colerobertson
commited on
Commit
•
bac01b2
1
Parent(s):
02160f9
Training in progress, epoch 1
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- model.safetensors +1 -1
- run-4/checkpoint-144/config.json +80 -0
- run-4/checkpoint-144/model.safetensors +3 -0
- run-4/checkpoint-144/optimizer.pt +3 -0
- run-4/checkpoint-144/preprocessor_config.json +9 -0
- run-4/checkpoint-144/rng_state.pth +3 -0
- run-4/checkpoint-144/scheduler.pt +3 -0
- run-4/checkpoint-144/trainer_state.json +247 -0
- run-4/checkpoint-144/training_args.bin +3 -0
- run-4/checkpoint-192/config.json +80 -0
- run-4/checkpoint-192/model.safetensors +3 -0
- run-4/checkpoint-192/optimizer.pt +3 -0
- run-4/checkpoint-192/preprocessor_config.json +9 -0
- run-4/checkpoint-192/rng_state.pth +3 -0
- run-4/checkpoint-192/scheduler.pt +3 -0
- run-4/checkpoint-192/trainer_state.json +326 -0
- run-4/checkpoint-192/training_args.bin +3 -0
- run-4/checkpoint-240/config.json +80 -0
- run-4/checkpoint-240/model.safetensors +3 -0
- run-4/checkpoint-240/optimizer.pt +3 -0
- run-4/checkpoint-240/preprocessor_config.json +9 -0
- run-4/checkpoint-240/rng_state.pth +3 -0
- run-4/checkpoint-240/scheduler.pt +3 -0
- run-4/checkpoint-240/trainer_state.json +405 -0
- run-4/checkpoint-240/training_args.bin +3 -0
- run-4/checkpoint-288/config.json +80 -0
- run-4/checkpoint-288/model.safetensors +3 -0
- run-4/checkpoint-288/optimizer.pt +3 -0
- run-4/checkpoint-288/preprocessor_config.json +9 -0
- run-4/checkpoint-288/rng_state.pth +3 -0
- run-4/checkpoint-288/scheduler.pt +3 -0
- run-4/checkpoint-288/trainer_state.json +477 -0
- run-4/checkpoint-288/training_args.bin +3 -0
- run-4/checkpoint-336/config.json +80 -0
- run-4/checkpoint-336/model.safetensors +3 -0
- run-4/checkpoint-336/optimizer.pt +3 -0
- run-4/checkpoint-336/preprocessor_config.json +9 -0
- run-4/checkpoint-336/rng_state.pth +3 -0
- run-4/checkpoint-336/scheduler.pt +3 -0
- run-4/checkpoint-336/trainer_state.json +556 -0
- run-4/checkpoint-336/training_args.bin +3 -0
- run-4/checkpoint-384/config.json +80 -0
- run-4/checkpoint-384/model.safetensors +3 -0
- run-4/checkpoint-384/optimizer.pt +3 -0
- run-4/checkpoint-384/preprocessor_config.json +9 -0
- run-4/checkpoint-384/rng_state.pth +3 -0
- run-4/checkpoint-384/scheduler.pt +3 -0
- run-4/checkpoint-384/trainer_state.json +628 -0
- run-4/checkpoint-384/training_args.bin +3 -0
- run-4/checkpoint-432/config.json +80 -0
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 94763496
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e1ed6ffdb684cc537f47f54766448db627495bfcb076b58f5bd92f9f75833010
|
3 |
size 94763496
|
run-4/checkpoint-144/config.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "NOT_WORD",
|
53 |
+
"1": "WORD"
|
54 |
+
},
|
55 |
+
"initializer_range": 0.02,
|
56 |
+
"intermediate_size": 3072,
|
57 |
+
"label2id": {
|
58 |
+
"NOT_WORD": "0",
|
59 |
+
"WORD": "1"
|
60 |
+
},
|
61 |
+
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
+
"mask_feature_length": 10,
|
64 |
+
"mask_feature_min_masks": 0,
|
65 |
+
"mask_feature_prob": 0.0,
|
66 |
+
"mask_time_length": 10,
|
67 |
+
"mask_time_min_masks": 2,
|
68 |
+
"mask_time_prob": 0.05,
|
69 |
+
"model_type": "hubert",
|
70 |
+
"num_attention_heads": 12,
|
71 |
+
"num_conv_pos_embedding_groups": 16,
|
72 |
+
"num_conv_pos_embeddings": 128,
|
73 |
+
"num_feat_extract_layers": 7,
|
74 |
+
"num_hidden_layers": 2,
|
75 |
+
"pad_token_id": 0,
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.38.1",
|
78 |
+
"use_weighted_layer_sum": false,
|
79 |
+
"vocab_size": 32
|
80 |
+
}
|
run-4/checkpoint-144/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:84be90551bec8e32bb108dda4349a49a0a709d76c0e31826f05686036125618e
|
3 |
+
size 94763496
|
run-4/checkpoint-144/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8185b0c73404b8fffa8583fd719233443888c59c9660c9da10c6e75c9be8a36e
|
3 |
+
size 189552570
|
run-4/checkpoint-144/preprocessor_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000
|
9 |
+
}
|
run-4/checkpoint-144/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:87b8bdd7b4355fd23f0b8256efb0158e4240e11263e992a13d50944c37692500
|
3 |
+
size 14244
|
run-4/checkpoint-144/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ab5c28f22d5574a7076ae76732817bd92c67e71ad78a2880a292f813c3e7a8d1
|
3 |
+
size 1064
|
run-4/checkpoint-144/trainer_state.json
ADDED
@@ -0,0 +1,247 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.7326732673267327,
|
3 |
+
"best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-4/checkpoint-96",
|
4 |
+
"epoch": 3.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 144,
|
7 |
+
"is_hyper_param_search": true,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1,
|
13 |
+
"grad_norm": 1.2733114957809448,
|
14 |
+
"learning_rate": 2.2702186710865246e-07,
|
15 |
+
"loss": 0.7025,
|
16 |
+
"step": 5
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21,
|
20 |
+
"grad_norm": 1.243804931640625,
|
21 |
+
"learning_rate": 4.5404373421730493e-07,
|
22 |
+
"loss": 0.6974,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.31,
|
27 |
+
"grad_norm": 1.7711552381515503,
|
28 |
+
"learning_rate": 6.810656013259573e-07,
|
29 |
+
"loss": 0.696,
|
30 |
+
"step": 15
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.42,
|
34 |
+
"grad_norm": 1.1453403234481812,
|
35 |
+
"learning_rate": 9.080874684346099e-07,
|
36 |
+
"loss": 0.6989,
|
37 |
+
"step": 20
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.52,
|
41 |
+
"grad_norm": 1.2729355096817017,
|
42 |
+
"learning_rate": 1.1351093355432624e-06,
|
43 |
+
"loss": 0.6968,
|
44 |
+
"step": 25
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.62,
|
48 |
+
"grad_norm": 1.1592165231704712,
|
49 |
+
"learning_rate": 1.3621312026519146e-06,
|
50 |
+
"loss": 0.6959,
|
51 |
+
"step": 30
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.73,
|
55 |
+
"grad_norm": 1.1798148155212402,
|
56 |
+
"learning_rate": 1.589153069760567e-06,
|
57 |
+
"loss": 0.6952,
|
58 |
+
"step": 35
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.83,
|
62 |
+
"grad_norm": 2.1216671466827393,
|
63 |
+
"learning_rate": 1.8161749368692197e-06,
|
64 |
+
"loss": 0.6886,
|
65 |
+
"step": 40
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.94,
|
69 |
+
"grad_norm": 1.3416370153427124,
|
70 |
+
"learning_rate": 2.043196803977872e-06,
|
71 |
+
"loss": 0.6864,
|
72 |
+
"step": 45
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.0,
|
76 |
+
"eval_f1": 0.72,
|
77 |
+
"eval_loss": 0.688262939453125,
|
78 |
+
"eval_runtime": 1.3468,
|
79 |
+
"eval_samples_per_second": 47.521,
|
80 |
+
"eval_steps_per_second": 5.94,
|
81 |
+
"step": 48
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 1.04,
|
85 |
+
"grad_norm": 2.1856281757354736,
|
86 |
+
"learning_rate": 2.169320063482679e-06,
|
87 |
+
"loss": 0.6917,
|
88 |
+
"step": 50
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.15,
|
92 |
+
"grad_norm": 1.4077153205871582,
|
93 |
+
"learning_rate": 2.1440954115817176e-06,
|
94 |
+
"loss": 0.6884,
|
95 |
+
"step": 55
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 1.25,
|
99 |
+
"grad_norm": 2.1792664527893066,
|
100 |
+
"learning_rate": 2.1188707596807562e-06,
|
101 |
+
"loss": 0.6668,
|
102 |
+
"step": 60
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.35,
|
106 |
+
"grad_norm": 1.0386197566986084,
|
107 |
+
"learning_rate": 2.093646107779795e-06,
|
108 |
+
"loss": 0.6694,
|
109 |
+
"step": 65
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"epoch": 1.46,
|
113 |
+
"grad_norm": 2.0565919876098633,
|
114 |
+
"learning_rate": 2.0684214558788335e-06,
|
115 |
+
"loss": 0.6561,
|
116 |
+
"step": 70
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 1.56,
|
120 |
+
"grad_norm": 1.2978509664535522,
|
121 |
+
"learning_rate": 2.043196803977872e-06,
|
122 |
+
"loss": 0.6789,
|
123 |
+
"step": 75
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.67,
|
127 |
+
"grad_norm": 2.058328628540039,
|
128 |
+
"learning_rate": 2.0179721520769108e-06,
|
129 |
+
"loss": 0.6633,
|
130 |
+
"step": 80
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 1.77,
|
134 |
+
"grad_norm": 0.6023226976394653,
|
135 |
+
"learning_rate": 1.9927475001759494e-06,
|
136 |
+
"loss": 0.6655,
|
137 |
+
"step": 85
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.88,
|
141 |
+
"grad_norm": 0.5510762929916382,
|
142 |
+
"learning_rate": 1.967522848274988e-06,
|
143 |
+
"loss": 0.6622,
|
144 |
+
"step": 90
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.98,
|
148 |
+
"grad_norm": 1.098602533340454,
|
149 |
+
"learning_rate": 1.9422981963740267e-06,
|
150 |
+
"loss": 0.6633,
|
151 |
+
"step": 95
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 2.0,
|
155 |
+
"eval_f1": 0.7326732673267327,
|
156 |
+
"eval_loss": 0.6816024780273438,
|
157 |
+
"eval_runtime": 1.3765,
|
158 |
+
"eval_samples_per_second": 46.493,
|
159 |
+
"eval_steps_per_second": 5.812,
|
160 |
+
"step": 96
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.08,
|
164 |
+
"grad_norm": 0.9589098691940308,
|
165 |
+
"learning_rate": 1.9170735444730654e-06,
|
166 |
+
"loss": 0.659,
|
167 |
+
"step": 100
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 2.19,
|
171 |
+
"grad_norm": 1.070695161819458,
|
172 |
+
"learning_rate": 1.8918488925721038e-06,
|
173 |
+
"loss": 0.6313,
|
174 |
+
"step": 105
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 2.29,
|
178 |
+
"grad_norm": 0.9913639426231384,
|
179 |
+
"learning_rate": 1.8666242406711424e-06,
|
180 |
+
"loss": 0.6652,
|
181 |
+
"step": 110
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 2.4,
|
185 |
+
"grad_norm": 1.0632878541946411,
|
186 |
+
"learning_rate": 1.841399588770181e-06,
|
187 |
+
"loss": 0.673,
|
188 |
+
"step": 115
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 2.5,
|
192 |
+
"grad_norm": 2.1036579608917236,
|
193 |
+
"learning_rate": 1.8161749368692197e-06,
|
194 |
+
"loss": 0.6451,
|
195 |
+
"step": 120
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 2.6,
|
199 |
+
"grad_norm": 1.08384108543396,
|
200 |
+
"learning_rate": 1.7909502849682583e-06,
|
201 |
+
"loss": 0.6322,
|
202 |
+
"step": 125
|
203 |
+
},
|
204 |
+
{
|
205 |
+
"epoch": 2.71,
|
206 |
+
"grad_norm": 0.9407000541687012,
|
207 |
+
"learning_rate": 1.765725633067297e-06,
|
208 |
+
"loss": 0.6755,
|
209 |
+
"step": 130
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 2.81,
|
213 |
+
"grad_norm": 0.9016568660736084,
|
214 |
+
