transformers

transformers copied to clipboard

`is_torch_tpu_available() got an unexpected keyword argument 'check_device'`

1

### System Info transformers==4.18.0 pytorch==1.12.0 pytorch-lightning==1.6.0 ### Who can help? @patrickvonplaten @sgugger @SaulLu When I run `examples/pytorch/question-answering/run_qa.py`, it reports a bug that `is_torch_tpu_available() got an unexpected keyword argument 'check_device'`. I...

Sakurakdx

Enable ONNX export for `VisionDecoderEncoderModel`

10

# 🚀 Feature request As discussed on the [forums](https://discuss.huggingface.co/t/can-i-export-a-visionencoderdecoder-checkpoint-to-onnx/12885), it would be nice if one could export `VisionDecoderEncoderModel` classes using the `transformers.onnx` package. ## Motivation It is currently not possible...

lewtun

Flax vs torch benchmark on Wav2vec2

6

So my question is, should FlaxWav2Vec2ForCTC generally be faster than Wav2Vec2ForCTC? 1.14 s ± 138 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) -> FlaxWav2Vec2ForCTC...

ZurabDz

Cannot replicate T5 performance on WMT14

6

### System Info I am trying to replicate T5 finetuning on WMT with the following hyperparameters (as close as possible to the paper https://www.jmlr.org/papers/volume21/20-074/20-074.pdf): --model_name_or_path t5-small --source_lang en --target_lang de...

ekurtulus

document-question-answering pipeline does not work with some models

3

### System Info Colab, latest release ### Who can help? @NielsRogge ### Information - [ ] The official example scripts - [ ] My own modified scripts ### Tasks -...

osanseviero

Fix led #19019

2

# What does this PR do? fixes #19019 by replacing the construction of the `eos_mask` in the `SequenceClassification`. Also adds a test to make sure that long sequence are properly...

ArthurZucker

Flax BART training fails when evaluating

5

### System Info - `transformers` version: 4.22.0.dev0 - Platform: Linux-5.15.0-41-generic-x86_64-with-glibc2.29 - Python version: 3.8.10 - Huggingface_hub version: 0.9.1 - Flax version (CPU?/GPU?/TPU?): 0.6.0 (gpu) - Jax version: 0.3.17 - JaxLib...

BramVanroy

Added OnnxConfig for MPNet models

5

# What does this PR do? This PR Adds `OnnxConfig` for MPNet based models. In order for the conversion to be compatible with some older versions of PyTorch (In my...

grafail

### System Info - `transformers` version: 4.22.0 - Platform: Linux-4.19.157-1.20201118.el7.x86_64-x86_64-with-glibc2.17 - Python version: 3.8.3 - Huggingface_hub version: 0.9.1 - PyTorch version (GPU?): 1.10.2+cu111 (True) - Tensorflow version (GPU?): 2.10.0 (True)...

lsz05

### Feature request Allow the **compute_metrics()** in the **Trainer** to take into account the **original input** in addition to the predictions and labels. ### Motivation It is currently possible to...

slvcsl