TensorProbe (code name: kj600) is a LLM pretrain debugger with model's torch module , optimizer status, collective communication tensor collection and aggregation. It also supports rule-based alerts.
最近更新: 接近2年前TensorProbe (code name: kj600) is a LLM pretrain debugger with model's torch module , optimizer status, collective communication tensor collection and aggregation. It also supports rule-based alerts.
最近更新: 接近2年前
