File size: 1,452 Bytes
1eeb028 d1d0c7f 1eeb028 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 |
# QuReTec: query resolution model
QuReTeC is a query resolution model. It finds the relevant terms in a question history.
It is based on **bert-large-uncased** with a max sequence length of 300.
# Config details
Training and evaluation was done using the following BertConfig:
```json
BertConfig {
"_name_or_path": "uva-irlab/quretec",
"architectures": ["BertForMaskedLM"],
"attention_probs_dropout_prob": 0.1,
"finetuning_task": "ner",
"gradient_checkpointing": false,
"hidden_act": "gelu",
"hidden_dropout_prob": 0.4,
"hidden_size": 1024,
"id2label": {
"0": "[PAD]",
"1": "O",
"2": "REL",
"3": "[CLS]",
"4": "[SEP]"
},
"initializer_range": 0.02,
"intermediate_size": 4096,
"label2id": {
"O": 1,
"REL": 2,
"[CLS]": 3,
"[PAD]": 0,
"[SEP]": 4
},
"layer_norm_eps": 1e-12,
"max_position_embeddings": 512,
"model_type": "bert",
"num_attention_heads": 16,
"num_hidden_layers": 24,
"pad_token_id": 0,
"position_embedding_type": "absolute",
"transformers_version": "4.6.1",
"type_vocab_size": 2,
"use_cache": true,
"vocab_size": 30522
}
```
# Original authors
QuReTeC model from the published SIGIR 2020 paper: Query Resolution for Conversational Search with Limited Supervision by N. Voskarides, D. Li, P. Ren, E. Kanoulas and M. de Rijke. [[pdf]](https://arxiv.org/abs/2005.11723).
# Contributions
Uploaded by G. Scheuer ([website](https://giguruscheuer.com)) |