# Fast-Inference with Ctranslate2

Speedup inference by 2x-8x using int8 inference in C++

quantized version of Helsinki-NLP/opus-mt-de-en

pip install hf-hub-ctranslate2>=1.0.0 ctranslate2>=3.13.0

Converted using

ct2-transformers-converter --model Helsinki-NLP/opus-mt-de-en --output_dir /home/michael/tmp-ct2fast-opus-mt-de-en --force --copy_files README.md generation_config.json tokenizer_config.json vocab.json source.spm .gitattributes target.spm --quantization float16

Checkpoint compatible to ctranslate2 and hf-hub-ctranslate2

• compute_type=int8_float16 for device="cuda"
• compute_type=int8 for device="cpu"

from hf_hub_ctranslate2 import TranslatorCT2fromHfHub, GeneratorCT2fromHfHub
from transformers import AutoTokenizer

model_name = "michaelfeil/ct2fast-opus-mt-de-en"
# use either TranslatorCT2fromHfHub or GeneratorCT2fromHfHub here, depending on model.
model = TranslatorCT2fromHfHub(
        # load in int8 on CUDA
        model_name_or_path=model_name, 
        device="cuda",
        compute_type="int8_float16",
        tokenizer=AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-de-en")
)
outputs = model.generate(
    text=["How do you call a fast Flan-ingo?", "User: How are you doing?"],
)
print(outputs)

Licence and other remarks:

This is just a quantized version. Licence conditions are intended to be idential to original huggingface repo.

Original description

opus-mt-de-en

• source languages: de

• target languages: en

• OPUS readme: de-en

• dataset: opus

• model: transformer-align

• pre-processing: normalization + SentencePiece

• download original weights: opus-2020-02-26.zip

• test set translations: opus-2020-02-26.test.txt

• test set scores: opus-2020-02-26.eval.txt

Benchmarks