Patreon
Home / Videos / No Powerful GPU? Train a FLUX LoRA in the Cloud in Under 1 H

No Powerful GPU? Train a FLUX LoRA in the Cloud in Under 1 Hour (2024 Kohya Guide)

Kohya SS GUI FLUX LoRA training step by step, with the best optimized configs for 4 GB to 48 GB GPUs and a 68 minute Windows tutorial with 74 chapters.

Published Sep 04, 2024 · 19,983 views · Watch on YouTube

Get the app and files

Kohya FLUX LoRA Training: Full Tutorial & Configs for Windows & Cloud

Kohya FLUX LoRA Training: Full Tutorial & Configs for Windows & Cloud

Kohya SS GUI FLUX LoRA training step by step, with the best optimized configs for 4 GB to 48 GB GPUs and a 68 minute Windows tutorial with 74 chapters.

Kohya FLUX Fine Tuning / DreamBooth Training Tutorial: Windows & Cloud

Kohya FLUX Fine Tuning / DreamBooth Training Tutorial: Windows & Cloud

Kohya FLUX full fine tuning / DreamBooth trains full checkpoints with configs for GPUs with as low as 8 GB VRAM, all using Torch Compile on Windows and Linux.

Decoding the Timestep Shift in Diffusion Training: Qwen Image & FLUX

Decoding the Timestep Shift in Diffusion Training: Qwen Image & FLUX

Timestep sampling shift explained for diffusion models: Qwen Image uses an exponential shift and FLUX a linear one, and 2.205 matches Qwen at 1664x928.

Kohya GUI Kaggle Notebook: Free SD 1.5 & SDXL Training on Dual T4 GPUs

Kohya GUI Kaggle Notebook: Free SD 1.5 & SDXL Training on Dual T4 GPUs

Kohya GUI Kaggle notebook trains Stable Diffusion 1.5 and SDXL LoRA or DreamBooth models for free on a Kaggle account, on a single T4 or dual T4 GPUs.

FLUX LoRA vs Fine Tuning / DreamBooth Comparison: Batch Size 1 vs 7

FLUX LoRA vs Fine Tuning / DreamBooth Comparison: Batch Size 1 vs 7

FLUX LoRA and Fine Tuning / DreamBooth compared with 15 and 256 training images at batch size 1 and 7, for both realism and stylization.

Kohya GUI: Full FLUX Dev DreamBooth on 6 GB GPUs Without Quality Loss

Kohya GUI: Full FLUX Dev DreamBooth on 6 GB GPUs Without Quality Loss

Kohya GUI can now fully fine-tune FLUX Dev on 6 GB GPUs with the same quality as 48 GB GPUs, and fine-tuning gives better results than any LoRA training.

