We follow a different training procedure: instead of using a two-phase approach, that pre-trains the model for 90% with 128
sequence length and 10% with 512 sequence length, we pre-train the model with 512 sequence length for 1M steps on a v3-32 TPU.
Stats
The current version of the model is trained on a filtered and sentence
segmented version of the Turkish
OSCAR corpus
,
a recent Wikipedia dump, various
OPUS corpora
and a
special corpus provided by
Kemal Oflazer
.
The final training corpus has a size of 35GB and 44,04,976,662 tokens.
Thanks to Google's TensorFlow Research Cloud (TFRC) we could train a cased model
on a TPU v3-32!
Usage
With Transformers >= 4.3 our cased ConvBERT model can be loaded like:
from transformers import AutoModel, AutoTokenizer
model_name = "dbmdz/convbert-base-turkish-cased"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModel.from_pretrained(model_name)
Results
For results on PoS tagging, NER and Question Answering downstream tasks, please refer to
this repository
.
For questions about our DBMDZ BERT models in general, just open an issue
here
🤗
Acknowledgments
Thanks to
Kemal Oflazer
for providing us
additional large corpora for Turkish. Many thanks to Reyyan Yeniterzi for providing
us the Turkish NER dataset for evaluation.
Research supported with Cloud TPUs from Google's TensorFlow Research Cloud (TFRC).
Thanks for providing access to the TFRC ❤️
Thanks to the generous support from the
Hugging Face
team,
it is possible to download both cased and uncased models from their S3 storage 🤗
Runs of dbmdz convbert-base-turkish-cased on huggingface.co
2.7K
Total runs
0
24-hour runs
56
3-day runs
1.0K
7-day runs
2.4K
30-day runs
More Information About convbert-base-turkish-cased huggingface.co Model
More convbert-base-turkish-cased license Visit here:
convbert-base-turkish-cased huggingface.co is an AI model on huggingface.co that provides convbert-base-turkish-cased's model effect (), which can be used instantly with this dbmdz convbert-base-turkish-cased model. huggingface.co supports a free trial of the convbert-base-turkish-cased model, and also provides paid use of the convbert-base-turkish-cased. Support call convbert-base-turkish-cased model through api, including Node.js, Python, http.
convbert-base-turkish-cased huggingface.co is an online trial and call api platform, which integrates convbert-base-turkish-cased's modeling effects, including api services, and provides a free online trial of convbert-base-turkish-cased, you can try convbert-base-turkish-cased online for free by clicking the link below.
dbmdz convbert-base-turkish-cased online free url in huggingface.co:
convbert-base-turkish-cased is an open source model from GitHub that offers a free installation service, and any user can find convbert-base-turkish-cased on GitHub to install. At the same time, huggingface.co provides the effect of convbert-base-turkish-cased install, users can directly use convbert-base-turkish-cased installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
convbert-base-turkish-cased install url in huggingface.co: