Upload 3 files
Browse files- config.json +110 -0
- pytorch_model.bin +3 -0
- results.txt +92 -0
config.json
ADDED
@@ -0,0 +1,110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "facebook/hubert-base-ls960",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"adapter_kernel_size": 3,
|
5 |
+
"adapter_stride": 2,
|
6 |
+
"add_adapter": false,
|
7 |
+
"apply_spec_augment": true,
|
8 |
+
"architectures": [
|
9 |
+
"Wav2Vec2SAP"
|
10 |
+
],
|
11 |
+
"attention_dropout": 0.1,
|
12 |
+
"bos_token_id": 1,
|
13 |
+
"classifier_proj_size": 256,
|
14 |
+
"codevector_dim": 256,
|
15 |
+
"contrastive_logits_temperature": 0.1,
|
16 |
+
"conv_bias": false,
|
17 |
+
"conv_dim": [
|
18 |
+
512,
|
19 |
+
512,
|
20 |
+
512,
|
21 |
+
512,
|
22 |
+
512,
|
23 |
+
512,
|
24 |
+
512
|
25 |
+
],
|
26 |
+
"conv_kernel": [
|
27 |
+
10,
|
28 |
+
3,
|
29 |
+
3,
|
30 |
+
3,
|
31 |
+
3,
|
32 |
+
2,
|
33 |
+
2
|
34 |
+
],
|
35 |
+
"conv_stride": [
|
36 |
+
5,
|
37 |
+
2,
|
38 |
+
2,
|
39 |
+
2,
|
40 |
+
2,
|
41 |
+
2,
|
42 |
+
2
|
43 |
+
],
|
44 |
+
"ctc_loss_reduction": "sum",
|
45 |
+
"ctc_zero_infinity": false,
|
46 |
+
"diversity_loss_weight": 0.1,
|
47 |
+
"do_stable_layer_norm": false,
|
48 |
+
"eos_token_id": 2,
|
49 |
+
"feat_extract_activation": "gelu",
|
50 |
+
"feat_extract_dropout": 0.0,
|
51 |
+
"feat_extract_norm": "group",
|
52 |
+
"feat_proj_dropout": 0.1,
|
53 |
+
"feat_quantizer_dropout": 0.0,
|
54 |
+
"final_dropout": 0.1,
|
55 |
+
"gradient_checkpointing": false,
|
56 |
+
"hidden_act": "gelu",
|
57 |
+
"hidden_dropout": 0.1,
|
58 |
+
"hidden_dropout_prob": 0.1,
|
59 |
+
"hidden_size": 768,
|
60 |
+
"initializer_range": 0.02,
|
61 |
+
"intermediate_size": 3072,
|
62 |
+
"layer_norm_eps": 1e-05,
|
63 |
+
"layerdrop": 0.1,
|
64 |
+
"mask_feature_length": 10,
|
65 |
+
"mask_feature_min_masks": 0,
|
66 |
+
"mask_feature_prob": 0.0,
|
67 |
+
"mask_time_length": 10,
|
68 |
+
"mask_time_min_masks": 2,
|
69 |
+
"mask_time_prob": 0.05,
|
70 |
+
"model_type": "wav2vec2",
|
71 |
+
"num_adapter_layers": 3,
|
72 |
+
"num_attention_heads": 12,
|
73 |
+
"num_codevector_groups": 2,
|
74 |
+
"num_codevectors_per_group": 320,
|
75 |
+
"num_conv_pos_embedding_groups": 16,
|
76 |
+
"num_conv_pos_embeddings": 128,
|
77 |
+
"num_feat_extract_layers": 7,
|
78 |
+
"num_hidden_layers": 12,
|
79 |
+
"num_negatives": 100,
|
80 |
+
"output_hidden_size": 768,
|
81 |
+
"pad_token_id": 0,
|
82 |
+
"proj_codevector_dim": 256,
|
83 |
+
"tdnn_dilation": [
|
84 |
+
1,
|
85 |
+
2,
|
86 |
+
3,
|
87 |
+
1,
|
88 |
+
1
|
89 |
