Instructions to use doc2query/msmarco-14langs-mt5-base-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use doc2query/msmarco-14langs-mt5-base-v1 with Transformers:
# Load model directly from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("doc2query/msmarco-14langs-mt5-base-v1") model = AutoModelForSeq2SeqLM.from_pretrained("doc2query/msmarco-14langs-mt5-base-v1") - Notebooks
- Google Colab
- Kaggle
Control language of generated sequence
#1
by iknoor - opened
Hi,
Can we control the language of generated queries in the mT5 model? For eg. if I have an input document in English language, can I generate the queries in Hindi in the zero-shot setting using mT5 (i.e. there is no train data for English-Hindi document-queries)? Basically, the aim would be to limit the vocabulary of the mt5 output sequence to a specific language, wdyt?
Thanks
Iknoor