Helsinki-NLP/opus-mt-tc-bible-big-deu_eng_fra_por_spa-gmw (revision 0a10a15), a multilingual model, exported to ONNX for on-device translation from French to Dutch. Put the target token >>nld<< before the input.
--task text2text-generation-with-past): one encoder and one merged decoder that handles both the first step and the cached steps (use_cache_branch input), so the decoder weights are stored once.quantize_dynamic, per_channel=True, EnableSubgraph so the layers inside the merged decoder's branches are quantised too).| File | Purpose |
|---|---|
encoder_model.onnx |
Encoder |
decoder_model_merged.onnx |
Decoder with KV cache |
source.spm, target.spm |
SentencePiece models |
vocab.json |
Shared vocabulary (includes the target-language tokens) |
config.json, generation_config.json |
Token IDs (eos, pad, decoder start) and generation settings |
With greedy decoding, the int8 files give the same output as the original PyTorch model on 25 of 30 short test phrases; the rest differ in wording (quantisation).
The model is by the Language Technology Research Group at the University of Helsinki (OPUS-MT, Tiedemann & Thottingal), released under Apache-2.0; this conversion is released under the same licence.
decoder_model_merged.onnx 202 MBencoder_model.onnx 127 MB