stephen37

20 ms·

1.
Martin Scorsese Is Embracing A.I.(nytimes.com)
52 points by stephen37 4mo ago|56 comments
2.
by stephen37 4mo ago
Martin Scorsese is backing Black Forest Labs, the company famous for FLUX models.
3.
by stephen37 1y ago
I got it working when I provided an image of a Model F keyboard. This is the strength of the model, provide it an input image and it will do some magic Disclaimer: I work for BFL
4.
FLUX.1 Kontext – Edit Images with Prompt(bfl.ai)
4 points by stephen37 1y ago|0 comments
5.
Don't Shard Your DB Manually(milvus.io)
1 points by stephen37 1y ago|0 comments
6.
DeepSearcher: A local open-source Deep Research(milvus.io)
229 points by stephen37 2y ago|26 comments
7.
Mistral Saba(mistral.ai)
145 points by stephen37 2y ago|21 comments
8.
I Built a Deep Research with Open Source–and So Can You(milvus.io)
2 points by stephen37 2y ago|0 comments
9.
Full Text Search in Milvus 2.5(milvus.io)
1 points by stephen37 2y ago|0 comments
10.
Visualize HNSW, FAISS and other ANNs index(github.com)
2 points by stephen37 2y ago|0 comments
11.
Semantic Search Is Cool, but What Else?(zilliz.com)
1 points by stephen37 2y ago|0 comments
12.
Milvus Lite: The Lightweight Version of Milvus(milvus.io)
56 points by stephen37 2y ago|17 comments
13.
by stephen37 2y ago
Oh yes good point, it's because at first I was using a local embedding model and then decided to use JinaAI as they have good embeddings. Jina AI is doing different things but one of them is having embeddings and I use their English&#x
14.
by stephen37 2y ago
I love working with Ollama, I was really surprised at how easy it is to build a simple RAG system with it. For example: https://github.com/stephen37/ollama_local_rag
15.
Picking the Right Vector Index for Your Project(thesequence.substack.com)
2 points by stephen37 2y ago|0 comments
16.
by stephen37 2y ago
On top of my head, pgvector only supports 2 indexes, those are running in memory only. They don't support GPU indexing, nor Disk based indexing, they also don't have separation of query and insertions. Also with different people I
17.
by stephen37 2y ago
It's nice to begin with but you usually quickly hit a scale issue from what I tried and talked to different people using it. Also they only offer index stored in Memory as far as I know, also lack the support of different index or more
18.
by stephen37 2y ago
Cool to see that Supabase is adding S3 protocol! Nice to see more and more storage solutions available. We, at Milvus, I've integrated S3, Parquet and other ones to make it possible for developers to use their data no matter what they
19.
Nvidia's Cagra: Game-Changer for GPU-Accelerated Graph Indexing in Vector DBs(zilliz.com)
4 points by stephen37 2y ago|0 comments
20.
RAG Reimagined: LLMs Lead the Charge in AI Retrieval Accuracy(zilliz.com)
2 points by stephen37 2y ago|0 comments
21.
Elevating Vector Search with GPU-accelerated indexes(zilliz.com)
1 points by stephen37 2y ago|0 comments
22.
by stephen37 3y ago
I am French originally and on my side what worked well was to speak to myself in English all the time. At home I would actually speak English out loud and it helped a lot with practicing my pronunciation and also to learn new words. Also re
23.
by stephen37 3y ago
What is Character.AI doing special to have that much more engagement than other apps?
24.
by stephen37 3y ago
It might be the first time that I see a phone number on a Github Repo, I wonder how many phone calls they will actually receive. It's probably the last thing I'd like to do when looking at an open source repository.