"learning_rate": 1.7405009811663356e-06,
|
215 |
+
"loss": 0.5985,
|
216 |
+
"step": 135
|
217 |
+
},
|
218 |
+
{
|
219 |
+
"epoch": 2.92,
|
220 |
+
"grad_norm": 1.1134448051452637,
|
221 |
+
"learning_rate": 1.7152763292653743e-06,
|
222 |
+
"loss": 0.603,
|
223 |
+
"step": 140
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 3.0,
|
227 |
+
"eval_f1": 0.7326732673267327,
|
228 |
+
"eval_loss": 0.6800689697265625,
|
229 |
+
"eval_runtime": 1.3861,
|
230 |
+
"eval_samples_per_second": 46.173,
|
231 |
+
"eval_steps_per_second": 5.772,
|
232 |
+
"step": 144
|
233 |
+
}
|
234 |
+
],
|
235 |
+
"logging_steps": 5,
|
236 |
+
"max_steps": 480,
|
237 |
+
"num_input_tokens_seen": 0,
|
238 |
+
"num_train_epochs": 10,
|
239 |
+
"save_steps": 500,
|
240 |
+
"total_flos": 2121874430755872.0,
|
241 |
+
"train_batch_size": 4,
|
242 |
+
"trial_name": null,
|
243 |
+
"trial_params": {
|
244 |
+
"learning_rate": 2.1794099242430636e-06,
|
245 |
+
"per_device_train_batch_size": 4
|
246 |
+
}
|
247 |
+
}
|
run-4/checkpoint-144/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6eadd8415a707c8e4de7440233358b89448cac114a2ada36aef1d9186553c0c3
|
3 |
+
size 4920
|
run-4/checkpoint-192/config.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "NOT_WORD",
|
53 |
+
"1": "WORD"
|
54 |
+
},
|
55 |
+
"initializer_range": 0.02,
|
56 |
+
"intermediate_size": 3072,
|
57 |
+
"label2id": {
|
58 |
+
"NOT_WORD": "0",
|
59 |
+
"WORD": "1"
|
60 |
+
},
|
61 |
+
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
+
"mask_feature_length": 10,
|
64 |
+
"mask_feature_min_masks": 0,
|
65 |
+
"mask_feature_prob": 0.0,
|
66 |
+
"mask_time_length": 10,
|
67 |
+
"mask_time_min_masks": 2,
|
68 |
+
"mask_time_prob": 0.05,
|
69 |
+
"model_type": "hubert",
|
70 |
+
"num_attention_heads": 12,
|
71 |
+
"num_conv_pos_embedding_groups": 16,
|
72 |
+
"num_conv_pos_embeddings": 128,
|
73 |
+
"num_feat_extract_layers": 7,
|
74 |
+
"num_hidden_layers": 2,
|
75 |
+
"pad_token_id": 0,
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.38.1",
|
78 |
+
"use_weighted_layer_sum": false,
|
79 |
+
"vocab_size": 32
|
80 |
+
}
|
run-4/checkpoint-192/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:04a662be668d43e34b3ede26defff9905cfe78fa0158e7f3bca9b7849a684f21
|
3 |
+
size 94763496
|
run-4/checkpoint-192/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:989d42785f68e982f9cc1de749a83d539a71c120933d2aae451e2785a66e335b
|
3 |
+
size 189552570
|
run-4/checkpoint-192/preprocessor_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000
|
9 |
+
}
|
run-4/checkpoint-192/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d5aef020ca2df517540ac9ff4e195e1c41a7b85939e93195d118078f119bc949
|
3 |
+
size 14244
|
run-4/checkpoint-192/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2463c4896d69602722ffa95084240326443a5f0698cb239a50fde88c55bcd421
|
3 |
+
size 1064
|
run-4/checkpoint-192/trainer_state.json
ADDED
@@ -0,0 +1,326 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.7326732673267327,
|
3 |
+
"best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-4/checkpoint-96",
|
4 |
+
"epoch": 4.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 192,
|
7 |
+
"is_hyper_param_search": true,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1,
|
13 |
+
"grad_norm": 1.2733114957809448,
|
14 |
+
"learning_rate": 2.2702186710865246e-07,
|
15 |
+
"loss": 0.7025,
|
16 |
+
"step": 5
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21,
|
20 |
+
"grad_norm": 1.243804931640625,
|
21 |
+
"learning_rate": 4.5404373421730493e-07,
|
22 |
+
"loss": 0.6974,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.31,
|
27 |
+
"grad_norm": 1.7711552381515503,
|
28 |
+
"learning_rate": 6.810656013259573e-07,
|
29 |
+
"loss": 0.696,
|
30 |
+
"step": 15
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.42,
|
34 |
+
"grad_norm": 1.1453403234481812,
|
35 |
+
"learning_rate": 9.080874684346099e-07,
|
36 |
+
"loss": 0.6989,
|
37 |
+
"step": 20
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.52,
|
41 |
+
"grad_norm": 1.2729355096817017,
|
42 |
+
"learning_rate": 1.1351093355432624e-06,
|
43 |
+
"loss": 0.6968,
|
44 |
+
"step": 25
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.62,
|
48 |
+
"grad_norm": 1.1592165231704712,
|
49 |
+
"learning_rate": 1.3621312026519146e-06,
|
50 |
+
"loss": 0.6959,
|
51 |
+
"step": 30
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.73,
|
55 |
+
"grad_norm": 1.1798148155212402,
|
56 |
+
"learning_rate": 1.589153069760567e-06,
|
57 |
+
"loss": 0.6952,
|
58 |
+
"step": 35
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.83,
|
62 |
+
"grad_norm": 2.1216671466827393,
|
63 |
+
"learning_rate": 1.8161749368692197e-06,
|
64 |
+
"loss": 0.6886,
|
65 |
+
"step": 40
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.94,
|
69 |
+
"grad_norm": 1.3416370153427124,
|
70 |
+
"learning_rate": 2.043196803977872e-06,
|
71 |
+
"loss": 0.6864,
|
72 |
+
"step": 45
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.0,
|
76 |
+
"eval_f1": 0.72,
|
77 |
+
"eval_loss": 0.688262939453125,
|
78 |
+
"eval_runtime": 1.3468,
|
79 |
+
"eval_samples_per_second": 47.521,
|
80 |
+
"eval_steps_per_second": 5.94,
|
81 |
+
"step": 48
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 1.04,
|
85 |
+
"grad_norm": 2.1856281757354736,
|
86 |
+
"learning_rate": 2.169320063482679e-06,
|
87 |
+
"loss": 0.6917,
|
88 |
+
"step": 50
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.15,
|
92 |
+
"grad_norm": 1.4077153205871582,
|
93 |
+
"learning_rate": 2.1440954115817176e-06,
|
94 |
+
"loss": 0.6884,
|
95 |
+
"step": 55
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 1.25,
|
99 |
+
"grad_norm": 2.1792664527893066,
|
100 |
+
"learning_rate": 2.1188707596807562e-06,
|
101 |
+
"loss": 0.6668,
|
102 |
+
"step": 60
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.35,
|
106 |
+
"grad_norm": 1.0386197566986084,
|
107 |
+
"learning_rate": 2.093646107779795e-06,
|
108 |
+
"loss": 0.6694,
|
109 |
+
"step": 65
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"epoch": 1.46,
|
113 |
+
"grad_norm": 2.0565919876098633,
|
114 |
+
"learning_rate": 2.0684214558788335e-06,
|
115 |
+
"loss": 0.6561,
|
116 |
+
"step": 70
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 1.56,
|
120 |
+
"grad_norm": 1.2978509664535522,
|
121 |
+
"learning_rate": 2.043196803977872e-06,
|
122 |
+
"loss": 0.6789,
|
123 |
+
"step": 75
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.67,
|
127 |
+
"grad_norm": 2.058328628540039,
|
128 |
+
"learning_rate": 2.0179721520769108e-06,
|
129 |
+
"loss": 0.6633,
|
130 |
+
"step": 80
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 1.77,
|
134 |
+
"grad_norm": 0.6023226976394653,
|
135 |
+
"learning_rate": 1.9927475001759494e-06,
|
136 |
+
"loss": 0.6655,
|
137 |
+
"step": 85
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.88,
|
141 |
+
"grad_norm": 0.5510762929916382,
|
142 |
+
"learning_rate": 1.967522848274988e-06,
|
143 |
+
"loss": 0.6622,
|
144 |
+
"step": 90
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.98,
|
148 |
+
"grad_norm": 1.098602533340454,
|
149 |
+
"learning_rate": 1.9422981963740267e-06,
|
150 |
+
"loss": 0.6633,
|
151 |
+
"step": 95
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 2.0,
|
155 |
+
"eval_f1": 0.7326732673267327,
|
156 |
+
"eval_loss": 0.6816024780273438,
|
157 |
+
"eval_runtime": 1.3765,
|
158 |
+
"eval_samples_per_second": 46.493,
|
159 |
+
"eval_steps_per_second": 5.812,
|
160 |
+
"step": 96
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.08,
|
164 |
+
"grad_norm": 0.9589098691940308,
|
165 |
+
"learning_rate": 1.9170735444730654e-06,
|
166 |
+
"loss": 0.659,
|
167 |
+
"step": 100
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 2.19,
|
171 |
+
"grad_norm": 1.070695161819458,
|
172 |
+
"learning_rate": 1.8918488925721038e-06,
|
173 |
+
"loss": 0.6313,
|
174 |
+
"step": 105
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 2.29,
|
178 |
+
"grad_norm": 0.9913639426231384,
|
179 |
+
"learning_rate": 1.8666242406711424e-06,
|
180 |
+
"loss": 0.6652,
|
181 |
+
"step": 110
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 2.4,
|
185 |
+
"grad_norm": 1.0632878541946411,
|
186 |
+
"learning_rate": 1.841399588770181e-06,
|
187 |
+
"loss": 0.673,
|
188 |
+
"step": 115
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 2.5,
|
192 |
+
"grad_norm": 2.1036579608917236,
|
193 |
+
"learning_rate": 1.8161749368692197e-06,
|
194 |
+
"loss": 0.6451,
|
195 |
+
"step": 120
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 2.6,
|
199 |
+
"grad_norm": 1.08384108543396,
|
200 |
+
"learning_rate": 1.7909502849682583e-06,
|
201 |
+
"loss": 0.6322,
|
202 |
+
"step": 125
|
203 |
+
},
|
204 |
+
{
|
205 |
+
"epoch": 2.71,
|
206 |
+
"grad_norm": 0.9407000541687012,
|
207 |
+
"learning_rate": 1.765725633067297e-06,
|
208 |
+
"loss": 0.6755,
|
209 |
+
"step": 130
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 2.81,
|
213 |
+
"grad_norm": 0.9016568660736084,
|
214 |
+
"learning_rate": 1.7405009811663356e-06,
|
215 |
+
"loss": 0.5985,
|
216 |
+
"step": 135
|
217 |
+
},
|
218 |
+
{
|
219 |
+
"epoch": 2.92,
|
220 |
+
"grad_norm": 1.1134448051452637,
|
221 |
+
"learning_rate": 1.7152763292653743e-06,
|
222 |
+
"loss": 0.603,
|
223 |
+
"step": 140
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 3.0,
|
227 |
+
"eval_f1": 0.7326732673267327,
|
228 |
+
"eval_loss": 0.6800689697265625,
|
229 |
+
"eval_runtime": 1.3861,
|
230 |
+
"eval_samples_per_second": 46.173,
|
231 |
+
"eval_steps_per_second": 5.772,
|
232 |
+
"step": 144
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 3.02,
|
236 |
+
"grad_norm": 0.7627719640731812,
|
237 |
+
"learning_rate": 1.6900516773644127e-06,
|
238 |
+
"loss": 0.6557,
|
239 |
+
"step": 145
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 3.12,
|
243 |
+
"grad_norm": 0.9291415214538574,
|
244 |
+
"learning_rate": 1.6648270254634511e-06,
|
245 |
+
"loss": 0.6219,
|
246 |
+
"step": 150
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 3.23,
|
250 |
+
"grad_norm": 0.9248765707015991,
|
251 |
+
"learning_rate": 1.6396023735624898e-06,
|
252 |
+
"loss": 0.6325,
|
253 |
+
"step": 155
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 3.33,
|
257 |
+
"grad_norm": 0.9842573404312134,
|
258 |
+
"learning_rate": 1.6143777216615284e-06,
|
259 |
+
"loss": 0.6521,
|
260 |
+
"step": 160
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 3.44,
|
264 |
+
"grad_norm": 0.8689214587211609,
