transformers

transformers copied to clipboard

[Tracker] [bnb] Supporting `device_map` containing GPU and CPU devices

1

### Feature request We should be able to provide custom `device_map` when using 8-bit models using `bitsandbytes`. This would enable users having more control over the modules they want to...

younesbelkada

v4.22.1 ErnieForMaskedLM Bug

3

### System Info  @LysandreJik ### Who can help? _No response_ ### Information - [ ] The official example scripts - [ ] My own modified scripts ### Tasks -...

wzjj98

### System Info - transformers version: 4.18.1 - Platform: Linux Jupyter Notebook, TF2.3 Python 3.6, 2 GPU - Python version: '1.7.1+cu101' - Using GPU in script?: yes - Using distributed...

pratikchhapolika

Deberta MaskedLM Corrections

11

# What does this PR do? The current implementations of DebertaForMaskedLM and DebertaV2ForMaskedLM do not load all of the weights from the checkpoints. After consulting the [original repo](https://github.com/microsoft/DeBERTa/blob/master/DeBERTa/deberta/bert.py), I modified...

nbroad1881

Add missing type hints

124

### This issue is part of our **Great Code Cleanup 2022**. If you're interested in helping out, take a look at [this thread](https://twitter.com/carrigmat/status/1502319813510766599), or come [join us on Discord](https://t.co/kS42XBvpWH) and...

Rocketknight1

XGLM - Fix Softmax NaNs when using FP16

6

# What does this PR do? Fixes #18049 following the exact same procedure used in #17437. Beside the added test, I also evaluated the fix on my personal use-case and...

gsarti

dalle mega

2

# What does this PR do? This PR adds the DalleMega model from [dalle-mini](https://github.com/borisdayma/dalle-mini) for text-2-image generation. The VQGAN model required for converting the tokens to image is in this...

patil-suraj

Error while loading a pre-trained wav2vec2 model

2

### System Info - `transformers` version: 4.20.1 - Platform: Linux-5.4.0-1085-azure-x86_64-with-glibc2.10 - Python version: 3.8.13 - Huggingface_hub version: 0.8.1 - PyTorch version (GPU?): 1.9.1+cu111 (True) - Tensorflow version (GPU?): not installed...

Aaryan369

GPT-J evaluation with multiple GPUs crashes

3

### System Info - `transformers` version: 4.21.0 - Platform: Linux-3.10.0-1160.71.1.el7.x86_64-x86_64-with-glibc2.17 - Python version: 3.10.4 - Huggingface_hub version: 0.8.1 - PyTorch version (GPU?): 1.12.0+cu116 (True) - Tensorflow version (GPU?): not installed...

manuelciosici

Padding offsets mapping via `tokenizer.pad`

8

### Feature request While preparing the dataset for the Named Entity Recognition task, I noticed that `tokenizer.pad` does not apply padding for `offset_mapping`, which is necessary not only for the...

vadimirtlach