Commit
•
ffc6710
1
Parent(s):
7c2ec3a
Upload 6 files
Browse files- README.md +77 -0
- all_results.json +14 -0
- eval_results.json +10 -0
- train_results.json +7 -0
- trainer_state.json +229 -0
- training_args.bin +3 -0
README.md
ADDED
@@ -0,0 +1,77 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- audio-classification
|
4 |
+
- generated_from_trainer
|
5 |
+
metrics:
|
6 |
+
- accuracy
|
7 |
+
- precision
|
8 |
+
- f1
|
9 |
+
model-index:
|
10 |
+
- name: wavlm-base
|
11 |
+
results: []
|
12 |
+
---
|
13 |
+
|
14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
15 |
+
should probably proofread and complete it, then remove this comment. -->
|
16 |
+
|
17 |
+
# wavlm-base
|
18 |
+
|
19 |
+
This model is a fine-tuned version of [microsoft/wavlm-base](https://huggingface.co/microsoft/wavlm-base) on the galsenai/waxal_dataset dataset.
|
20 |
+
It achieves the following results on the evaluation set:
|
21 |
+
- Loss: 2.1345
|
22 |
+
- Accuracy: 0.6783
|
23 |
+
- Precision: 0.8774
|
24 |
+
- F1: 0.7615
|
25 |
+
|
26 |
+
## Model description
|
27 |
+
|
28 |
+
More information needed
|
29 |
+
|
30 |
+
## Intended uses & limitations
|
31 |
+
|
32 |
+
More information needed
|
33 |
+
|
34 |
+
## Training and evaluation data
|
35 |
+
|
36 |
+
More information needed
|
37 |
+
|
38 |
+
## Training procedure
|
39 |
+
|
40 |
+
### Training hyperparameters
|
41 |
+
|
42 |
+
The following hyperparameters were used during training:
|
43 |
+
- learning_rate: 3e-05
|
44 |
+
- train_batch_size: 30
|
45 |
+
- eval_batch_size: 30
|
46 |
+
- seed: 0
|
47 |
+
- gradient_accumulation_steps: 4
|
48 |
+
- total_train_batch_size: 120
|
49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
50 |
+
- lr_scheduler_type: linear
|
51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
52 |
+
- num_epochs: 32.0
|
53 |
+
|
54 |
+
### Training results
|
55 |
+
|
56 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | F1 |
|
57 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|
|
58 |
+
| 4.4506 | 2.53 | 500 | 4.8601 | 0.0224 | 0.0136 | 0.0066 |
|
59 |
+
| 3.0523 | 5.05 | 1000 | 4.6674 | 0.0720 | 0.0460 | 0.0394 |
|
60 |
+
| 1.949 | 7.58 | 1500 | 4.1533 | 0.1156 | 0.1847 | 0.1064 |
|
61 |
+
| 1.3427 | 10.1 | 2000 | 3.8173 | 0.1448 | 0.2382 | 0.1347 |
|
62 |
+
| 1.0064 | 12.63 | 2500 | 3.5546 | 0.2183 | 0.4464 | 0.2385 |
|
63 |
+
| 0.7985 | 15.15 | 3000 | 3.1172 | 0.3842 | 0.6336 | 0.4258 |
|
64 |
+
| 0.6505 | 17.68 | 3500 | 2.9231 | 0.5165 | 0.7677 | 0.5995 |
|
65 |
+
| 0.5367 | 20.2 | 4000 | 2.4935 | 0.5961 | 0.8182 | 0.6755 |
|
66 |
+
| 0.465 | 22.73 | 4500 | 2.2411 | 0.6412 | 0.8624 | 0.7272 |
|
67 |
+
| 0.4075 | 25.25 | 5000 | 2.1345 | 0.6783 | 0.8774 | 0.7615 |
|
68 |
+
| 0.3793 | 27.78 | 5500 | 2.2535 | 0.6681 | 0.8792 | 0.7543 |
|
69 |
+
| 0.3418 | 30.3 | 6000 | 2.3390 | 0.6662 | 0.8905 | 0.7576 |
|
70 |
+
|
71 |
+
|
