The goal of this repository is to provide a simple, modular, and extendable toolkit for IndicTrans2 and be compatible with the HuggingFace models released.
- The repository has been renamed to
IndicTransToolkit
. - The custom tokenizer is now removed from the repository. Please revert to a previous commit (v1.0.1) to use it (strongly discouraged). The official (and only tokenizer) is available on HF along with the models.
- The PreTrainedTokenizer for IndicTrans2 is now available on HF 🎉🎉 Note that, you still need the
IndicProcessor
to pre-process the sentences before tokenization. - In favor of the standard PreTrainedTokenizer, we deprecated the custom tokenizer. However, this custom tokenizer will still be available here for backward compatibility, but no further updates/bug-fixes will be provided.
- The
indic_evaluate
function is now consolidated into a concreteIndicEvaluator
class. - The data collation function for training is consolidated into a concrete
IndicDataCollator
class. - A simple batching method is now available in the
IndicProcessor
.
Python 3.8+
- Indic NLP Library
- Other requirements as listed in
requirements.txt
- Editable installation (Note, this may take a while):
git clone https://github.com/VarunGumma/IndicTransToolkit
cd IndicTransToolkit
pip install --editable ./
For the training usecase, please refer here.
import torch
from IndicTransToolkit import IndicProcessor
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
ip = IndicProcessor(inference=True)
tokenizer = AutoTokenizer.from_pretrained("ai4bharat/indictrans2-en-indic-dist-200M", trust_remote_code=True)
model = AutoModelForSeq2SeqLM.from_pretrained("ai4bharat/indictrans2-en-indic-dist-200M", trust_remote_code=True)
sentences = [
"This is a test sentence.",
"This is another longer different test sentence.",
"Please send an SMS to 9876543210 and an email on newemail123@xyz.com by 15th October, 2023.",
]
batch = ip.preprocess_batch(sentences, src_lang="eng_Latn", tgt_lang="hin_Deva")
batch = tokenizer(batch, padding="longest", truncation=True, max_length=256, return_tensors="pt")
with torch.inference_mode():
outputs = model.generate(**batch, num_beams=5, num_return_sequences=1, max_length=256)
with tokenizer.as_target_tokenizer():
# This scoping is absolutely necessary, as it will instruct the tokenizer to tokenize using the target vocabulary.
# Failure to use this scoping will result in gibberish/unexpected predictions as the output will be de-tokenized with the source vocabulary instead.
outputs = tokenizer.batch_decode(outputs, skip_special_tokens=True, clean_up_tokenization_spaces=True)
outputs = ip.postprocess_batch(outputs, lang="hin_Deva")
print(outputs)
>>> ['यह एक परीक्षण वाक्य है।', 'यह एक और लंबा अलग परीक्षण वाक्य है।', 'कृपया 9876543210 पर एक एस. एम. एस. भेजें और 15 अक्टूबर, 2023 तक newemail123@xyz.com पर एक ईमेल भेजें।']
IndicEvaluator
is a python implementation of compute_metrics.sh.- We have found that this python implementation gives slightly lower scores than the original
compute_metrics.sh
. So, please use this function cautiously, and feel free to raise a PR if you have found the bug/fix.
from IndicTransToolkit import IndicEvaluator
# this method returns a dictionary with BLEU and ChrF2++ scores with appropriate signatures
evaluator = IndicEvaluator()
scores = evaluator.evaluate(tgt_lang=tgt_lang, preds=pred_file, refs=ref_file)
# alternatively, you can pass the list of predictions and references instead of files
# scores = evaluator.evaluate(tgt_lang=tgt_lang, preds=preds, refs=refs)
ip = IndicProcessor(inference=True)
for batch in ip.get_batches(source_sentences, batch_size=32):
# perform necessary operations on the batch
# ... pre-processing
# ... tokenization
# ... generation
# ... decoding
- For
Python >= 3.12
, you can use the inbuilt batching function,itertools.batched
, instead of theget_batches
method. (docs)
- Varun Gumma (varun230999@gmail.com)
- Jay Gala (jaygala24@gmail.com)
- Pranjal Agadh Chitale (pranjalchitale@gmail.com)
- Raj Dabre (prajdabre@gmail.com)
Since this a bleeding-edge module, you may encounter broken stuff and import issues once in a while. In case you encounter any bugs or want additional functionalities, please feel free to raise Issues
/Pull Requests
or contact the authors.
If you use our codebase, or models, please do cite the following paper:
@article{
gala2023indictrans,
title={IndicTrans2: Towards High-Quality and Accessible Machine Translation Models for all 22 Scheduled Indian Languages},
author={Jay Gala and Pranjal A Chitale and A K Raghavan and Varun Gumma and Sumanth Doddapaneni and Aswanth Kumar M and Janki Atul Nawale and Anupama Sujatha and Ratish Puduppully and Vivek Raghavan and Pratyush Kumar and Mitesh M Khapra and Raj Dabre and Anoop Kunchukuttan},
journal={Transactions on Machine Learning Research},
issn={2835-8856},
year={2023},
url={https://openreview.net/forum?id=vfT4YuzAYA},
note={}
}