optimum

optimum copied to clipboard

# What does this PR do? Add the folding of `nn.BatchNorm2d` into `nn.Conv2d`, and of `nn.BatchNorm1d` into `nn.Linear`. The former is useful for example for resnet, beit, regnet, dpt, levit,...

fxmarty

Fix the way the linear modules are removed from the parent module

1

# What does this PR do? Fixes # (issue) ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks...

michaelbenayoun

add mt5 to ORTConfigManager conf list

24

# What does this PR do? Add MT5 to `ORTConfigManager`. Fixes #321 I re-arranged in alphabetical order all available models. I can put it back like it was if needed....

chainyo

ValueError on attention mask size check for ORTModelForSeq2SeqLM with blenderbot-400M-distill

1

The line ORTModelForSeq2SeqLM.from_pretrained("facebook/blenderbot-400M-distill", from_transformers=True) causes the following error: ``` blenderbot/modeling_blenderbot.py", line 228, in forward f"Attention mask should be of size {(bsz, 1, tgt_len, src_len)}, but is {attention_mask.size()}" ValueError: Attention mask...

guillermo-gabrielli-fer

Optimum model not using all available CPU threads

3

### System Info ```shell - python 3.8.13 - Ubuntu 22.04 - optimum 1.3.0 - torch 1.11.0 (cpu version) - transformers 4.21.0 - datasets 2.4.0 - onnxruntime 1.12.0 ``` ### Who...

jessecambon

Quantisation of BigBirdForTokenClassification suffers significant performance drop

12

### System Info ```shell Apple M1 Pro macOS 12.5 Monterey optimum[onnxruntime]==1.3.0 python==3.9.11 ``` ### Who can help? @JingyaHuang ### Information - [ ] The official example scripts - [X] My...

lewisbails

Fix CPU device parameter

2

# What does this PR do? Fixes a crash in Trainer using CPU. This parameter is expected to be of type `torch.device` not `str`. Otherwise this crashes with: ``` ../../.mamba/envs/score/lib/python3.7/site-packages/optimum/onnxruntime/trainer.py:1315:...

mausch

Fix typo in variable name

1

# What does this PR do? Fixes a small typo ## Before submitting - [x] This PR fixes a typo or improves the docs (you can dismiss the other checks...

mausch

Optimum's DeBERTa-V2 behavior strange when training with ORT (training hangs or takes impossibly long)

6

### System Info Running with CUDA 11.5, Python 3.8, and torch 11.11. I installed the Python dependencies from requirements.txt in the text-classification example folder. I installed transformers from source, and...

carzh

MT5 Support

4

Big Thanks for the awesome Optimum library, MT5 is a Massive Transformers And Supports Many Languages, but it has a slow inference time for speedup MT5 inference time, We need...

MiladMolazadeh