|
265 |
+
"learning_rate": 1.589153069760567e-06,
|
266 |
+
"loss": 0.5929,
|
267 |
+
"step": 165
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 3.54,
|
271 |
+
"grad_norm": 1.0012000799179077,
|
272 |
+
"learning_rate": 1.5639284178596057e-06,
|
273 |
+
"loss": 0.584,
|
274 |
+
"step": 170
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 3.65,
|
278 |
+
"grad_norm": 0.7438368797302246,
|
279 |
+
"learning_rate": 1.5387037659586443e-06,
|
280 |
+
"loss": 0.6813,
|
281 |
+
"step": 175
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 3.75,
|
285 |
+
"grad_norm": 1.8603870868682861,
|
286 |
+
"learning_rate": 1.513479114057683e-06,
|
287 |
+
"loss": 0.6099,
|
288 |
+
"step": 180
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 3.85,
|
292 |
+
"grad_norm": 0.9918416738510132,
|
293 |
+
"learning_rate": 1.4882544621567216e-06,
|
294 |
+
"loss": 0.6192,
|
295 |
+
"step": 185
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 3.96,
|
299 |
+
"grad_norm": 1.9146322011947632,
|
300 |
+
"learning_rate": 1.4630298102557603e-06,
|
301 |
+
"loss": 0.6472,
|
302 |
+
"step": 190
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 4.0,
|
306 |
+
"eval_f1": 0.7326732673267327,
|
307 |
+
"eval_loss": 0.6818161010742188,
|
308 |
+
"eval_runtime": 1.3841,
|
309 |
+
"eval_samples_per_second": 46.239,
|
310 |
+
"eval_steps_per_second": 5.78,
|
311 |
+
"step": 192
|
312 |
+
}
|
313 |
+
],
|
314 |
+
"logging_steps": 5,
|
315 |
+
"max_steps": 480,
|
316 |
+
"num_input_tokens_seen": 0,
|
317 |
+
"num_train_epochs": 10,
|
318 |
+
"save_steps": 500,
|
319 |
+
"total_flos": 2891755054954176.0,
|
320 |
+
"train_batch_size": 4,
|
321 |
+
"trial_name": null,
|
322 |
+
"trial_params": {
|
323 |
+
"learning_rate": 2.1794099242430636e-06,
|
324 |
+
"per_device_train_batch_size": 4
|
325 |
+
}
|
326 |
+
}
|
run-4/checkpoint-192/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6eadd8415a707c8e4de7440233358b89448cac114a2ada36aef1d9186553c0c3
|
3 |
+
size 4920
|
run-4/checkpoint-240/config.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "NOT_WORD",
|
53 |
+
"1": "WORD"
|
54 |
+
},
|
55 |
+
"initializer_range": 0.02,
|
56 |
+
"intermediate_size": 3072,
|
57 |
+
"label2id": {
|
58 |
+
"NOT_WORD": "0",
|
59 |
+
"WORD": "1"
|
60 |
+
},
|
61 |
+
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
+
"mask_feature_length": 10,
|
64 |
+
"mask_feature_min_masks": 0,
|
65 |
+
"mask_feature_prob": 0.0,
|
66 |
+
"mask_time_length": 10,
|
67 |
+
"mask_time_min_masks": 2,
|
68 |
+
"mask_time_prob": 0.05,
|
69 |
+
"model_type": "hubert",
|
70 |
+
"num_attention_heads": 12,
|
71 |
+
"num_conv_pos_embedding_groups": 16,
|
72 |
+
"num_conv_pos_embeddings": 128,
|
73 |
+
"num_feat_extract_layers": 7,
|
74 |
+
"num_hidden_layers": 2,
|
75 |
+
"pad_token_id": 0,
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.38.1",
|
78 |
+
"use_weighted_layer_sum": false,
|
79 |
+
"vocab_size": 32
|
80 |
+
}
|
run-4/checkpoint-240/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:336f5153e14e700e4c4e5e048754a4434d79bdee3267bc3a53cbb003b875ea7e
|
3 |
+
size 94763496
|
run-4/checkpoint-240/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:22b1916abc98dec4fd4f1436e9d8dd462ba5c4401dddd0941a36031f3bc679ff
|
3 |
+
size 189552570
|
run-4/checkpoint-240/preprocessor_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000
|
9 |
+
}
|
run-4/checkpoint-240/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f8eb071c49709b6f4047e7f48105f0dd51daaf73e0a11fd742255aa4c3526f42
|
3 |
+
size 14244
|
run-4/checkpoint-240/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e2e35d3649cf053f4cc488647dacb8d6e20774271115c5998276bc7752ca7e23
|
3 |
+
size 1064
|
run-4/checkpoint-240/trainer_state.json
ADDED
@@ -0,0 +1,405 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.7326732673267327,
|
3 |
+
"best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-4/checkpoint-96",
|
4 |
+
"epoch": 5.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 240,
|
7 |
+
"is_hyper_param_search": true,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1,
|
13 |
+
"grad_norm": 1.2733114957809448,
|
14 |
+
"learning_rate": 2.2702186710865246e-07,
|
15 |
+
"loss": 0.7025,
|
16 |
+
"step": 5
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21,
|
20 |
+
"grad_norm": 1.243804931640625,
|
21 |
+
"learning_rate": 4.5404373421730493e-07,
|
22 |
+
"loss": 0.6974,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.31,
|
27 |
+
"grad_norm": 1.7711552381515503,
|
28 |
+
"learning_rate": 6.810656013259573e-07,
|
29 |
+
"loss": 0.696,
|
30 |
+
"step": 15
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.42,
|
34 |
+
"grad_norm": 1.1453403234481812,
|
35 |
+
"learning_rate": 9.080874684346099e-07,
|
36 |
+
"loss": 0.6989,
|
37 |
+
"step": 20
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.52,
|
41 |
+
"grad_norm": 1.2729355096817017,
|
42 |
+
"learning_rate": 1.1351093355432624e-06,
|
43 |
+
"loss": 0.6968,
|
44 |
+
"step": 25
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.62,
|
48 |
+
"grad_norm": 1.1592165231704712,
|
49 |
+
"learning_rate": 1.3621312026519146e-06,
|
50 |
+
"loss": 0.6959,
|
51 |
+
"step": 30
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.73,
|
55 |
+
"grad_norm": 1.1798148155212402,
|
56 |
+
"learning_rate": 1.589153069760567e-06,
|
57 |
+
"loss": 0.6952,
|
58 |
+
"step": 35
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.83,
|
62 |
+
"grad_norm": 2.1216671466827393,
|
63 |
+
"learning_rate": 1.8161749368692197e-06,
|
64 |
+
"loss": 0.6886,
|
65 |
+
"step": 40
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.94,
|
69 |
+
"grad_norm": 1.3416370153427124,
|
70 |
+
"learning_rate": 2.043196803977872e-06,
|
71 |
+
"loss": 0.6864,
|
72 |
+
"step": 45
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.0,
|
76 |
+
"eval_f1": 0.72,
|
77 |
+
"eval_loss": 0.688262939453125,
|
78 |
+
"eval_runtime": 1.3468,
|
79 |
+
"eval_samples_per_second": 47.521,
|
80 |
+
"eval_steps_per_second": 5.94,
|
81 |
+
"step": 48
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 1.04,
|
85 |
+
"grad_norm": 2.1856281757354736,
|
86 |
+
"learning_rate": 2.169320063482679e-06,
|
87 |
+
"loss": 0.6917,
|
88 |
+
"step": 50
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.15,
|
92 |
+
"grad_norm": 1.4077153205871582,
|
93 |
+
"learning_rate": 2.1440954115817176e-06,
|
94 |
+
"loss": 0.6884,
|
95 |
+
"step": 55
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 1.25,
|
99 |
+
"grad_norm": 2.1792664527893066,
|
100 |
+
"learning_rate": 2.1188707596807562e-06,
|
101 |
+
"loss": 0.6668,
|
102 |
+
"step": 60
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.35,
|
106 |
+
"grad_norm": 1.0386197566986084,
|
107 |
+
"learning_rate": 2.093646107779795e-06,
|
108 |
+
"loss": 0.6694,
|
109 |
+
"step": 65
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"epoch": 1.46,
|
113 |
+
"grad_norm": 2.0565919876098633,
|
114 |
+
"learning_rate": 2.0684214558788335e-06,
|
115 |
+
"loss": 0.6561,
|
116 |
+
"step": 70
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 1.56,
|
120 |
+
"grad_norm": 1.2978509664535522,
|
121 |
+
"learning_rate": 2.043196803977872e-06,
|
122 |
+
"loss": 0.6789,
|
123 |
+
"step": 75
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.67,
|
127 |
+
"grad_norm": 2.058328628540039,
|
128 |
+
"learning_rate": 2.0179721520769108e-06,
|
129 |
+
"loss": 0.6633,
|
130 |
+
"step": 80
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 1.77,
|
134 |
+
"grad_norm": 0.6023226976394653,
|
135 |
+
"learning_rate": 1.9927475001759494e-06,
|
136 |
+
"loss": 0.6655,
|
137 |
+
"step": 85
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.88,
|
141 |
+
"grad_norm": 0.5510762929916382,
|
142 |
+
"learning_rate": 1.967522848274988e-06,
|
143 |
+
"loss": 0.6622,
|
144 |
+
"step": 90
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.98,
|
148 |
+
"grad_norm": 1.098602533340454,
|
149 |
+
"learning_rate": 1.9422981963740267e-06,
|
150 |
+
"loss": 0.6633,
|
151 |
+
"step": 95
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 2.0,
|
155 |
+
"eval_f1": 0.7326732673267327,
|
156 |
+
"eval_loss": 0.6816024780273438,
|
157 |
+
"eval_runtime": 1.3765,
|
158 |
+
"eval_samples_per_second": 46.493,
|
159 |
+
"eval_steps_per_second": 5.812,
|
160 |
+
"step": 96
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.08,
|
164 |
+
"grad_norm": 0.9589098691940308,
|
165 |
+
"learning_rate": 1.9170735444730654e-06,
|
166 |
+
"loss": 0.659,
|
167 |
+
"step": 100
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 2.19,
|
171 |
+
"grad_norm": 1.070695161819458,
|
172 |
+
"learning_rate": 1.8918488925721038e-06,
|
173 |
+
"loss": 0.6313,
|
174 |
+
"step": 105
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 2.29,
|
178 |
+
"grad_norm": 0.9913639426231384,
|
179 |
+
"learning_rate": 1.8666242406711424e-06,
|
180 |
+
"loss": 0.6652,
|
181 |
+
"step": 110
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 2.4,
|
185 |
+
"grad_norm": 1.0632878541946411,
|
186 |
+
"learning_rate": 1.841399588770181e-06,
|
187 |
+
"loss": 0.673,
|
188 |
+
"step": 115
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 2.5,
|
192 |
+
"grad_norm": 2.1036579608917236,
|
193 |
+
"learning_rate": 1.8161749368692197e-06,
|
194 |
+
"loss": 0.6451,
|
195 |
+
"step": 120
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 2.6,
|
199 |
+
"grad_norm": 1.08384108543396,
|
200 |
+
"learning_rate": 1.7909502849682583e-06,
|
201 |
+
"loss": 0.6322,
|
202 |
+
"step": 125
|
203 |
+
},
|
204 |
+
{
|
205 |
+
"epoch": 2.71,
|
206 |
+
"grad_norm": 0.9407000541687012,
|
207 |
+
"learning_rate": 1.765725633067297e-06,
|
208 |
+
"loss": 0.6755,
|
209 |
+
"step": 130
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 2.81,
|
213 |
+
"grad_norm": 0.9016568660736084,
|
214 |
+
"learning_rate": 1.7405009811663356e-06,
|
215 |
+
"loss": 0.5985,
|
216 |
+
"step": 135
|
217 |
+
},
|
218 |
+
{
|
219 |
+
"epoch": 2.92,
|
220 |
+
"grad_norm": 1.1134448051452637,
|
221 |
+
"learning_rate": 1.7152763292653743e-06,
|
222 |
+
"loss": 0.603,
|
223 |
+
"step": 140
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 3.0,
|
227 |
+
"eval_f1": 0.7326732673267327,
|
228 |
+
"eval_loss": 0.6800689697265625,
|
229 |
+
"eval_runtime": 1.3861,
|
230 |
+
"eval_samples_per_second": 46.173,
|
231 |
+
"eval_steps_per_second": 5.772,
|
232 |
+
"step": 144
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 3.02,
|
236 |
+
