From 1c0e10ec5cb700840ef3957fc15e6594bdd657ff Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Thu, 10 Oct 2024 21:26:24 -0700 Subject: [PATCH] Fix ray docs (#761) --- docs/Getting-Started-TPU-VM.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/Getting-Started-TPU-VM.md b/docs/Getting-Started-TPU-VM.md index d0728d1c1..20fdaa765 100644 --- a/docs/Getting-Started-TPU-VM.md +++ b/docs/Getting-Started-TPU-VM.md @@ -233,7 +233,7 @@ Then, **in a separate terminal**, you can submit a job to the cluster. To replic ```bash export RAY_ADDRESS=http://localhost:8265 # tell ray where the cluster is -python infra/launch_on_ray.py --tpu_type v4-32 --foreground --config_path config/gpt2_small.yaml --trainer.checkpointer.base_path gs://' +python infra/launch_on_ray.py --tpu_type v4-32 --foreground -- python src/levanter/main/train_lm.py --config_path config/gpt2_small.yaml --trainer.checkpointer.base_path gs://' ``` Even without `--foreground`, the job will be restarted if it fails. The `--tpu_type` flag is required, and should be