hummingbird
/
fastNLP
Not watched
2
0
Fork 0
Code
Releases 13 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain
Browse Source

* rename DataSet.get_fields() into get_all_fields() 

* add DataSet.get_field(), to fetch a FieldArray based on its name
* remove old tutorials & add new tutorials
tags/v0.2.0^2
FengZiYjun 7 years ago
parent
447746d9f5
commit
720a264eb3
13 changed files with 2215 additions and 1004 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +2
    -2
      fastNLP/api/processor.py
  2. +1
    -1
      fastNLP/core/batch.py
  3. +6
    -1
      fastNLP/core/dataset.py
  4. +1
    -1
      fastNLP/core/sampler.py
  5. +3
    -3
      fastNLP/models/cnn_text_classification.py
  6. +16
    -0
      test/core/test_dataset.py
  7. +40
    -1
      test/data_for_tests/tutorial_sample_dataset.csv
  8. +911
    -0
      tutorials/fastnlp_10min_tutorial_v2.ipynb
  9. +860
    -0
      tutorials/fastnlp_10tmin_tutorial.ipynb
  10. +333
    -0
      tutorials/fastnlp_1_minute_tutorial.ipynb
  11. +42
    -22
      tutorials/fastnlp_advanced_tutorial.ipynb
  12. +0
    -526
      tutorials/fastnlp_tutorial_1203.ipynb
  13. +0
    -447
      tutorials/fastnlp_tutorial_1204.ipynb

+ 2
- 2
fastNLP/api/processor.py View File

@@ -268,7 +268,7 @@ class SetTensorProcessor(Processor):
self.default = default

def process(self, dataset):
set_dict = {name: self.default for name in dataset.get_fields().keys()}
set_dict = {name: self.default for name in dataset.get_all_fields().keys()}
set_dict.update(self.field_dict)
dataset._set_need_tensor(**set_dict)
return dataset
@@ -282,7 +282,7 @@ class SetIsTargetProcessor(Processor):
self.default = default

def process(self, dataset):
set_dict = {name: self.default for name in dataset.get_fields().keys()}
set_dict = {name: self.default for name in dataset.get_all_fields().keys()}
set_dict.update(self.field_dict)
dataset.set_target(**set_dict)
return dataset

+ 1
- 1
fastNLP/core/batch.py View File

@@ -43,7 +43,7 @@ class Batch(object):

indices = self.idx_list[self.curidx:endidx]

for field_name, field in self.dataset.get_fields().items():
for field_name, field in self.dataset.get_all_fields().items():
if field.is_target or field.is_input:
batch = field.get(indices)
if not self.as_numpy:


+ 6
- 1
fastNLP/core/dataset.py View File

@@ -157,7 +157,12 @@ class DataSet(object):
"""
self.field_arrays.pop(name)

def get_fields(self):
def get_field(self, field_name):
if field_name not in self.field_arrays:
raise KeyError("Field name {} not found in DataSet".format(field_name))
return self.field_arrays[field_name]

def get_all_fields(self):
"""Return all the fields with their names.

:return dict field_arrays: the internal data structure of DataSet.


+ 1
- 1
fastNLP/core/sampler.py View File

@@ -55,7 +55,7 @@ class BucketSampler(BaseSampler):

def __call__(self, data_set):

seq_lens = data_set.get_fields()[self.seq_lens_field_name].content
seq_lens = data_set.get_all_fields()[self.seq_lens_field_name].content
total_sample_num = len(seq_lens)

bucket_indexes = []


+ 3
- 3
fastNLP/models/cnn_text_classification.py View File

@@ -44,7 +44,7 @@ class CNNText(torch.nn.Module):
x = self.conv_pool(x) # [N,L,C] -> [N,C]
x = self.dropout(x)
x = self.fc(x) # [N,C] -> [N, N_class]
return {'output': x}
return {'pred': x}

def predict(self, word_seq):
"""
@@ -53,5 +53,5 @@ class CNNText(torch.nn.Module):
:return predict: dict of torch.LongTensor, [batch_size, seq_len]
"""
output = self(word_seq)
_, predict = output['output'].max(dim=1)
return {'predict': predict}
_, predict = output['pred'].max(dim=1)
return {'pred': predict}

+ 16
- 0
test/core/test_dataset.py View File

@@ -2,6 +2,7 @@ import os
import unittest

from fastNLP.core.dataset import DataSet
from fastNLP.core.fieldarray import FieldArray
from fastNLP.core.instance import Instance


@@ -162,6 +163,21 @@ class TestDataSet(unittest.TestCase):
ds_1 = DataSet.load("./my_ds.pkl")
os.remove("my_ds.pkl")

def test_get_all_fields(self):
ds = DataSet({"x": [[1, 2, 3, 4]] * 10, "y": [[5, 6]] * 10})
ans = ds.get_all_fields()
self.assertEqual(ans["x"].content, [[1, 2, 3, 4]] * 10)
self.assertEqual(ans["y"].content, [[5, 6]] * 10)

def test_get_field(self):
ds = DataSet({"x": [[1, 2, 3, 4]] * 10, "y": [[5, 6]] * 10})
ans = ds.get_field("x")
self.assertTrue(isinstance(ans, FieldArray))
self.assertEqual(ans.content, [[1, 2, 3, 4]] * 10)
ans = ds.get_field("y")
self.assertTrue(isinstance(ans, FieldArray))
self.assertEqual(ans.content, [[5, 6]] * 10)


class TestDataSetIter(unittest.TestCase):
def test__repr__(self):


+ 40
- 1
test/data_for_tests/tutorial_sample_dataset.csv View File

@@ -35,4 +35,43 @@ There 's very little sense to what 's going on here , but the makers serve up th
Cattaneo should have followed the runaway success of his first film , The Full Monty , with something different . 2
They 're the unnamed , easily substitutable forces that serve as whatever terror the heroes of horror movies try to avoid . 1
It almost feels as if the movie is more interested in entertaining itself than in amusing us . 1
The movie 's progression into rambling incoherence gives new meaning to the phrase ` fatal script error . ' 0
The movie 's progression into rambling incoherence gives new meaning to the phrase ` fatal script error . ' 0
I still like Moonlight Mile , better judgment be damned . 3
A welcome relief from baseball movies that try too hard to be mythic , this one is a sweet and modest and ultimately winning story . 3
a bilingual charmer , just like the woman who inspired it 3
Like a less dizzily gorgeous companion to Mr. Wong 's In the Mood for Love -- very much a Hong Kong movie despite its mainland setting . 2
As inept as big-screen remakes of The Avengers and The Wild Wild West . 1
It 's everything you 'd expect -- but nothing more . 2
Best indie of the year , so far . 4
Hatfield and Hicks make the oddest of couples , and in this sense the movie becomes a study of the gambles of the publishing world , offering a case study that exists apart from all the movie 's political ramifications . 3
It 's like going to a house party and watching the host defend himself against a frothing ex-girlfriend . 1
That the Chuck Norris `` grenade gag '' occurs about 7 times during Windtalkers is a good indication of how serious-minded the film is . 2
The plot is romantic comedy boilerplate from start to finish . 2
It arrives with an impeccable pedigree , mongrel pep , and almost indecipherable plot complications . 2
A film that clearly means to preach exclusively to the converted . 2
I still like Moonlight Mile , better judgment be damned . 3
A welcome relief from baseball movies that try too hard to be mythic , this one is a sweet and modest and ultimately winning story . 3
a bilingual charmer , just like the woman who inspired it 3
Like a less dizzily gorgeous companion to Mr. Wong 's In the Mood for Love -- very much a Hong Kong movie despite its mainland setting . 2
As inept as big-screen remakes of The Avengers and The Wild Wild West . 1
It 's everything you 'd expect -- but nothing more . 2
Best indie of the year , so far . 4
Hatfield and Hicks make the oddest of couples , and in this sense the movie becomes a study of the gambles of the publishing world , offering a case study that exists apart from all the movie 's political ramifications . 3
It 's like going to a house party and watching the host defend himself against a frothing ex-girlfriend . 1
That the Chuck Norris `` grenade gag '' occurs about 7 times during Windtalkers is a good indication of how serious-minded the film is . 2
The plot is romantic comedy boilerplate from start to finish . 2
It arrives with an impeccable pedigree , mongrel pep , and almost indecipherable plot complications . 2
A film that clearly means to preach exclusively to the converted . 2
I still like Moonlight Mile , better judgment be damned . 3
A welcome relief from baseball movies that try too hard to be mythic , this one is a sweet and modest and ultimately winning story . 3
a bilingual charmer , just like the woman who inspired it 3
Like a less dizzily gorgeous companion to Mr. Wong 's In the Mood for Love -- very much a Hong Kong movie despite its mainland setting . 2
As inept as big-screen remakes of The Avengers and The Wild Wild West . 1
It 's everything you 'd expect -- but nothing more . 2
Best indie of the year , so far . 4
Hatfield and Hicks make the oddest of couples , and in this sense the movie becomes a study of the gambles of the publishing world , offering a case study that exists apart from all the movie 's political ramifications . 3
It 's like going to a house party and watching the host defend himself against a frothing ex-girlfriend . 1
That the Chuck Norris `` grenade gag '' occurs about 7 times during Windtalkers is a good indication of how serious-minded the film is . 2
The plot is romantic comedy boilerplate from start to finish . 2
It arrives with an impeccable pedigree , mongrel pep , and almost indecipherable plot complications . 2
A film that clearly means to preach exclusively to the converted . 2