"grad_norm": 0.7627719640731812,
|
237 |
+
"learning_rate": 1.6900516773644127e-06,
|
238 |
+
"loss": 0.6557,
|
239 |
+
"step": 145
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 3.12,
|
243 |
+
"grad_norm": 0.9291415214538574,
|
244 |
+
"learning_rate": 1.6648270254634511e-06,
|
245 |
+
"loss": 0.6219,
|
246 |
+
"step": 150
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 3.23,
|
250 |
+
"grad_norm": 0.9248765707015991,
|
251 |
+
"learning_rate": 1.6396023735624898e-06,
|
252 |
+
"loss": 0.6325,
|
253 |
+
"step": 155
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 3.33,
|
257 |
+
"grad_norm": 0.9842573404312134,
|
258 |
+
"learning_rate": 1.6143777216615284e-06,
|
259 |
+
"loss": 0.6521,
|
260 |
+
"step": 160
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 3.44,
|
264 |
+
"grad_norm": 0.8689214587211609,
|
265 |
+
"learning_rate": 1.589153069760567e-06,
|
266 |
+
"loss": 0.5929,
|
267 |
+
"step": 165
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 3.54,
|
271 |
+
"grad_norm": 1.0012000799179077,
|
272 |
+
"learning_rate": 1.5639284178596057e-06,
|
273 |
+
"loss": 0.584,
|
274 |
+
"step": 170
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 3.65,
|
278 |
+
"grad_norm": 0.7438368797302246,
|
279 |
+
"learning_rate": 1.5387037659586443e-06,
|
280 |
+
"loss": 0.6813,
|
281 |
+
"step": 175
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 3.75,
|
285 |
+
"grad_norm": 1.8603870868682861,
|
286 |
+
"learning_rate": 1.513479114057683e-06,
|
287 |
+
"loss": 0.6099,
|
288 |
+
"step": 180
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 3.85,
|
292 |
+
"grad_norm": 0.9918416738510132,
|
293 |
+
"learning_rate": 1.4882544621567216e-06,
|
294 |
+
"loss": 0.6192,
|
295 |
+
"step": 185
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 3.96,
|
299 |
+
"grad_norm": 1.9146322011947632,
|
300 |
+
"learning_rate": 1.4630298102557603e-06,
|
301 |
+
"loss": 0.6472,
|
302 |
+
"step": 190
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 4.0,
|
306 |
+
"eval_f1": 0.7326732673267327,
|
307 |
+
"eval_loss": 0.6818161010742188,
|
308 |
+
"eval_runtime": 1.3841,
|
309 |
+
"eval_samples_per_second": 46.239,
|
310 |
+
"eval_steps_per_second": 5.78,
|
311 |
+
"step": 192
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 4.06,
|
315 |
+
"grad_norm": 0.9502781629562378,
|
316 |
+
"learning_rate": 1.437805158354799e-06,
|
317 |
+
"loss": 0.6447,
|
318 |
+
"step": 195
|
319 |
+
},
|
320 |
+
{
|
321 |
+
"epoch": 4.17,
|
322 |
+
"grad_norm": 0.8570067286491394,
|
323 |
+
"learning_rate": 1.4125805064538375e-06,
|
324 |
+
"loss": 0.5306,
|
325 |
+
"step": 200
|
326 |
+
},
|
327 |
+
{
|
328 |
+
"epoch": 4.27,
|
329 |
+
"grad_norm": 0.8097484111785889,
|
330 |
+
"learning_rate": 1.3873558545528762e-06,
|
331 |
+
"loss": 0.6202,
|
332 |
+
"step": 205
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 4.38,
|
336 |
+
"grad_norm": 2.0106472969055176,
|
337 |
+
"learning_rate": 1.3621312026519146e-06,
|
338 |
+
"loss": 0.6705,
|
339 |
+
"step": 210
|
340 |
+
},
|
341 |
+
{
|
342 |
+
"epoch": 4.48,
|
343 |
+
"grad_norm": 1.090775489807129,
|
344 |
+
"learning_rate": 1.3369065507509533e-06,
|
345 |
+
"loss": 0.6297,
|
346 |
+
"step": 215
|
347 |
+
},
|
348 |
+
{
|
349 |
+
"epoch": 4.58,
|
350 |
+
"grad_norm": 0.8988145589828491,
|
351 |
+
"learning_rate": 1.311681898849992e-06,
|
352 |
+
"loss": 0.5896,
|
353 |
+
"step": 220
|
354 |
+
},
|
355 |
+
{
|
356 |
+
"epoch": 4.69,
|
357 |
+
"grad_norm": 0.9149978756904602,
|
358 |
+
"learning_rate": 1.2864572469490305e-06,
|
359 |
+
"loss": 0.6156,
|
360 |
+
"step": 225
|
361 |
+
},
|
362 |
+
{
|
363 |
+
"epoch": 4.79,
|
364 |
+
"grad_norm": 1.9398412704467773,
|
365 |
+
"learning_rate": 1.2612325950480692e-06,
|
366 |
+
"loss": 0.6305,
|
367 |
+
"step": 230
|
368 |
+
},
|
369 |
+
{
|
370 |
+
"epoch": 4.9,
|
371 |
+
"grad_norm": 0.9217966794967651,
|
372 |
+
"learning_rate": 1.2360079431471078e-06,
|
373 |
+
"loss": 0.5943,
|
374 |
+
"step": 235
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 5.0,
|
378 |
+
"grad_norm": 0.9083653688430786,
|
379 |
+
"learning_rate": 1.2107832912461465e-06,
|
380 |
+
"loss": 0.6386,
|
381 |
+
"step": 240
|
382 |
+
},
|
383 |
+
{
|
384 |
+
"epoch": 5.0,
|
385 |
+
"eval_f1": 0.7326732673267327,
|
386 |
+
"eval_loss": 0.6846389770507812,
|
387 |
+
"eval_runtime": 1.4094,
|
388 |
+
"eval_samples_per_second": 45.409,
|
389 |
+
"eval_steps_per_second": 5.676,
|
390 |
+
"step": 240
|
391 |
+
}
|
392 |
+
],
|
393 |
+
"logging_steps": 5,
|
394 |
+
"max_steps": 480,
|
395 |
+
"num_input_tokens_seen": 0,
|
396 |
+
"num_train_epochs": 10,
|
397 |
+
"save_steps": 500,
|
398 |
+
"total_flos": 3654362860415712.0,
|
399 |
+
"train_batch_size": 4,
|
400 |
+
"trial_name": null,
|
401 |
+
"trial_params": {
|
402 |
+
"learning_rate": 2.1794099242430636e-06,
|
403 |
+
"per_device_train_batch_size": 4
|
404 |
+
}
|
405 |
+
}
|
run-4/checkpoint-240/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6eadd8415a707c8e4de7440233358b89448cac114a2ada36aef1d9186553c0c3
|
3 |
+
size 4920
|
run-4/checkpoint-288/config.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "NOT_WORD",
|
53 |
+
"1": "WORD"
|
54 |
+
},
|
55 |
+
"initializer_range": 0.02,
|
56 |
+
"intermediate_size": 3072,
|
57 |
+
"label2id": {
|
58 |
+
"NOT_WORD": "0",
|
59 |
+
"WORD": "1"
|
60 |
+
},
|
61 |
+
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
+
"mask_feature_length": 10,
|
64 |
+
"mask_feature_min_masks": 0,
|
65 |
+
"mask_feature_prob": 0.0,
|
66 |
+
"mask_time_length": 10,
|
67 |
+
"mask_time_min_masks": 2,
|
68 |
+
"mask_time_prob": 0.05,
|
69 |
+
"model_type": "hubert",
|
70 |
+
"num_attention_heads": 12,
|
71 |
+
"num_conv_pos_embedding_groups": 16,
|
72 |
+
"num_conv_pos_embeddings": 128,
|
73 |
+
"num_feat_extract_layers": 7,
|
74 |
+
"num_hidden_layers": 2,
|
75 |
+
"pad_token_id": 0,
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.38.1",
|
78 |
+
"use_weighted_layer_sum": false,
|
79 |
+
"vocab_size": 32
|
80 |
+
}
|
run-4/checkpoint-288/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:be27592b57346e1ba7fb5ac856d5e685d4d7d316009088a8a4a6b512cb22d9a4
|
3 |
+
size 94763496
|
run-4/checkpoint-288/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7c627abefd26f649e146eb3f7adc646b23cae944e9576351a11327cb5a485e7d
|
3 |
+
size 189552570
|
run-4/checkpoint-288/preprocessor_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000
|
9 |
+
}
|
run-4/checkpoint-288/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9903236b654011babeaee26ea70e1c6278fa670549b900c6df1d64732428a642
|
3 |
+
size 14244
|
run-4/checkpoint-288/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2163f47f6ce948a3a821c2032fa9ddfd17ee0992318acd25664f5287b23dd105
|
3 |
+
size 1064
|
run-4/checkpoint-288/trainer_state.json
ADDED
@@ -0,0 +1,477 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.7326732673267327,
|
3 |
+
"best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-4/checkpoint-96",
|
4 |
+
"epoch": 6.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 288,
|
7 |
+
"is_hyper_param_search": true,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1,
|
13 |
+
"grad_norm": 1.2733114957809448,
|
14 |
+
"learning_rate": 2.2702186710865246e-07,
|
15 |
+
"loss": 0.7025,
|
16 |
+
"step": 5
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21,
|
20 |
+
"grad_norm": 1.243804931640625,
|
21 |
+
"learning_rate": 4.5404373421730493e-07,
|
22 |
+
"loss": 0.6974,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.31,
|
27 |
+
"grad_norm": 1.7711552381515503,
|
28 |
+
"learning_rate": 6.810656013259573e-07,
|
29 |
+
"loss": 0.696,
|
30 |
+
"step": 15
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.42,
|
34 |
+
"grad_norm": 1.1453403234481812,
|
35 |
+
"learning_rate": 9.080874684346099e-07,
|
36 |
+
"loss": 0.6989,
|
37 |
+
"step": 20
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.52,
|
41 |
+
"grad_norm": 1.2729355096817017,
|
42 |
+
"learning_rate": 1.1351093355432624e-06,
|
43 |
+
"loss": 0.6968,
|
44 |
+
"step": 25
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.62,
|
48 |
+
"grad_norm": 1.1592165231704712,
|
49 |
+
"learning_rate": 1.3621312026519146e-06,
|
50 |
+
"loss": 0.6959,
|
51 |
+
"step": 30
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.73,
|
55 |
+
"grad_norm": 1.1798148155212402,
|
56 |
+
"learning_rate": 1.589153069760567e-06,
|
57 |
+
"loss": 0.6952,
|
58 |
+
"step": 35
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.83,
|
62 |
+
"grad_norm": 2.1216671466827393,
|
63 |
+
"learning_rate": 1.8161749368692197e-06,
|
64 |
+
"loss": 0.6886,
|
65 |
+
"step": 40
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.94,
|
69 |
+
"grad_norm": 1.3416370153427124,
|
70 |
+
"learning_rate": 2.043196803977872e-06,
|
71 |
+
"loss": 0.6864,
|
72 |
+
"step": 45
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.0,
|
76 |
+
"eval_f1": 0.72,
|
77 |
+
"eval_loss": 0.688262939453125,
|
78 |
+
"eval_runtime": 1.3468,
|
79 |
+
"eval_samples_per_second": 47.521,
|
80 |
+
"eval_steps_per_second": 5.94,
|
81 |
+
"step": 48
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 1.04,
|
85 |
+
"grad_norm": 2.1856281757354736,
|
86 |
+
"learning_rate": 2.169320063482679e-06,
|
87 |
+
"loss": 0.6917,
|
88 |
+
"step": 50
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.15,
|
92 |
+
"grad_norm": 1.4077153205871582,
|
93 |
+
"learning_rate": 2.1440954115817176e-06,
|
94 |
+
"loss": 0.6884,
|
95 |
+
"step": 55
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 1.25,
|
99 |
+
"grad_norm": 2.1792664527893066,
|
100 |
+
"learning_rate": 2.1188707596807562e-06,
|
101 |
+
"loss": 0.6668,
|
102 |
+
"step": 60
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.35,
|
106 |
+
"grad_norm": 1.0386197566986084,
|
107 |
+
"learning_rate": 2.093646107779795e-06,
|
108 |
+
"loss": 0.6694,
|
109 |
+
"step": 65
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"epoch": 1.46,
|
113 |
+
"grad_norm": 2.0565919876098633,
|
114 |
+
"learning_rate": 2.0684214558788335e-06,
|
115 |
+
"loss": 0.6561,
|
116 |
+
"step": 70
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 1.56,
|
120 |
+
"grad_norm": 1.2978509664535522,
|
121 |
+
"learning_rate": 2.043196803977872e-06,
|
122 |
+
"loss": 0.6789,
|
123 |
+
"step": 75
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.67,
|
127 |
+
"grad_norm": 2.058328628540039,
|
128 |
+
"learning_rate": 2.0179721520769108e-06,
|
129 |
+
"loss": 0.6633,
