trained-tokenizer/fertility_test/multiLingualFertility.py

import sentencepiece as spm
from datasets import concatenate_datasets, load_dataset, load_from_disk
import logging
from datasets import DatasetDict, Dataset

sp = spm.SentencePieceProcessor()
sp.load('ta_te_kan_ml_50kspm_tokenizer.model')
sentence = "The Indian cricket fans were very disappointed after India's loss against Australia in the World Cup."

tokens = sp.encode(sentence, out_type=str)
print(len(tokens) , tokens)

sentence = "विश्व कप में ऑस्ट्रेलिया के खिलाफ भारत की हार के बाद भारतीय क्रिकेट प्रशंसक काफी निराश थे।"

tokens = sp.encode(sentence, out_type=str)
print(len(tokens) , tokens)

decoded_sentence = sp.decode(tokens)

# exit()
# print("Tokens:", tokens)
# print("Decoded sentence:", decoded_sentence)
# exit()
def initialize_logger(log_file):
    logging.basicConfig(filename=log_file, level=logging.INFO, format='%(asctime)s: %(message)s')

def log_parameters(vocab_size, pretrained_model, en_fertility_score, hi_fertility_score , kn_fertility_score ,ml_fertility_score,ta_fertility_score,te_fertility_score, log_file='parameters.log'):
    initialize_logger(log_file)
    logging.info(f"Vocabulary Size: {vocab_size}, Tokenizer type: {pretrained_model}, English Fertility Score: {en_fertility_score} , Hindi Fertility Score: {hi_fertility_score}, Kannada Fertility Score: {kn_fertility_score}, <Malayalam Fertility Score: {ml_fertility_score}, Tamil Fertility Score: {ta_fertility_score}, Telugu Fertility Score: {te_fertility_score}")

dataset_hi= load_dataset('ai4bharat/samanantar', 'hi', split='train', cache_dir='/sml1/atul/CENTRAL_CACHE')
dataset_kn= load_dataset('ai4bharat/samanantar', 'kn', split='train', cache_dir='/sml1/atul/CENTRAL_CACHE')
dataset_ml= load_dataset('ai4bharat/samanantar', 'ml', split='train', cache_dir='/sml1/atul/CENTRAL_CACHE')
dataset_ta= load_dataset('ai4bharat/samanantar', 'ta', split='train', cache_dir='/sml1/atul/CENTRAL_CACHE')
dataset_te= load_dataset('ai4bharat/samanantar', 'te', split='train', cache_dir='/sml1/atul/CENTRAL_CACHE')

test_en = dataset_hi['src'][:10000]
test_hi = dataset_hi['tgt'][:10000]
test_kn = dataset_kn['tgt'][:10000]
test_ml = dataset_ml['tgt'][:10000]
test_ta = dataset_ta['tgt'][:10000]
test_te = dataset_te['tgt'][:10000]

en_fertility_score=0
hi_fertility_score=0
kn_fertility_score=0
ml_fertility_score=0
ta_fertility_score=0
te_fertility_score=0

for data in test_en:
    tok=sp.encode(data, out_type=str)
    en_fertility_score += (len(tok)) / len(data.split())
en_fertility_score=en_fertility_score/10000

for data in test_hi:
    tok=sp.encode(data, out_type=str)
    hi_fertility_score += (len(tok))  / len(data.split())
hi_fertility_score=hi_fertility_score/10000

for data in test_kn:
    tok=sp.encode(data, out_type=str)
    kn_fertility_score += (len(tok))  / len(data.split())
kn_fertility_score=kn_fertility_score/10000

for data in test_ml:
    tok=sp.encode(data, out_type=str)
    ml_fertility_score += (len(tok))  / len(data.split())
ml_fertility_score=ml_fertility_score/10000

for data in test_ta:
    tok=sp.encode(data, out_type=str)
    ta_fertility_score += (len(tok))  / len(data.split())
ta_fertility_score=ta_fertility_score/10000

for data in test_te:
    tok=sp.encode(data, out_type=str)
    te_fertility_score += (len(tok))  / len(data.split())
te_fertility_score=te_fertility_score/10000

log_parameters(64000, "kn-ml-ta-te-only-spm", en_fertility_score, hi_fertility_score , kn_fertility_score, ml_fertility_score,ta_fertility_score,te_fertility_score )