mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-05-27 22:45:57 +00:00
c00e13a291
* remove implicit linking of tensorrt and dnnl ep shared libs (#5262) * Update DirectML Nuget to 1.3.0 (#5274) * Update PyTorch TransformerModel sample (#5275) * Insert telemetry template into GPU build, add telemry build switches. (#5278) * Synchronize training dependency versions between Docker image and Python wheel (#5261) * Downgrade GCC (#5269) * Remove --enable_symbolic_shape_infer_tests to fix linux ci pipeline build error. Co-authored-by: Edward Chen Co-authored-by: George Wu <jywu@microsoft.com> Co-authored-by: Dwayne Robinson <dwayner@microsoft.com> Co-authored-by: Thiago Crepaldi <thiago.crepaldi@microsoft.com> Co-authored-by: Dmitri Smirnov <yuslepukhin@users.noreply.github.com> Co-authored-by: edgchen1 <18449977+edgchen1@users.noreply.github.com> Co-authored-by: Changming Sun <chasun@microsoft.com> |
||
|---|---|---|
| .. | ||
| ort_train.py | ||
| ort_utils.py | ||
| pt_model.py | ||
| pt_train.py | ||
| README.md | ||
| utils.py | ||
TransformerModel example
This example was adapted from Pytorch's Sequence-to-Sequence Modeling with nn.Transformer and TorchText tutorial
Requirements
- PyTorch 1.6+
- TorchText 0.6+
- ONNX Runtime 1.5+
Running PyTorch version
python pt_model.py
Running ONNX Runtime version
python ort_model.py
Optional arguments
| Argument | Description | Default |
|---|---|---|
| --batch-size | input batch size for training | 20 |
| --test-batch-size | input batch size for testing | 20 |
| --epochs | number of epochs to train | 2 |
| --lr | learning rate | 0.001 |
| --no-cuda | disables CUDA training | False |
| --seed | random seed | 1 |
| --log-interval | how many batches to wait before logging training status | 200 |