TEST SETUP: rented cloud GPUs (RunPod / Vast.ai / Modal), one methodology for every card
LLMs: llama.cpp llama-bench, tokens/sec · IMAGE / VIDEO: diffusers pipelines, images/min or frames/sec
MODELS IN THIS VIDEO: Llama 3.1 8B (Q4_K_M) · LTX-Video (distilled) (30 steps) · Qwen2.5-Coder 14B (Q4_K_M) · Qwen3 32B · Qwen3 4B (Q4_K_M) · Stable Diffusion XL (30 steps) · Wan 2.2 5B (720p) · Z-Image Turbo (8 steps)
DATA: gpubattle.com — every result is labeled measured or estimated on the page
Today, we're going to be looking at the RTX 4080 Super against the RTX 4090 for AI workloads. Check out the full web page here.
https://gpubattle.com/ai/geforce-rtx-4090?utm_source=youtube
LLMs: llama.cpp llama-bench, tokens/sec · IMAGE / VIDEO: diffusers pipelines, images/min or frames/sec
MODELS IN THIS VIDEO: Llama 3.1 8B (Q4_K_M) · LTX-Video (distilled) (30 steps) · Qwen2.5-Coder 14B (Q4_K_M) · Qwen3 32B · Qwen3 4B (Q4_K_M) · Stable Diffusion XL (30 steps) · Wan 2.2 5B (720p) · Z-Image Turbo (8 steps)
DATA: gpubattle.com — every result is labeled measured or estimated on the page
Today, we're going to be looking at the RTX 4080 Super against the RTX 4090 for AI workloads. Check out the full web page here.
https://gpubattle.com/ai/geforce-rtx-4090?utm_source=youtube
- Категория
- AI Картинки
Выполните вход или зарегистрируйтесь для отправки комментария.
Комментариев нет.
