
- <a href="https://hopit.ai">hopit.ai</a>
27
- ·
28
  <a href="https://calendly.com/arkid_/new-meeting?back=1">Book a call</a>
29
  </p>
 
 
 
30
  </section>
31
 
32
  <section>
 
10
  ---
11
 
12
  <section>
13
+ <h1>Hopit AI</h1>
14
+ <p><b>Fashion AI infrastructure.</b> We build the MODA family of fashion retrieval
15
+ models and benchmark them the way we wish everyone did: <b>full corpus, one shared
16
+ harness, every competitor under identical protocol, including the cells we lose.</b></p>
17
  <p>
18
+ <a href="https://hopit.ai">hopit.ai</a> ·
19
+ <a href="https://hopit-ai.github.io/Moda/">Benchmarks</a> ·
20
+ <a href="https://github.com/hopit-ai/Moda">Code</a> ·
21
+ <a href="https://huggingface.co/spaces/HopitAI/moda-fashion-search">Demo</a> ·
 
 
 
 
 
 
 
 
 
22
  <a href="https://calendly.com/arkid_/new-meeting?back=1">Book a call</a>
23
  </p>
24
+ <p><b>Start here.</b> Text search: <a href="https://huggingface.co/HopitAI/moda-pro-lite">MODA Pro Lite</a> (213M, open weights).
25
+ Image search: <a href="https://huggingface.co/HopitAI/moda-fashion-distilled">MODA-SigLIP-Distilled</a> (203M, #1 open on LookBench).
26
+ Both are plain bi-encoders that work with any vector database.</p>
27
  </section>
28
 
29
  <section>