- Researched and utilized preprocessing techniques and word embedding methods, including Bag-of-Words, N-gram, GloVe, Word2Vec, etc.
- Implemented models such as CNN, RNN, LSTM for tasks like sentiment classification, text matching, etc.
torch 1.5.1
torchtext 0.6.0
torchvision 0.6.0
spacy 3.3.0
jsonlines
在https://featurize.cn/ 服务器租借平台上测试,使用CUDA 11.2 + torch 1.10 + torchvision 0.11 + torchtext 0.6.0 + spacy 3.3.1也可以跑通
通过python -m spacy download en_core_web_sm
下载en_core_web_sm
torchtext第一次根据glove权重构造Vector时候时间会比较长,可根据进度条等待
torchtext—ValueError: Fan in and fan out can not be computed for tensor with fewer than 2 dimensions 可参考https://blog.csdn.net/qq_23262411/article/details/100173224?spm=1001.2101.3001.6650.1&utm_medium=distribute.pc_relevant.none-task-blog-2%7Edefault%7ECTRLIST%7Edefault-1-100173224-blog-114952212.pc_relevant_aa&depth_1-utm_source=distribute.pc_relevant.none-task-blog-2%7Edefault%7ECTRLIST%7Edefault-1-100173224-blog-114952212.pc_relevant_aa&utm_relevant_index=2