Chapters

  1. 0:00 Introduction to FLUX Training on Cloud Services (Massed Compute and RunPod)
  2. 0:45 Overview of Platform Differences and Why Massed Compute is Preferred for FLUX Training
  3. 2:01 Using FLUX, Kohya GUI, and Using 4x GPUs for Fast Training
  4. 3:08 Exploring Massed Compute Coupons and Discounts: How to Save on GPU Costs
  5. 5:35 Detailed Setup for Training FLUX on Massed Compute: Account Creation, Billing, and Deploying Instances
  6. 6:59 Deploying Multiple GPUs on Massed Compute for Faster Training
  7. 8:53 Setting Up ThinLinc Client for File Transfers Between Local Machine and Cloud
  8. 9:04 Troubleshooting ThinLinc File Transfer Issues on Massed Compute
  9. 9:25 Preparing to Install Kohya GUI and Download Necessary Models on Massed Compute
  10. 10:02 Upgrading to the Latest Version of Kohya for FLUX Training
  11. 11:02 Downloading FLUX Training Models and Preparing the Dataset
  12. 11:53 Checking VRAM Usage with nvitop: Real-Time Monitoring During FLUX Training
  13. 13:33 Speed Optimization Tips: Disabling T5 Attention Mask for Faster Training
  14. 17:44 Understanding the Trade-offs: Applying T5 Attention Mask vs. Training Speed
  15. 18:40 Setting Up Multi-GPU Training for FLUX on Massed Compute
  16. 18:52 Adjusting Epochs and Learning Rate for Multi-GPU Training
  17. 22:24 Achieving Near-Linear Speed Gain with 4x GPUs on Massed Compute
  18. 24:34 Uploading FLUX LoRAs to Hugging Face for Easy Access and Sharing
  19. 24:56 Using SwarmUI on Your Local Machine via Cloudflare for Image Generation
  20. 26:04 Moving Models to the Correct Folders in SwarmUI for FLUX Image Generation
  21. 27:07 Setting Up and Running Grid Generation to Compare Different Checkpoints
  22. 30:43 Downloading and Managing LoRAs and Models on Hugging Face
  23. 33:35 Generating Images with FLUX on SwarmUI and Finding the Best Checkpoints
  24. 38:22 Advanced Configurations in SwarmUI for Optimized Image Generation
  25. 39:25 How to Use Forge Web UI with FLUX Models on Massed Compute
  26. 39:33 Setting Up and Configuring Forge Web UI for FLUX on Massed Compute
  27. 40:03 Moving Models and LoRAs to Forge Web UI for Image Generation
  28. 41:15 Generating Images with LoRAs on Forge Web UI
  29. 44:38 Transition to RunPod: Setting Up FLUX Training and Using SwarmUI/Forge Web UI
  30. 45:13 RunPod Network Volume Storage: Setup and Integration with FLUX Training
  31. 45:49 Differences Between Massed Compute and RunPod: Speed, Cost, and Hardware
  32. 47:19 Deploying Instances on RunPod and Setting Up JupyterLab
  33. 48:05 Installing Kohya GUI and Downloading Models for FLUX Training on RunPod
  34. 48:48 Preparing Datasets and Starting FLUX Training on RunPod
  35. 51:55 Monitoring VRAM and Training Speed on RunPod’s A40 GPUs
  36. 56:42 Optimizing Training Speed by Disabling T5 Attention Mask on RunPod
  37. 58:20 Comparing GPU Performance Across Platforms: A6000 vs A40 in FLUX Training
  38. 58:38 Setting Up Multi-GPU Training on RunPod for Faster FLUX Training
  39. 58:58 Adjusting Learning Rate and Epochs for Multi-GPU Training on RunPod
  40. 63:41 Achieving Near-Linear Speed Gain with Multi-GPU FLUX Training on RunPod
  41. 65:46 Completing FLUX Training on RunPod and Preparing Models for Use
  42. 65:52 Managing Multiple Checkpoints: Best Practices for FLUX Training
  43. 66:04 Using SwarmUI on RunPod for Image Generation with FLUX LoRAs
  44. 68:18 Setting Up Multiple Backends on SwarmUI for Multi-GPU Image Generation
  45. 70:50 Generating Images and Comparing Checkpoints on SwarmUI on RunPod
  46. 71:55 Uploading FLUX LoRAs to Hugging Face from RunPod for Easy Access
  47. 72:08 Advanced Download Techniques: Using Hugging Face CLI for Batch Downloads
  48. 75:16 Fast Download and Upload of Models and LoRAs on Hugging Face
  49. 77:14 Using Forge Web UI on RunPod for Image Generation with FLUX LoRAs
  50. 78:01 Troubleshooting Installation Issues with Forge Web UI on RunPod
  51. 83:25 Generating Images on Forge Web UI with FLUX Models and LoRAs
  52. 84:20 Conclusion and Upcoming Research on Fine-Tuning FLUX with CLIP Large Models

