瀏覽代碼

Update examples/albert/README.md

Co-authored-by: Alexander Borzunov <hxrussia@gmail.com>
justheuristic 3 年之前
父節點
當前提交
7eff0cccac
共有 1 個文件被更改,包括 1 次插入1 次删除
  1. 1 1
      examples/albert/README.md

+ 1 - 1
examples/albert/README.md

@@ -55,7 +55,7 @@ To join the collaboration with a GPU trainer,
   (see [default paths](./arguments.py#L117-L134) for reference)
 - Run:
   ```bash
-  ./run_trainer.py  --initial_peers ONE_OR_MORE_PEERS --per_device_train_batch_size BATCH_SIZE_FOR_YOUR_GPU_DEFAULT_4
+  ./run_trainer.py  --initial_peers ONE_OR_MORE_PEERS --per_device_train_batch_size BATCH_SIZE_FOR_YOUR_GPU
   ```
 
   Here, `ONE_OR_MORE_PEERS` stands for multiaddresses of one or multiple existing peers (training monitors or existing