Wildly differing results compared to crfsuite-rs

[decahedron1]

I'm using crfs for part-of-speech tagging and I noticed the output is very different compared to crfsuite-rs.

With the input The quick brown fox jumped over the lazy dog:

crfsuite-rs: DT JJ NN RB VBN IN DT JJ NN
crfs: DT JJ NN JJ NN IN DT JJ NN

Tagging code using crfs
Model: pos.crf
Test script:

use std::fs;

use crfs::Model;
use pos_test::PartOfSpeechTagger;

fn main() {
	let model = fs::read("./pos.crf").unwrap();
	let model = Model::new(&model).unwrap();
	let mut tagger = PartOfSpeechTagger::new(&model, Default::default()).unwrap();
	let sentence = "The quick brown fox jumped over the lazy dog.";
	let results = tagger.tag(&sentence.split_ascii_whitespace().collect::<Vec<_>>()).unwrap();
	for part in results {
		print!("{part} ");
	}
	println!();
}