+
],
|
90 |
+
"tdnn_dim": [
|
91 |
+
512,
|
92 |
+
512,
|
93 |
+
512,
|
94 |
+
512,
|
95 |
+
1500
|
96 |
+
],
|
97 |
+
"tdnn_kernel": [
|
98 |
+
5,
|
99 |
+
3,
|
100 |
+
3,
|
101 |
+
1,
|
102 |
+
1
|
103 |
+
],
|
104 |
+
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
105 |
+
"torch_dtype": "float32",
|
106 |
+
"transformers_version": "4.18.0",
|
107 |
+
"use_weighted_layer_sum": false,
|
108 |
+
"vocab_size": 32,
|
109 |
+
"xvector_output_dim": 512
|
110 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3a8a4919673d0ad55668a2c81e896366a60db835c4a629d09d36e7672247ab84
|
3 |
+
size 379922579
|
results.txt
ADDED
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
Iteration: 0 - Loss: 5.867379665374756
|
2 |
+
Pearsonr: 0.0028354616317359032; P: 0.9516919355087852
|
3 |
+
Spearmanr: 0.012299765484740878; P: 0.7927008078133582
|
4 |
+
|
5 |
+
Iteration: 200 - Loss: 3.7480859756469727
|
6 |
+
Pearsonr: 0.6822427720997155; P: 3.8114867974009106e-64
|
7 |
+
Spearmanr: 0.6987927552459419; P: 1.714898855652711e-68
|
8 |
+
|
9 |
+
Iteration: 400 - Loss: 2.0871903896331787
|
10 |
+
Pearsonr: 0.6943990472247945; P: 2.612534323701347e-67
|
11 |
+
Spearmanr: 0.698542095171836; P: 2.0058476660977182e-68
|
12 |
+
|
13 |
+
Iteration: 600 - Loss: 1.3873023986816406
|
14 |
+
Pearsonr: 0.7382703009781203; P: 3.539646706959335e-80
|
15 |
+
Spearmanr: 0.7317829917482052; P: 4.0871994881328883e-78
|
16 |
+
|
17 |
+
Iteration: 800 - Loss: 1.6505166292190552
|
18 |
+
Pearsonr: 0.7465803710876929; P: 6.538924589337557e-83
|
19 |
+
Spearmanr: 0.7414975687829316; P: 3.1616087401744213e-81
|
20 |
+
|
21 |
+
Iteration: 1000 - Loss: 1.5760451555252075
|
22 |
+
Pearsonr: 0.757935030756042; P: 7.998492014911909e-87
|
23 |
+
Spearmanr: 0.7598607705301921; P: 1.6517174271374772e-87
|
24 |
+
|
25 |
+
Iteration: 1200 - Loss: 1.286901831626892
|
26 |
+
Pearsonr: 0.7878125446249106; P: 3.060342803100379e-98
|
27 |
+
Spearmanr: 0.7823517522386733; P: 5.090941237732941e-96
|
28 |
+
|
29 |
+
Iteration: 1400 - Loss: 3.440028190612793
|
30 |
+
Pearsonr: 0.7594830946088903; P: 2.2532070980525375e-87
|
31 |
+
Spearmanr: 0.758575530094818; P: 4.740985098161106e-87
|
32 |
+
|
33 |
+
Iteration: 1600 - Loss: 0.8019772171974182
|
34 |
+
Pearsonr: 0.7768805717302603; P: 7.38705989563506e-94
|
35 |
+
Spearmanr: 0.7702167368907785; P: 2.628935078242846e-91
|
36 |
+
|
37 |
+
Iteration: 1800 - Loss: 1.4795581102371216
|
38 |
+
Pearsonr: 0.7696546871144265; P: 4.2757988951980333e-91
|
39 |
+
Spearmanr: 0.7647337765893464; P: 2.8518454357595734e-89
|
40 |
+
|
41 |
+
Iteration: 2000 - Loss: 0.7853589057922363
|
42 |
+
Pearsonr: 0.7958349190270064; P: 1.2618719971811567e-101
|
43 |
+
Spearmanr: 0.7880092395740745; P: 2.5383201459914505e-98
|
44 |
+
|
45 |
+
Iteration: 2200 - Loss: 0.7508143782615662
|
46 |
+
Pearsonr: 0.7832399027977113; P: 2.2385521795276424e-96
|
47 |
+
Spearmanr: 0.7752700843832545; P: 3.1124469433138995e-93
|
48 |
+
|
49 |
+
Iteration: 2400 - Loss: 0.6243414878845215
|
50 |
+
Pearsonr: 0.7965960366731291; P: 5.915798260829097e-102
|
51 |
+
Spearmanr: 0.7945271660875569; P: 4.6031071332605635e-101
|
52 |
+
|
53 |
+
Iteration: 2600 - Loss: 0.8454068303108215
|
54 |
+
Pearsonr: 0.7812743191623615; P: 1.3722313859908278e-95
|
55 |
+
Spearmanr: 0.7722127384529298; P: 4.6208024170937416e-92
|
56 |
+
|
57 |
+
Iteration: 2800 - Loss: 0.43072304129600525
|
58 |
+
Pearsonr: 0.7763269153914357; P: 1.2128208094472161e-93
|
59 |
+
Spearmanr: 0.763392420111192; P: 8.802552267016054e-89
|
60 |
+
|
61 |
+
Iteration: 3000 - Loss: 0.9466016292572021
|
62 |
+
Pearsonr: 0.7794136428091348; P: 7.505224540929025e-95
|
63 |
+
Spearmanr: 0.7732777439352554; P: 1.814177327074164e-92
|
64 |
+
|
65 |
+
Iteration: 3200 - Loss: 0.6584780216217041
|
66 |
+
Pearsonr: 0.7914478464882384; P: 9.342569140562221e-100
|
67 |
+
Spearmanr: 0.7821939455941844; P: 5.888762177139021e-96
|
68 |
+
|
69 |
+
Iteration: 3400 - Loss: 0.6548047065734863
|
70 |
+
Pearsonr: 0.8005467401053116; P: 1.1001062577181383e-103
|
71 |
+
Spearmanr: 0.8003169025058752; P: 1.3905231416379682e-103
|
72 |
+
|
73 |
+
Iteration: 3600 - Loss: 0.435733437538147
|
74 |
+
Pearsonr: 0.7806550155780387; P: 2.420157587523084e-95
|
75 |
+
Spearmanr: 0.768956619273207; P: 7.80798530040152e-91
|
76 |
+
|
77 |
+
Iteration: 3800 - Loss: 1.8730363845825195
|
78 |
+
Pearsonr: 0.7884560152829609; P: 1.658562138153901e-98
|
79 |
+
Spearmanr: 0.7875980025733722; P: 3.752018888637302e-98
|
80 |
+
|
81 |
+
Iteration: 4000 - Loss: 0.6273457407951355
|
82 |
+
Pearsonr: 0.7902370138145393; P: 3.009679742052806e-99
|
83 |
+
Spearmanr: 0.7902340678855697; P: 3.018229553013081e-99
|
84 |
+
|
85 |
+
Iteration: 4200 - Loss: 0.3261318504810333
|
86 |
+
Pearsonr: 0.7951340031012234; P: 2.5278939659288376e-101
|
87 |
+
Spearmanr: 0.7879911936087802; P: 2.582271764255765e-98
|
88 |
+
|
89 |
+
Iteration: 4400 - Loss: 0.9046502113342285
|
90 |
+
Pearsonr: 0.802561513418009; P: 1.392517582493545e-104
|
91 |
+
Spearmanr: 0.795848197069901; P: 1.245339358156881e-101
|
92 |
+
|