72 |
+
### Framework versions
|
73 |
+
|
74 |
+
- Transformers 4.27.0.dev0
|
75 |
+
- Pytorch 1.11.0+cu113
|
76 |
+
- Datasets 2.9.1.dev0
|
77 |
+
- Tokenizers 0.13.2
|
all_results.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 32.0,
|
3 |
+
"eval_accuracy": 0.6782872300113679,
|
4 |
+
"eval_f1": 0.76152721933526,
|
5 |
+
"eval_loss": 2.134546995162964,
|
6 |
+
"eval_precision": 0.8773557553113618,
|
7 |
+
"eval_runtime": 70.2134,
|
8 |
+
"eval_samples_per_second": 37.585,
|
9 |
+
"eval_steps_per_second": 1.253,
|
10 |
+
"train_loss": 1.2320244793940072,
|
11 |
+
"train_runtime": 19932.9865,
|
12 |
+
"train_samples_per_second": 38.125,
|
13 |
+
"train_steps_per_second": 0.318
|
14 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 32.0,
|
3 |
+
"eval_accuracy": 0.6782872300113679,
|
4 |
+
"eval_f1": 0.76152721933526,
|
5 |
+
"eval_loss": 2.134546995162964,
|
6 |
+
"eval_precision": 0.8773557553113618,
|
7 |
+
"eval_runtime": 70.2134,
|
8 |
+
"eval_samples_per_second": 37.585,
|
9 |
+
"eval_steps_per_second": 1.253
|
10 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 32.0,
|
3 |
+
"train_loss": 1.2320244793940072,
|
4 |
+
"train_runtime": 19932.9865,
|
5 |
+
"train_samples_per_second": 38.125,
|
6 |
+
"train_steps_per_second": 0.318
|
7 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,229 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.6782872300113679,
|
3 |
+
"best_model_checkpoint": "wavlm-base/checkpoint-5000",
|
4 |
+
"epoch": 32.0,
|
5 |
+
"global_step": 6336,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 2.53,
|
12 |
+
"learning_rate": 2.3659305993690853e-05,
|
13 |
+
"loss": 4.4506,
|
14 |
+
"step": 500
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 2.53,
|
18 |
+
"eval_accuracy": 0.02235695339143615,
|
19 |
+
"eval_f1": 0.006616186351103852,
|
20 |
+
"eval_loss": 4.86008358001709,
|
21 |
+
"eval_precision": 0.013553337387865622,
|
22 |
+
"eval_runtime": 71.9324,
|
23 |
+
"eval_samples_per_second": 36.687,
|
24 |
+
"eval_steps_per_second": 1.223,
|
25 |
+
"step": 500
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"epoch": 5.05,
|
29 |
+
"learning_rate": 2.8074359873728517e-05,
|
30 |
+
"loss": 3.0523,
|
31 |
+
"step": 1000
|
32 |
+
},
|
33 |
+
{
|
34 |
+
"epoch": 5.05,
|
35 |
+
"eval_accuracy": 0.07199696854869268,
|
36 |
+
"eval_f1": 0.03943960465213343,
|
37 |
+
"eval_loss": 4.667360305786133,
|
38 |
+
"eval_precision": 0.0460329111931258,
|
39 |
+
"eval_runtime": 71.0687,
|
40 |
+
"eval_samples_per_second": 37.133,
|
41 |
+
"eval_steps_per_second": 1.238,
|
42 |
+
"step": 1000
|
43 |
+
},
|
44 |
+
{
|
45 |
+
"epoch": 7.58,
|
46 |
+
"learning_rate": 2.5443703963521573e-05,
|
47 |
+
"loss": 1.949,
|
48 |
+
"step": 1500
|
49 |
+
},
|
50 |
+
{
|
51 |
+
"epoch": 7.58,
|
52 |
+
"eval_accuracy": 0.11557408109132247,
|
53 |
+
"eval_f1": 0.10642928153043356,