About this video

Unlock the power of FLUX LoRA training, even if you're short on GPUs or looking to boost speed and scale! This comprehensive guide takes you from novice to expert, showing you how to use Kohya GUI for creating top-notch FLUX LoRAs in the cloud. We'll cover everything: maximizing quality, optimizing speed, and finding the best deals. With our exclusive Massed Compute discount, you can rent 4x RTX A6000 GPUs for just $1.25 per hour, supercharging your training process. Learn how to leverage RunPod for both cost-effective computing and permanent storage. We'll also dive into lightning-fast uploads of your training checkpoints to Hugging Face, seamless downloads, and integrating LoRAs with popular tools like SwarmUI and Forge Web UI. Get ready to master the art of efficient, high-quality AI model training!

🔗 Full Instructions and Links Written Post (the one used in the tutorial) ⤵️

▶️ https://www.patreon.com/posts/click-to-open-post-used-in-tutorial-110879657

0:00 Introduction to FLUX Training on Cloud Services (Massed Compute and RunPod)

0:45 Overview of Platform Differences and Why Massed Compute is Preferred for FLUX Training

2:01 Using FLUX, Kohya GUI, and Using 4x GPUs for Fast Training

3:08 Exploring Massed Compute Coupons and Discounts: How to Save on GPU Costs

5:35 Detailed Setup for Training FLUX on Massed Compute: Account Creation, Billing, and Deploying Instances

6:59 Deploying Multiple GPUs on Massed Compute for Faster Training

8:53 Setting Up ThinLinc Client for File Transfers Between Local Machine and Cloud

9:04 Troubleshooting ThinLinc File Transfer Issues on Massed Compute

9:25 Preparing to Install Kohya GUI and Download Necessary Models on Massed Compute

10:02 Upgrading to the Latest Version of Kohya for FLUX Training

11:02 Downloading FLUX Training Models and Preparing the Dataset

11:53 Checking VRAM Usage with nvitop: Real-Time Monitoring During FLUX Training

13:33 Speed Optimization Tips: Disabling T5 Attention Mask for Faster Training

17:44 Understanding the Trade-offs: Applying T5 Attention Mask vs. Training Speed

18:40 Setting Up Multi-GPU Training for FLUX on Massed Compute

18:52 Adjusting Epochs and Learning Rate for Multi-GPU Training

22:24 Achieving Near-Linear Speed Gain with 4x GPUs on Massed Compute

24:34 Uploading FLUX LoRAs to Hugging Face for Easy Access and Sharing

24:56 Using SwarmUI on Your Local Machine via Cloudflare for Image Generation

26:04 Moving Models to the Correct Folders in SwarmUI for FLUX Image Generation

27:07 Setting Up and Running Grid Generation to Compare Different Checkpoints

30:43 Downloading and Managing LoRAs and Models on Hugging Face

33:35 Generating Images with FLUX on SwarmUI and Finding the Best Checkpoints

38:22 Advanced Configurations in SwarmUI for Optimized Image Generation

39:25 How to Use Forge Web UI with FLUX Models on Massed Compute

39:33 Setting Up and Configuring Forge Web UI for FLUX on Massed Compute

40:03 Moving Models and LoRAs to Forge Web UI for Image Generation

41:15 Generating Images with LoRAs on Forge Web UI

44:38 Transition to RunPod: Setting Up FLUX Training and Using SwarmUI/Forge Web UI

45:13 RunPod Network Volume Storage: Setup and Integration with FLUX Training

45:49 Differences Between Massed Compute and RunPod: Speed, Cost, and Hardware

47:19 Deploying Instances on RunPod and Setting Up JupyterLab

48:05 Installing Kohya GUI and Downloading Models for FLUX Training on RunPod

48:48 Preparing Datasets and Starting FLUX Training on RunPod

51:55 Monitoring VRAM and Training Speed on RunPod’s A40 GPUs

56:42 Optimizing Training Speed by Disabling T5 Attention Mask on RunPod

58:20 Comparing GPU Performance Across Platforms: A6000 vs A40 in FLUX Training