From 292005b6b54d502be5ace20350dca48325006d53 Mon Sep 17 00:00:00 2001 From: Zeyu Chen Date: Sun, 21 Apr 2019 21:16:37 +0800 Subject: [PATCH] Update README.md --- demo/text-classification/README.md | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) diff --git a/demo/text-classification/README.md b/demo/text-classification/README.md index 4987f103..c1dd36b5 100644 --- a/demo/text-classification/README.md +++ b/demo/text-classification/README.md @@ -85,19 +85,17 @@ ClassifyReader中的`data_generator`会自动按照模型对应词表对数据 ### Step3: 构建网络并创建分类迁移任务 ```python # NOTE: 必须使用fluid.program_guard接口传入Module返回的预训练模型program -with fluid.program_guard(program): - label = fluid.layers.data(name="label", shape=[1], dtype='int64') - pooled_output = outputs["pooled_output"] +pooled_output = outputs["pooled_output"] - # feed_list的Tensor顺序不可以调整 - feed_list = [ - inputs["input_ids"].name, inputs["position_ids"].name, - inputs["segment_ids"].name, inputs["input_mask"].name, label.name - ] +# feed_list的Tensor顺序不可以调整 +feed_list = [ + inputs["input_ids"].name, inputs["position_ids"].name, + inputs["segment_ids"].name, inputs["input_mask"].name, label.name +] - cls_task = hub.create_text_cls_task( - feature=pooled_output, label=label, num_classes=reader.get_num_labels()) +cls_task = hub.create_text_cls_task( + feature=pooled_output, num_classes=dataset.num_labels) ``` **NOTE:** 基于预训练模型的迁移学习网络搭建，必须在`with fluid.program_gurad()`作用域内组件网络 1. `outputs["pooled_output"]`返回了ERNIE/BERT模型对应的[CLS]向量,可以用于句子或句对的特征表达。 -- GitLab