|
54 |
+
"eval_loss": 4.153301239013672,
|
55 |
+
"eval_precision": 0.18473852089235457,
|
56 |
+
"eval_runtime": 71.0445,
|
57 |
+
"eval_samples_per_second": 37.146,
|
58 |
+
"eval_steps_per_second": 1.239,
|
59 |
+
"step": 1500
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 10.1,
|
63 |
+
"learning_rate": 2.2813048053314627e-05,
|
64 |
+
"loss": 1.3427,
|
65 |
+
"step": 2000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 10.1,
|
69 |
+
"eval_accuracy": 0.14475179992421372,
|
70 |
+
"eval_f1": 0.1346763066246049,
|
71 |
+
"eval_loss": 3.8172836303710938,
|
72 |
+
"eval_precision": 0.23822785193835883,
|
73 |
+
"eval_runtime": 74.8027,
|
74 |
+
"eval_samples_per_second": 35.279,
|
75 |
+
"eval_steps_per_second": 1.176,
|
76 |
+
"step": 2000
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"epoch": 12.63,
|
80 |
+
"learning_rate": 2.0182392143107683e-05,
|
81 |
+
"loss": 1.0064,
|
82 |
+
"step": 2500
|
83 |
+
},
|
84 |
+
{
|
85 |
+
"epoch": 12.63,
|
86 |
+
"eval_accuracy": 0.2182644941265631,
|
87 |
+
"eval_f1": 0.23851786199348127,
|
88 |
+
"eval_loss": 3.5546185970306396,
|
89 |
+
"eval_precision": 0.44639974594463316,
|
90 |
+
"eval_runtime": 71.1082,
|
91 |
+
"eval_samples_per_second": 37.112,
|
92 |
+
"eval_steps_per_second": 1.238,
|
93 |
+
"step": 2500
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 15.15,
|
97 |
+
"learning_rate": 1.7551736232900737e-05,
|
98 |
+
"loss": 0.7985,
|
99 |
+
"step": 3000
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 15.15,
|
103 |
+
"eval_accuracy": 0.3842364532019704,
|
104 |
+
"eval_f1": 0.4258289880546809,
|
105 |
+
"eval_loss": 3.117234230041504,
|
106 |
+
"eval_precision": 0.6336488091054493,
|
107 |
+
"eval_runtime": 70.3629,
|
108 |
+
"eval_samples_per_second": 37.506,
|
109 |
+
"eval_steps_per_second": 1.251,
|
110 |
+
"step": 3000
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 17.68,
|
114 |
+
"learning_rate": 1.4921080322693792e-05,
|
115 |
+
"loss": 0.6505,
|
116 |
+
"step": 3500
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 17.68,
|
120 |
+
"eval_accuracy": 0.5164835164835165,
|
121 |
+
"eval_f1": 0.5994643376487049,
|
122 |
+
"eval_loss": 2.9230926036834717,
|
123 |
+
"eval_precision": 0.7677367523432096,
|
124 |
+
"eval_runtime": 70.3236,
|
125 |
+
"eval_samples_per_second": 37.527,
|
126 |
+
"eval_steps_per_second": 1.251,
|
127 |
+
"step": 3500
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 20.2,
|
131 |
+
"learning_rate": 1.2290424412486847e-05,
|
132 |
+
"loss": 0.5367,
|
133 |
+
"step": 4000
|
134 |
+
},
|
135 |
+
{
|
136 |
+
"epoch": 20.2,
|
137 |
+
"eval_accuracy": 0.5960591133004927,
|
138 |
+
"eval_f1": 0.6755266051820856,
|
139 |
+
"eval_loss": 2.4934628009796143,
|
140 |
+
"eval_precision": 0.818165400699518,
|
141 |
+
"eval_runtime": 69.6955,
|
142 |
+
"eval_samples_per_second": 37.865,
