File size: 2,147 Bytes
2b81c45 1f4bec9 2b81c45 2fc1b77 1f4bec9 2b81c45 2fc1b77 1f4bec9 2b81c45 2fc1b77 1f4bec9 2b81c45 2fc1b77 2b81c45 b34faf5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 |
import streamlit as st
from transformers import MBartForConditionalGeneration, MBart50TokenizerFast
from transformers import AutoModelForMaskedLM, AutoTokenizer
# Load the models and tokenizers
model_translation = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-50-one-to-many-mmt")
model_masked_lm = AutoModelForMaskedLM.from_pretrained("alabnii/jmedroberta-base-sentencepiece")
model_translation.eval()
tokenizer_translation = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50-one-to-many-mmt", src_lang="en_XX")
tokenizer_masked_lm = AutoTokenizer.from_pretrained("alabnii/jmedroberta-base-sentencepiece")
text = st.text_area('Enter the text:')
if text:
model_inputs = tokenizer_translation(text, return_tensors="pt")
lg = st.text_input("Select Language: hi.Hindi, te.Telugu, gu.Gujarati, bn.Bengali")
if lg=='hi':
generated_tokens = model_translation.generate(
**model_inputs,
forced_bos_token_id=tokenizer_translation.lang_code_to_id["hi_IN"]
)
translation = tokenizer_translation.batch_decode(generated_tokens, skip_special_tokens=True)
st.write(translation)
elif lg=='te':
generated_tokens = model_translation.generate(
**model_inputs,
forced_bos_token_id=tokenizer_translation.lang_code_to_id["te_IN"]
)
translation = tokenizer_translation.batch_decode(generated_tokens, skip_special_tokens=True)
st.write(translation)
elif lg=='gu':
generated_tokens = model_translation.generate(
**model_inputs,
forced_bos_token_id=tokenizer_translation.lang_code_to_id["gu_IN"]
)
translation = tokenizer_translation.batch_decode(generated_tokens, skip_special_tokens=True)
st.write(translation)
elif lg=='bn':
generated_tokens = model_translation.generate(
**model_inputs,
forced_bos_token_id=tokenizer_translation.lang_code_to_id["bn_IN"]
)
translation = tokenizer_translation.batch_decode(generated_tokens, skip_special_tokens=True)
st.json(translation)
else:
st.write('invalid choice!') |