From cf8a7c2490783c079717a1c816c4657ec5b3f4c0 Mon Sep 17 00:00:00 2001 From: Mishig Davaadorj Date: Wed, 27 Apr 2022 16:46:55 +0200 Subject: [PATCH] Update custom_models.mdx (#16964) BertModelForSequenceClassification -> BertForSequenceClassification --- docs/source/en/custom_models.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/source/en/custom_models.mdx b/docs/source/en/custom_models.mdx index 20bf93c68b..05d0a6c2ac 100644 --- a/docs/source/en/custom_models.mdx +++ b/docs/source/en/custom_models.mdx @@ -106,7 +106,7 @@ directly upload your config to the Hub. Now that we have our ResNet configuration, we can go on writing the model. We will actually write two: one that extracts the hidden features from a batch of images (like [`BertModel`]) and one that is suitable for image -classification (like [`BertModelForSequenceClassification`]). +classification (like [`BertForSequenceClassification`]). As we mentioned before, we'll only write a loose wrapper of the model to keep it simple for this example. The only thing we need to do before writing this class is a map between the block types and actual block classes. Then the