+ 911
- 0
tutorials/fastnlp_10min_tutorial_v2.ipynb View File

@@ -0,0 +1,911 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"fastNLP上手教程\n",
"-------\n",
"\n",
"fastNLP提供方便的数据预处理，训练和测试模型的功能"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"DataSet & Instance\n",
"------\n",
"\n",
"fastNLP用DataSet和Instance保存和处理数据。每个DataSet表示一个数据集，每个Instance表示一个数据样本。一个DataSet存有多个Instance，每个Instance可以自定义存哪些内容。\n",
"\n",
"有一些read_*方法，可以轻松从文件读取数据，存成DataSet。"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"8529"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"from fastNLP import DataSet\n",
"from fastNLP import Instance\n",
"\n",
"# 从csv读取数据到DataSet\n",
"dataset = DataSet.read_csv('../sentence.csv', headers=('raw_sentence', 'label'), sep='\\t')\n",
"print(len(dataset))"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': A series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story .,\n'label': 1}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 使用数字索引[k]，获取第k个样本\n",
"print(dataset[0])\n",
"\n",
"# 索引也可以是负数\n",
"print(dataset[-3])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Instance\n",
"Instance表示一个样本，由一个或多个field（域，属性，特征）组成，每个field有名字和值。\n",
"\n",
"在初始化Instance时即可定义它包含的域，使用 \"field_name=field_value\"的写法。"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'raw_sentence': fake data,\n'label': 0}"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# DataSet.append(Instance)加入新数据\n",
"dataset.append(Instance(raw_sentence='fake data', label='0'))\n",
"dataset[-1]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## DataSet.apply方法\n",
"数据预处理利器"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': a series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story .,\n'label': 1}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 将所有数字转为小写\n",
"dataset.apply(lambda x: x['raw_sentence'].lower(), new_field_name='raw_sentence')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': a series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story .,\n'label': 1}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# label转int\n",
"dataset.apply(lambda x: int(x['label']), new_field_name='label')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': a series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story .,\n'label': 1,\n'words': ['a', 'series', 'of', 'escapades', 'demonstrating', 'the', 'adage', 'that', 'what', 'is', 'good', 'for', 'the', 'goose', 'is', 'also', 'good', 'for', 'the', 'gander', ',', 'some', 'of', 'which', 'occasionally', 'amuses', 'but', 'none', 'of', 'which', 'amounts', 'to', 'much', 'of', 'a', 'story', '.']}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 使用空格分割句子\n",
"def split_sent(ins):\n",
" return ins['raw_sentence'].split()\n",
"dataset.apply(split_sent, new_field_name='words')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': a series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story .,\n'label': 1,\n'words': ['a', 'series', 'of', 'escapades', 'demonstrating', 'the', 'adage', 'that', 'what', 'is', 'good', 'for', 'the', 'goose', 'is', 'also', 'good', 'for', 'the', 'gander', ',', 'some', 'of', 'which', 'occasionally', 'amuses', 'but', 'none', 'of', 'which', 'amounts', 'to', 'much', 'of', 'a', 'story', '.'],\n'seq_len': 37}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 增加长度信息\n",
"dataset.apply(lambda x: len(x['words']), new_field_name='seq_len')\n",
"print(dataset[0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## DataSet.drop\n",
"筛选数据"
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"8358"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"dataset.drop(lambda x: x['seq_len'] <= 3)\n",
"print(len(dataset))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 配置DataSet\n",
"1. 哪些域是特征，哪些域是标签\n",
"2. 切分训练集/验证集"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [],
"source": [
"# 设置DataSet中，哪些field要转为tensor\n",
"\n",
"# set target，loss或evaluate中的golden，计算loss，模型评估时使用\n",
"dataset.set_target(\"label\")\n",
"# set input，模型forward时使用\n",
"dataset.set_input(\"words\")"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"5851"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"2507"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 分出测试集、训练集\n",
"\n",
"test_data, train_data = dataset.split(0.3)\n",
"print(len(test_data))\n",
"print(len(train_data))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Vocabulary\n",
"------\n",
"\n",
"fastNLP中的Vocabulary轻松构建词表，将词转成数字"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': the project 's filmmakers forgot to include anything even halfway scary as they poorly rejigger fatal attraction into a high school setting .,\n'label': 0,\n'words': [4, 423, 9, 316, 1, 8, 1, 312, 72, 1478, 885, 14, 86, 725, 1, 1913, 1431, 53, 5, 455, 736, 1, 2],\n'seq_len': 23}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"from fastNLP import Vocabulary\n",
"\n",
"# 构建词表, Vocabulary.add(word)\n",
"vocab = Vocabulary(min_freq=2)\n",
"train_data.apply(lambda x: [vocab.add(word) for word in x['words']])\n",
"vocab.build_vocab()\n",
"\n",
"# index句子, Vocabulary.to_index(word)\n",
"train_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='words')\n",
"test_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='words')\n",
"\n",
"\n",
"print(test_data[0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Model\n",
"定义一个PyTorch模型"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"CNNText(\n (embed): Embedding(\n (embed): Embedding(3459, 50, padding_idx=0)\n (dropout): Dropout(p=0.0)\n )\n (conv_pool): ConvMaxpool(\n (convs): ModuleList(\n (0): Conv1d(50, 3, kernel_size=(3,), stride=(1,), padding=(2,))\n (1): Conv1d(50, 4, kernel_size=(4,), stride=(1,), padding=(2,))\n (2): Conv1d(50, 5, kernel_size=(5,), stride=(1,), padding=(2,))\n )\n )\n (dropout): Dropout(p=0.1)\n (fc): Linear(\n (linear): Linear(in_features=12, out_features=5, bias=True)\n )\n)"
]
},
"execution_count": 20,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from fastNLP.models import CNNText\n",
"model = CNNText(embed_num=len(vocab), embed_dim=50, num_classes=5, padding=2, dropout=0.1)\n",
"model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"这是上述模型的forward方法。如果你不知道什么是forward方法，请参考我们的PyTorch教程。\n",
"\n",
"注意两点：\n",
"1. forward参数名字叫**word_seq**，请记住。\n",
"2. forward的返回值是一个**dict**，其中有个key的名字叫**output**。\n",
"\n",
"```Python\n",
" def forward(self, word_seq):\n",
" \"\"\"\n",
"\n",
" :param word_seq: torch.LongTensor, [batch_size, seq_len]\n",
" :return output: dict of torch.LongTensor, [batch_size, num_classes]\n",
" \"\"\"\n",
" x = self.embed(word_seq) # [N,L] -> [N,L,C]\n",
" x = self.conv_pool(x) # [N,L,C] -> [N,C]\n",
" x = self.dropout(x)\n",
" x = self.fc(x) # [N,C] -> [N, N_class]\n",
" return {'output': x}\n",
"```"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"这是上述模型的predict方法，是用来直接输出该任务的预测结果，与forward目的不同。\n",
"\n",
"注意两点：\n",
"1. predict参数名也叫**word_seq**。\n",
"2. predict的返回值是也一个**dict**，其中有个key的名字叫**predict**。\n",
"\n",
"```\n",
" def predict(self, word_seq):\n",
" \"\"\"\n",
"\n",
" :param word_seq: torch.LongTensor, [batch_size, seq_len]\n",
" :return predict: dict of torch.LongTensor, [batch_size, seq_len]\n",
" \"\"\"\n",
" output = self(word_seq)\n",
" _, predict = output['output'].max(dim=1)\n",
" return {'predict': predict}\n",
"```"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Trainer & Tester\n",
"------\n",
"\n",
"使用fastNLP的Trainer训练模型"
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import Trainer\n",
"from copy import deepcopy\n",
"from fastNLP.core.losses import CrossEntropyLoss\n",
"from fastNLP.core.metrics import AccuracyMetric\n",
"\n",
"\n",
"# 更改DataSet中对应field的名称，与模型的forward的参数名一致\n",
"# 因为forward的参数叫word_seq, 所以要把原本叫words的field改名为word_seq\n",
"# 这里的演示是让你了解这种**命名规则**\n",
"train_data.rename_field('words', 'word_seq')\n",
"test_data.rename_field('words', 'word_seq')\n",
"\n",
"# 顺便把label换名为label_seq\n",
"train_data.rename_field('label', 'label_seq')\n",
"test_data.rename_field('label', 'label_seq')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### loss\n",
"训练模型需要提供一个损失函数\n",
"\n",
"下面提供了一个在分类问题中常用的交叉熵损失。注意它的**初始化参数**。\n",
"\n",
"pred参数对应的是模型的forward返回的dict的一个key的名字，这里是\"output\"。\n",
"\n",
"target参数对应的是dataset作为标签的field的名字，这里是\"label_seq\"。"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [],
"source": [
"loss = CrossEntropyLoss(pred=\"output\", target=\"label_seq\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Metric\n",
"定义评价指标\n",
"\n",
"这里使用准确率。参数的“命名规则”跟上面类似。\n",
"\n",
"pred参数对应的是模型的predict方法返回的dict的一个key的名字，这里是\"predict\"。\n",
"\n",
"target参数对应的是dataset作为标签的field的名字，这里是\"label_seq\"。"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [],
"source": [
"metric = AccuracyMetric(pred=\"predict\", target=\"label_seq\")"
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training epochs started 2018-12-07 14:11:31"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"data": {
"text/plain": [
"HBox(children=(IntProgress(value=0, layout=Layout(flex='2'), max=915), HTML(value='')), layout=Layout(display=…"
]
},
"execution_count": 0,
"metadata": {},
"output_type": "execute_result"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/5. Step:183/915. AccuracyMetric: acc=0.350367"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 2/5. Step:366/915. AccuracyMetric: acc=0.409332"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 3/5. Step:549/915. AccuracyMetric: acc=0.572552"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 4/5. Step:732/915. AccuracyMetric: acc=0.711331"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 5/5. Step:915/915. AccuracyMetric: acc=0.801572"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
}
],
"source": [
"# 实例化Trainer，传入模型和数据，进行训练\n",
"# 先在test_data拟合\n",
"copy_model = deepcopy(model)\n",
"overfit_trainer = Trainer(model=copy_model, train_data=test_data, dev_data=test_data,\n",
" loss=loss,\n",
" metrics=metric,\n",
" save_path=None,\n",
" batch_size=32,\n",
" n_epochs=5)\n",
"overfit_trainer.train()"
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training epochs started 2018-12-07 14:12:21"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"data": {
"text/plain": [
"HBox(children=(IntProgress(value=0, layout=Layout(flex='2'), max=395), HTML(value='')), layout=Layout(display=…"
]
},
"execution_count": 0,
"metadata": {},
"output_type": "execute_result"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/5. Step:79/395. AccuracyMetric: acc=0.250043"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 2/5. Step:158/395. AccuracyMetric: acc=0.280807"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 3/5. Step:237/395. AccuracyMetric: acc=0.280978"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 4/5. Step:316/395. AccuracyMetric: acc=0.285592"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 5/5. Step:395/395. AccuracyMetric: acc=0.278927"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
}
],
"source": [
"# 用train_data训练，在test_data验证\n",
"trainer = Trainer(model=model, train_data=train_data, dev_data=test_data,\n",
" loss=CrossEntropyLoss(pred=\"output\", target=\"label_seq\"),\n",
" metrics=AccuracyMetric(pred=\"predict\", target=\"label_seq\"),\n",
" save_path=None,\n",
" batch_size=32,\n",
" n_epochs=5)\n",
"trainer.train()\n",
"print('Train finished!')"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[tester] \nAccuracyMetric: acc=0.280636"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'AccuracyMetric': {'acc': 0.280636}}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 调用Tester在test_data上评价效果\n",
"from fastNLP import Tester\n",
"\n",
"tester = Tester(data=test_data, model=model, metrics=AccuracyMetric(pred=\"predict\", target=\"label_seq\"),\n",
" batch_size=4)\n",
"acc = tester.test()\n",
"print(acc)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}

