Tags¶

Pages grouped by hardware, backend and workload. Hardware tags mark pages about specific hardware; general setup guides carry only their backend and workload tags.

AMD iGPU

• AMD Ryzen AI 300 NPU versus Radeon 890M
• Socket power and energy on Strix Point
• Speculative decoding on Radeon 890M
• Whisper accuracy and speed on FLEURS

CPU

• Whisper accuracy and speed on FLEURS
• llama.cpp, Ollama & llama-swap

CUDA

• Check GPU hardware
• ComfyUI + TRELLIS
• Whisper accuracy and speed on FLEURS
• llama.cpp, Ollama & llama-swap
• vLLM

Intel iGPU

• Whisper accuracy and speed on FLEURS

LLM inference

• AMD Ryzen AI 300 NPU versus Radeon 890M
• Fine-tunes on the NPU
• Ryzen AI NPU
• Socket power and energy on Strix Point
• Speculative decoding on Radeon 890M
• llama.cpp, Ollama & llama-swap
• vLLM

NPU

• AMD Ryzen AI 300 NPU versus Radeon 890M
• Fine-tunes on the NPU
• Ryzen AI NPU
• Socket power and energy on Strix Point
• Whisper accuracy and speed on FLEURS

NVIDIA GPU

• Whisper accuracy and speed on FLEURS

ROCm

• Check GPU hardware
• ComfyUI + TRELLIS
• Fine-tuning, Unsloth & RAG
• Socket power and energy on Strix Point
• Whisper accuracy and speed on FLEURS
• llama.cpp, Ollama & llama-swap
• vLLM

Vulkan

• AMD Ryzen AI 300 NPU versus Radeon 890M
• Socket power and energy on Strix Point
• Speculative decoding on Radeon 890M
• Whisper accuracy and speed on FLEURS
• llama.cpp, Ollama & llama-swap

fine-tuning

• Fine-tunes on the NPU
• Fine-tuning, Unsloth & RAG

image and 3D

• ComfyUI + TRELLIS

speech-to-text

• Ryzen AI NPU
• Socket power and energy on Strix Point
• Whisper accuracy and speed on FLEURS