Skip to content

Instantly share code, notes, and snippets.

@kristjan-eljand
Last active April 14, 2021 08:32
Show Gist options
  • Save kristjan-eljand/76a1f032dab3f4920dab2261453e28d6 to your computer and use it in GitHub Desktop.
Save kristjan-eljand/76a1f032dab3f4920dab2261453e28d6 to your computer and use it in GitHub Desktop.
Text generation with hugging face pre-trained models on non-English language
from pprint import pprint #for user-friendly output printing
# 1. Create a pipeline for text-generation task
generator = pipeline('text-generation', model='distilgpt2')
# 2. Translate the sentence beginnings from Est to Eng
beginnings_origin = [
"Eesti toodab elektrienergiat peamiselt",
"Taastuvenergia on oluline, sest"
]
translated_beginnings = [translate(b, EST_TO_ENG)[0]['translation_text'] for b in beginnings_origin]
print("Sentence beginnings that are translated to English:\n")
pprint(translated_beginnings)
# 3. Generate texts in English based on the beginnings
generated_english = []
for b in translated_beginnings:
result = generator(b, max_length=50, num_return_sequences=1)
text = result[0]['generated_text']
generated_english.append(text)
print("Generated text in English:\n")
pprint(generated_english)
# 4. Translate generated texts back to Estonian
generated_estonian = [translate(g, ENG_TO_EST)[0]['translation_text'] for g in generated_english]
print("Generated text that is translated back to Estonian:\n")
pprint(generated_estonian)
# Output:
# Sentence beginnings that are translated to English:
# 'Estonia produces electricity mainly'
# 'Renewable energy is important because:'
#
# Generated text in English:
# 'Estonia produces electricity mainly from natural gas-fired hydroelectric
# power plants, according to a report last summer by The World
# Bank.'
#
# 'Renewable energy is important because: The more we know about fossil fuels,
# the more we realize that we need to reduce greenhouse gas emissions (about 10%).'
#
# Generated text that is translated back to Estonian:
# 'Eesti toodab elektrit peamiselt maagaasil töötavatest hüdroelektrijaamadest,
# vastavalt Maailmapanga möödunud suve aruandele.'
#
# 'Taastuvenergia on oluline, sest mida rohkem me teame fossiilkütustest, seda
# enam mõistame, et peame vähendama kasvuhoonegaaside heitkoguseid (umbes 10%).'
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment