Muhtasham··on Benchmarking Llama 3.1 405B on 8x AMD MI300X GPUsThanks for detailed analysis, would be curios to see FP8 comparison too, given vllm has some custom kernels
Muhtasham··on Show HN: Dstack – an open-source engine for running GPU workloadsSuper useful for spinning LLM fine-tuning jobs!