It is now possible to generate 16 Megapixel (4096x4096) raw images with SANA 4K model using under 8GB VRAM, 4 Megapixel (2048x2048) images using under 6GB VRAM, and 1 Megapixel (1024x1024) images using under 4GB VRAM thanks to new optimizations

Tried Expressions with FLUX LoRA training with my new training dataset (includes expressions and used 256 images (image 19) as experiment) - even learnt body shape perfectly - prompts, workflow and more information at the oldest comment

20 New SDXL Fine Tuning Tests and Their Results

Most Awaited Full Fine Tuning (with DreamBooth effect) Tutorial Generated Images - Full Workflow Shared In The Comments - NO Paywall This Time - Explained OneTrainer - Cumulative Experience of 16 Months Stable Diffusion

Wow this is what I am looking forward to. Driverless flying cars just got approval from regulators in China! 🤯

So I upgraded to a 5090

How can i make cursor to automate browser when developing a web based app?

Unpopular Opinion: The Cursor hate is MOSTLY unjustified

Is Nvidia 5090 Python incompatible with CUDA,

Open Sourcing TripoSG: High-Fidelity 3D Generation from Single Images using Large-Scale Flow Models (1.5B Model Released!)

General humanoid robots started to getting deployed. They took our jobs!

I have Compared Kohya vs OneTrainer for FLUX Dev Finetuning / DreamBooth Training

The most advanced DeepFake / FaceSwap app FaceFusion 3.1.1 now working perfect on RTX 5000 series as well - Our installers upgraded to support RTX 5000 series

Blown away by item arrangement and text in GPT4o - seems like nothing compares

To Pro 6000 or not to Pro 6000

gpt 4o image generator is amazing, any chance we are getting something similar open source?

Can vertical GPU panel mounted right side of the HAF 700 Evo instead of left? I don't need hdmi cable output gonna use rtx 5090 + 3090

The real and authentic useage case of ChatGPT image generation. We definitely need such model from China as open source

I use the best cooler available but CPU throttling is real for AMD 9950X

Working on improved Trellis App with Batch Features and support for RTX 5000 series

Thumbnail designers are COOKED (X: @theJosephBlaze)

Is is true some repos on Github contains malware?

Speeding up ComfyUI workflows using TeaCache and Model Compiling - experimental results

Testing OneTrainer FLUX training - 4x config. Also training with our best Kohya settings to compare.

Rented 1000 GB machine to pre-compile Flash Attention and Sage Attention for Linux to increase my scripts installation speed on RunPod and Massed Compute for my followers.