+ 860
- 0
tutorials/fastnlp_10tmin_tutorial.ipynb View File

@@ -0,0 +1,860 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"fastNLP上手教程\n",
"-------\n",
"\n",
"fastNLP提供方便的数据预处理，训练和测试模型的功能"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"DataSet & Instance\n",
"------\n",
"\n",
"fastNLP用DataSet和Instance保存和处理数据。每个DataSet表示一个数据集，每个Instance表示一个数据样本。一个DataSet存有多个Instance，每个Instance可以自定义存哪些内容。\n",
"\n",
"有一些read_*方法，可以轻松从文件读取数据，存成DataSet。"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': A series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story .,\n'label': 1}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"from fastNLP import DataSet\n",
"from fastNLP import Instance\n",
"\n",
"# 从csv读取数据到DataSet\n",
"win_path = \"C:\\\\Users\\zyfeng\\Desktop\\FudanNLP\\\\fastNLP\\\\test\\\\data_for_tests\\\\tutorial_sample_dataset.csv\"\n",
"dataset = DataSet.read_csv(win_path, headers=('raw_sentence', 'label'), sep='\\t')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'raw_sentence': fake data,\n'label': 0}"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# DataSet.append(Instance)加入新数据\n",
"\n",
"dataset.append(Instance(raw_sentence='fake data', label='0'))\n",
"dataset[-1]"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"# DataSet.apply(func, new_field_name)对数据预处理\n",
"\n",
"# 将所有数字转为小写\n",
"dataset.apply(lambda x: x['raw_sentence'].lower(), new_field_name='raw_sentence')\n",
"# label转int\n",
"dataset.apply(lambda x: int(x['label']), new_field_name='label_seq', is_target=True)\n",
"# 使用空格分割句子\n",
"dataset.drop(lambda x: len(x['raw_sentence'].split()) == 0)\n",
"def split_sent(ins):\n",
" return ins['raw_sentence'].split()\n",
"dataset.apply(split_sent, new_field_name='words', is_input=True)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"# DataSet.drop(func)筛除数据\n",
"# 删除低于某个长度的词语\n",
"dataset.drop(lambda x: len(x['words']) <= 3)"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Train size: "
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" "
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"54"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Test size: "
]
}
],
"source": [
"# 分出测试集、训练集\n",
"\n",
"test_data, train_data = dataset.split(0.3)\n",
"print(\"Train size: \", len(test_data))\n",
"print(\"Test size: \", len(train_data))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Vocabulary\n",
"------\n",
"\n",
"fastNLP中的Vocabulary轻松构建词表，将词转成数字"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': the plot is romantic comedy boilerplate from start to finish .,\n'label': 2,\n'label_seq': 2,\n'words': ['the', 'plot', 'is', 'romantic', 'comedy', 'boilerplate', 'from', 'start', 'to', 'finish', '.'],\n'word_seq': [2, 13, 9, 24, 25, 26, 15, 27, 11, 28, 3]}"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"from fastNLP import Vocabulary\n",
"\n",
"# 构建词表, Vocabulary.add(word)\n",
"vocab = Vocabulary(min_freq=2)\n",
"train_data.apply(lambda x: [vocab.add(word) for word in x['words']])\n",
"vocab.build_vocab()\n",
"\n",
"# index句子, Vocabulary.to_index(word)\n",
"train_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='word_seq', is_input=True)\n",
"test_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='word_seq', is_input=True)\n",
"\n",
"\n",
"print(test_data[0])"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"batch_x has: {'words': array([list(['this', 'kind', 'of', 'hands-on', 'storytelling', 'is', 'ultimately', 'what', 'makes', 'shanghai', 'ghetto', 'move', 'beyond', 'a', 'good', ',', 'dry', ',', 'reliable', 'textbook', 'and', 'what', 'allows', 'it', 'to', 'rank', 'with', 'its', 'worthy', 'predecessors', '.']),\n",
" list(['the', 'entire', 'movie', 'is', 'filled', 'with', 'deja', 'vu', 'moments', '.'])],\n",
" dtype=object), 'word_seq': tensor([[ 19, 184, 6, 1, 481, 9, 206, 50, 91, 1210, 1609, 1330,\n",
" 495, 5, 63, 4, 1269, 4, 1, 1184, 7, 50, 1050, 10,\n",
" 8, 1611, 16, 21, 1039, 1, 2],\n",
" [ 3, 711, 22, 9, 1282, 16, 2482, 2483, 200, 2, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0]])}\n",
"batch_y has: {'label_seq': tensor([3, 2])}\n"
]
}
],
"source": [
"# 假设你们需要做强化学习或者gan之类的项目，也许你们可以使用这里的dataset\n",
"from fastNLP.core.batch import Batch\n",
"from fastNLP.core.sampler import RandomSampler\n",
"\n",
"batch_iterator = Batch(dataset=train_data, batch_size=2, sampler=RandomSampler())\n",
"for batch_x, batch_y in batch_iterator:\n",
" print(\"batch_x has: \", batch_x)\n",
" print(\"batch_y has: \", batch_y)\n",
" break"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Model\n"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"CNNText(\n (embed): Embedding(\n (embed): Embedding(77, 50, padding_idx=0)\n (dropout): Dropout(p=0.0)\n )\n (conv_pool): ConvMaxpool(\n (convs): ModuleList(\n (0): Conv1d(50, 3, kernel_size=(3,), stride=(1,), padding=(2,))\n (1): Conv1d(50, 4, kernel_size=(4,), stride=(1,), padding=(2,))\n (2): Conv1d(50, 5, kernel_size=(5,), stride=(1,), padding=(2,))\n )\n )\n (dropout): Dropout(p=0.1)\n (fc): Linear(\n (linear): Linear(in_features=12, out_features=5, bias=True)\n )\n)"
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# 定义一个简单的Pytorch模型\n",
"\n",
"from fastNLP.models import CNNText\n",
"model = CNNText(embed_num=len(vocab), embed_dim=50, num_classes=5, padding=2, dropout=0.1)\n",
"model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Trainer & Tester\n",
"------\n",
"\n",
"使用fastNLP的Trainer训练模型"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import Trainer\n",
"from copy import deepcopy\n",
"from fastNLP import CrossEntropyLoss\n",
