Timpal0l
Rank:Average Model Cost: $0.0000
Number of Runs: 37,490
Models by this creator
mdeberta-v3-base-squad2
mdeberta-v3-base-squad2
The mdeberta-v3-base-squad2 model is a question-answering model that has been fine-tuned for 3 epochs on the SQuAD2.0 dataset. It is based on DeBERTa, which improves upon BERT and RoBERTa models using disentangled attention and enhanced mask decoder techniques. In the mDeBERTa V3 version, ELECTRA-Style pre-training with Gradient Disentangled Embedding Sharing is used to further improve efficiency and model performance on downstream tasks. The model has 12 layers, a hidden size of 768, and a vocabulary containing 250K tokens. It was trained using the 2.5T CC100 multilingual data.
$-/run
37.0K
Huggingface
xlm-roberta-base-faq-extractor
$-/run
368
Huggingface
bert_large_sv_nordic_pile_150
bert_large_sv_nordic_pile_150
Platform did not provide a description for this model.
$-/run
44
Huggingface
bert_large_sv_nordic_pile_320
bert_large_sv_nordic_pile_320
Platform did not provide a description for this model.
$-/run
31
Huggingface
gpt-sw3-356m
$-/run
5
Huggingface
test-distilgpt2-finetuned-common-voice
test-distilgpt2-finetuned-common-voice
Platform did not provide a description for this model.
$-/run
4
Huggingface
distilgpt2-finetuned-common-voice
distilgpt2-finetuned-common-voice
Platform did not provide a description for this model.
$-/run
0
Huggingface