|
143 |
+
"eval_steps_per_second": 1.263,
|
144 |
+
"step": 4000
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 22.73,
|
148 |
+
"learning_rate": 9.659768502279902e-06,
|
149 |
+
"loss": 0.465,
|
150 |
+
"step": 4500
|
151 |
+
},
|
152 |
+
{
|
153 |
+
"epoch": 22.73,
|
154 |
+
"eval_accuracy": 0.6411519514967791,
|
155 |
+
"eval_f1": 0.7272288000165434,
|
156 |
+
"eval_loss": 2.241117238998413,
|
157 |
+
"eval_precision": 0.86241474957474,
|
158 |
+
"eval_runtime": 70.29,
|
159 |
+
"eval_samples_per_second": 37.544,
|
160 |
+
"eval_steps_per_second": 1.252,
|
161 |
+
"step": 4500
|
162 |
+
},
|
163 |
+
{
|
164 |
+
"epoch": 25.25,
|
165 |
+
"learning_rate": 7.029112592072957e-06,
|
166 |
+
"loss": 0.4075,
|
167 |
+
"step": 5000
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 25.25,
|
171 |
+
"eval_accuracy": 0.6782872300113679,
|
172 |
+
"eval_f1": 0.76152721933526,
|
173 |
+
"eval_loss": 2.134546995162964,
|
174 |
+
"eval_precision": 0.8773557553113618,
|
175 |
+
"eval_runtime": 70.2432,
|
176 |
+
"eval_samples_per_second": 37.569,
|
177 |
+
"eval_steps_per_second": 1.253,
|
178 |
+
"step": 5000
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"epoch": 27.78,
|
182 |
+
"learning_rate": 4.398456681866012e-06,
|
183 |
+
"loss": 0.3793,
|
184 |
+
"step": 5500
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 27.78,
|
188 |
+
"eval_accuracy": 0.6680560818491853,
|
189 |
+
"eval_f1": 0.7542953137768733,
|
190 |
+
"eval_loss": 2.2534902095794678,
|
191 |
+
"eval_precision": 0.8791790950841867,
|
192 |
+
"eval_runtime": 70.1811,
|
193 |
+
"eval_samples_per_second": 37.603,
|
194 |
+
"eval_steps_per_second": 1.254,
|
195 |
+
"step": 5500
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 30.3,
|
199 |
+
"learning_rate": 1.767800771659067e-06,
|
200 |
+
"loss": 0.3418,
|
201 |
+
"step": 6000
|
202 |
+
},
|
203 |
+
{
|
204 |
+
"epoch": 30.3,
|
205 |
+
"eval_accuracy": 0.6661614247821145,
|
206 |
+
"eval_f1": 0.7575999245015367,
|
207 |
+
"eval_loss": 2.3390462398529053,
|
208 |
+
"eval_precision": 0.8904662594400684,
|
209 |
+
"eval_runtime": 70.3334,
|
210 |
+
"eval_samples_per_second": 37.521,
|
211 |
+
"eval_steps_per_second": 1.251,
|
212 |
+
"step": 6000
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 32.0,
|
216 |
+
"step": 6336,
|
217 |
+
"total_flos": 3.804690514724178e+19,
|
218 |
+
"train_loss": 1.2320244793940072,
|
219 |
+
"train_runtime": 19932.9865,
|
220 |
+
"train_samples_per_second": 38.125,
|
221 |
+
"train_steps_per_second": 0.318
|
222 |
+
}
|
223 |
+
],
|
224 |
+
"max_steps": 6336,
|
225 |
+
"num_train_epochs": 32,
|
226 |
+
"total_flos": 3.804690514724178e+19,
|
227 |
+
"trial_name": null,
|
228 |
+
"trial_params": null
|
229 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9d78303a058fc6cb98946f3d25ed93eee9a4848daff7fe0a1f307e959024839c
|
3 |
+
size 3503
|