"from fastNLP import AccuracyMetric"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training epochs started 2018-12-07 14:07:20"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"data": {
"text/plain": [
"HBox(children=(IntProgress(value=0, layout=Layout(flex='2'), max=20), HTML(value='')), layout=Layout(display='…"
]
},
"execution_count": 0,
"metadata": {},
"output_type": "execute_result"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/10. Step:2/20. AccuracyMetric: acc=0.037037"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 2/10. Step:4/20. AccuracyMetric: acc=0.296296"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 3/10. Step:6/20. AccuracyMetric: acc=0.333333"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 4/10. Step:8/20. AccuracyMetric: acc=0.555556"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 5/10. Step:10/20. AccuracyMetric: acc=0.611111"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 6/10. Step:12/20. AccuracyMetric: acc=0.481481"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 7/10. Step:14/20. AccuracyMetric: acc=0.62963"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 8/10. Step:16/20. AccuracyMetric: acc=0.685185"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 9/10. Step:18/20. AccuracyMetric: acc=0.722222"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 10/10. Step:20/20. AccuracyMetric: acc=0.777778"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
}
],
"source": [
"# 进行overfitting测试\n",
"copy_model = deepcopy(model)\n",
"overfit_trainer = Trainer(model=copy_model, \n",
" train_data=test_data, \n",
" dev_data=test_data,\n",
" loss=CrossEntropyLoss(pred=\"output\", target=\"label_seq\"),\n",
" metrics=AccuracyMetric(),\n",
" n_epochs=10,\n",
" save_path=None)\n",
"overfit_trainer.train()"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training epochs started 2018-12-07 14:08:10"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"data": {
"text/plain": [
"HBox(children=(IntProgress(value=0, layout=Layout(flex='2'), max=5), HTML(value='')), layout=Layout(display='i…"
]
},
"execution_count": 0,
"metadata": {},
"output_type": "execute_result"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/5. Step:1/5. AccuracyMetric: acc=0.037037"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 2/5. Step:2/5. AccuracyMetric: acc=0.037037"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 3/5. Step:3/5. AccuracyMetric: acc=0.037037"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 4/5. Step:4/5. AccuracyMetric: acc=0.185185"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 5/5. Step:5/5. AccuracyMetric: acc=0.240741"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Train finished!"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 实例化Trainer，传入模型和数据，进行训练\n",
"trainer = Trainer(model=model, \n",
" train_data=train_data, \n",
" dev_data=test_data,\n",
" loss=CrossEntropyLoss(pred=\"output\", target=\"label_seq\"),\n",
" metrics=AccuracyMetric(),\n",
" n_epochs=5)\n",
"trainer.train()\n",
"print('Train finished!')"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[tester] \nAccuracyMetric: acc=0.240741"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"from fastNLP import Tester\n",
"\n",
"tester = Tester(data=test_data, model=model, metrics=AccuracyMetric())\n",
"acc = tester.test()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# In summary\n",
"\n",
"## fastNLP Trainer的伪代码逻辑\n",
"### 1. 准备DataSet，假设DataSet中共有如下的fields\n",
" ['raw_sentence', 'word_seq1', 'word_seq2', 'raw_label','label']\n",
" 通过\n",
" DataSet.set_input('word_seq1', word_seq2', flag=True)将'word_seq1', 'word_seq2'设置为input\n",
" 通过\n",
" DataSet.set_target('label', flag=True)将'label'设置为target\n",
"### 2. 初始化模型\n",
" class Model(nn.Module):\n",
" def __init__(self):\n",
" xxx\n",
" def forward(self, word_seq1, word_seq2):\n",
" # (1) 这里使用的形参名必须和DataSet中的input field的名称对应。因为我们是通过形参名, 进行赋值的\n",
" # (2) input field的数量可以多于这里的形参数量。但是不能少于。\n",
" xxxx\n",
" # 输出必须是一个dict\n",
"### 3. Trainer的训练过程\n",
" (1) 从DataSet中按照batch_size取出一个batch，调用Model.forward\n",
" (2) 将 Model.forward的结果 与 标记为target的field 传入Losser当中。\n",
" 由于每个人写的Model.forward的output的dict可能key并不一样，比如有人是{'pred':xxx}, {'output': xxx}; \n",
" 另外每个人将target可能也会设置为不同的名称, 比如有人是label, 有人设置为target；\n",
" 为了解决以上的问题，我们的loss提供映射机制\n",
" 比如CrossEntropyLosser的需要的输入是(prediction, target)。但是forward的output是{'output': xxx}; 'label'是target\n",
" 那么初始化losser的时候写为CrossEntropyLosser(prediction='output', target='label')即可\n",
" (3) 对于Metric是同理的\n",
" Metric计算也是从 forward的结果中取值 与 设置target的field中取值。 也是可以通过映射找到对应的值 \n",
" \n",
" \n",
"\n",
"## 一些问题.\n",
"### 1. DataSet中为什么需要设置input和target\n",
" 只有被设置为input或者target的数据才会在train的过程中被取出来\n",
" (1.1) 我们只会在设置为input的field中寻找传递给Model.forward的参数。\n",
" (1.2) 我们在传递值给losser或者metric的时候会使用来自: \n",
" (a)Model.forward的output\n",
" (b)被设置为target的field\n",
" \n",
"\n",
"### 2. 我们是通过forwad中的形参名将DataSet中的field赋值给对应的参数\n",
" (1.1) 构建模型过程中，\n",
" 例如:\n",
" DataSet中x，seq_lens是input，那么forward就应该是\n",
" def forward(self, x, seq_lens):\n",
" pass\n",
" 我们是通过形参名称进行匹配的field的\n",
" \n",
"\n",
"\n",
"### 1. 加载数据到DataSet\n",
"### 2. 使用apply操作对DataSet进行预处理\n",
" (2.1) 处理过程中将某些field设置为input，某些field设置为target\n",
"### 3. 构建模型\n",
" (3.1) 构建模型过程中，需要注意forward函数的形参名需要和DataSet中设置为input的field名称是一致的。\n",
" 例如:\n",
" DataSet中x，seq_lens是input，那么forward就应该是\n",
" def forward(self, x, seq_lens):\n",
" pass\n",
" 我们是通过形参名称进行匹配的field的\n",
" (3.2) 模型的forward的output需要是dict类型的。\n",
" 建议将输出设置为{\"pred\": xx}.\n",
" \n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}

