diff --git a/notebook/experiment/rwkv-x-exp/multi-size-train/v5-L12-D2048-part1.ipynb b/notebook/experiment/rwkv-x-exp/multi-size-train/v5-L12-D2048-part1.ipynb index 01b3bfc4..ae8b955f 100644 --- a/notebook/experiment/rwkv-x-exp/multi-size-train/v5-L12-D2048-part1.ipynb +++ b/notebook/experiment/rwkv-x-exp/multi-size-train/v5-L12-D2048-part1.ipynb @@ -36,7 +36,7 @@ "metadata": {}, "outputs": [], "source": [ - "DEEPSPEED_STRAT=\"deepspeed_stage_1\"\n", + "DEEPSPEED_STRAT=\"deepspeed_stage_2_offload\"\n", "GPU_DEVICES=\"auto\"\n", "ENABLE_WANDB=True\n", "\n",