|
130 |
+
"step": 80
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 1.77,
|
134 |
+
"grad_norm": 0.6023226976394653,
|
135 |
+
"learning_rate": 1.9927475001759494e-06,
|
136 |
+
"loss": 0.6655,
|
137 |
+
"step": 85
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.88,
|
141 |
+
"grad_norm": 0.5510762929916382,
|
142 |
+
"learning_rate": 1.967522848274988e-06,
|
143 |
+
"loss": 0.6622,
|
144 |
+
"step": 90
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.98,
|
148 |
+
"grad_norm": 1.098602533340454,
|
149 |
+
"learning_rate": 1.9422981963740267e-06,
|
150 |
+
"loss": 0.6633,
|
151 |
+
"step": 95
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 2.0,
|
155 |
+
"eval_f1": 0.7326732673267327,
|
156 |
+
"eval_loss": 0.6816024780273438,
|
157 |
+
"eval_runtime": 1.3765,
|
158 |
+
"eval_samples_per_second": 46.493,
|
159 |
+
"eval_steps_per_second": 5.812,
|
160 |
+
"step": 96
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.08,
|
164 |
+
"grad_norm": 0.9589098691940308,
|
165 |
+
"learning_rate": 1.9170735444730654e-06,
|
166 |
+
"loss": 0.659,
|
167 |
+
"step": 100
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 2.19,
|
171 |
+
"grad_norm": 1.070695161819458,
|
172 |
+
"learning_rate": 1.8918488925721038e-06,
|
173 |
+
"loss": 0.6313,
|
174 |
+
"step": 105
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 2.29,
|
178 |
+
"grad_norm": 0.9913639426231384,
|
179 |
+
"learning_rate": 1.8666242406711424e-06,
|
180 |
+
"loss": 0.6652,
|
181 |
+
"step": 110
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 2.4,
|
185 |
+
"grad_norm": 1.0632878541946411,
|
186 |
+
"learning_rate": 1.841399588770181e-06,
|
187 |
+
"loss": 0.673,
|
188 |
+
"step": 115
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 2.5,
|
192 |
+
"grad_norm": 2.1036579608917236,
|
193 |
+
"learning_rate": 1.8161749368692197e-06,
|
194 |
+
"loss": 0.6451,
|
195 |
+
"step": 120
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 2.6,
|
199 |
+
"grad_norm": 1.08384108543396,
|
200 |
+
"learning_rate": 1.7909502849682583e-06,
|
201 |
+
"loss": 0.6322,
|
202 |
+
"step": 125
|
203 |
+
},
|
204 |
+
{
|
205 |
+
"epoch": 2.71,
|
206 |
+
"grad_norm": 0.9407000541687012,
|
207 |
+
"learning_rate": 1.765725633067297e-06,
|
208 |
+
"loss": 0.6755,
|
209 |
+
"step": 130
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 2.81,
|
213 |
+
"grad_norm": 0.9016568660736084,
|
214 |
+
"learning_rate": 1.7405009811663356e-06,
|
215 |
+
"loss": 0.5985,
|
216 |
+
"step": 135
|
217 |
+
},
|
218 |
+
{
|
219 |
+
"epoch": 2.92,
|
220 |
+
"grad_norm": 1.1134448051452637,
|
221 |
+
"learning_rate": 1.7152763292653743e-06,
|
222 |
+
"loss": 0.603,
|
223 |
+
"step": 140
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 3.0,
|
227 |
+
"eval_f1": 0.7326732673267327,
|
228 |
+
"eval_loss": 0.6800689697265625,
|
229 |
+
"eval_runtime": 1.3861,
|
230 |
+
"eval_samples_per_second": 46.173,
|
231 |
+
"eval_steps_per_second": 5.772,
|
232 |
+
"step": 144
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 3.02,
|
236 |
+
"grad_norm": 0.7627719640731812,
|
237 |
+
"learning_rate": 1.6900516773644127e-06,
|
238 |
+
"loss": 0.6557,
|
239 |
+
"step": 145
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 3.12,
|
243 |
+
"grad_norm": 0.9291415214538574,
|
244 |
+
"learning_rate": 1.6648270254634511e-06,
|
245 |
+
"loss": 0.6219,
|
246 |
+
"step": 150
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 3.23,
|
250 |
+
"grad_norm": 0.9248765707015991,
|
251 |
+
"learning_rate": 1.6396023735624898e-06,
|
252 |
+
"loss": 0.6325,
|
253 |
+
"step": 155
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 3.33,
|
257 |
+
"grad_norm": 0.9842573404312134,
|
258 |
+
"learning_rate": 1.6143777216615284e-06,
|
259 |
+
"loss": 0.6521,
|
260 |
+
"step": 160
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 3.44,
|
264 |
+
"grad_norm": 0.8689214587211609,
|
265 |
+
"learning_rate": 1.589153069760567e-06,
|
266 |
+
"loss": 0.5929,
|
267 |
+
"step": 165
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 3.54,
|
271 |
+
"grad_norm": 1.0012000799179077,
|
272 |
+
"learning_rate": 1.5639284178596057e-06,
|
273 |
+
"loss": 0.584,
|
274 |
+
"step": 170
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 3.65,
|
278 |
+
"grad_norm": 0.7438368797302246,
|
279 |
+
"learning_rate": 1.5387037659586443e-06,
|
280 |
+
"loss": 0.6813,
|
281 |
+
"step": 175
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 3.75,
|
285 |
+
"grad_norm": 1.8603870868682861,
|
286 |
+
"learning_rate": 1.513479114057683e-06,
|
287 |
+
"loss": 0.6099,
|
288 |
+
"step": 180
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 3.85,
|
292 |
+
"grad_norm": 0.9918416738510132,
|
293 |
+
"learning_rate": 1.4882544621567216e-06,
|
294 |
+
"loss": 0.6192,
|
295 |
+
"step": 185
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 3.96,
|
299 |
+
"grad_norm": 1.9146322011947632,
|
300 |
+
"learning_rate": 1.4630298102557603e-06,
|
301 |
+
"loss": 0.6472,
|
302 |
+
"step": 190
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 4.0,
|
306 |
+
"eval_f1": 0.7326732673267327,
|
307 |
+
"eval_loss": 0.6818161010742188,
|
308 |
+
"eval_runtime": 1.3841,
|
309 |
+
"eval_samples_per_second": 46.239,
|
310 |
+
"eval_steps_per_second": 5.78,
|
311 |
+
"step": 192
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 4.06,
|
315 |
+
"grad_norm": 0.9502781629562378,
|
316 |
+
"learning_rate": 1.437805158354799e-06,
|
317 |
+
"loss": 0.6447,
|
318 |
+
"step": 195
|
319 |
+
},
|
320 |
+
{
|
321 |
+
"epoch": 4.17,
|
322 |
+
"grad_norm": 0.8570067286491394,
|
323 |
+
"learning_rate": 1.4125805064538375e-06,
|
324 |
+
"loss": 0.5306,
|
325 |
+
"step": 200
|
326 |
+
},
|
327 |
+
{
|
328 |
+
"epoch": 4.27,
|
329 |
+
"grad_norm": 0.8097484111785889,
|
330 |
+
"learning_rate": 1.3873558545528762e-06,
|
331 |
+
"loss": 0.6202,
|
332 |
+
"step": 205
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 4.38,
|
336 |
+
"grad_norm": 2.0106472969055176,
|
337 |
+
"learning_rate": 1.3621312026519146e-06,
|
338 |
+
"loss": 0.6705,
|
339 |
+
"step": 210
|
340 |
+
},
|
341 |
+
{
|
342 |
+
"epoch": 4.48,
|
343 |
+
"grad_norm": 1.090775489807129,
|
344 |
+
"learning_rate": 1.3369065507509533e-06,
|
345 |
+
"loss": 0.6297,
|
346 |
+
"step": 215
|
347 |
+
},
|
348 |
+
{
|
349 |
+
"epoch": 4.58,
|
350 |
+
"grad_norm": 0.8988145589828491,
|
351 |
+
"learning_rate": 1.311681898849992e-06,
|
352 |
+
"loss": 0.5896,
|
353 |
+
"step": 220
|
354 |
+
},
|
355 |
+
{
|
356 |
+
"epoch": 4.69,
|
357 |
+
"grad_norm": 0.9149978756904602,
|
358 |
+
"learning_rate": 1.2864572469490305e-06,
|
359 |
+
"loss": 0.6156,
|
360 |
+
"step": 225
|
361 |
+
},
|
362 |
+
{
|
363 |
+
"epoch": 4.79,
|
364 |
+
"grad_norm": 1.9398412704467773,
|
365 |
+
"learning_rate": 1.2612325950480692e-06,
|
366 |
+
"loss": 0.6305,
|
367 |
+
"step": 230
|
368 |
+
},
|
369 |
+
{
|
370 |
+
"epoch": 4.9,
|
371 |
+
"grad_norm": 0.9217966794967651,
|
372 |
+
"learning_rate": 1.2360079431471078e-06,
|
373 |
+
"loss": 0.5943,
|
374 |
+
"step": 235
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 5.0,
|
378 |
+
"grad_norm": 0.9083653688430786,
|
379 |
+
"learning_rate": 1.2107832912461465e-06,
|
380 |
+
"loss": 0.6386,
|
381 |
+
"step": 240
|
382 |
+
},
|
383 |
+
{
|
384 |
+
"epoch": 5.0,
|
385 |
+
"eval_f1": 0.7326732673267327,
|
386 |
+
"eval_loss": 0.6846389770507812,
|
387 |
+
"eval_runtime": 1.4094,
|
388 |
+
"eval_samples_per_second": 45.409,
|
389 |
+
"eval_steps_per_second": 5.676,
|
390 |
+
"step": 240
|
391 |
+
},
|
392 |
+
{
|
393 |
+
"epoch": 5.1,
|
394 |
+
"grad_norm": 0.9323675036430359,
|
395 |
+
"learning_rate": 1.1855586393451851e-06,
|
396 |
+
"loss": 0.5779,
|
397 |
+
"step": 245
|
398 |
+
},
|
399 |
+
{
|
400 |
+
"epoch": 5.21,
|
401 |
+
"grad_norm": 0.7549787163734436,
|
402 |
+
"learning_rate": 1.1603339874442238e-06,
|
403 |
+
"loss": 0.5948,
|
404 |
+
"step": 250
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 5.31,
|
408 |
+
"grad_norm": 0.8535837531089783,
|
409 |
+
"learning_rate": 1.1351093355432624e-06,
|
410 |
+
"loss": 0.6928,
|
411 |
+
"step": 255
|
412 |
+
},
|
413 |
+
{
|
414 |
+
"epoch": 5.42,
|
415 |
+
"grad_norm": 1.2038137912750244,
|
416 |
+
"learning_rate": 1.109884683642301e-06,
|
417 |
+
"loss": 0.5887,
|
418 |
+
"step": 260
|
419 |
+
},
|
420 |
+
{
|
421 |
+
"epoch": 5.52,
|
422 |
+
"grad_norm": 0.9501279592514038,
|
423 |
+
"learning_rate": 1.0846600317413395e-06,
|
424 |
+
"loss": 0.5776,
|
425 |
+
"step": 265
|
426 |
+
},
|
427 |
+
{
|
428 |
+
"epoch": 5.62,
|
429 |
+
"grad_norm": 0.7421719431877136,
|
430 |
+
"learning_rate": 1.0594353798403781e-06,
|
431 |
+
"loss": 0.6734,
|
432 |
+
"step": 270
|
433 |
+
},
|
434 |
+
{
|
435 |
+
"epoch": 5.73,
|
436 |
+
"grad_norm": 0.8555863499641418,
|
437 |
+
"learning_rate": 1.0342107279394168e-06,
|
438 |
+
"loss": 0.6399,
|
439 |
+
"step": 275
|
440 |
+
},
|
441 |
+
{
|
442 |
+
"epoch": 5.83,
|
443 |
+
"grad_norm": 0.8841156363487244,
|
444 |
+
"learning_rate": 1.0089860760384554e-06,
|
445 |
+
"loss": 0.6173,
|
446 |
+
"step": 280
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 5.94,
|
450 |
+
"grad_norm": 0.8565478324890137,
|
451 |
+
"learning_rate": 9.83761424137494e-07,
|
452 |
+
"loss": 0.5537,
|
453 |
+
"step": 285
|
454 |
+
},
|
455 |
+
{
|
456 |
+
"epoch": 6.0,
|
457 |
+
"eval_f1": 0.7326732673267327,
|
458 |
+
"eval_loss": 0.6864242553710938,
|
459 |
+
"eval_runtime": 1.3622,
|
460 |
+
"eval_samples_per_second": 46.983,
|
461 |
+
"eval_steps_per_second": 5.873,
|
462 |
+
"step": 288
|
463 |
+
}
|
464 |
+
],
|
465 |
+
"logging_steps": 5,
|
466 |
+
"max_steps": 480,
|
467 |
+
"num_input_tokens_seen": 0,
|
468 |
+
"num_train_epochs": 10,
|
469 |
+
"save_steps": 500,
|
470 |
+
"total_flos": 4356539523260784.0,
|
471 |
+
"train_batch_size": 4,
|
472 |
+
"trial_name": null,
|
473 |
+
"trial_params": {
|
474 |
+
"learning_rate": 2.1794099242430636e-06,
|
475 |
+
"per_device_train_batch_size": 4
|
476 |
+
}
|
477 |
+
}
|
run-4/checkpoint-288/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6eadd8415a707c8e4de7440233358b89448cac114a2ada36aef1d9186553c0c3
|
3 |
+
size 4920
|
run-4/checkpoint-336/config.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "NOT_WORD",
|
53 |
+
"1": "WORD"
|
54 |
+
},
|
55 |
+
"initializer_range": 0.02,
|
56 |
+
"intermediate_size": 3072,
|
57 |
+
"label2id": {
|
58 |
+
"NOT_WORD": "0",
|
59 |
+
"WORD": "1"
|
60 |
+
},
|
61 |
+