+ 333
- 0
tutorials/fastnlp_1_minute_tutorial.ipynb View File

@@ -0,0 +1,333 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"collapsed": true
},
"source": [
"# FastNLP 1分钟上手教程"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## step 1\n",
"读取数据集"
]
},
{
"cell_type": "code",
"execution_count": 50,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import DataSet\n",
"# linux_path = \"../test/data_for_tests/tutorial_sample_dataset.csv\"\n",
"win_path = \"C:\\\\Users\\zyfeng\\Desktop\\FudanNLP\\\\fastNLP\\\\test\\\\data_for_tests\\\\tutorial_sample_dataset.csv\"\n",
"ds = DataSet.read_csv(win_path, headers=('raw_sentence', 'label'), sep='\\t')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## step 2\n",
"数据预处理\n",
"1. 类型转换\n",
"2. 切分验证集\n",
"3. 构建词典"
]
},
{
"cell_type": "code",
"execution_count": 52,
"metadata": {},
"outputs": [],
"source": [
"# 将所有数字转为小写\n",
"ds.apply(lambda x: x['raw_sentence'].lower(), new_field_name='raw_sentence')\n",
"# label转int\n",
"ds.apply(lambda x: int(x['label']), new_field_name='label_seq', is_target=True)\n",
"\n",
"def split_sent(ins):\n",
" return ins['raw_sentence'].split()\n",
"ds.apply(split_sent, new_field_name='words', is_input=True)\n"
]
},
{
"cell_type": "code",
"execution_count": 60,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Train size: "
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" "
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"54"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Test size: "
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" "
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"23"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# 分割训练集/验证集\n",
"train_data, dev_data = ds.split(0.3)\n",
"print(\"Train size: \", len(train_data))\n",
"print(\"Test size: \", len(dev_data))"
]
},
{
"cell_type": "code",
"execution_count": 61,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import Vocabulary\n",
"vocab = Vocabulary(min_freq=2)\n",
"train_data.apply(lambda x: [vocab.add(word) for word in x['words']])\n",
"\n",
"# index句子, Vocabulary.to_index(word)\n",
"train_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='word_seq', is_input=True)\n",
"dev_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='word_seq', is_input=True)\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## step 3\n",
" 定义模型"
]
},
{
"cell_type": "code",
"execution_count": 62,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP.models import CNNText\n",
"model = CNNText(embed_num=len(vocab), embed_dim=50, num_classes=5, padding=2, dropout=0.1)\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## step 4\n",
"开始训练"
]
},
{
"cell_type": "code",
"execution_count": 63,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training epochs started 2018-12-07 14:03:41"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"data": {
"text/plain": [
"HBox(children=(IntProgress(value=0, layout=Layout(flex='2'), max=6), HTML(value='')), layout=Layout(display='i…"
]
},
"execution_count": 0,
"metadata": {},
"output_type": "execute_result"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/3. Step:2/6. AccuracyMetric: acc=0.26087"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 2/3. Step:4/6. AccuracyMetric: acc=0.347826"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 3/3. Step:6/6. AccuracyMetric: acc=0.608696"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Train finished!"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"from fastNLP import Trainer, CrossEntropyLoss, AccuracyMetric\n",
"trainer = Trainer(model=model, \n",
" train_data=train_data, \n",
" dev_data=dev_data,\n",
" loss=CrossEntropyLoss(),\n",
" metrics=AccuracyMetric()\n",
" )\n",
"trainer.train()\n",
"print('Train finished!')\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 本教程结束。更多操作请参考进阶教程。"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 2",
"language": "python",
"name": "python2"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.6"
}
},
"nbformat": 4,
"nbformat_minor": 0
}

tutorials/fastnlp_in_six_lines.ipynb → tutorials/fastnlp_advanced_tutorial.ipynb View File

@@ -6,48 +6,68 @@
"collapsed": true
},
"source": [
"# 六行代码搞定FastNLP"
"## FastNLP 进阶教程\n",
"本教程阅读时间平均30分钟"
]
},
{
"cell_type": "code",
"execution_count": null,
"cell_type": "markdown",
"metadata": {},
"outputs": [],
"source": [
"from fastNLP.core.dataset import DataSet\n",
"import fastNLP.io.dataset_loader"
"## 数据部分\n",
"### DataSet\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"cell_type": "markdown",
"metadata": {},
"outputs": [],
"source": [
"ds = DataSet.read_naive(\"../test/data_for_tests/tutorial_sample_dataset.csv\")"
"### Instance"
]
},
{
"cell_type": "code",
"execution_count": null,
"cell_type": "markdown",
"metadata": {},
"outputs": [],
"source": []
"source": [
"### Vocabulary"
]
},
{
"cell_type": "code",
"execution_count": null,
"cell_type": "markdown",
"metadata": {},
"outputs": [],
"source": []
"source": [
"## 模型部分\n",
"### model"
]
},
{
"cell_type": "code",
"execution_count": null,
"cell_type": "markdown",
"metadata": {},
"outputs": [],
"source": []
"source": [
"## 训练测试部分\n",
"### Loss"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Metric"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Trainer"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Tester"
]
},
{
"cell_type": "code",

