khmer_word_tokenizer / config.json
phonsobon's picture
Upload Khmer BiLSTM-CRF tokenizer
44b751f verified
Raw
History Blame Contribute Delete
440 Bytes
{
"model_type": "bilstm-crf",
"language": "km",
"task": "word-segmentation",
"embed_dim": 128,
"hidden_dim": 256,
"num_layers": 2,
"dropout": 0.3,
"vocab_size": 364,
"num_tags": 4,
"tag2idx": {
"PAD": 0,
"B": 1,
"I": 2,
"O": 3
},
"datasets": [
"ye-kyaw-thu/khPOS (corpus-draft-ver-1.0)",
"Asian-Language-Treebank/ALT-Parallel-Corpus",
"phylypo/segmentation-crf-khmer (kh_data_1000)"
]
}