Instructions to use tonythethompson/opus-mt-en-ROMANCE with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use tonythethompson/opus-mt-en-ROMANCE with Transformers:
# Use a pipeline as a high-level helper # Warning: Pipeline type "translation" is no longer supported in transformers v5. # You must load the model directly (see below) or downgrade to v4.x with: # 'pip install "transformers<5.0.0' from transformers import pipeline pipe = pipeline("translation", model="tonythethompson/opus-mt-en-ROMANCE")# Load model directly from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("tonythethompson/opus-mt-en-ROMANCE") model = AutoModelForSeq2SeqLM.from_pretrained("tonythethompson/opus-mt-en-ROMANCE") - Notebooks
- Google Colab
- Kaggle
opus-mt-en-ROMANCE
Overview
This repository contains an ONNX export of Helsinki-NLP/opus-mt-en-ROMANCE, a Marian machine translation model for translating English into Romance languages.
Source
ONNX source repository: onnx-community/opus-mt-en-ROMANCE
Base model: Helsinki-NLP/opus-mt-en-ROMANCE
The upstream base model is a Transformers-compatible Marian model with source language en and target language group roa.
Files
| File | Description |
|---|---|
onnx/encoder_model.onnx |
ONNX encoder model |
onnx/encoder_model_fp16.onnx |
FP16 ONNX encoder model |
onnx/decoder_model.onnx |
ONNX decoder model |
onnx/decoder_model_fp16.onnx |
FP16 ONNX decoder model |
onnx/decoder_model_merged.onnx |
Merged ONNX decoder model |
onnx/decoder_model_merged_fp16.onnx |
FP16 merged ONNX decoder model |
onnx/decoder_with_past_model.onnx |
ONNX decoder with past-key-value inputs |
onnx/decoder_with_past_model_fp16.onnx |
FP16 ONNX decoder with past-key-value inputs |
config.json |
Marian model configuration |
generation_config.json |
Text generation configuration |
quantize_config.json |
ONNX quantization configuration |
source.spm |
Source SentencePiece model |
target.spm |
Target SentencePiece model |
tokenizer.json |
Tokenizer JSON |
tokenizer_config.json |
Tokenizer configuration |
special_tokens_map.json |
Special token mapping |
vocab.json |
Vocabulary |
Intended Use
Use these ONNX model files for English-to-Romance-language machine translation experiments or integration work.
Target Languages
The tokenizer configuration lists Romance-language targets including French, Walloon, Franco-Provencal, Occitan, Catalan, Romansh, Ladin, Friulian, Ligurian, Lombard, Spanish variants, Portuguese variants, Galician, Ladino, Aragonese, Mirandese, Italian, Corsican, Neapolitan, Sicilian, Venetian, Sardinian, Romanian, and Latin.
Training Data
Not documented in this ONNX repository. See the upstream Helsinki-NLP base model documentation for training details.
Evaluation
No ONNX-specific evaluation or parity results are documented in this repository.
Limitations
- This is not a Spanish-to-Portuguese-only model; it translates from English to a Romance-language target group.
- ONNX runtime compatibility and parity with the upstream Transformers model are not documented here.
- Translation quality can vary by target language, domain, and decoding configuration.
License
Apache-2.0, matching the upstream Helsinki-NLP base model metadata.
- Downloads last month
- 16
Model tree for tonythethompson/opus-mt-en-ROMANCE
Base model
onnx-community/opus-mt-en-ROMANCE