# Fast-Inference with Ctranslate2
Speedup inference by 2x-8x using int8 inference in C++
quantized version of Helsinki-NLP/opus-mt-ROMANCE-en
pip install hf-hub-ctranslate2>=1.0.0 ctranslate2>=3.13.0
Converted using
ct2-transformers-converter --model Helsinki-NLP/opus-mt-ROMANCE-en --output_dir /home/michael/tmp-ct2fast-opus-mt-ROMANCE-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_float16fordevice="cuda"compute_type=int8fordevice="cpu"
from hf_hub_ctranslate2 import TranslatorCT2fromHfHub, GeneratorCT2fromHfHub
from transformers import AutoTokenizer
model_name = "michaelfeil/ct2fast-opus-mt-ROMANCE-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-ROMANCE-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-ROMANCE-en
source languages: fr,fr_BE,fr_CA,fr_FR,wa,frp,oc,ca,rm,lld,fur,lij,lmo,es,es_AR,es_CL,es_CO,es_CR,es_DO,es_EC,es_ES,es_GT,es_HN,es_MX,es_NI,es_PA,es_PE,es_PR,es_SV,es_UY,es_VE,pt,pt_br,pt_BR,pt_PT,gl,lad,an,mwl,it,it_IT,co,nap,scn,vec,sc,ro,la
target languages: en
dataset: opus
model: transformer
pre-processing: normalization + SentencePiece
download original weights: opus-2020-04-01.zip
test set translations: opus-2020-04-01.test.txt
test set scores: opus-2020-04-01.eval.txt
Benchmarks
| testset | BLEU | chr-F |
|---|---|---|
| Tatoeba.fr.en | 62.2 | 0.750 |