The above data is obtained from 7303 training sets from the reuters data set, and sklearn is used to extract the tfidf features. The results obtained are all 0. What's going on? 
When I take part of this data, I can get the correct tfidf results for these small parts of the data. 
![[Web front-end] Node.js quick start](https://img.php.cn/upload/course/000/000/067/662b5d34ba7c0227.png)
The above code may be caused by your accuracy being too low or min_count
For example, if the word frequency is 1 and the total number of words is 1e9, the corresponding tf is 1e-9, which is ignored.