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
+
"mask_feature_length": 10,
|
64 |
+
"mask_feature_min_masks": 0,
|
65 |
+
"mask_feature_prob": 0.0,
|
66 |
+
"mask_time_length": 10,
|
67 |
+
"mask_time_min_masks": 2,
|
68 |
+
"mask_time_prob": 0.05,
|
69 |
+
"model_type": "hubert",
|
70 |
+
"num_attention_heads": 12,
|
71 |
+
"num_conv_pos_embedding_groups": 16,
|
72 |
+
"num_conv_pos_embeddings": 128,
|
73 |
+
"num_feat_extract_layers": 7,
|
74 |
+
"num_hidden_layers": 2,
|
75 |
+
"pad_token_id": 0,
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.38.1",
|
78 |
+
"use_weighted_layer_sum": false,
|
79 |
+
"vocab_size": 32
|
80 |
+
}
|
run-4/checkpoint-336/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3425a160cf28659111e87b6ab601bd4f64cd88b8baa924505a0ed42513beb151
|
3 |
+
size 94763496
|
run-4/checkpoint-336/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a80e51af62d25c253503b816ec7b98c85cc9ba062954c975249337e8ce646afe
|
3 |
+
size 189552570
|
run-4/checkpoint-336/preprocessor_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000
|
9 |
+
}
|
run-4/checkpoint-336/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0d12884ae20f0c926a355fda8650edc055a398d4c7c42545ccdb7d60bd202452
|
3 |
+
size 14244
|
run-4/checkpoint-336/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e0d819fb0c1d431c458fec913aa5469b8a7ed1313a0add2ae71b76c0b9a0a219
|
3 |
+
size 1064
|
run-4/checkpoint-336/trainer_state.json
ADDED
@@ -0,0 +1,556 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.7326732673267327,
|
3 |
+
"best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-4/checkpoint-96",
|
4 |
+
"epoch": 7.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 336,
|
7 |
+
"is_hyper_param_search": true,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1,
|
13 |
+
"grad_norm": 1.2733114957809448,
|
14 |
+
"learning_rate": 2.2702186710865246e-07,
|
15 |
+
"loss": 0.7025,
|
16 |
+
"step": 5
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21,
|
20 |
+
"grad_norm": 1.243804931640625,
|
21 |
+
"learning_rate": 4.5404373421730493e-07,
|
22 |
+
"loss": 0.6974,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.31,
|
27 |
+
"grad_norm": 1.7711552381515503,
|
28 |
+
"learning_rate": 6.810656013259573e-07,
|
29 |
+
"loss": 0.696,
|
30 |
+
"step": 15
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.42,
|
34 |
+
"grad_norm": 1.1453403234481812,
|
35 |
+
"learning_rate": 9.080874684346099e-07,
|
36 |
+
"loss": 0.6989,
|
37 |
+
"step": 20
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.52,
|
41 |
+
"grad_norm": 1.2729355096817017,
|
42 |
+
"learning_rate": 1.1351093355432624e-06,
|
43 |
+
"loss": 0.6968,
|
44 |
+
"step": 25
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.62,
|
48 |
+
"grad_norm": 1.1592165231704712,
|
49 |
+
"learning_rate": 1.3621312026519146e-06,
|
50 |
+
"loss": 0.6959,
|
51 |
+
"step": 30
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.73,
|
55 |
+
"grad_norm": 1.1798148155212402,
|
56 |
+
"learning_rate": 1.589153069760567e-06,
|
57 |
+
"loss": 0.6952,
|
58 |
+
"step": 35
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.83,
|
62 |
+
"grad_norm": 2.1216671466827393,
|
63 |
+
"learning_rate": 1.8161749368692197e-06,
|
64 |
+
"loss": 0.6886,
|
65 |
+
"step": 40
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.94,
|
69 |
+
"grad_norm": 1.3416370153427124,
|
70 |
+
"learning_rate": 2.043196803977872e-06,
|
71 |
+
"loss": 0.6864,
|
72 |
+
"step": 45
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.0,
|
76 |
+
"eval_f1": 0.72,
|
77 |
+
"eval_loss": 0.688262939453125,
|
78 |
+
"eval_runtime": 1.3468,
|
79 |
+
"eval_samples_per_second": 47.521,
|
80 |
+
"eval_steps_per_second": 5.94,
|
81 |
+
"step": 48
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 1.04,
|
85 |
+
"grad_norm": 2.1856281757354736,
|
86 |
+
"learning_rate": 2.169320063482679e-06,
|
87 |
+
"loss": 0.6917,
|
88 |
+
"step": 50
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.15,
|
92 |
+
"grad_norm": 1.4077153205871582,
|
93 |
+
"learning_rate": 2.1440954115817176e-06,
|
94 |
+
"loss": 0.6884,
|
95 |
+
"step": 55
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 1.25,
|
99 |
+
"grad_norm": 2.1792664527893066,
|
100 |
+
"learning_rate": 2.1188707596807562e-06,
|
101 |
+
"loss": 0.6668,
|
102 |
+
"step": 60
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.35,
|
106 |
+
"grad_norm": 1.0386197566986084,
|
107 |
+
"learning_rate": 2.093646107779795e-06,
|
108 |
+
"loss": 0.6694,
|
109 |
+
"step": 65
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"epoch": 1.46,
|
113 |
+
"grad_norm": 2.0565919876098633,
|
114 |
+
"learning_rate": 2.0684214558788335e-06,
|
115 |
+
"loss": 0.6561,
|
116 |
+
"step": 70
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 1.56,
|
120 |
+
"grad_norm": 1.2978509664535522,
|
121 |
+
"learning_rate": 2.043196803977872e-06,
|
122 |
+
"loss": 0.6789,
|
123 |
+
"step": 75
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.67,
|
127 |
+
"grad_norm": 2.058328628540039,
|
128 |
+
"learning_rate": 2.0179721520769108e-06,
|
129 |
+
"loss": 0.6633,
|
130 |
+
"step": 80
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 1.77,
|
134 |
+
"grad_norm": 0.6023226976394653,
|
135 |
+
"learning_rate": 1.9927475001759494e-06,
|
136 |
+
"loss": 0.6655,
|
137 |
+
"step": 85
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.88,
|
141 |
+
"grad_norm": 0.5510762929916382,
|
142 |
+
"learning_rate": 1.967522848274988e-06,
|
143 |
+
"loss": 0.6622,
|
144 |
+
"step": 90
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.98,
|
148 |
+
"grad_norm": 1.098602533340454,
|
149 |
+
"learning_rate": 1.9422981963740267e-06,
|
150 |
+
"loss": 0.6633,
|
151 |
+
"step": 95
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 2.0,
|
155 |
+
"eval_f1": 0.7326732673267327,
|
156 |
+
"eval_loss": 0.6816024780273438,
|
157 |
+
"eval_runtime": 1.3765,
|
158 |
+
"eval_samples_per_second": 46.493,
|
159 |
+
"eval_steps_per_second": 5.812,
|
160 |
+
"step": 96
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.08,
|
164 |
+
"grad_norm": 0.9589098691940308,
|
165 |
+
"learning_rate": 1.9170735444730654e-06,
|
166 |
+
"loss": 0.659,
|
167 |
+
"step": 100
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 2.19,
|
171 |
+
"grad_norm": 1.070695161819458,
|
172 |
+
"learning_rate": 1.8918488925721038e-06,
|
173 |
+
"loss": 0.6313,
|
174 |
+
"step": 105
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 2.29,
|
178 |
+
"grad_norm": 0.9913639426231384,
|
179 |
+
"learning_rate": 1.8666242406711424e-06,
|
180 |
+
"loss": 0.6652,
|
181 |
+
"step": 110
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 2.4,
|
185 |
+
"grad_norm": 1.0632878541946411,
|
186 |
+
"learning_rate": 1.841399588770181e-06,
|
187 |
+
"loss": 0.673,
|
188 |
+
"step": 115
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 2.5,
|
192 |
+
"grad_norm": 2.1036579608917236,
|
193 |
+
"learning_rate": 1.8161749368692197e-06,
|
194 |
+
"loss": 0.6451,
|
195 |
+
"step": 120
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 2.6,
|
199 |
+
"grad_norm": 1.08384108543396,
|
200 |
+
"learning_rate": 1.7909502849682583e-06,
|
201 |
+
"loss": 0.6322,
|
202 |
+
"step": 125
|
203 |
+
},
|
204 |
+
{
|
205 |
+
"epoch": 2.71,
|
206 |
+
"grad_norm": 0.9407000541687012,
|
207 |
+
"learning_rate": 1.765725633067297e-06,
|
208 |
+
"loss": 0.6755,
|
209 |
+
"step": 130
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 2.81,
|
213 |
+
"grad_norm": 0.9016568660736084,
|
214 |
+
"learning_rate": 1.7405009811663356e-06,
|
215 |
+
"loss": 0.5985,
|
216 |
+
"step": 135
|
217 |
+
},
|
218 |
+
{
|
219 |
+
"epoch": 2.92,
|
220 |
+
"grad_norm": 1.1134448051452637,
|
221 |
+
"learning_rate": 1.7152763292653743e-06,
|
222 |
+
"loss": 0.603,
|
223 |
+
"step": 140
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 3.0,
|
227 |
+
"eval_f1": 0.7326732673267327,
|
228 |
+
"eval_loss": 0.6800689697265625,
|
229 |
+
"eval_runtime": 1.3861,
|
230 |
+
"eval_samples_per_second": 46.173,
|
231 |
+
"eval_steps_per_second": 5.772,
|
232 |
+
"step": 144
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 3.02,
|
236 |
+
"grad_norm": 0.7627719640731812,
|
237 |
+
"learning_rate": 1.6900516773644127e-06,
|
238 |
+
"loss": 0.6557,
|
239 |
+
"step": 145
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 3.12,
|
243 |
+
"grad_norm": 0.9291415214538574,
|
244 |
+
"learning_rate": 1.6648270254634511e-06,
|
245 |
+
"loss": 0.6219,
|
246 |
+
"step": 150
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 3.23,
|
250 |
+
"grad_norm": 0.9248765707015991,
|
251 |
+
"learning_rate": 1.6396023735624898e-06,
|
252 |
+
"loss": 0.6325,
|
253 |
+
"step": 155
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 3.33,
|
257 |
+
"grad_norm": 0.9842573404312134,
|
258 |
+
"learning_rate": 1.6143777216615284e-06,
|
259 |
+
"loss": 0.6521,
|
260 |
+
"step": 160
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 3.44,
|
264 |
+
"grad_norm": 0.8689214587211609,
|
265 |
+
"learning_rate": 1.589153069760567e-06,
|
266 |
+
"loss": 0.5929,
|
267 |
+
"step": 165
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 3.54,
|
271 |
+
"grad_norm": 1.0012000799179077,
|
272 |
+
"learning_rate": 1.5639284178596057e-06,
|
273 |
+
"loss": 0.584,
|
274 |
+
"step": 170
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 3.65,
|
278 |
+
"grad_norm": 0.7438368797302246,
|
279 |
+
"learning_rate": 1.5387037659586443e-06,
|
280 |
+
"loss": 0.6813,
|
281 |
+
"step": 175
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 3.75,
|
285 |
+
"grad_norm": 1.8603870868682861,
|
286 |
+
"learning_rate": 1.513479114057683e-06,
|
287 |
+
"loss": 0.6099,
|
288 |
+
"step": 180
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 3.85,
|
292 |
+
"grad_norm": 0.9918416738510132,
|
293 |
+
"learning_rate": 1.4882544621567216e-06,
|
294 |
+
"loss": 0.6192,
|
295 |
+
"step": 185
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 3.96,
|
299 |
+
"grad_norm": 1.9146322011947632,
|
300 |
+
"learning_rate": 1.4630298102557603e-06,
|
301 |
+
"loss": 0.6472,
|
302 |
+
"step": 190
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 4.0,
|
306 |
+
"eval_f1": 0.7326732673267327,
|
307 |
+
"eval_loss": 0.6818161010742188,
|
308 |
+
"eval_runtime": 1.3841,
|
309 |
+
"eval_samples_per_second": 46.239,
|
310 |
+
"eval_steps_per_second": 5.78,
|
311 |
+
"step": 192
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 4.06,
|
315 |
+
