Survey of Small Language Models from Penn State, ...

Llama3.1 learns to Listen

Hindi-LLM

Resource list for generating JSON using LLMs via function calling, tools, CFG. Libraries, Models, Notebooks, etc.

Collection of Snowflake Notebook demos, tutorials, and examples

Composio equip's your AI agents & LLMs with 100+ high-quality integrations via function calling

Official inference framework for 1-bit LLMs

Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation

Arch is an intelligent prompt gateway. Engineered with (fast) LLMs for the secure handling, robust observability, and seamless integration of prompts with APIs - all outside business logic. Built b…

🔍 An LLM-based Multi-agent Framework of Web Search Engine (like Perplexity.ai Pro and SearchGPT)

The Enterprise-Grade Production-Ready Multi-Agent Orchestration Framework Join our Community: https://discord.com/servers/agora-999382051935506503

Entropy Based Sampling and Parallel CoT Decoding

🔥🕷️ Crawl4AI: Open-source LLM Friendly Web Crawler & Scrapper

OCR, layout analysis, reading order, table recognition in 90+ languages

[CCS'24] A dataset consists of 15,140 ChatGPT prompts from Reddit, Discord, websites, and open-source datasets (including 1,405 jailbreak prompts).

Composable building blocks to build Llama Apps

Utilities intended for use with Llama models.

Finetune Llama 3.2, Mistral, Phi, Qwen & Gemma LLMs 2-5x faster with 80% less memory

The Programmable Cypher-based Neuro-Symbolic AGI that lets you program its behavior using Graph-based Prompt Programming: for people who want AI to behave as expected

Kolmogorov-Arnold Transformer: A PyTorch Implementation with CUDA kernel

Implementation on how to use Kolmogorov-Arnold Networks (KANs) for classification and regression tasks.

High accuracy RAG for answering questions from scientific documents with citations

A framework for few-shot evaluation of language models.

An Open Source Toolkit For LLM Distillation

A collection of projects designed to help developers quickly get started with building deployable applications using the Anthropic API