Hi guys, Welcome to the Advanced Video Search Engine, and you are watching I spent 3800$ on a new PC to use AI... Do I regret it?. That video is uploaded by Not4Talent at 2023-11-28T03:35:11-08:00. We are promoting this video only for entertainment and educational purpose only. So, I hope you like & enjoy it. [wQmVXnrBbrU]
| Name | I spent 3800$ on a new PC to use AI... Do I regret it? |
| Video Uploader | Video From Not4Talent |
| Upload Date | This Video Uploaded At 28-11-2023 11:35:11 |
| Video Discription | My old GTX 1080Ti was falling behind on AI. That's why I bought a new RTX 4090, that needed a fully new PC. Will the more than 3000 dollars spent be worth the cost? Well, let's live test and react to it. Sorry for the off-topic video, and with a poorly set up camera xD. Hope it helps someone out there but it mainly serves as an anouncement. Now my pc won't be a set back anymore, so expect more videos :3 Join our discrod server: https://discord.gg/8TSMNHvGCH to learn about this and more! Thanks to everyone that helped with this video! (Lefourbe did a lot more testing that I didn't know where to add. Ask on discord if you are interested) ------------- Links used in the VIDEO ---------- GTX 1080 TI vs RTX 4090: https://technical.city/es/video/GeForce-GTX-1080-Ti-vs-GeForce-RTX-4090 -----Optimizing your pc: --DarkStoorm's gide (check pinned comment too): https://gist.github.com/DarkStoorM/159d723c93dde06b17b6f251c46cd827 --Kenjin's tips: If you have not so much VRAM and want to optimize your upscaling without using --medvram or --lowvram You can check Tiled diffusion with Controlnet Tiles to optimize furthermore, it's a very good tool. You'll need to find the best parameters for you though to get the best speed and quality: https://github.com/pkuliyi2015/multidiffusion-upscaler-for-automatic1111. Removing --medvram or --lowvram if you have enough memory and they are active. If you have --no-half, you can replace it by --upcast-sampling (see https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/8782) Adding set PYTORCH_CUDA_ALLOC_CONF= garbage_collection_threshold:0.9,max_split_size_mb:512 in webui-user.bat to reduce memory fragmentation if you run the application for a long period of time. His Sources: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Optimizations --Another option you have is just installing ComfyUI, as it is more optimized and runs SDXL better than A1111. It's a bit more complex tho: https://github.com/comfyanonymous/ComfyUI -----GPU studies: Irendering & Puget Systems: https://irendering.net/the-best-gpus-for-stable-diffusion/ https://www.pugetsystems.com/labs/articles/stable-diffusion-performance-nvidia-geforce-vs-amd-radeon/ MSI's word: https://www.msi.com/blog/stable-diffusion-xl-best-value-rtx-graphics-card ------------- Useful stuff for this ---------- NEW PC SPECS: Tower: Cougar Duoface PRO Power Supply: FSP Hydro Pro Series 1000W 80+ Gold or similar Processor: Intel® Core™ i9-13900KF (24 cores, frequency from 2.2GHz to 5.80GHz, 36 MB Intel® Smart Cache) Motherboard: ASUS Z790-P or similar depending on availability (equal or superior capabilities will always be maintained) Cooling: Abysm Artic 360mm liquid cooling Hard Drives: 2TB SSD M.2 NVMe 4TB HDD SATA3 Memory: 64GB DDR5 5200MHz Graphics: ASUS Nvidia GeForce RTX 4090 24GB OC 3 x DisplayPort 2 x HDMI Wi-Fi: ASUS WiFi 6 (or similar, depending on availability) Bluetooth: Bluetooth 5.0 OLD PC SPECS: Tower: Cooler Master MasterBox TD500L USB 3.0 with Window Power Supply: Cooler Master V Series 750W 80 Plus Gold Full Modular Processor: Intel Core i7-8700 3.2GHz BOX Motherboard: Gigabyte Aorus GA-Z370 Gaming K3 CPU Cooler: Corsair Cooling Hydro Series H110I Hard Drive: Samsung SSD 970 EVO NVMe M.2 500 GB Memory: G.Skill Ripjaws V Red DDR4 2400 PC4-19200 16GB 2x8GB CL15 Graphics Card: MSI GeForce GTX 1080Ti Gaming X Trio 11GB GDDR5X ------------- Social Media ---------- -Instagram: https://www.instagram.com/not4talent_ai/ -Twitter: https://twitter.com/not4talent Make sure to subscribe if you want to learn about AI and grow with the community as we surf the AI wave :3 0:00 intro 0:17 Text2Image SD1.5 1:36 Text2Image SDXL 2:35 Extra Test with SDXL 3:15 Planned image 5:18 Batch Size of 8 6:40 Noise comparison 7:24 Max upscale 9:32 Max upscale XL 9:43 LORA training 11:40 LORAXL training 12:16 XYZ plot 12:42 Editing while Generating 13:46 Gaming While generating 15:02 Conclusion and resources 17:08 Thanks for watching wQmVXnrBbrU |
| Category | Education |
| Tags | Education Download MP4 | Education Download MP3 | Education Download MP4 360p | Education Download MP4 480p | Education Download MP4 720p | Education Download MP4 1080p |
