add config
Browse files- config.json +95 -0
config.json
ADDED
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"dataloader": {
|
3 |
+
"kwargs": {
|
4 |
+
"batch_size": 1024
|
5 |
+
},
|
6 |
+
"type": "pytorch_dataloader"
|
7 |
+
},
|
8 |
+
"dataset": {
|
9 |
+
"kwargs": {
|
10 |
+
"feature_columns": [],
|
11 |
+
"label_columns": [
|
12 |
+
"label1"
|
13 |
+
],
|
14 |
+
"mask": {
|
15 |
+
"kwargs": {},
|
16 |
+
"type": "pad_mask"
|
17 |
+
}
|
18 |
+
},
|
19 |
+
"type": "TransformerDatasetWithDict"
|
20 |
+
},
|
21 |
+
"loss": {
|
22 |
+
"kwargs": {},
|
23 |
+
"type": "CorrLoss"
|
24 |
+
},
|
25 |
+
"max_epochs": 30,
|
26 |
+
"min_epochs": 10,
|
27 |
+
"model": {
|
28 |
+
"kwargs": {
|
29 |
+
"embedding_config": {
|
30 |
+
"kwargs": {
|
31 |
+
"dropout": 0.2,
|
32 |
+
"fixed": false,
|
33 |
+
"inlinear": true
|
34 |
+
},
|
35 |
+
"type": "Exp2dEmbedding"
|
36 |
+
},
|
37 |
+
"embedding_size": 256,
|
38 |
+
"encoder_config": {
|
39 |
+
"kwargs": {
|
40 |
+
"activation": "gelu",
|
41 |
+
"d_ff": 256,
|
42 |
+
"d_k": 32,
|
43 |
+
"d_v": 32,
|
44 |
+
"layer_num": 4,
|
45 |
+
"n_heads": 32,
|
46 |
+
"res_dropout": 0.1
|
47 |
+
},
|
48 |
+
"type": "TSTPlusEncoder"
|
49 |
+
},
|
50 |
+
"feature_size": 256,
|
51 |
+
"head_config": {
|
52 |
+
"kwargs": {
|
53 |
+
"dropout": 0.5,
|
54 |
+
"hidden_size": 256,
|
55 |
+
"num_layers": 1
|
56 |
+
},
|
57 |
+
"type": "FCHead"
|
58 |
+
},
|
59 |
+
"input_size": 0,
|
60 |
+
"pooling_config": {
|
61 |
+
"kwargs": {},
|
62 |
+
"type": "LastPooling"
|
63 |
+
},
|
64 |
+
"y_len": 1
|
65 |
+
},
|
66 |
+
"type": "TransformerModel"
|
67 |
+
},
|
68 |
+
"model_name": "TransformerTest",
|
69 |
+
"monitor": "val_loss",
|
70 |
+
"optimizer": {
|
71 |
+
"kwargs": {
|
72 |
+
"lr": 5e-05,
|
73 |
+
"weight_decay": 1e-06
|
74 |
+
},
|
75 |
+
"type": "Adam"
|
76 |
+
},
|
77 |
+
"patience": 10,
|
78 |
+
"pp": "test_0",
|
79 |
+
"scheduler": {
|
80 |
+
"kwargs": {
|
81 |
+
"gamma": 0.5,
|
82 |
+
"milestones": [
|
83 |
+
10,
|
84 |
+
15
|
85 |
+
]
|
86 |
+
},
|
87 |
+
"type": "MultiStepLR"
|
88 |
+
},
|
89 |
+
"seq_length": 10,
|
90 |
+
"test_dataset": "toy_dataset/test.csv",
|
91 |
+
"train_dataset": "toy_dataset/train.csv",
|
92 |
+
"transformers_version": "4.10.0.dev0",
|
93 |
+
"use_cache": true,
|
94 |
+
"valid_dataset": "toy_dataset/valid.csv"
|
95 |
+
}
|