Канал: code_your_own_AI

Q* explained: Complex Multi-Step AI Reasoning

Q* explained: Complex Multi-Step AI Reasoning

5 Easy Ways to help LLMs to Reason

5 Easy Ways to help LLMs to Reason

Claude 3.5 SONNET hallucinates w/ a Logic Bug?

Claude 3.5 SONNET hallucinates w/ a Logic Bug?

Decoding AI's Blind Spots: Solving Causal Reasoning

Decoding AI's Blind Spots: Solving Causal Reasoning

NEW Multi-Modal AI by APPLE

NEW Multi-Modal AI by APPLE

Financial AI Brilliance: 7 Children at Stanford? 😆

Financial AI Brilliance: 7 Children at Stanford? 😆

Text-to-GRAPH w/ LGGM: Generative Graph Models

Text-to-GRAPH w/ LGGM: Generative Graph Models

NEW TextGrad by Stanford: Better than DSPy

NEW TextGrad by Stanford: Better than DSPy

Adversarial Questions Test Multimodal MED AI sys

Adversarial Questions Test Multimodal MED AI sys

BEST RAG you can buy: LAW AI (Stanford)

BEST RAG you can buy: LAW AI (Stanford)

RAG explained step-by-step up to GROKKED RAG sys

RAG explained step-by-step up to GROKKED RAG sys

GROKKED LLM beats RAG Reasoning (Part 3)

GROKKED LLM beats RAG Reasoning (Part 3)

LLM - Reasoning SOLVED (new research)

LLM - Reasoning SOLVED (new research)

New Discovery: LLMs have a Performance Phase

New Discovery: LLMs have a Performance Phase

One Thought on the Future of AI Agents World Model

One Thought on the Future of AI Agents World Model

NEW ChatGPT EDU for AI Universities: Unique SALE

NEW ChatGPT EDU for AI Universities: Unique SALE

Agentic AI: The Future is here?

Agentic AI: The Future is here?

LongRoPE & Theta Scaling to 1 Mio Token (2/2)

LongRoPE & Theta Scaling to 1 Mio Token (2/2)

RoPE Rotary Position Embedding to 100K context length

RoPE Rotary Position Embedding to 100K context length

Many-Shot VISUAL ICL is amazing! (Stanford)

Many-Shot VISUAL ICL is amazing! (Stanford)

In-Context Learning: EXTREME vs Fine-Tuning, RAG

In-Context Learning: EXTREME vs Fine-Tuning, RAG

Warning GPT-4o: DON'T translate to Chinese (MIT)

Warning GPT-4o: DON'T translate to Chinese (MIT)

CODE Fine-Tune Vision Language VLM eg PaliGemma-3B

CODE Fine-Tune Vision Language VLM eg PaliGemma-3B

GPT-4 Turbo vs GPT-4o in Reasoning TEST #gpt4o

GPT-4 Turbo vs GPT-4o in Reasoning TEST #gpt4o

New Trick for Fine-Tuning LLMs #airesearch

New Trick for Fine-Tuning LLMs #airesearch

From Dating Apps to AI: Gen Z Edition 😆

From Dating Apps to AI: Gen Z Edition 😆

Do not use Llama-3 70B for these tasks ...

Do not use Llama-3 70B for these tasks ...

New xLSTM explained: Better than Transformer LLMs?

New xLSTM explained: Better than Transformer LLMs?

GPT-4o in stealth as im-a-good-gpt2-chatbot

GPT-4o in stealth as im-a-good-gpt2-chatbot

Understand DSPy: Programming AI Pipelines

Understand DSPy: Programming AI Pipelines

Latest Insights in AI Performance Models

Latest Insights in AI Performance Models

New Discovery: Retrieval Heads for Long Context

New Discovery: Retrieval Heads for Long Context

Multi-Token Prediction (forget next token LLM?)

Multi-Token Prediction (forget next token LLM?)

NEW LLM Test: Reasoning & gpt2-chatbot

NEW LLM Test: Reasoning & gpt2-chatbot

LLMs: Rewriting Our Tomorrow (plus code) #ai

LLMs: Rewriting Our Tomorrow (plus code) #ai

Autonomous AI Agents: 14% MAX Performance

Autonomous AI Agents: 14% MAX Performance

Snowflake Arctic 480B LLM as 128x4B MoE? WHY?