+ 0
- 526
tutorials/fastnlp_tutorial_1203.ipynb View File

@@ -1,526 +0,0 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"fastNLP上手教程\n",
"-------\n",
"\n",
"fastNLP提供方便的数据预处理，训练和测试模型的功能"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/Users/yh/miniconda2/envs/python3/lib/python3.6/site-packages/tqdm/autonotebook/__init__.py:14: TqdmExperimentalWarning: Using `tqdm.autonotebook.tqdm` in notebook mode. Use `tqdm.tqdm` instead to force console mode (e.g. in jupyter console)\n",
" \" (e.g. in jupyter console)\", TqdmExperimentalWarning)\n"
]
}
],
"source": [
"import sys\n",
"sys.path.append('/Users/yh/Desktop/fastNLP/fastNLP/')\n",
"\n",
"import fastNLP as fnlp"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"DataSet & Instance\n",
"------\n",
"\n",
"fastNLP用DataSet和Instance保存和处理数据。每个DataSet表示一个数据集，每个Instance表示一个数据样本。一个DataSet存有多个Instance，每个Instance可以自定义存哪些内容。\n",
"\n",
"有一些read_*方法，可以轻松从文件读取数据，存成DataSet。"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': A series of escapades demonstrating the adage that what is good for the goose is also good for the gander , some of which occasionally amuses but none of which amounts to much of a story .,\n",
"'label': 1}\n"
]
}
],
"source": [
"from fastNLP import DataSet\n",
"from fastNLP import Instance\n",
"\n",
"# 从csv读取数据到DataSet\n",
"dataset = DataSet.read_csv('sentence.csv', headers=('raw_sentence', 'label'), sep='\\t')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'raw_sentence': fake data,\n",
"'label': 0}"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# DataSet.append(Instance)加入新数据\n",
"\n",
"dataset.append(Instance(raw_sentence='fake data', label='0'))\n",
"dataset[-1]"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"# DataSet.apply(func, new_field_name)对数据预处理\n",
"\n",
"# 将所有数字转为小写\n",
"dataset.apply(lambda x: x['raw_sentence'].lower(), new_field_name='raw_sentence')\n",
"# label转int\n",
"dataset.apply(lambda x: int(x['label']), new_field_name='label_seq', is_target=True)\n",
"# 使用空格分割句子\n",
"dataset.drop(lambda x:len(x['raw_sentence'].split())==0)\n",
"def split_sent(ins):\n",
" return ins['raw_sentence'].split()\n",
"dataset.apply(split_sent, new_field_name='words', is_input=True)"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"# DataSet.drop(func)筛除数据\n",
"# 删除低于某个长度的词语\n",
"# dataset.drop(lambda x: len(x['words']) <= 3)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Train size: 5971\n",
"Test size: 2558\n"
]
}
],
"source": [
"# 分出测试集、训练集\n",
"\n",
"test_data, train_data = dataset.split(0.3)\n",
"print(\"Train size: \", len(test_data))\n",
"print(\"Test size: \", len(train_data))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Vocabulary\n",
"------\n",
"\n",
"fastNLP中的Vocabulary轻松构建词表，将词转成数字"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'raw_sentence': gussied up with so many distracting special effects and visual party tricks that it 's not clear whether we 're supposed to shriek or laugh .,\n",
"'label': 1,\n",
"'label_seq': 1,\n",
"'words': ['gussied', 'up', 'with', 'so', 'many', 'distracting', 'special', 'effects', 'and', 'visual', 'party', 'tricks', 'that', 'it', \"'s\", 'not', 'clear', 'whether', 'we', \"'re\", 'supposed', 'to', 'shriek', 'or', 'laugh', '.'],\n",
"'word_seq': [1, 65, 16, 43, 108, 1, 329, 433, 7, 319, 1313, 1, 12, 10, 11, 27, 1428, 567, 86, 134, 1949, 8, 1, 49, 506, 2]}\n"
]
}
],
"source": [
"from fastNLP import Vocabulary\n",
"\n",
"# 构建词表, Vocabulary.add(word)\n",
"vocab = Vocabulary(min_freq=2)\n",
"train_data.apply(lambda x: [vocab.add(word) for word in x['words']])\n",
"vocab.build_vocab()\n",
"\n",
"# index句子, Vocabulary.to_index(word)\n",
"train_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='word_seq', is_input=True)\n",
"test_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='word_seq', is_input=True)\n",
"\n",
"\n",
"print(test_data[0])"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"batch_x has: {'words': array([list(['this', 'kind', 'of', 'hands-on', 'storytelling', 'is', 'ultimately', 'what', 'makes', 'shanghai', 'ghetto', 'move', 'beyond', 'a', 'good', ',', 'dry', ',', 'reliable', 'textbook', 'and', 'what', 'allows', 'it', 'to', 'rank', 'with', 'its', 'worthy', 'predecessors', '.']),\n",
" list(['the', 'entire', 'movie', 'is', 'filled', 'with', 'deja', 'vu', 'moments', '.'])],\n",
" dtype=object), 'word_seq': tensor([[ 19, 184, 6, 1, 481, 9, 206, 50, 91, 1210, 1609, 1330,\n",
" 495, 5, 63, 4, 1269, 4, 1, 1184, 7, 50, 1050, 10,\n",
" 8, 1611, 16, 21, 1039, 1, 2],\n",
" [ 3, 711, 22, 9, 1282, 16, 2482, 2483, 200, 2, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0]])}\n",
"batch_y has: {'label_seq': tensor([3, 2])}\n"
]
}
],
"source": [
"# 假设你们需要做强化学习或者gan之类的项目，也许你们可以使用这里的dataset\n",
"from fastNLP.core.batch import Batch\n",
"from fastNLP.core.sampler import RandomSampler\n",
"\n",
"batch_iterator = Batch(dataset=train_data, batch_size=2, sampler=RandomSampler())\n",
"for batch_x, batch_y in batch_iterator:\n",
" print(\"batch_x has: \", batch_x)\n",
" print(\"batch_y has: \", batch_y)\n",
" break"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Model\n"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"CNNText(\n",
" (embed): Embedding(\n",
" (embed): Embedding(3470, 50, padding_idx=0)\n",
" (dropout): Dropout(p=0.0)\n",
" )\n",
" (conv_pool): ConvMaxpool(\n",
" (convs): ModuleList(\n",
" (0): Conv1d(50, 3, kernel_size=(3,), stride=(1,), padding=(2,))\n",
" (1): Conv1d(50, 4, kernel_size=(4,), stride=(1,), padding=(2,))\n",
" (2): Conv1d(50, 5, kernel_size=(5,), stride=(1,), padding=(2,))\n",
" )\n",
" )\n",
" (dropout): Dropout(p=0.1)\n",
" (fc): Linear(\n",
" (linear): Linear(in_features=12, out_features=5, bias=True)\n",
" )\n",
")"
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# 定义一个简单的Pytorch模型\n",
"\n",
"from fastNLP.models import CNNText\n",
"model = CNNText(embed_num=len(vocab), embed_dim=50, num_classes=5, padding=2, dropout=0.1)\n",
"model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Trainer & Tester\n",
"------\n",
"\n",
"使用fastNLP的Trainer训练模型"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import Trainer\n",
"from copy import deepcopy\n",
"from fastNLP.core.losses import CrossEntropyLoss\n",
"from fastNLP.core.metrics import AccuracyMetric"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training epochs started 2018-12-05 15:37:15\n"
]
},
{
"data": {
"text/plain": [
"HBox(children=(IntProgress(value=0, layout=Layout(flex='2'), max=1870), HTML(value='')), layout=Layout(display…"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/10. Step:187/1870. AccuracyMetric: acc=0.351365\n",
"Epoch 2/10. Step:374/1870. AccuracyMetric: acc=0.470943\n",
"Epoch 3/10. Step:561/1870. AccuracyMetric: acc=0.600402\n",
"Epoch 4/10. Step:748/1870. AccuracyMetric: acc=0.702227\n",
"Epoch 5/10. Step:935/1870. AccuracyMetric: acc=0.79099\n",
"Epoch 6/10. Step:1122/1870. AccuracyMetric: acc=0.846424\n",
"Epoch 7/10. Step:1309/1870. AccuracyMetric: acc=0.874058\n",
"Epoch 8/10. Step:1496/1870. AccuracyMetric: acc=0.898844\n",
"Epoch 9/10. Step:1683/1870. AccuracyMetric: acc=0.910568\n",
"Epoch 10/10. Step:1870/1870. AccuracyMetric: acc=0.921286\n",
"\r"
]
}
],
"source": [
"# 进行overfitting测试\n",
"copy_model = deepcopy(model)\n",
"overfit_trainer = Trainer(model=copy_model, \n",
" train_data=test_data, \n",
" dev_data=test_data,\n",
" losser=CrossEntropyLoss(pred=\"output\", target=\"label_seq\"),\n",
" metrics=AccuracyMetric(),\n",
" n_epochs=10,\n",
" save_path=None)\n",
"overfit_trainer.train()"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"training epochs started 2018-12-05 15:37:41\n"
]
},
{
"data": {
"text/plain": [
"HBox(children=(IntProgress(value=0, layout=Layout(flex='2'), max=400), HTML(value='')), layout=Layout(display=…"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\r"
]
},
{
"ename": "AttributeError",
"evalue": "'NoneType' object has no attribute 'squeeze'",
"traceback": [
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
"\u001b[0;31mAttributeError\u001b[0m Traceback (most recent call last)",