"grad_norm": 0.9502781629562378,
|
316 |
+
"learning_rate": 1.437805158354799e-06,
|
317 |
+
"loss": 0.6447,
|
318 |
+
"step": 195
|
319 |
+
},
|
320 |
+
{
|
321 |
+
"epoch": 4.17,
|
322 |
+
"grad_norm": 0.8570067286491394,
|
323 |
+
"learning_rate": 1.4125805064538375e-06,
|
324 |
+
"loss": 0.5306,
|
325 |
+
"step": 200
|
326 |
+
},
|
327 |
+
{
|
328 |
+
"epoch": 4.27,
|
329 |
+
"grad_norm": 0.8097484111785889,
|
330 |
+
"learning_rate": 1.3873558545528762e-06,
|
331 |
+
"loss": 0.6202,
|
332 |
+
"step": 205
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 4.38,
|
336 |
+
"grad_norm": 2.0106472969055176,
|
337 |
+
"learning_rate": 1.3621312026519146e-06,
|
338 |
+
"loss": 0.6705,
|
339 |
+
"step": 210
|
340 |
+
},
|
341 |
+
{
|
342 |
+
"epoch": 4.48,
|
343 |
+
"grad_norm": 1.090775489807129,
|
344 |
+
"learning_rate": 1.3369065507509533e-06,
|
345 |
+
"loss": 0.6297,
|
346 |
+
"step": 215
|
347 |
+
},
|
348 |
+
{
|
349 |
+
"epoch": 4.58,
|
350 |
+
"grad_norm": 0.8988145589828491,
|
351 |
+
"learning_rate": 1.311681898849992e-06,
|
352 |
+
"loss": 0.5896,
|
353 |
+
"step": 220
|
354 |
+
},
|
355 |
+
{
|
356 |
+
"epoch": 4.69,
|
357 |
+
"grad_norm": 0.9149978756904602,
|
358 |
+
"learning_rate": 1.2864572469490305e-06,
|
359 |
+
"loss": 0.6156,
|
360 |
+
"step": 225
|
361 |
+
},
|
362 |
+
{
|
363 |
+
"epoch": 4.79,
|
364 |
+
"grad_norm": 1.9398412704467773,
|
365 |
+
"learning_rate": 1.2612325950480692e-06,
|
366 |
+
"loss": 0.6305,
|
367 |
+
"step": 230
|
368 |
+
},
|
369 |
+
{
|
370 |
+
"epoch": 4.9,
|
371 |
+
"grad_norm": 0.9217966794967651,
|
372 |
+
"learning_rate": 1.2360079431471078e-06,
|
373 |
+
"loss": 0.5943,
|
374 |
+
"step": 235
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 5.0,
|
378 |
+
"grad_norm": 0.9083653688430786,
|
379 |
+
"learning_rate": 1.2107832912461465e-06,
|
380 |
+
"loss": 0.6386,
|
381 |
+
"step": 240
|
382 |
+
},
|
383 |
+
{
|
384 |
+
"epoch": 5.0,
|
385 |
+
"eval_f1": 0.7326732673267327,
|
386 |
+
"eval_loss": 0.6846389770507812,
|
387 |
+
"eval_runtime": 1.4094,
|
388 |
+
"eval_samples_per_second": 45.409,
|
389 |
+
"eval_steps_per_second": 5.676,
|
390 |
+
"step": 240
|
391 |
+
},
|
392 |
+
{
|
393 |
+
"epoch": 5.1,
|
394 |
+
"grad_norm": 0.9323675036430359,
|
395 |
+
"learning_rate": 1.1855586393451851e-06,
|
396 |
+
"loss": 0.5779,
|
397 |
+
"step": 245
|
398 |
+
},
|
399 |
+
{
|
400 |
+
"epoch": 5.21,
|
401 |
+
"grad_norm": 0.7549787163734436,
|
402 |
+
"learning_rate": 1.1603339874442238e-06,
|
403 |
+
"loss": 0.5948,
|
404 |
+
"step": 250
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 5.31,
|
408 |
+
"grad_norm": 0.8535837531089783,
|
409 |
+
"learning_rate": 1.1351093355432624e-06,
|
410 |
+
"loss": 0.6928,
|
411 |
+
"step": 255
|
412 |
+
},
|
413 |
+
{
|
414 |
+
"epoch": 5.42,
|
415 |
+
"grad_norm": 1.2038137912750244,
|
416 |
+
"learning_rate": 1.109884683642301e-06,
|
417 |
+
"loss": 0.5887,
|
418 |
+
"step": 260
|
419 |
+
},
|
420 |
+
{
|
421 |
+
"epoch": 5.52,
|
422 |
+
"grad_norm": 0.9501279592514038,
|
423 |
+
"learning_rate": 1.0846600317413395e-06,
|
424 |
+
"loss": 0.5776,
|
425 |
+
"step": 265
|
426 |
+
},
|
427 |
+
{
|
428 |
+
"epoch": 5.62,
|
429 |
+
"grad_norm": 0.7421719431877136,
|
430 |
+
"learning_rate": 1.0594353798403781e-06,
|
431 |
+
"loss": 0.6734,
|
432 |
+
"step": 270
|
433 |
+
},
|
434 |
+
{
|
435 |
+
"epoch": 5.73,
|
436 |
+
"grad_norm": 0.8555863499641418,
|
437 |
+
"learning_rate": 1.0342107279394168e-06,
|
438 |
+
"loss": 0.6399,
|
439 |
+
"step": 275
|
440 |
+
},
|
441 |
+
{
|
442 |
+
"epoch": 5.83,
|
443 |
+
"grad_norm": 0.8841156363487244,
|
444 |
+
"learning_rate": 1.0089860760384554e-06,
|
445 |
+
"loss": 0.6173,
|
446 |
+
"step": 280
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 5.94,
|
450 |
+
"grad_norm": 0.8565478324890137,
|
451 |
+
"learning_rate": 9.83761424137494e-07,
|
452 |
+
"loss": 0.5537,
|
453 |
+
"step": 285
|
454 |
+
},
|
455 |
+
{
|
456 |
+
"epoch": 6.0,
|
457 |
+
"eval_f1": 0.7326732673267327,
|
458 |
+
"eval_loss": 0.6864242553710938,
|
459 |
+
"eval_runtime": 1.3622,
|
460 |
+
"eval_samples_per_second": 46.983,
|
461 |
+
"eval_steps_per_second": 5.873,
|
462 |
+
"step": 288
|
463 |
+
},
|
464 |
+
{
|
465 |
+
"epoch": 6.04,
|
466 |
+
"grad_norm": 0.8750139474868774,
|
467 |
+
"learning_rate": 9.585367722365327e-07,
|
468 |
+
"loss": 0.5531,
|
469 |
+
"step": 290
|
470 |
+
},
|
471 |
+
{
|
472 |
+
"epoch": 6.15,
|
473 |
+
"grad_norm": 1.0445302724838257,
|
474 |
+
"learning_rate": 9.333121203355712e-07,
|
475 |
+
"loss": 0.638,
|
476 |
+
"step": 295
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 6.25,
|
480 |
+
"grad_norm": 0.7958914637565613,
|
481 |
+
"learning_rate": 9.080874684346099e-07,
|
482 |
+
"loss": 0.547,
|
483 |
+
"step": 300
|
484 |
+
},
|
485 |
+
{
|
486 |
+
"epoch": 6.35,
|
487 |
+
"grad_norm": 0.9992254376411438,
|
488 |
+
"learning_rate": 8.828628165336485e-07,
|
489 |
+
"loss": 0.6425,
|
490 |
+
"step": 305
|
491 |
+
},
|
492 |
+
{
|
493 |
+
"epoch": 6.46,
|
494 |
+
"grad_norm": 0.8400682806968689,
|
495 |
+
"learning_rate": 8.576381646326871e-07,
|
496 |
+
"loss": 0.6955,
|
497 |
+
"step": 310
|
498 |
+
},
|
499 |
+
{
|
500 |
+
"epoch": 6.56,
|
501 |
+
"grad_norm": 0.742438793182373,
|
502 |
+
"learning_rate": 8.324135127317256e-07,
|
503 |
+
"loss": 0.6473,
|
504 |
+
"step": 315
|
505 |
+
},
|
506 |
+
{
|
507 |
+
"epoch": 6.67,
|
508 |
+
"grad_norm": 0.6693254113197327,
|
509 |
+
"learning_rate": 8.071888608307642e-07,
|
510 |
+
"loss": 0.603,
|
511 |
+
"step": 320
|
512 |
+
},
|
513 |
+
{
|
514 |
+
"epoch": 6.77,
|
515 |
+
"grad_norm": 1.0816401243209839,
|
516 |
+
"learning_rate": 7.819642089298028e-07,
|
517 |
+
"loss": 0.6053,
|
518 |
+
"step": 325
|
519 |
+
},
|
520 |
+
{
|
521 |
+
"epoch": 6.88,
|
522 |
+
"grad_norm": 0.7275277376174927,
|
523 |
+
"learning_rate": 7.567395570288415e-07,
|
524 |
+
"loss": 0.612,
|
525 |
+
"step": 330
|
526 |
+
},
|
527 |
+
{
|
528 |
+
"epoch": 6.98,
|
529 |
+
"grad_norm": 0.7834873795509338,
|
530 |
+
"learning_rate": 7.315149051278801e-07,
|
531 |
+
"loss": 0.55,
|
532 |
+
"step": 335
|
533 |
+
},
|
534 |
+
{
|
535 |
+
"epoch": 7.0,
|
536 |
+
"eval_f1": 0.7326732673267327,
|
537 |
+
"eval_loss": 0.6889228820800781,
|
538 |
+
"eval_runtime": 1.3769,
|
539 |
+
"eval_samples_per_second": 46.483,
|
540 |
+
"eval_steps_per_second": 5.81,
|
541 |
+
"step": 336
|
542 |
+
}
|
543 |
+
],
|
544 |
+
"logging_steps": 5,
|
545 |
+
"max_steps": 480,
|
546 |
+
"num_input_tokens_seen": 0,
|
547 |
+
"num_train_epochs": 10,
|
548 |
+
"save_steps": 500,
|
549 |
+
"total_flos": 5116387614670704.0,
|
550 |
+
"train_batch_size": 4,
|
551 |
+
"trial_name": null,
|
552 |
+
"trial_params": {
|
553 |
+
"learning_rate": 2.1794099242430636e-06,
|
554 |
+
"per_device_train_batch_size": 4
|
555 |
+
}
|
556 |
+
}
|
run-4/checkpoint-336/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6eadd8415a707c8e4de7440233358b89448cac114a2ada36aef1d9186553c0c3
|
3 |
+
size 4920
|
run-4/checkpoint-384/config.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "NOT_WORD",
|
53 |
+
"1": "WORD"
|
54 |
+
},
|
55 |
+
"initializer_range": 0.02,
|
56 |
+
"intermediate_size": 3072,
|
57 |
+
"label2id": {
|
58 |
+
"NOT_WORD": "0",
|
59 |
+
"WORD": "1"
|
60 |
+
},
|
61 |
+
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
+
"mask_feature_length": 10,
|
64 |
+
"mask_feature_min_masks": 0,
|
65 |
+
"mask_feature_prob": 0.0,
|
66 |
+
"mask_time_length": 10,
|
67 |
+
"mask_time_min_masks": 2,
|
68 |
+
"mask_time_prob": 0.05,
|
69 |
+
"model_type": "hubert",
|
70 |
+
"num_attention_heads": 12,
|
71 |
+
"num_conv_pos_embedding_groups": 16,
|
72 |
+
"num_conv_pos_embeddings": 128,
|
73 |
+
"num_feat_extract_layers": 7,
|
74 |
+
"num_hidden_layers": 2,
|
75 |
+
"pad_token_id": 0,
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.38.1",
|
78 |
+
"use_weighted_layer_sum": false,
|
79 |
+
"vocab_size": 32
|
80 |
+
}
|
run-4/checkpoint-384/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f03572b2dcb34be269d7a8392a1db64aa78ab3c991d84591fe98b5b2f300eea6
|
3 |
+
size 94763496
|
run-4/checkpoint-384/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:370faca0c4d8022d7391f5e21d5c7cf5baa3cc1575da59ece8f1da8e073be1d6
|
3 |
+
size 189552570
|
run-4/checkpoint-384/preprocessor_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000
|
9 |
+
}
|
run-4/checkpoint-384/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5326b9611b4fb9dc5dc0b29580e7e48abf50913e44071592799c052bebfbacd7
|
3 |
+
size 14244
|
run-4/checkpoint-384/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:53cf1db083d766e2eeb0b68e04895788e327c7d113d516f642c4fc0792596377
|
3 |
+
size 1064
|
run-4/checkpoint-384/trainer_state.json
ADDED
@@ -0,0 +1,628 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.7326732673267327,
|
3 |
+
"best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-4/checkpoint-96",
|
4 |
+
"epoch": 8.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 384,
|
7 |
+
"is_hyper_param_search": true,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1,
|
13 |
+
"grad_norm": 1.2733114957809448,
|
14 |
+
"learning_rate": 2.2702186710865246e-07,
|
15 |
+
"loss": 0.7025,
|
16 |
+
"step": 5
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21,
|
20 |
+
"grad_norm": 1.243804931640625,
|
21 |
+
"learning_rate": 4.5404373421730493e-07,
|
22 |
+
"loss": 0.6974,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.31,
|
27 |
+
"grad_norm": 1.7711552381515503,
|
28 |
+
"learning_rate": 6.810656013259573e-07,
|
29 |
+
"loss": 0.696,
|
30 |
+
"step": 15
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.42,
|
34 |
+
"grad_norm": 1.1453403234481812,
|
35 |
+
"learning_rate": 9.080874684346099e-07,
|
36 |
+
"loss": 0.6989,
|
37 |
+
"step": 20
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.52,
|
41 |
+
"grad_norm": 1.2729355096817017,
|
42 |
+
"learning_rate": 1.1351093355432624e-06,
|
43 |
+
"loss": 0.6968,
|
44 |
+
"step": 25
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.62,
|
48 |
+
"grad_norm": 1.1592165231704712,
|
49 |
+
"learning_rate": 1.3621312026519146e-06,
|
50 |
+
"loss": 0.6959,
|
51 |
+
"step": 30
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.73,
|
55 |
+
"grad_norm": 1.1798148155212402,
|
56 |
+
"learning_rate": 1.589153069760567e-06,
|
57 |
+
"loss": 0.6952,
|
58 |
+
"step": 35
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.83,
|
62 |
+
"grad_norm": 2.1216671466827393,
|
63 |
+
