patrickvonplaten Follow

Patrick von Platen patrickvonplaten

340 followers · 30 following · 147

Achievements

Highlights

Organizations

Block or Report

Block or report patrickvonplaten

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse

Popular repositories

notebooks Public

Some notebooks for NLP

Jupyter Notebook 85 21
AdvancedAutomaticSpeechRecognition Public

Python 6
AdvertisementDetection Public

Python 3
TRexGameRL Public

JavaScript 2
scientific_images Public

2
files_to_link_to Public

Python 2 1

1,639 contributions in the last year

Learn how we count contributions

Less More

Contribution activity

October 2021

Created 56 commits in 9 repositories

Created 1 repository

patrickvonplaten/UniSpeech Python Oct 26

Created a pull request in huggingface/datasets that received 9 comments

Oct 14

[Audio datasets] Adapting all audio datasets

This PR adds the new Audio(...) features - see: #2324 to the most important audio datasets: Librispeech Timit Common Voice AMI ... (others I'm for…

+99 −31 • 9 comments

Opened 27 other pull requests in 6 repositories

huggingface/transformers 20 merged 1 open

[Speech Recognition] - Distributed training: Make sure vocab file removal and creation don't interfer Oct 26
[Speech Recognition CTC] Add auth token to fine-tune private models Oct 26
[TPU tests] Enable first TPU examples pytorch Oct 22
[tests] fix hubert test sort Oct 22
[ASR] Small fix model card creation Oct 21
[ASR] Make speech recognition example more general to load any tokenizer Oct 20
Add ASR colabs Oct 19
Add TF<>PT and Flax<>PT everywhere Oct 18
[Flax] Clip fix test Oct 18
[Speech] Move all examples to new audio feature Oct 18
[Speech] Refactor Examples Oct 17
[Docs] More general docstrings Oct 15
[Speech Examples] Add new audio feature Oct 15
[Testing] Move speech datasets to hf-internal testing ... Oct 14
Allow single byte decoding Oct 13
[Wav2Vec2] Make sure tensors are always bool for mask_indices Oct 12
Add Unispeech & Unispeech-SAT Oct 11
[Gradient checkpoining] Correct disabling find_unused_parameters in Trainer when gradient checkpointing is enabled Oct 11
[Wav2Vec2] Fix mask_feature_prob Oct 7
[Generation] Fix max_new_tokens Oct 7
[Speech Examples] Add pytorch speech pretraining Oct 5

huggingface/notebooks 2 merged

Add English Speech ASR fine-tuning Oct 19
Add multi-lingual Speech ASR fine-tuning Oct 19

microsoft/UniSpeech 1 merged

Add links to Unispeech-SAT models on HF Oct 26

huggingface/blog 1 merged

[Wav2Vec2] Update blog posts with new audio feature and other Oct 19

google/flax 1 merged

Add link to Transformers examples Oct 8

deepmind/optax 1 closed

Update CONTRIBUTING.md Oct 2

Reviewed 44 pull requests in 6 repositories

huggingface/transformers 38 pull requests

Add SEW CTC models Oct 26
Remove n_ctx from configs Oct 26
Add Unispeech & Unispeech-SAT Oct 26
[TPU tests] Enable first TPU examples pytorch Oct 25
BartEnocder add set_input_embeddings Oct 25
Add ESM to huggingface Oct 25
Add FlaxVisionEncoderDecoderModel Oct 25
Add TF<>PT and Flax<>PT everywhere Oct 24
Adding handle_long_generation paramters for text-generation pipeline. Oct 24
Fix lazy init to stop hiding errors in import Oct 24
Replace assert statements with exceptions (#13871) Oct 22
Fix ignore_mismatched_sizes Oct 21
Fix weight loading issue Oct 20
Add ASR colabs Oct 19
[Examples] Use Audio feature in speech classification Oct 18
[Speech] Refactor Examples Oct 18
[Docs] More general docstrings Oct 15
Add the SEW and SEW-D speech models Oct 15
Add REALM Oct 12
[Speech Examples] Add pytorch speech pretraining Oct 11
Add TrOCR + VisionEncoderDecoderModel Oct 11
[Gradient checkpoining] Correct disabling find_unused_parameters in Trainer when gradient checkpointing is enabled Oct 11
Raise ValueError instead of asserts in src/transformers/benchmark/benchmark.py Oct 11
Honor existing attention mask in tokenzier.pad Oct 11
Add TFEncoderDecoderModel + Add cross-attention to some TF models Oct 8
Some pull request reviews not shown.

huggingface/datasets 2 pull requests

Update SUPERB to use Audio features Oct 18
Create Audio feature Oct 11

huggingface/notebooks 1 pull request

Add audio classification notebook Oct 21

huggingface/blog 1 pull request

[Wav2Vec2] Update blog posts with new audio feature and other Oct 19

google/flax 1 pull request

Add link to Transformers examples Oct 11

pytorch/fairseq 1 pull request

[Hubert] Fix padding mask Oct 11

Created an issue in huggingface/datasets that received 5 comments

Oct 6

[save_to_disk] Using `select()` followed by `save_to_disk` saves complete dataset making it hard to create dummy dataset

Describe the bug When only keeping a dummy size of a dataset (say the first 100 samples), and then saving it to disk to upload it in the following …

5 comments

Opened 8 other issues in 5 repositories

microsoft/UniSpeech 2 open 1 closed

Fine-tuned Unispeech models cannot be downloaded Oct 22
Access Required for UniSpeech-SAT models Oct 11
Release vocab.json for Common Voice Oct 11

huggingface/datasets 1 closed 1 open

huggingface/transformers 1 open

[Trainer] Push to hub takes too much space for local .git folder Oct 26

bootphon/phonemizer 1 closed

Decoder - Convert phoneme sequence back to string Oct 21

pytorch/pytorch 1 closed

Backwards compatibility is broken between 1.9.1 and nightly for torch.cosine_similarity Oct 4

Patrick von Platen patrickvonplaten

Achievements

Achievements

Highlights

Organizations

Block or report patrickvonplaten

Popular repositories

1,639 contributions in the last year

Contribution activity

October 2021

Created a pull request in huggingface/datasets that received 9 comments

[Audio datasets] Adapting all audio datasets

Created an issue in huggingface/datasets that received 5 comments

[save_to_disk] Using select() followed by save_to_disk saves complete dataset making it hard to create dummy dataset

[save_to_disk] Using `select()` followed by `save_to_disk` saves complete dataset making it hard to create dummy dataset