Snowflake Arctic 480B LLM as 128x4B MoE? WHY?

No more Fine-Tuning: Unsupervised ICL+

No more Fine-Tuning: Unsupervised ICL+

NEW Phi-3 mini 3.8B LLM for Your PHONE: 1st TEST

NEW Phi-3 mini 3.8B LLM for Your PHONE: 1st TEST

BEST LLMs for Coding, Long Context, Overall Perform

BEST LLMs for Coding, Long Context, Overall Perform

Next-Gen AI: RecurrentGemma (Long Context Length)

Next-Gen AI: RecurrentGemma (Long Context Length)

Gemini 1.5 PRO vs Lllama3-70B-Instruct: TEST

Gemini 1.5 PRO vs Lllama3-70B-Instruct: TEST

Llama 3 70B TESTED: FAILS Logical Reasoning

Llama 3 70B TESTED: FAILS Logical Reasoning

Mighty New TransformerFAM (Feedback Attention Mem)

Mighty New TransformerFAM (Feedback Attention Mem)

NEW: INFINI Attention w/ 1 Mio Context Length

NEW: INFINI Attention w/ 1 Mio Context Length

RING Attention explained: 1 Mio Context Length

RING Attention explained: 1 Mio Context Length

NEW WizardLM-2 8x22B: Fine-tune & Stage-DPO align

NEW WizardLM-2 8x22B: Fine-tune & Stage-DPO align

Catching Up w/ AI: Past 5 Days

Catching Up w/ AI: Past 5 Days

New View on AI: Monad Algebra in Category Theory

New View on AI: Monad Algebra in Category Theory

RAG ipynb: CRAG, LlamaIndex, Ollama, ReAct Agent

RAG ipynb: CRAG, LlamaIndex, Ollama, ReAct Agent

LLM Prompt Personalization: Easy Tutorial #ai

LLM Prompt Personalization: Easy Tutorial #ai

On-Device: Functional Tokens (Octopus v2)

On-Device: Functional Tokens (Octopus v2)

SUPERHUMAN RAG #ai

SUPERHUMAN RAG #ai

NVIDIA's lack of an AI PC - Good for Investors?

NVIDIA's lack of an AI PC - Good for Investors?

NEW Challenge for LLM: CONSISTENCY ALIGNMENT

NEW Challenge for LLM: CONSISTENCY ALIGNMENT

JAMBA MoE: Open Source MAMBA w/ Transformer: CODE

JAMBA MoE: Open Source MAMBA w/ Transformer: CODE

DBRX: My First Performance TEST - Causal Reasoning

DBRX: My First Performance TEST - Causal Reasoning

DBRX: MOST POWERFUL Open Source LLM - NEW @Databricks

DBRX: MOST POWERFUL Open Source LLM - NEW @Databricks

LLM2LLM: Synthetic Data for Fine-Tuning (UC Berkeley)

LLM2LLM: Synthetic Data for Fine-Tuning (UC Berkeley)

ORPO: NEW DPO Alignment and SFT Method for LLM

ORPO: NEW DPO Alignment and SFT Method for LLM

NVIDIA NIM RAG Optimization: QuietSTAR (Stanford)

NVIDIA NIM RAG Optimization: QuietSTAR (Stanford)

NIM-RAG Hallucinations: Understand WHY

NIM-RAG Hallucinations: Understand WHY

MAMBA LLM for Personalized Medicine?

MAMBA LLM for Personalized Medicine?

Symmetry Breaking in AI Diffusion: New Discovery

Symmetry Breaking in AI Diffusion: New Discovery

Diffusion Transformer explained (Stable Diffusion 3)

Diffusion Transformer explained (Stable Diffusion 3)

New LLM Benchmark Leaderboard: WildBench

New LLM Benchmark Leaderboard: WildBench

The Perfect AI Prompt (and how to create it)

The Perfect AI Prompt (and how to create it)

AI Applicant Tracking ATS for Recruiting

AI Applicant Tracking ATS for Recruiting

All of AI Explained: From Data to AI World Models

All of AI Explained: From Data to AI World Models

CLAUDE 3 OPUS: First Causal Reasoning Test

CLAUDE 3 OPUS: First Causal Reasoning Test

GenAI WORMS Breach RAG Systems: AI Cybersecurity

GenAI WORMS Breach RAG Systems: AI Cybersecurity

Your NEW Job in AI

Your NEW Job in AI

NEW AI generates Video Games: GENIE AI explained

NEW AI generates Video Games: GENIE AI explained

META AI attacks OpenAI: V-JEPA vs SORA (VIDEO AI)

META AI attacks OpenAI: V-JEPA vs SORA (VIDEO AI)

AGI: Humanity's Last Invention?

