| import streamlit as st | |
| import datasets | |
| from functools import partial | |
| data = datasets.load_dataset("json", data_files="small_test_data.jsonl")["train"].select(range(100)) | |
| bad_cutoff = st.slider('Bad words cutoff', 0, 1) | |
| stp_cutoff = st.slider('Stop words cutoff', 0, 1) | |
| ppl_cutoff = st.slider('ppl cutoff', 0, 1) | |
| def filter_ppl(examples, invert=False): | |
| return [ppl < ppl_cutoff for ppl in examples["ppl"]] | |
| def filter_bad(examples, invert=False): | |
| return [bad < bad_cutoff for bad in examples["bad_words"]] | |
| def filter_stp(examples, invert=False): | |
| return [stp > stp_cutoff for stp in examples["stop_words"]] | |
| st.table(data) | |