fastNLP.core.sampler

sampler 子类实现了 fastNLP 所需的各种采样器。

class fastNLP.core.sampler.Sampler

别名 fastNLP.Sampler fastNLP.core.sampler.Sampler

Sampler 类的基类. 规定以何种顺序取出data中的元素

子类必须实现 __call__ 方法. 输入 DataSet 对象, 返回其中元素的下标序列

__init__

Initialize self. See help(type(self)) for accurate signature.

class fastNLP.core.sampler.BucketSampler(num_buckets=10, batch_size=None, seq_len_field_name='seq_len')

基类 fastNLP.Sampler

别名 fastNLP.BucketSampler fastNLP.core.sampler.BucketSampler

带Bucket的 Random Sampler. 可以随机地取出长度相似的元素

__init__(num_buckets=10, batch_size=None, seq_len_field_name='seq_len')

参数:

• num_buckets (int) -- bucket的数量
• batch_size (int) -- batch的大小. 默认为None，Trainer在调用BucketSampler时，会将该值正确设置，如果是非Trainer场景使用，需 要显示传递该值
• seq_len_field_name (str) -- 对应序列长度的 field 的名字

set_batch_size(batch_size)

参数:batch_size (int) -- 每个batch的大小 返回:

class fastNLP.core.sampler.SequentialSampler

基类 fastNLP.Sampler

别名 fastNLP.SequentialSampler fastNLP.core.sampler.SequentialSampler

顺序取出元素的 Sampler

__init__

Initialize self. See help(type(self)) for accurate signature.

class fastNLP.core.sampler.RandomSampler

基类 fastNLP.Sampler

别名 fastNLP.RandomSampler fastNLP.core.sampler.RandomSampler

随机化取元素的 Sampler

__init__

Initialize self. See help(type(self)) for accurate signature.