On February 15, OpenAI launched Sora, a tool that uses AI to generate videos based on written prompts. Despite earlier demonstrations of similar tools by Google, Meta and Runaway, the quality of the videos generated by Sora is amazing. The quality, duration and authenticity of the videos are among the best available. Frankly it’s hard to tell if it’s AI generated without careful screening. And with a 60-second video length there is no doubt that quality creations on video apps like TikTok will explode.
Until 2023, if you want to get started with AIGC model training, you’ll still be limited to PCs with NVIDIA graphics, which are the only ones that have both a CUDA-accelerated instruction and large video memory. Only NVIDIA cards have both CUDA-accelerated instruction sets and large video memory, and these PCs typically start at $800.
With the release of AMD first processor with an integrated AI compute engine, the 7840HS, in 2023, the bar for AIGC has been lowered to start at $500. the 7840HS has a built-in NPU for the AMD XDNA™ architecture. Up to 10 TOPS of arithmetic can be delivered through the Ryzen™ AI engine. PyTorch, TensorFlow and ONNX models are supported. The built-in Radeon™ 780M supports calling memory as video memory. Also AMD ROCm™ software supports emulating CUDA.
Nowadays, the most cost-effective way to get started with AIGC model training is to use the 7840HS mini PC for local deployment. Compared with traditional AI cloud processing computing, it is not only cheaper but also can be carried out without network. PELADN HA-4 mini host built-in 7840HS cores provide powerful and reliable performance to help training generative AI models.
$559
AMD Ryzen 7 7840HS Processor
32GB DDR4 3200 MHz
1TB M.2 2280 SSD
Radeon™ 780M
AMD XDNA™
Ryzen™ AI
AMD ROCm™
Leave a comment
This site is protected by hCaptcha and the hCaptcha Privacy Policy and Terms of Service apply.