AGI: Humanity's Last Invention?

Rethinking AI: Solutions for Logical Reasoning?

Rethinking AI: Solutions for Logical Reasoning?

LLM (w/ RAG) need a new Logic Layer (Stanford)

LLM (w/ RAG) need a new Logic Layer (Stanford)

NEW TECH for SORA (OpenAI)

NEW TECH for SORA (OpenAI)

16GB GPUs: Enough for Text-to-Video AI?

16GB GPUs: Enough for Text-to-Video AI?

SECRET Text-to-VIDEO AI: 1Mio Token, Open-Source

SECRET Text-to-VIDEO AI: 1Mio Token, Open-Source

FRACTALS to power our LLMs

FRACTALS to power our LLMs

New AI Dev by Apple & Google

New AI Dev by Apple & Google

LLM Phase Transition: New Discovery

LLM Phase Transition: New Discovery

How to Create an Effective PROMPT - no code guide

How to Create an Effective PROMPT - no code guide

Simple ideas to improve your RAG (Stanford, Google)

Simple ideas to improve your RAG (Stanford, Google)

NEW DSPyG: DSPy combined w/ Graph Optimizer in PyG

NEW DSPyG: DSPy combined w/ Graph Optimizer in PyG

DSPy on ICL RAG Classification: Code explained

DSPy on ICL RAG Classification: Code explained

DSPy explained: No more LangChain PROMPT Templates

DSPy explained: No more LangChain PROMPT Templates

AI DSP: LLM Pipeline to Retriever Model (Stanford)

AI DSP: LLM Pipeline to Retriever Model (Stanford)

AI State Machines - State Agents - State Spaces explained

AI State Machines - State Agents - State Spaces explained

AI Domain Generalisation: ICRM and ContextViT

AI Domain Generalisation: ICRM and ContextViT

FASTER Code for SFT + DPO Training: UNSLOTH

FASTER Code for SFT + DPO Training: UNSLOTH

AutoRT: Auto Robotics Transformers LEARN

AutoRT: Auto Robotics Transformers LEARN

LLM Prompt FORMATS make or break your LLM and RAG

LLM Prompt FORMATS make or break your LLM and RAG

Self-Extend LLM: Upgrade your context length

Self-Extend LLM: Upgrade your context length

My Response to Recent Criticism: Facts and Clarifications

My Response to Recent Criticism: Facts and Clarifications

No LLM can solve this - Lost in the Middle

No LLM can solve this - Lost in the Middle

LLama-2 7B: 400K context length - Beyond Limits?

LLama-2 7B: 400K context length - Beyond Limits?

LLMs FAIL at 2K context length - Yours Too?

LLMs FAIL at 2K context length - Yours Too?

Supercharge Multi-LLM Intelligence w/ CALM

Supercharge Multi-LLM Intelligence w/ CALM

Rise of AI Companions

Rise of AI Companions

Modular RAG: Essential for AI in 2024?

Modular RAG: Essential for AI in 2024?

A NEW TRUTH: AI in 2024

A NEW TRUTH: AI in 2024

2024 - and now?

2024 - and now?

Mistral MEDIUM vs Mixtral 8x7B: 4x more powerful?

Mistral MEDIUM vs Mixtral 8x7B: 4x more powerful?

