System Hardware Monitoring (Supports Ascend)
During the tracking of experiments, SwanLab automatically monitors the hardware resources of the machine and records them in the "System" charts.
Currently, SwanLab supports monitoring hardware resources for two types of AI computing chips (Huawei Ascend and NVIDIA), covering metrics such as GPU utilization, VRAM usage, GPU temperature, and GPU power consumption.
In addition, SwanLab also supports monitoring hardware resources such as CPU, memory, and disk.
We are thrilled to collaborate with the Ascend Computing team to offer trainers more opportunities to experience domestic computing power.
NPU Monitoring Charts:
AI Chip Environment Records:
