Re:Laion-400M: open-source dataset of 400M image-text pairs
by wongarsu on Monday September 13, 2021 @06:40AM
(#28509778)
Not only does it produce really good results, the interface with it's ability to start another search with either the description or image of any of the results is great for complex searches.
[ Parent
| on HN ]
Re:Laion-400M: open-source dataset of 400M image-text pairs
by mkl on Monday September 13, 2021 @07:21AM
(#28510072)
Those "search by description" links reveal a strange problem with the index:
1. Search for "monitor": https://rom1504.github.io/clip-retrieval/?back=https%3A%2F%2...
2. Note that there are several results with AOC model E2260SWDN.
3. Click the search button beside one of them.
4. Note that none of the results contain "E2260SWDN". Searching for "E2260SWDN" by itself is the same.
The back button doesn't work in this interface either. The URL changes, but it doesn't actually show that previous search.
Re:Laion-400M: open-source dataset of 400M image-text pairs
by rom1504 on Tuesday September 14, 2021 @04:17AM
(#28521766)
That's interesting indeed!
Note that the description search as well as the image search are both using a knn index on embeddings and not exact search. That helps for finding semantically close by items but indeed for exact reference match it might not be the best solution.
Re-indexing the dataset with something like elastic search would give the reference search results you expect.