MAMBA (S6) Fine-Tuned + DPO-Aligned: TEST

MAMBA (S6) Fine-Tuned + DPO-Aligned: TEST

BEYOND MAMBA AI (S6): Vector FIELDS

BEYOND MAMBA AI (S6): Vector FIELDS

New: AI Agent Self-Improvement + Self-Fine-Tune

New: AI Agent Self-Improvement + Self-Fine-Tune

MEDICAL Prompt Engineering PRO + Clinical RAG

MEDICAL Prompt Engineering PRO + Clinical RAG

2024's EASY GUIDE to PROMPT Engineering #ai

2024's EASY GUIDE to PROMPT Engineering #ai

MAMBA AI (S6): Better than Transformers?

MAMBA AI (S6): Better than Transformers?

How to START with AI: Real-time DATA / RAG!

How to START with AI: Real-time DATA / RAG!

How to START with AI: Beyond CODE?

How to START with AI: Beyond CODE?

The BEST 12 LLM on this Planet: ONE is Open-Source

The BEST 12 LLM on this Planet: ONE is Open-Source

GEMINI PRO Outclassed by Mixtral 8x7B? Live Test!

GEMINI PRO Outclassed by Mixtral 8x7B? Live Test!

MIXTRAL 8x7B MoE Instruct: LIVE Performance Test

MIXTRAL 8x7B MoE Instruct: LIVE Performance Test

Mixture of Experts LLM - MoE explained in simple terms

Mixture of Experts LLM - MoE explained in simple terms

Open Source AI Under the Shadow of META

Open Source AI Under the Shadow of META

Best 7B LLM: Notus (w/ LoRA) better than ZEPHYR?

Best 7B LLM: Notus (w/ LoRA) better than ZEPHYR?

BEST 2 out of 7 LLMs for Causal Reasoning (7B-70B)

BEST 2 out of 7 LLMs for Causal Reasoning (7B-70B)

INTEL LLM: Neural-chat-7b vs PPLX-7B-online

INTEL LLM: Neural-chat-7b vs PPLX-7B-online

The Future of Shopping - on Amazon w/ AWS TITAN AI

The Future of Shopping - on Amazon w/ AWS TITAN AI

New PPLX-7B vs Mistral-7B vs PPLX-70B: Performance LLMs

New PPLX-7B vs Mistral-7B vs PPLX-70B: Performance LLMs

Q* (Q-star) by OpenAI Explained (no AGI)

Q* (Q-star) by OpenAI Explained (no AGI)

The NEXT Q* moment - From Hype to Substance

The NEXT Q* moment - From Hype to Substance

From LLM to complete System AI (Sensor AI)

From LLM to complete System AI (Sensor AI)

AI for Climate Tech: UN COP28 w/ JAIS CLIMATE AI

AI for Climate Tech: UN COP28 w/ JAIS CLIMATE AI

NEW Inflection-2: The 2nd best LLM in the world?

NEW Inflection-2: The 2nd best LLM in the world?

Stable Diffusion x10: LCM-LoRA (CODE & Theory)

Stable Diffusion x10: LCM-LoRA (CODE & Theory)

Will OpenAI cease to exist in 24h? Open Letter to Board

Will OpenAI cease to exist in 24h? Open Letter to Board

Sam's Legacy: ChatGPT 4 counteracts an AGI (AI Safety)

Sam's Legacy: ChatGPT 4 counteracts an AGI (AI Safety)

After RAG, Vector & GPT Store: NEW AI Breakthrough UNFOLDS

After RAG, Vector & GPT Store: NEW AI Breakthrough UNFOLDS

RAG's Collapse: Uncovering Deep Flaws in LLM External Knowledge Retrieval

RAG's Collapse: Uncovering Deep Flaws in LLM External Knowledge Retrieval

OpenAI's new ASSISTANTS API: How To Create GPTs

OpenAI's new ASSISTANTS API: How To Create GPTs

Fine-tune my Coding-LLM w/ PEFT LoRA Quantization - PART 2

Fine-tune my Coding-LLM w/ PEFT LoRA Quantization - PART 2

Fine-tune my Coding-LLM w/ PEFT LoRA Quantization

Fine-tune my Coding-LLM w/ PEFT LoRA Quantization

New GPT-4 Turbo Explains PYTHON Code, sums PDFs

New GPT-4 Turbo Explains PYTHON Code, sums PDFs

Dev Days of OpenAI and GitHub compared: CODE DEV AI 2024

Dev Days of OpenAI and GitHub compared: CODE DEV AI 2024

PEFT w/ Multi LoRA explained (LLM fine-tuning)

