Explore a detailed comparison of logical and mathematical reasoning capabilities across three language models - DeepSeek R1 1.5B, Microsoft Phi3.5 3.8B, and LLama 3.2 3B - in this 26-minute technical evaluation video. Learn how to implement Chain of Thought (CoT) prompting techniques using Ollama and Langchain frameworks, with all models quantized to 4Int for performance analysis. Access the complete evaluation process through the provided Jupyter notebook to understand the comparative strengths and limitations of these models in handling complex reasoning tasks.
Evaluating LLM Performance: DeepSeek, Phi-3.5, and LLaMA 3 Using Chain of Thought Reasoning
The Machine Learning Engineer via YouTube
Overview
Syllabus
LLM,s: CoT Evaluation DeepSeek R1 1.B, Phi3.5 3.8B and LLama 3.2 3B #datascience #machinelearning
Taught by
The Machine Learning Engineer
Related Courses
-
Evaluación de Razonamiento y Cálculo en LLMs: DeepSeek, Phi-3.5 y LLama 3.2
-
Designing Prompt Templates for DeepSeek and Phi-3.5 Language Models
-
Diseñar Prompt Templates para DeepSeek 1.5 y Phi3.5 3.8 - GuÃa en Español
-
Building Multi-Agent RAG Systems with LangGraph and LangServe - Local LLM Implementation
-
Building RAG Applications with LangGraph: Multiple Local LLMs and LangServe Authentication
-
LangChain Multimodal Chat with Ollama and LLama 3.2 11B Vision
