Usage Sample
''''''''''''
.. code:: python
import torch
from sklearn.model_selection import train_test_split
from nlpx.text_token import Tokenizer
from nlpx.model.classifier import TextCNNClassifier
from nlpx.model.wrapper import ClassModelWrapper
from nlpx.dataset import TokenDataset, PaddingTokenCollator
if __name__ == '__main__':
classes = ['class1', 'class2', 'class3'...]
texts = [[str],]
labels = [0, 0, 1, 2, 1...]
tokenizer = Tokenizer.from_texts(texts, min_freq=5)
sent = 'I love you'
tokens = tokenizer.encode(sent, max_length=6)
# [101, 66, 88, 99, 102, 0]
sent = tokenizer.decode(tokens)
# ['<BOS>', 'I', 'love', 'you', '<EOS>', '<PAD>']
tokens = tokenizer.batch_encode(texts, padding=False)
X_train, X_test, y_train, y_test = train_test_split(tokens, labels, test_size=0.2)
train_set = TokenDataset(X_train, y_train)
val_set = TokenDataset(X_test, y_test)
model = TextCNNClassifier(embed_dim=128, vocab_size=tokenizer.vocab_size, num_classes=len(classes))
model_wrapper = ClassModelWrapper(model, classes=classes)
model_wrapper.train(train_set, val_set, show_progress=True, collate_fn=PaddingTokenCollator(tokenizer.pad))
result = model_wrapper.evaluate(val_set, collate_fn=PaddingTokenCollator(tokenizer.pad))
# 0.953125
test_inputs = torch.tensor(test_tokens, dtype=torch.long)
result = model_wrapper.predict(test_inputs)
# [0, 1]
result = model_wrapper.predict_classes(test_inputs)
# ['class1', 'class2']
result = model_wrapper.predict_proba(test_inputs)
# ([0, 1], array([0.99439645, 0.99190724], dtype=float32))
result = model_wrapper.predict_classes_proba(test_inputs)
# (['class1', 'class2'], array([0.99439645, 0.99190724], dtype=float32))
Raw data
{
"_id": null,
"home_page": "https://gitee.com/summry/nlpx",
"name": "NLPX",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.6",
"maintainer_email": null,
"keywords": "NLP, nlp, AI, llm, GPT, Machine learning, Deep learning, tokenize, torch",
"author": "summy",
"author_email": "xiazhongbiao@126.com",
"download_url": "https://files.pythonhosted.org/packages/30/72/eef6a8441b478e1b30d6beeaf9ac6863a0da01518838f438b54e45bbbb80/NLPX-1.8.1.tar.gz",
"platform": null,
"description": "Usage Sample\n''''''''''''\n\n.. code:: python\n\n import torch\n from sklearn.model_selection import train_test_split\n from nlpx.text_token import Tokenizer\n from nlpx.model.classifier import TextCNNClassifier\n from nlpx.model.wrapper import ClassModelWrapper\n from nlpx.dataset import TokenDataset, PaddingTokenCollator\n\n if __name__ == '__main__':\n classes = ['class1', 'class2', 'class3'...]\n texts = [[str],]\n labels = [0, 0, 1, 2, 1...]\n tokenizer = Tokenizer.from_texts(texts, min_freq=5)\n sent = 'I love you'\n tokens = tokenizer.encode(sent, max_length=6)\n # [101, 66, 88, 99, 102, 0]\n sent = tokenizer.decode(tokens)\n # ['<BOS>', 'I', 'love', 'you', '<EOS>', '<PAD>']\n\n tokens = tokenizer.batch_encode(texts, padding=False)\n X_train, X_test, y_train, y_test = train_test_split(tokens, labels, test_size=0.2)\n train_set = TokenDataset(X_train, y_train)\n val_set = TokenDataset(X_test, y_test)\n\n model = TextCNNClassifier(embed_dim=128, vocab_size=tokenizer.vocab_size, num_classes=len(classes))\n model_wrapper = ClassModelWrapper(model, classes=classes)\n model_wrapper.train(train_set, val_set, show_progress=True, collate_fn=PaddingTokenCollator(tokenizer.pad))\n\n result = model_wrapper.evaluate(val_set, collate_fn=PaddingTokenCollator(tokenizer.pad))\n # 0.953125\n\n test_inputs = torch.tensor(test_tokens, dtype=torch.long)\n result = model_wrapper.predict(test_inputs)\n # [0, 1]\n\n result = model_wrapper.predict_classes(test_inputs)\n # ['class1', 'class2']\n\n result = model_wrapper.predict_proba(test_inputs)\n # ([0, 1], array([0.99439645, 0.99190724], dtype=float32))\n\n result = model_wrapper.predict_classes_proba(test_inputs)\n # (['class1', 'class2'], array([0.99439645, 0.99190724], dtype=float32))\n\n\n",
"bugtrack_url": null,
"license": null,
"summary": "A tool set for NLP. Text classification. Trainer. Tokenizer",
"version": "1.8.1",
"project_urls": {
"Homepage": "https://gitee.com/summry/nlpx"
},
"split_keywords": [
"nlp",
" nlp",
" ai",
" llm",
" gpt",
" machine learning",
" deep learning",
" tokenize",
" torch"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "3072eef6a8441b478e1b30d6beeaf9ac6863a0da01518838f438b54e45bbbb80",
"md5": "2a8755751b0bcb5a4da8750f51eff952",
"sha256": "e139e42e82678d07ba3fbf2beae32dc0e5daaa2c1f2a833ae6e100ccca09abc3"
},
"downloads": -1,
"filename": "NLPX-1.8.1.tar.gz",
"has_sig": false,
"md5_digest": "2a8755751b0bcb5a4da8750f51eff952",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.6",
"size": 29254,
"upload_time": "2024-12-18T16:36:48",
"upload_time_iso_8601": "2024-12-18T16:36:48.460358Z",
"url": "https://files.pythonhosted.org/packages/30/72/eef6a8441b478e1b30d6beeaf9ac6863a0da01518838f438b54e45bbbb80/NLPX-1.8.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-12-18 16:36:48",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "nlpx"
}
JSON