swedish / config.json
HarikaR's picture
End of training
535dbb6 verified
{
"_name_or_path": "KBLab/megatron-bert-large-swedish-cased-165-zero-shot",
"architectures": [
"MegatronBertForSequenceClassification"
],
"attention_probs_dropout_prob": 0.1,
"finetuning_task": "mnli",
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 1024,
"id2label": {
"0": "informational",
"1": "transactional",
"2": "navigational_informational",
"3": "navigational_transactional",
"4": "transactional_download"
},
"initializer_range": 0.02,
"intermediate_size": 4096,
"label2id": {
"informational": 0,
"navigational_informational": 2,
"navigational_transactional": 3,
"transactional": 1,
"transactional_download": 4
},
"layer_norm_eps": 1e-12,
"max_position_embeddings": 512,
"model_type": "megatron-bert",
"num_attention_heads": 16,
"num_hidden_layers": 24,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"problem_type": "single_label_classification",
"tokenizer_type": "BertWordPieceCase",
"torch_dtype": "float32",
"transformers_version": "4.35.2",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 64128
}