Instructions to use stefan-it/umt5-small with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use stefan-it/umt5-small with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("stefan-it/umt5-small") model = AutoModelForSeq2SeqLM.from_pretrained("stefan-it/umt5-small", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Commit History
-
Adding `safetensors` variant of this model
This is an automated PR created with https://huggingface.co/spaces/safetensors/convert This new file is equivalent to `pytorch_model.bin` but safe in the sense that no arbitrary code can be put into it. These files also happen to load much faster than their pytorch counterpart: https://colab.research.google.com/github/huggingface/notebooks/blob/main/safetensors_doc/en/speed.ipynb The widgets on your model page will run using this model even if this is not merged making sure the file actually works. If you find any issues: please report here: https://huggingface.co/spaces/safetensors/convert/discussions Feel free to ignore this PR.5cf508b
-
tokenizer: fix number of extra ids
It's 300, as specified here: https://github.com/google-research/t5x/blob/main/t5x/examples/scalable_t5/umt5/vocab.gin#L5ad60d7b
-
readme: add some interesting details about umT5
0685c7d -
readme: mention conversion script
a67d2ab -
readme: add initial version
19d7d84 -
model: add latest PyTorch version
f8d09e1 -
tokenizer: bump number of extra ids
82caf5f -
tokenizer: add config
e166915 -
tokenizer: add special tokens map (imported from mT5)
347c39d -
model: fix copy paste bug in conversion script!!
https://gist.github.com/stefan-it/5d6a4ec89e7ad97181983881434cb4eb/revisions6f56448 -
readme: currently do not maintain PyTorch weights (but they are coming back!)
5e438c9 -
model: update latest version
Transformers version: https://github.com/huggingface/transformers/pull/22626 @ 9dfc612 Conversion script: https://gist.github.com/stefan-it/5d6a4ec89e7ad97181983881434cb4eb/revisionscf6ff25 -
config: adjust architecture
97679e8 -
config: prepare for upcoming umT5 integration
dad4406 -
model: revert transpose change
533e5ef -
model: transpose changes
3849401 -
config: adjust vocab size (slightly larger than mT5)
9e5b4bd -
model: add converted PyTorch model
c93fb7e -
vocab: add original spm model
Taken from: https://console.cloud.google.com/storage/browser/_details/t5-data/vocabs/umt5.256000/sentencepiece.model;tab=live_object488bf51 -
model: add converted model
Using: https://gist.github.com/stefan-it/5d6a4ec89e7ad97181983881434cb4ebbc9fe7c -
initial commit
cba65cb