|
|
from os import path
|
|
|
import streamlit as st
|
|
|
import tensorflow as tf
|
|
|
import random
|
|
|
from transformers import ElectraTokenizerFast, TFElectraForQuestionAnswering
|
|
|
from datasets import Dataset, DatasetDict, load_dataset
|
|
|
|
|
|
model_hf = "nguyennghia0902/bestfailed_electra-small-discriminator_5e-05_16"
|
|
|
tokenizer = ElectraTokenizerFast.from_pretrained(model_hf)
|
|
|
reload_model = TFElectraForQuestionAnswering.from_pretrained(model_hf)
|
|
|
|
|
|
@st.cache_resource
|
|
|
def predict(question, context):
|
|
|
inputs = tokenizer(question, context, return_offsets_mapping=True,return_tensors="tf",max_length=512, truncation=True)
|
|
|
offset_mapping = inputs.pop("offset_mapping")
|
|
|
outputs = reload_model(**inputs)
|
|
|
answer_start_index = int(tf.math.argmax(outputs.start_logits, axis=-1)[0])
|
|
|
answer_end_index = int(tf.math.argmax(outputs.end_logits, axis=-1)[0])
|
|
|
start_char = offset_mapping[0][answer_start_index][0]
|
|
|
end_char = offset_mapping[0][answer_end_index][1]
|
|
|
predicted_answer_text = context[start_char:end_char]
|
|
|
|
|
|
return predicted_answer_text
|
|
|
|
|
|
def main():
|
|
|
st.set_page_config(page_title="Sample in Dataset", page_icon="📝")
|
|
|
|
|
|
|
|
|
col1, col2 = st.columns([2, 1])
|
|
|
col1.title("Sample in Dataset")
|
|
|
|
|
|
new_data = load_dataset("nguyennghia0902/project02_textming_dataset", data_files={'train': 'raw_newformat_data/traindata-00000-of-00001.arrow', 'test': 'raw_newformat_data/testdata-00000-of-00001.arrow'})
|
|
|
|
|
|
sampleQ = ""
|
|
|
sampleC = ""
|
|
|
sampleA = ""
|
|
|
if st.button("Sample"):
|
|
|
sample = random.choice(new_data['test'])
|
|
|
sampleQ = sample['question']
|
|
|
sampleC = sample['context']
|
|
|
sampleA = sample['answers']["text"][0]
|
|
|
|
|
|
question = st.text_area(
|
|
|
"Sample QUESTION: ",
|
|
|
sampleQ,
|
|
|
height=15,
|
|
|
)
|
|
|
text = st.text_area(
|
|
|
"Sample CONTEXT:",
|
|
|
sampleC,
|
|
|
height=100,
|
|
|
)
|
|
|
answer = st.text_area(
|
|
|
"True ANSWER:",
|
|
|
sampleA,
|
|
|
height=20,
|
|
|
)
|
|
|
|
|
|
|
|
|
if st.button("Predict"):
|
|
|
prediction = ""
|
|
|
stripped_text = text.strip()
|
|
|
if not stripped_text:
|
|
|
st.error("Please enter a context.")
|
|
|
return
|
|
|
stripped_question = question.strip()
|
|
|
if not stripped_question:
|
|
|
st.error("Please enter a question.")
|
|
|
return
|
|
|
|
|
|
prediction = predict(stripped_question, stripped_text)
|
|
|
st.success(prediction)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
main() |