"\u001b[0;32m<ipython-input-12-5603b8b11a82>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m 7\u001b[0m \u001b[0mn_epochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 8\u001b[0m save_path='save/')\n\u001b[0;32m----> 9\u001b[0;31m \u001b[0mtrainer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 10\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Train finished!'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/Desktop/fastNLP/fastNLP/fastNLP/core/trainer.py\u001b[0m in \u001b[0;36mtrain\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m 163\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_summary_writer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mSummaryWriter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 164\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_tqdm\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 165\u001b[0;31m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_tqdm_train\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 166\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 167\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_print_train\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/Desktop/fastNLP/fastNLP/fastNLP/core/trainer.py\u001b[0m in \u001b[0;36m_tqdm_train\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m 206\u001b[0m \u001b[0mpbar\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwrite\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0meval_str\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 207\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalidate_every\u001b[0m \u001b[0;34m<\u001b[0m \u001b[0;36m0\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdev_data\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 208\u001b[0;31m \u001b[0meval_res\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_do_validation\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 209\u001b[0m \u001b[0meval_str\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"Epoch {}/{}. Step:{}/{}. \"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mepoch\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mn_epochs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtotal_steps\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m\\\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 210\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtester\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_format_eval_results\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0meval_res\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/Desktop/fastNLP/fastNLP/fastNLP/core/trainer.py\u001b[0m in \u001b[0;36m_do_validation\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m 265\u001b[0m \u001b[0mres\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtester\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtest\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 266\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mres\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 267\u001b[0;31m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_summary_writer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_scalar\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"valid_{}\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mglobal_step\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 268\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave_path\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_better_eval_result\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mres\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 269\u001b[0m \u001b[0mmetric_key\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetric_key\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetric_key\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;34m\"None\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/miniconda2/envs/python3/lib/python3.6/site-packages/tensorboardX/writer.py\u001b[0m in \u001b[0;36madd_scalar\u001b[0;34m(self, tag, scalar_value, global_step, walltime)\u001b[0m\n\u001b[1;32m 332\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_check_caffe2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mscalar_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 333\u001b[0m \u001b[0mscalar_value\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mworkspace\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mFetchBlob\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mscalar_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 334\u001b[0;31m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfile_writer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_summary\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mscalar\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtag\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscalar_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mglobal_step\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwalltime\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 335\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 336\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0madd_scalars\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmain_tag\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtag_scalar_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mglobal_step\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwalltime\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;32m~/miniconda2/envs/python3/lib/python3.6/site-packages/tensorboardX/summary.py\u001b[0m in \u001b[0;36mscalar\u001b[0;34m(name, scalar, collections)\u001b[0m\n\u001b[1;32m 115\u001b[0m \u001b[0mname\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_clean_tag\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 116\u001b[0m \u001b[0mscalar\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmake_np\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mscalar\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 117\u001b[0;31m \u001b[0;32massert\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mscalar\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msqueeze\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mndim\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'scalar should be 0D'\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 118\u001b[0m \u001b[0mscalar\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfloat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mscalar\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 119\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mSummary\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mSummary\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mValue\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtag\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msimple_value\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mscalar\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
"\u001b[0;31mAttributeError\u001b[0m: 'NoneType' object has no attribute 'squeeze'"
],
"output_type": "error"
}
],
"source": [
"# 实例化Trainer，传入模型和数据，进行训练\n",
"trainer = Trainer(model=model, \n",
" train_data=train_data, \n",
" dev_data=test_data,\n",
" losser=CrossEntropyLoss(pred=\"output\", target=\"label_seq\"),\n",
" metrics=AccuracyMetric(),\n",
" n_epochs=5,\n",
" save_path='save/')\n",
"trainer.train()\n",
"print('Train finished!')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import Tester\n",
"\n",
"tester = Tester(data=test_data, model=model, metrics=AccuracyMetric())\n",
"acc = tester.test()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# In summary\n",
"\n",
"## fastNLP Trainer的伪代码逻辑\n",
"### 1. 准备DataSet，假设DataSet中共有如下的fields\n",
" ['raw_sentence', 'word_seq1', 'word_seq2', 'raw_label','label']\n",
" 通过\n",
" DataSet.set_input('word_seq1', word_seq2', flag=True)将'word_seq1', 'word_seq2'设置为input\n",
" 通过\n",
" DataSet.set_target('label', flag=True)将'label'设置为target\n",
"### 2. 初始化模型\n",
" class Model(nn.Module):\n",
" def __init__(self):\n",
" xxx\n",
" def forward(self, word_seq1, word_seq2):\n",
" # (1) 这里使用的形参名必须和DataSet中的input field的名称对应。因为我们是通过形参名, 进行赋值的\n",
" # (2) input field的数量可以多于这里的形参数量。但是不能少于。\n",
" xxxx\n",
" # 输出必须是一个dict\n",
"### 3. Trainer的训练过程\n",
" (1) 从DataSet中按照batch_size取出一个batch，调用Model.forward\n",
" (2) 将 Model.forward的结果 与 标记为target的field 传入Losser当中。\n",
" 由于每个人写的Model.forward的output的dict可能key并不一样，比如有人是{'pred':xxx}, {'output': xxx}; \n",
" 另外每个人将target可能也会设置为不同的名称, 比如有人是label, 有人设置为target；\n",
" 为了解决以上的问题，我们的loss提供映射机制\n",
" 比如CrossEntropyLosser的需要的输入是(prediction, target)。但是forward的output是{'output': xxx}; 'label'是target\n",
" 那么初始化losser的时候写为CrossEntropyLosser(prediction='output', target='label')即可\n",
" (3) 对于Metric是同理的\n",
" Metric计算也是从 forward的结果中取值 与 设置target的field中取值。 也是可以通过映射找到对应的值 \n",
" \n",
" \n",
"\n",
"## 一些问题.\n",
"### 1. DataSet中为什么需要设置input和target\n",
" 只有被设置为input或者target的数据才会在train的过程中被取出来\n",
" (1.1) 我们只会在设置为input的field中寻找传递给Model.forward的参数。\n",
" (1.2) 我们在传递值给losser或者metric的时候会使用来自: \n",
" (a)Model.forward的output\n",
" (b)被设置为target的field\n",
" \n",
"\n",
"### 2. 我们是通过forwad中的形参名将DataSet中的field赋值给对应的参数\n",
" (1.1) 构建模型过程中，\n",
" 例如:\n",
" DataSet中x，seq_lens是input，那么forward就应该是\n",
" def forward(self, x, seq_lens):\n",
" pass\n",
" 我们是通过形参名称进行匹配的field的\n",
" \n",
"\n",
"\n",
"### 1. 加载数据到DataSet\n",
"### 2. 使用apply操作对DataSet进行预处理\n",
" (2.1) 处理过程中将某些field设置为input，某些field设置为target\n",
"### 3. 构建模型\n",
" (3.1) 构建模型过程中，需要注意forward函数的形参名需要和DataSet中设置为input的field名称是一致的。\n",
" 例如:\n",
" DataSet中x，seq_lens是input，那么forward就应该是\n",
" def forward(self, x, seq_lens):\n",
" pass\n",
" 我们是通过形参名称进行匹配的field的\n",
" (3.2) 模型的forward的output需要是dict类型的。\n",
" 建议将输出设置为{\"pred\": xx}.\n",
" \n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}

