Using ONNX format of the facebook/mbart-large-50-many-to-many-mmt?

Intermediate
1

Hi all,

Used this code to export mentioned model to ONNX format:

import torch
from transformers import MBartForConditionalGeneration, MBart50TokenizerFast

model_path = "./mbart_large_50_model"
model = MBartForConditionalGeneration.from_pretrained(model_path)
tokenizer = MBart50TokenizerFast.from_pretrained(model_path)

input_text = "This is just simple text"
inputs = tokenizer(input_text, return_tensors="pt")

onnx_path = "./mbart_large_50.onnx"

# Export to ONNX format:
torch.onnx.export(
    model,                                 # PyTorch model
    (inputs["input_ids"],),               
    onnx_path,                            # the path for the  resulting ONNX file
    input_names=["input_ids"],            # input tensor names
    output_names=["logits"],              # output tensor names
    dynamic_axes={"input_ids": {0: "batch", 1: "sequence"}},
    opset_version=14                      
)

Got resulting mbart_large_50.onnx but I’m completely stuck on how to use it to translate between two languages. Any suggestions would be appreciated.

2

now i have converted my mbart50 trained model to .onnx but when I’m going to test the model it gives input sentences. why

3

It is recommended to use Optimum as much as possible when converting to ONNX format.