AI 工程训练营 - 使用 AWS SageMaker 构建、训练和部署模型
- 01. AI Engineering Bootcamp Learn AWS SageMaker with Patrik Szepesi
- 02. Course Introduction
- 03. Setting Up Our AWS Account
- 04. Set Up IAM Roles + Best Practices
- 05. AWS Security Best Practices
- 06. Set Up AWS SageMaker Domain
- 07. UI Domain Change
- 08. Setting Up SageMaker Environment
- 09. SageMaker Studio and Pricing
- 10. Setup SageMaker Server + PyTorch
- 11. HuggingFace Models, Sentiment Analysis, and AutoScaling
- 12. Get Dataset for Multiclass Text Classification
- 13. Creating Our AWS S3 Bucket
- 14. Uploading Our Training Data to S3
- 15. Exploratory Data Analysis - Part 1
- 16. Exploratory Data Analysis - Part 2
- 17. Data Visualization and Best Practices
- 18. Setting Up Our Training Job Notebook + Reasons to Use SageMaker
- 19. Python Script for HuggingFace Estimator
- 20. Creating Our Optional Experiment Notebook - Part 1
- 21. Creating Our Optional Experiment Notebook - Part 2
- 22. Encoding Categorical Labels to Numeric Values
- 23. Understanding the Tokenization Vocabulary
- 24. Encoding Tokens
- 25. Practical Example of Tokenization and Encoding
- 26. Creating Our Dataset Loader Class
- 27. Setting Pytorch DataLoader
- 28. Which Path Will You Take_
- 29. DistilBert vs. Bert Differences
- 30. Embeddings In A Continuous Vector Space
- 31. Introduction To Positional Encodings
- 32. Positional Encodings - Part 1
- 33. Positional Encodings - Part 2 (Even and Odd Indices)
- 34. Why Use Sine and Cosine Functions
- 35. Understanding the Nature of Sine and Cosine Functions
- 36. Visualizing Positional Encodings in Sine and Cosine Graphs
- 37. Solving the Equations to Get the Values for Positional Encodings
- 38. Introduction to Attention Mechanism
- 39. Query, Key and Value Matrix
- 40. Getting Started with Our Step by Step Attention Calculation
- 41. Calculating Key Vectors
- 42. Query Matrix Introduction
- 43. Calculating Raw Attention Scores
- 44. Understanding the Mathematics Behind Dot Products and Vector Alignment
- 45. Visualizing Raw Attention Scores in 2D
- 46. Converting Raw Attention Scores to Probability Distributions with Softmax
- 47. Normalization
- 48. Understanding the Value Matrix and Value Vector
- 49. Calculating the Final Context Aware Rich Representation for the Word _River_
- 50. Understanding the Output
- 51. Understanding Multi Head Attention
- 52. Multi Head Attention Example and Subsequent Layers
- 53. Masked Language Learning
- 54. Exercise Imposter Syndrome
- 55. Creating Our Custom Model Architecture with PyTorch
- 56. Adding the Dropout, Linear Layer, and ReLU to Our Model
- 57. Creating Our Accuracy Function
- 58. Creating Our Train Function
- 59. Finishing Our Train Function
- 60. Setting Up the Validation Function
- 61. Passing Parameters In SageMaker
- 62. Setting Up Model Parameters For Training
- 63. Understanding The Mathematics Behind Cross Entropy Loss
- 64. Finishing Our Script.py File
- 65. Quota Increase
- 66. Starting Our Training Job
- 67. Debugging Our Training Job With AWS CloudWatch
- 68. Analyzing Our Training Job Results
- 69. Creating Our Inference Script For Our PyTorch Model
- 70. Finishing Our PyTorch Inference Script
- 71. Setting Up Our Deployment
- 72. Deploying Our Model To A SageMaker Endpoint
- 73. Introduction to Endpoint Load Testing
- 74. Creating Our Test Data for Load Testing
- 75. Upload Testing Data to S3
- 76. Creating Our Model for Load Testing
- 77. Starting Our Load Test Job
- 78. Analyze Load Test Results
- 79. Deploying Our Endpoint
- 80. Creating Lambda Function to Call Our Endpoint
- 81. Setting Up Our AWS API Gateway
- 82. Testing Our Model with Postman, API Gateway and Lambda
- 83. Cleaning Up Resources
- 84. Thank You!