@UnfoldAI_com

In-depth tutorials and resources for ML/AI system design, optimization, and implementation.

unfoldai.com/
Joined August 2023
  • Tweets141
  • Following47
  • Followers4
AWS Competition alert! WorldMind Evidence is designed for regulated industries such as pharmaceutical research, clinical trials, and regulatory submissions, where every claim must be verifiable and traceable to authoritative documentation. Check more: builder.aws.com/content/3AhF…

AIdeas: WorldMind Evidence | AWS Builder Center

WorldMind Evidence is designed for regulated industries such as pharmaceutical research, clinical trials, and regulatory submissions, where every claim must be verifiable and traceable to authorita...

builder.aws.com
9
Vector search doesn't always require new infrastructure. pgvector enables semantic capabilities in existing Postgres deployments—simpler ops, lower costs, fewer failure points. Specialized solutions matter at scale. Most projects aren't there yet.
22
What is AI-as-Judge? Learn more: unfoldai.com/ai-as-judge/

AI as Judge — Using Language Models to evaluate AI outputs | UnfoldAI

Table of Contents hide 1 Understanding LLM-as-Judge 2 Effective evaluation strategies 3 Real-world applications 4 Current limitations 5 Looking forward The landscape of AI evaluation...

unfoldai.com
19
🤖 Pre-training: The expensive foundation phase where LLMs learn language 🎯 Fine-tuning: Specialized training for specific domains 🎓 Instruct-tuning: Teaching models to follow instructions better Understanding these training phases is key to working with LLMs models.
1
13
Read more: unfoldai.com/difference-betw…

What is the difference between pre-training, fine-tuning, and instruct-tuning | UnfoldAI

Discover the key differences between pre-training, fine-tuning, and instruct-tuning in LLMs. Learn how these training phases shape AI models' capabilities.

unfoldai.com
5
New deep dive into DeepSeek R1 just published. The model achieves o1 level reasoning through reinforcement learning. Even more interesting: their 1.5B distilled version maintains most capabilities. Time to rethink our assumptions?
1
64
unfoldai.com/deepseek-r1/

DeepSeek-R1 — Training Language Models to reason through Reinforcement Learning | UnfoldAI

Table of Contents hide 1 Model architecture 2 Implementation overview 3 Training pipeline 4 Performance analysis 5 Hardware considerations 6 Discussion 7 Conclusion In this...

unfoldai.com
7
Kokoro-82M is a new text-to-speech (TTS) model, delivering exceptional performance with only 82 million parameters. 🚀 🗣 The model achieves state-of-the-art performance with 82M parameters, outperforming 14x bigger models.
1
31
Learn more: 🔗 unfoldai.com/kokoro-82m/

Kokoro-82M — When smaller means better in text-to-speech | UnfoldAI

Table of Contents hide 1 Architecture overview 2 Performance metrics 3 Training 4 Usage 5 Example generation and voices 6 Voice mixing 7 Limitations 8...

unfoldai.com
6
💥 New query expansion models for RAG systems 💥 The collection addresses a common challenge in search and retrieval – bridging the gap between user queries and document content.
1
12
huggingface.co/collections/s…

Query expansion - a s-emanuilov Collection

A collection of models along with the training dataset, designed to improve search queries and retrieval in RAG systems.

huggingface.co
5
Query pgvector — Vector Operations in PostgreSQL using pgvector Pgvector enables vector similarity search operations in PostgreSQL. Here’s a practical guide to working with vector queries:
1
21
unfoldai.com/query-pgvector/

Query pgvector — Vector Operations in PostgreSQL using pgvector | UnfoldAI

Table of Contents hide 1 Basic vector queries 2 Common distance operators 3 Indexing for performance 4 Additional features 5 Performance tips Pgvector enables vector...

unfoldai.com
9
How to add a sitemap in Django? Learn in our short article, straight to the point!
1
32
unfoldai.com/how-to-add-site…

How to add sitemap to Django project | UnfoldAI

Table of Contents hide 1 Enable the Sitemap framework 2 Create the Sitemap class 3 Configure URLs 4 Model requirements 5 Testing A sitemap helps...

unfoldai.com
4
We started a new category "Shorts". There you can find some quick answer to common questions about Django, FastAPI and building AI systems.
1
34
A link to the category: unfoldai.com/category/shorts…

Short answers and tips for programming and AI | UnfoldAI

In this section, I will post short answers to specific questions about programming or AI concepts.

unfoldai.com
4
We have another app in our collection
Simeon Emanuilov@s_emanuilov1 Jan 2025
Going to sleep after a long day feels like a great start to the year. My new micro-SaaS is currently ranked second on Product Hunt. I hope it stays there until the end. If you have a minute, please support it: producthunt.com/products/mon… Thank you! 🙏

Monkt: Transform files and web pages into AI-ready Markdown or JSON | Product Hunt

Monkt convert PDFs, Word files, Excel sheets, PowerPoint presentations and web pages into structured Markdown or JSON while preserving semantic structure. Apply custom schemas, process in batches,...

producthunt.com
4
Converting docs to Markdown and JSON? Every AI practitioner has encountered this situation: You possess thousands of documents in various formats—PDFs, Word files, Excel sheets, and web pages—that must be converted into a uniform format for your LLM training or AI workflow.
1
1
9
more replies
Check out a new tool, which we released: monkt.com/ (Converting documents to Markdown and JSON at scale)

Monkt - Document Transformation Tool

Transform Documents into AI-Ready Markdown or structured JSON

monkt.com
1
6
Open-source tools like Microsoft’s MarkItDown (github.com/microsoft/markitd…), which provide solid foundations for document conversion. While these tools offer great processing capabilities, they present specific implementation challenges.
1
1
16
Load more