From c0a9c98ec41079aefed95c6af0d50dccd8757767 Mon Sep 17 00:00:00 2001 From: Ikko Eltociear Ashimine Date: Thu, 3 Aug 2023 02:59:45 +0900 Subject: [PATCH] [Docs] fix typo in migration.md splited -> splitted --- docs/en/migration.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/en/migration.md b/docs/en/migration.md index bdebdf6f5a9..bb7b22cb8ad 100644 --- a/docs/en/migration.md +++ b/docs/en/migration.md @@ -358,7 +358,7 @@ model = dict( Changes in **`data`**: -- The original `data` field is splited to `train_dataloader`, `val_dataloader` and +- The original `data` field is splitted to `train_dataloader`, `val_dataloader` and `test_dataloader`. This allows us to configure them in fine-grained. For example, you can specify different sampler and batch size during training and test. - The `samples_per_gpu` is renamed to `batch_size`. @@ -465,7 +465,7 @@ train_pipeline = [ Changes in **`evaluation`**: -- The **`evaluation`** field is splited to `val_evaluator` and `test_evaluator`. And it won't supports `interval` and `save_best` arguments. +- The **`evaluation`** field is splitted to `val_evaluator` and `test_evaluator`. And it won't supports `interval` and `save_best` arguments. The `interval` is moved to `train_cfg.val_interval`, see [the schedule settings](./user_guides/config.md#schedule-settings) and the `save_best` is moved to `default_hooks.checkpoint.save_best`, see [the runtime settings](./user_guides/config.md#runtime-settings). - The 'accuracy' metric is renamed to [`Accuracy`](mmpretrain.evaluation.Accuracy).