Главная
Study mode:
on
1
[] Introduction to Mark Kim-Huang
2
[] Join the LLMs in Production Conference Part 2 on June 15-16!
3
[] Fine-Tuning LLMs: Best Practices and When to Go Small
4
[] Model approaches
5
[] You might think that you could just use OpenAI but only older base models are available
6
[] Why custom LLMs over closed source models?
7
[] Small models work well for simple tasks
8
[] Types of Fine-Tuning
9
[] Strategies for improving fine-tuning performance
10
[] Challenges
11
[] Define your task
12
[] Task framework
13
[] Defining tasks
14
[] Clustering task diversifies training data and improves out-of-domain performance
15
[] Prompt engineering
16
[] Constructing a prompt
17
[] Synthesize more data
18
[] Constructing a prompt
19
[] Increase fine-tuning efficiency with LoRa
20
[] Naive data parallelism with mixed precision is inefficient
21
[] Further reading on mixed precision
22
[] Parameter efficient fine-tuning with LoRa
23
[] LoRa Data Parallel with Mixed Precision
24
[] Summary
25
[] Q&A
26
[] Mark's journey to LLMs
27
[] Task clustering mixing with existing data sets
28
[] LangChain Auto Evaluator evaluating LLMs
29
[] Cloud platforms costs
30
[] Vector database used at Preemo
31
[] Finding a reasoning path of a model on Prompting
32
[] When to fine-tune versus prompting with a context window
33
[] Wrap up
Description:
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only! Grab it Explore best practices for fine-tuning Large Language Models (LLMs) and learn when to opt for smaller models in this 54-minute talk by Mark Kim-Huang, Co-Founder and Head of AI at Preemo Inc. Dive into challenges and state-of-the-art techniques for fine-tuning LLMs, including strategies to improve performance, efficient data parallelism, and parameter-efficient fine-tuning with LoRa. Discover the benefits of custom LLMs over closed-source models, understand various types of fine-tuning, and learn how to define tasks and construct effective prompts. The talk also covers data synthesis, cloud platform costs, vector databases, and the trade-offs between fine-tuning and prompting with context windows. Gain insights into Mark's journey to LLMs and participate in a Q&A session addressing topics such as task clustering, LangChain Auto Evaluator, and finding reasoning paths in models.

Fine-Tuning LLMs: Best Practices and When to Go Small - Lecture 124

MLOps.community
Add to list
#Computer Science #Machine Learning #Fine-Tuning #Artificial Intelligence #MLOps #Deep Learning #Stable Diffusion #LoRA (Low-Rank Adaptation) #Model Optimization
0:00 / 0:00
1 of 33

[] Introduction to Mark Kim-Huang