Zarakun/ukrainian_news_classification

The model for Zarakun/ukrainian_news dataset classification task

This model is an instance of XLM-RoBERTa pretrained with the head with 15 classes

Loading the model

Loading script that loads the model from checkpoint

>>> config = XLMRobertaConfig.from_pretrained("FacebookAI/xlm-roberta-base")
>>> config.num_labels = 15
>>> model_pr = XLMRobertaForSequenceClassification(config)
>>> api = hub.HfApi()
>>> api.hf_hub_download(
...     repo_id="Zarakun/ukrainian_news_classification",
...     repo_type="model",
...     filename="checkpoints/{}".format(<checkpoint_name>),
...     local_dir="./"
... )
>>> model_pr.load_state_dict(torch.load("./checkpoints/{}".format(<checkpoint_name>), map_location="cpu"))