+ 0
- 447
tutorials/fastnlp_tutorial_1204.ipynb View File

@@ -1,447 +0,0 @@
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"fastNLP上手教程\n",
"-------\n",
"\n",
"fastNLP提供方便的数据预处理，训练和测试模型的功能"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import sys\n",
"sys.path.append('/Users/yh/Desktop/fastNLP/fastNLP/')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"DataSet & Instance\n",
"------\n",
"\n",
"fastNLP用DataSet和Instance保存和处理数据。每个DataSet表示一个数据集，每个Instance表示一个数据样本。一个DataSet存有多个Instance，每个Instance可以自定义存哪些内容。\n",
"\n",
"有一些read_*方法，可以轻松从文件读取数据，存成DataSet。"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import DataSet\n",
"from fastNLP import Instance\n",
"\n",
"# 从csv读取数据到DataSet\n",
"dataset = DataSet.read_csv('../sentence.csv', headers=('raw_sentence', 'label'), sep='\\t')\n",
"print(len(dataset))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 使用数字索引[k]，获取第k个样本\n",
"print(dataset[0])\n",
"\n",
"# 索引也可以是负数\n",
"print(dataset[-3])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Instance\n",
"Instance表示一个样本，由一个或多个field（域，属性，特征）组成，每个field有名字和值。\n",
"\n",
"在初始化Instance时即可定义它包含的域，使用 \"field_name=field_value\"的写法。"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# DataSet.append(Instance)加入新数据\n",
"dataset.append(Instance(raw_sentence='fake data', label='0'))\n",
"dataset[-1]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## DataSet.apply方法\n",
"数据预处理利器"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 将所有数字转为小写\n",
"dataset.apply(lambda x: x['raw_sentence'].lower(), new_field_name='raw_sentence')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# label转int\n",
"dataset.apply(lambda x: int(x['label']), new_field_name='label')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 使用空格分割句子\n",
"def split_sent(ins):\n",
" return ins['raw_sentence'].split()\n",
"dataset.apply(split_sent, new_field_name='words')\n",
"print(dataset[0])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 增加长度信息\n",
"dataset.apply(lambda x: len(x['words']), new_field_name='seq_len')\n",
"print(dataset[0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## DataSet.drop\n",
"筛选数据"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dataset.drop(lambda x: x['seq_len'] <= 3)\n",
"print(len(dataset))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 配置DataSet\n",
"1. 哪些域是特征，哪些域是标签\n",
"2. 切分训练集/验证集"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 设置DataSet中，哪些field要转为tensor\n",
"\n",
"# set target，loss或evaluate中的golden，计算loss，模型评估时使用\n",
"dataset.set_target(\"label\")\n",
"# set input，模型forward时使用\n",
"dataset.set_input(\"words\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 分出测试集、训练集\n",
"\n",
"test_data, train_data = dataset.split(0.3)\n",
"print(len(test_data))\n",
"print(len(train_data))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Vocabulary\n",
"------\n",
"\n",
"fastNLP中的Vocabulary轻松构建词表，将词转成数字"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import Vocabulary\n",
"\n",
"# 构建词表, Vocabulary.add(word)\n",
"vocab = Vocabulary(min_freq=2)\n",
"train_data.apply(lambda x: [vocab.add(word) for word in x['words']])\n",
"vocab.build_vocab()\n",
"\n",
"# index句子, Vocabulary.to_index(word)\n",
"train_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='words')\n",
"test_data.apply(lambda x: [vocab.to_index(word) for word in x['words']], new_field_name='words')\n",
"\n",
"\n",
"print(test_data[0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Model\n",
"定义一个PyTorch模型"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP.models import CNNText\n",
"model = CNNText(embed_num=len(vocab), embed_dim=50, num_classes=5, padding=2, dropout=0.1)\n",
"model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"这是上述模型的forward方法。如果你不知道什么是forward方法，请参考我们的PyTorch教程。\n",
"\n",
"注意两点：\n",
"1. forward参数名字叫**word_seq**，请记住。\n",
"2. forward的返回值是一个**dict**，其中有个key的名字叫**output**。\n",
"\n",
"```Python\n",
" def forward(self, word_seq):\n",
" \"\"\"\n",
"\n",
" :param word_seq: torch.LongTensor, [batch_size, seq_len]\n",
" :return output: dict of torch.LongTensor, [batch_size, num_classes]\n",
" \"\"\"\n",
" x = self.embed(word_seq) # [N,L] -> [N,L,C]\n",
" x = self.conv_pool(x) # [N,L,C] -> [N,C]\n",
" x = self.dropout(x)\n",
" x = self.fc(x) # [N,C] -> [N, N_class]\n",
" return {'output': x}\n",
"```"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"这是上述模型的predict方法，是用来直接输出该任务的预测结果，与forward目的不同。\n",
"\n",
"注意两点：\n",
"1. predict参数名也叫**word_seq**。\n",
"2. predict的返回值是也一个**dict**，其中有个key的名字叫**predict**。\n",
"\n",
"```\n",
" def predict(self, word_seq):\n",
" \"\"\"\n",
"\n",
" :param word_seq: torch.LongTensor, [batch_size, seq_len]\n",
" :return predict: dict of torch.LongTensor, [batch_size, seq_len]\n",
" \"\"\"\n",
" output = self(word_seq)\n",
" _, predict = output['output'].max(dim=1)\n",
" return {'predict': predict}\n",
"```"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Trainer & Tester\n",
"------\n",
"\n",
"使用fastNLP的Trainer训练模型"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from fastNLP import Trainer\n",
"from copy import deepcopy\n",
"from fastNLP.core.losses import CrossEntropyLoss\n",
"from fastNLP.core.metrics import AccuracyMetric\n",
"\n",
"\n",
"# 更改DataSet中对应field的名称，与模型的forward的参数名一致\n",
"# 因为forward的参数叫word_seq, 所以要把原本叫words的field改名为word_seq\n",
"# 这里的演示是让你了解这种**命名规则**\n",
"train_data.rename_field('words', 'word_seq')\n",
"test_data.rename_field('words', 'word_seq')\n",
"\n",
"# 顺便把label换名为label_seq\n",
"train_data.rename_field('label', 'label_seq')\n",
"test_data.rename_field('label', 'label_seq')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### loss\n",
"训练模型需要提供一个损失函数\n",
"\n",
"下面提供了一个在分类问题中常用的交叉熵损失。注意它的**初始化参数**。\n",
"\n",
"pred参数对应的是模型的forward返回的dict的一个key的名字，这里是\"output\"。\n",
"\n",
"target参数对应的是dataset作为标签的field的名字，这里是\"label_seq\"。"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"loss = CrossEntropyLoss(pred=\"output\", target=\"label_seq\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Metric\n",
"定义评价指标\n",
"\n",
"这里使用准确率。参数的“命名规则”跟上面类似。\n",
"\n",
"pred参数对应的是模型的predict方法返回的dict的一个key的名字，这里是\"predict\"。\n",
"\n",
"target参数对应的是dataset作为标签的field的名字，这里是\"label_seq\"。"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"metric = AccuracyMetric(pred=\"predict\", target=\"label_seq\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 实例化Trainer，传入模型和数据，进行训练\n",
"# 先在test_data拟合\n",
"copy_model = deepcopy(model)\n",
"overfit_trainer = Trainer(model=copy_model, train_data=test_data, dev_data=test_data,\n",
" losser=loss,\n",
" metrics=metric,\n",
" save_path=None,\n",
" batch_size=32,\n",
" n_epochs=5)\n",
"overfit_trainer.train()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 用train_data训练，在test_data验证\n",
"trainer = Trainer(model=model, train_data=train_data, dev_data=test_data,\n",
" losser=CrossEntropyLoss(pred=\"output\", target=\"label_seq\"),\n",
" metrics=AccuracyMetric(pred=\"predict\", target=\"label_seq\"),\n",
" save_path=None,\n",
" batch_size=32,\n",
" n_epochs=5)\n",
"trainer.train()\n",
"print('Train finished!')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# 调用Tester在test_data上评价效果\n",
"from fastNLP import Tester\n",
"\n",
"tester = Tester(data=test_data, model=model, metrics=AccuracyMetric(pred=\"predict\", target=\"label_seq\"),\n",
" batch_size=4)\n",
"acc = tester.test()\n",
"print(acc)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}

Write Preview
Loading…
Cancel
Save