PEFT w/ Multi LoRA explained (LLM fine-tuning)

New GPT-4 TURBO 128K & GPT Store by OpenAI

New GPT-4 TURBO 128K & GPT Store by OpenAI

RAG optimized PEFT-LoRA: Your Questions answered

RAG optimized PEFT-LoRA: Your Questions answered

LLama 2 + PEFT Docs: CODE interactive LLM w/ RAG

LLama 2 + PEFT Docs: CODE interactive LLM w/ RAG

AI Phone - LLM Quantization, Privacy, Fine-Tuning, Reasoning

AI Phone - LLM Quantization, Privacy, Fine-Tuning, Reasoning

New LLM-Quantization LoftQ outperforms QLoRA

New LLM-Quantization LoftQ outperforms QLoRA

LLM as Next-Gen Operating System (MemGPT, Berkeley)

LLM as Next-Gen Operating System (MemGPT, Berkeley)

Generative AI: Latest AI Insights

Generative AI: Latest AI Insights

Self-Reflective AI: Self-RAG for Multi-AI-Agents explained

Self-Reflective AI: Self-RAG for Multi-AI-Agents explained

NEW FUYU-8B LLM (multimodal)

NEW FUYU-8B LLM (multimodal)

NEFTune: NEW LLM Fine-Tuning plus 25% Performance

NEFTune: NEW LLM Fine-Tuning plus 25% Performance

Train MISTRAL 7B to outperform LLama 2 70B (ZEPHYR 7B Alpha)

Train MISTRAL 7B to outperform LLama 2 70B (ZEPHYR 7B Alpha)

20 LLMs: Pure Reasoning CHAOS! Are small LLM better?

20 LLMs: Pure Reasoning CHAOS! Are small LLM better?

New LLaVA AI explained: GPT-4 VISION's Little Brother

New LLaVA AI explained: GPT-4 VISION's Little Brother

Scientific Problem-Solving: GPT-4 VISION (multimodal)

Scientific Problem-Solving: GPT-4 VISION (multimodal)

NEW StreamingLLM by MIT & Meta: Code explained

NEW StreamingLLM by MIT & Meta: Code explained

Explore AI COHERE CORAL w/ RAG enabled

Explore AI COHERE CORAL w/ RAG enabled

Multi-AI Autogen RAG fails. Quantum Analogon explains WHY!

Multi-AI Autogen RAG fails. Quantum Analogon explains WHY!

Chain-of-Verification: How to fight AI Hallucination

Chain-of-Verification: How to fight AI Hallucination

The Future of AI: GPT-4 Video, autonomous AI Agent

The Future of AI: GPT-4 Video, autonomous AI Agent

AutoGen: Multi-Agent AI Conversation, complete code

AutoGen: Multi-Agent AI Conversation, complete code

MISTRAL 7B: Reasoning, ICL and Multi-Agent Testing

MISTRAL 7B: Reasoning, ICL and Multi-Agent Testing

MISTRAL 7B explained - Preview of LLama3 LLM

MISTRAL 7B explained - Preview of LLama3 LLM

How to code long-context LLM: LongLoRA explained on LLama 2 100K

How to code long-context LLM: LongLoRA explained on LLama 2 100K

Multimodal LLM: Microsoft's new KOSMOS-2.5 for Image Text

Multimodal LLM: Microsoft's new KOSMOS-2.5 for Image Text

Mathematics w/ Donut AI and Nougat AI - Swin Transformer

Mathematics w/ Donut AI and Nougat AI - Swin Transformer

US AirForce: Synthetic Data for Vision Language AI

US AirForce: Synthetic Data for Vision Language AI

NEW: Chain of Density Prompt (CoD), LLM + Knowledge Graph Prompt

NEW: Chain of Density Prompt (CoD), LLM + Knowledge Graph Prompt

Create for free: AI Photos (Text2Image SDXL)

Create for free: AI Photos (Text2Image SDXL)

Stable Diffusion XL T2I-Adapter, ControlLoRA w/ CODE & Demo

Stable Diffusion XL T2I-Adapter, ControlLoRA w/ CODE & Demo

AGI - InstructDiffusion by Microsoft: better than SDXL?

