Is there any plans to support torch.export?

描述

from fairseq.models import FairseqEncoderDecoderModel
import torch

translator= FairseqEncoderDecoderModel.from_pretrained(
    # ..
)
example_inputs = translator.dummy_inputs

with torch.no_grad():
    ep = torch.export.export(translator_model, example_inputs)

贡献者指南

技术栈

pythonpytorch

领域

machine learning

议题类型

feature

难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。

4

预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。

over 1 week

活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。

stale

清晰度议题是否清楚说明期望改动、验收标准和下一步。

unclear

前置要求

PythonPyTorchfairseq model internalstorch.export API

新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。

20

研究方向

Investigate whether fairseq currently has any export mechanism (e.g., ONNX, TorchScript) and evaluate compatibility with torch.export. Review the FairseqEncoderDecoderModel class in fairseq/models to understand the model architecture and dummy inputs method. Check the PyTorch torch.export documentation for constraints on dynamic shapes and control flow. Search the fairseq repository for any open PRs or discussions related to export or compilation. Provide a concrete proposal for integrating torch.export, noting potential challenges with fairseq's sequence to sequence models.