AGI - InstructDiffusion by Microsoft: better than SDXL?

SDXL explained w/ shape-shifting ControlNet (Stanford Univ)

SDXL explained w/ shape-shifting ControlNet (Stanford Univ)

Tutorial: SDXL Web UI - for Beginners

Tutorial: SDXL Web UI - for Beginners

When You Buy AI Products: LEGAL Implications - Gedankenexperiment

When You Buy AI Products: LEGAL Implications - Gedankenexperiment

New Tutorial on LLM Quantization w/ QLoRA, GPTQ and Llamacpp, LLama 2

New Tutorial on LLM Quantization w/ QLoRA, GPTQ and Llamacpp, LLama 2

LLama2 locally on Mac or PC with GGUF

LLama2 locally on Mac or PC with GGUF

AI in Academia: ChatGPT's Role in Enhancing Education

AI in Academia: ChatGPT's Role in Enhancing Education

NEW FALCON-180B: create 3 AGENTS simultaneously?

NEW FALCON-180B: create 3 AGENTS simultaneously?

FALCON-180B LLM: GPU configuration w/ Quantization QLoRA - GPTQ

FALCON-180B LLM: GPU configuration w/ Quantization QLoRA - GPTQ

NEW Falcon-180B: FINALLY intelligent?

NEW Falcon-180B: FINALLY intelligent?

Code Interpreter GPT-4: multiple files upload Adventure

Code Interpreter GPT-4: multiple files upload Adventure

AI for Teacher - Teaching with AI OpenAI, plus How to Create an AI Tutor

AI for Teacher - Teaching with AI OpenAI, plus How to Create an AI Tutor

Q: How to create an Instruction Dataset for Fine-tuning my LLM?

Q: How to create an Instruction Dataset for Fine-tuning my LLM?

Graph-of-Thoughts (GoT) for AI reasoning Agents

Graph-of-Thoughts (GoT) for AI reasoning Agents

GPT-4 Code Interpreter: QLoRA for RL in Robotics

GPT-4 Code Interpreter: QLoRA for RL in Robotics

How to Code RLHF on LLama2 w/ LoRA, 4-bit, TRL, DPO

How to Code RLHF on LLama2 w/ LoRA, 4-bit, TRL, DPO

GPT-4 Prompt to Optimize your Text for Flow and Clarity

GPT-4 Prompt to Optimize your Text for Flow and Clarity

WizardCoder 34B: Complex Fine-Tuning Explained

WizardCoder 34B: Complex Fine-Tuning Explained

Q: How put 1000 PDFs into my LLM?

Q: How put 1000 PDFs into my LLM?

WizardCoder - Python OUTPERFORMS Code LLama - Python 34B

WizardCoder - Python OUTPERFORMS Code LLama - Python 34B

Stack Overflow AI

Stack Overflow AI

UNNATURAL "Code LLAMA 2": The Secret Version

UNNATURAL "Code LLAMA 2": The Secret Version

Fine-tune LLama2 w/ PEFT, LoRA, 4bit, TRL, SFT code #llama2

Fine-tune LLama2 w/ PEFT, LoRA, 4bit, TRL, SFT code #llama2

NEW: CODE LLAMA - New Coding LLM by META

NEW: CODE LLAMA - New Coding LLM by META

Code to Fine-tune ChatGPT w/ synthetic GPT-4 dataset

Code to Fine-tune ChatGPT w/ synthetic GPT-4 dataset

AI Agents explained: An Introduction

AI Agents explained: An Introduction

META Speech Translation w/ new AI: SeamlessM4T

META Speech Translation w/ new AI: SeamlessM4T

META releases new Translation AI: SeamlessM4T for 100 languages

META releases new Translation AI: SeamlessM4T for 100 languages

TOOL LLM for API calls explained

TOOL LLM for API calls explained

GPT-4 Prompt to Create a Thematic Overview

GPT-4 Prompt to Create a Thematic Overview

LLama-2 controls Robots: The most beautiful explanation! #ai

LLama-2 controls Robots: The most beautiful explanation! #ai

How To Become a CREATIVE WRITER w/ ChatGPT Agents

How To Become a CREATIVE WRITER w/ ChatGPT Agents