"learning_rate": 1.8161749368692197e-06,
|
64 |
+
"loss": 0.6886,
|
65 |
+
"step": 40
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.94,
|
69 |
+
"grad_norm": 1.3416370153427124,
|
70 |
+
"learning_rate": 2.043196803977872e-06,
|
71 |
+
"loss": 0.6864,
|
72 |
+
"step": 45
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.0,
|
76 |
+
"eval_f1": 0.72,
|
77 |
+
"eval_loss": 0.688262939453125,
|
78 |
+
"eval_runtime": 1.3468,
|
79 |
+
"eval_samples_per_second": 47.521,
|
80 |
+
"eval_steps_per_second": 5.94,
|
81 |
+
"step": 48
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 1.04,
|
85 |
+
"grad_norm": 2.1856281757354736,
|
86 |
+
"learning_rate": 2.169320063482679e-06,
|
87 |
+
"loss": 0.6917,
|
88 |
+
"step": 50
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.15,
|
92 |
+
"grad_norm": 1.4077153205871582,
|
93 |
+
"learning_rate": 2.1440954115817176e-06,
|
94 |
+
"loss": 0.6884,
|
95 |
+
"step": 55
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 1.25,
|
99 |
+
"grad_norm": 2.1792664527893066,
|
100 |
+
"learning_rate": 2.1188707596807562e-06,
|
101 |
+
"loss": 0.6668,
|
102 |
+
"step": 60
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.35,
|
106 |
+
"grad_norm": 1.0386197566986084,
|
107 |
+
"learning_rate": 2.093646107779795e-06,
|
108 |
+
"loss": 0.6694,
|
109 |
+
"step": 65
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"epoch": 1.46,
|
113 |
+
"grad_norm": 2.0565919876098633,
|
114 |
+
"learning_rate": 2.0684214558788335e-06,
|
115 |
+
"loss": 0.6561,
|
116 |
+
"step": 70
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 1.56,
|
120 |
+
"grad_norm": 1.2978509664535522,
|
121 |
+
"learning_rate": 2.043196803977872e-06,
|
122 |
+
"loss": 0.6789,
|
123 |
+
"step": 75
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.67,
|
127 |
+
"grad_norm": 2.058328628540039,
|
128 |
+
"learning_rate": 2.0179721520769108e-06,
|
129 |
+
"loss": 0.6633,
|
130 |
+
"step": 80
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 1.77,
|
134 |
+
"grad_norm": 0.6023226976394653,
|
135 |
+
"learning_rate": 1.9927475001759494e-06,
|
136 |
+
"loss": 0.6655,
|
137 |
+
"step": 85
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.88,
|
141 |
+
"grad_norm": 0.5510762929916382,
|
142 |
+
"learning_rate": 1.967522848274988e-06,
|
143 |
+
"loss": 0.6622,
|
144 |
+
"step": 90
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.98,
|
148 |
+
"grad_norm": 1.098602533340454,
|
149 |
+
"learning_rate": 1.9422981963740267e-06,
|
150 |
+
"loss": 0.6633,
|
151 |
+
"step": 95
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 2.0,
|
155 |
+
"eval_f1": 0.7326732673267327,
|
156 |
+
"eval_loss": 0.6816024780273438,
|
157 |
+
"eval_runtime": 1.3765,
|
158 |
+
"eval_samples_per_second": 46.493,
|
159 |
+
"eval_steps_per_second": 5.812,
|
160 |
+
"step": 96
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.08,
|
164 |
+
"grad_norm": 0.9589098691940308,
|
165 |
+
"learning_rate": 1.9170735444730654e-06,
|
166 |
+
"loss": 0.659,
|
167 |
+
"step": 100
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 2.19,
|
171 |
+
"grad_norm": 1.070695161819458,
|
172 |
+
"learning_rate": 1.8918488925721038e-06,
|
173 |
+
"loss": 0.6313,
|
174 |
+
"step": 105
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 2.29,
|
178 |
+
"grad_norm": 0.9913639426231384,
|
179 |
+
"learning_rate": 1.8666242406711424e-06,
|
180 |
+
"loss": 0.6652,
|
181 |
+
"step": 110
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 2.4,
|
185 |
+
"grad_norm": 1.0632878541946411,
|
186 |
+
"learning_rate": 1.841399588770181e-06,
|
187 |
+
"loss": 0.673,
|
188 |
+
"step": 115
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 2.5,
|
192 |
+
"grad_norm": 2.1036579608917236,
|
193 |
+
"learning_rate": 1.8161749368692197e-06,
|
194 |
+
"loss": 0.6451,
|
195 |
+
"step": 120
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 2.6,
|
199 |
+
"grad_norm": 1.08384108543396,
|
200 |
+
"learning_rate": 1.7909502849682583e-06,
|
201 |
+
"loss": 0.6322,
|
202 |
+
"step": 125
|
203 |
+
},
|
204 |
+
{
|
205 |
+
"epoch": 2.71,
|
206 |
+
"grad_norm": 0.9407000541687012,
|
207 |
+
"learning_rate": 1.765725633067297e-06,
|
208 |
+
"loss": 0.6755,
|
209 |
+
"step": 130
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 2.81,
|
213 |
+
"grad_norm": 0.9016568660736084,
|
214 |
+
"learning_rate": 1.7405009811663356e-06,
|
215 |
+
"loss": 0.5985,
|
216 |
+
"step": 135
|
217 |
+
},
|
218 |
+
{
|
219 |
+
"epoch": 2.92,
|
220 |
+
"grad_norm": 1.1134448051452637,
|
221 |
+
"learning_rate": 1.7152763292653743e-06,
|
222 |
+
"loss": 0.603,
|
223 |
+
"step": 140
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 3.0,
|
227 |
+
"eval_f1": 0.7326732673267327,
|
228 |
+
"eval_loss": 0.6800689697265625,
|
229 |
+
"eval_runtime": 1.3861,
|
230 |
+
"eval_samples_per_second": 46.173,
|
231 |
+
"eval_steps_per_second": 5.772,
|
232 |
+
"step": 144
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 3.02,
|
236 |
+
"grad_norm": 0.7627719640731812,
|
237 |
+
"learning_rate": 1.6900516773644127e-06,
|
238 |
+
"loss": 0.6557,
|
239 |
+
"step": 145
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 3.12,
|
243 |
+
"grad_norm": 0.9291415214538574,
|
244 |
+
"learning_rate": 1.6648270254634511e-06,
|
245 |
+
"loss": 0.6219,
|
246 |
+
"step": 150
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 3.23,
|
250 |
+
"grad_norm": 0.9248765707015991,
|
251 |
+
"learning_rate": 1.6396023735624898e-06,
|
252 |
+
"loss": 0.6325,
|
253 |
+
"step": 155
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 3.33,
|
257 |
+
"grad_norm": 0.9842573404312134,
|
258 |
+
"learning_rate": 1.6143777216615284e-06,
|
259 |
+
"loss": 0.6521,
|
260 |
+
"step": 160
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 3.44,
|
264 |
+
"grad_norm": 0.8689214587211609,
|
265 |
+
"learning_rate": 1.589153069760567e-06,
|
266 |
+
"loss": 0.5929,
|
267 |
+
"step": 165
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 3.54,
|
271 |
+
"grad_norm": 1.0012000799179077,
|
272 |
+
"learning_rate": 1.5639284178596057e-06,
|
273 |
+
"loss": 0.584,
|
274 |
+
"step": 170
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 3.65,
|
278 |
+
"grad_norm": 0.7438368797302246,
|
279 |
+
"learning_rate": 1.5387037659586443e-06,
|
280 |
+
"loss": 0.6813,
|
281 |
+
"step": 175
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 3.75,
|
285 |
+
"grad_norm": 1.8603870868682861,
|
286 |
+
"learning_rate": 1.513479114057683e-06,
|
287 |
+
"loss": 0.6099,
|
288 |
+
"step": 180
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 3.85,
|
292 |
+
"grad_norm": 0.9918416738510132,
|
293 |
+
"learning_rate": 1.4882544621567216e-06,
|
294 |
+
"loss": 0.6192,
|
295 |
+
"step": 185
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 3.96,
|
299 |
+
"grad_norm": 1.9146322011947632,
|
300 |
+
"learning_rate": 1.4630298102557603e-06,
|
301 |
+
"loss": 0.6472,
|
302 |
+
"step": 190
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 4.0,
|
306 |
+
"eval_f1": 0.7326732673267327,
|
307 |
+
"eval_loss": 0.6818161010742188,
|
308 |
+
"eval_runtime": 1.3841,
|
309 |
+
"eval_samples_per_second": 46.239,
|
310 |
+
"eval_steps_per_second": 5.78,
|
311 |
+
"step": 192
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 4.06,
|
315 |
+
"grad_norm": 0.9502781629562378,
|
316 |
+
"learning_rate": 1.437805158354799e-06,
|
317 |
+
"loss": 0.6447,
|
318 |
+
"step": 195
|
319 |
+
},
|
320 |
+
{
|
321 |
+
"epoch": 4.17,
|
322 |
+
"grad_norm": 0.8570067286491394,
|
323 |
+
"learning_rate": 1.4125805064538375e-06,
|
324 |
+
"loss": 0.5306,
|
325 |
+
"step": 200
|
326 |
+
},
|
327 |
+
{
|
328 |
+
"epoch": 4.27,
|
329 |
+
"grad_norm": 0.8097484111785889,
|
330 |
+
"learning_rate": 1.3873558545528762e-06,
|
331 |
+
"loss": 0.6202,
|
332 |
+
"step": 205
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 4.38,
|
336 |
+
"grad_norm": 2.0106472969055176,
|
337 |
+
"learning_rate": 1.3621312026519146e-06,
|
338 |
+
"loss": 0.6705,
|
339 |
+
"step": 210
|
340 |
+
},
|
341 |
+
{
|
342 |
+
"epoch": 4.48,
|
343 |
+
"grad_norm": 1.090775489807129,
|
344 |
+
"learning_rate": 1.3369065507509533e-06,
|
345 |
+
"loss": 0.6297,
|
346 |
+
"step": 215
|
347 |
+
},
|
348 |
+
{
|
349 |
+
"epoch": 4.58,
|
350 |
+
"grad_norm": 0.8988145589828491,
|
351 |
+
"learning_rate": 1.311681898849992e-06,
|
352 |
+
"loss": 0.5896,
|
353 |
+
"step": 220
|
354 |
+
},
|
355 |
+
{
|
356 |
+
"epoch": 4.69,
|
357 |
+
"grad_norm": 0.9149978756904602,
|
358 |
+
"learning_rate": 1.2864572469490305e-06,
|
359 |
+
"loss": 0.6156,
|
360 |
+
"step": 225
|
361 |
+
},
|
362 |
+
{
|
363 |
+
"epoch": 4.79,
|
364 |
+
"grad_norm": 1.9398412704467773,
|
365 |
+
"learning_rate": 1.2612325950480692e-06,
|
366 |
+
"loss": 0.6305,
|
367 |
+
"step": 230
|
368 |
+
},
|
369 |
+
{
|
370 |
+
"epoch": 4.9,
|
371 |
+
"grad_norm": 0.9217966794967651,
|
372 |
+
"learning_rate": 1.2360079431471078e-06,
|
373 |
+
"loss": 0.5943,
|
374 |
+
"step": 235
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 5.0,
|
378 |
+
"grad_norm": 0.9083653688430786,
|
379 |
+
"learning_rate": 1.2107832912461465e-06,
|
380 |
+
"loss": 0.6386,
|
381 |
+
"step": 240
|
382 |
+
},
|
383 |
+
{
|
384 |
+
"epoch": 5.0,
|
385 |
+
"eval_f1": 0.7326732673267327,
|
386 |
+
"eval_loss": 0.6846389770507812,
|
387 |
+
"eval_runtime": 1.4094,
|
388 |
+
"eval_samples_per_second": 45.409,
|
389 |
+
"eval_steps_per_second": 5.676,
|
390 |
+
"step": 240
|
391 |
+
},
|
392 |
+
{
|
393 |
+
"epoch": 5.1,
|
394 |
+
"grad_norm": 0.9323675036430359,
|
395 |
+
"learning_rate": 1.1855586393451851e-06,
|
396 |
+
"loss": 0.5779,
|
397 |
+
"step": 245
|
398 |
+
},
|
399 |
+
{
|
400 |
+
"epoch": 5.21,
|
401 |
+
"grad_norm": 0.7549787163734436,
|
402 |
+
"learning_rate": 1.1603339874442238e-06,
|
403 |
+
"loss": 0.5948,
|
404 |
+
"step": 250
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 5.31,
|
408 |
+
"grad_norm": 0.8535837531089783,
|
409 |
+
"learning_rate": 1.1351093355432624e-06,
|
410 |
+
"loss": 0.6928,
|
411 |
+
"step": 255
|
412 |
+
},
|
413 |
+
{
|
414 |
+
"epoch": 5.42,
|
415 |
+
"grad_norm": 1.2038137912750244,
|
416 |
+
"learning_rate": 1.109884683642301e-06,
|
417 |
+
"loss": 0.5887,
|
418 |
+
"step": 260
|
419 |
+
},
|
420 |
+
{
|
421 |
+
"epoch": 5.52,
|
422 |
+
"grad_norm": 0.9501279592514038,
|
423 |
+
"learning_rate": 1.0846600317413395e-06,
|
424 |
+
"loss": 0.5776,
|
425 |
+
"step": 265
|
426 |
+
},
|
427 |
+
{
|
428 |
+
"epoch": 5.62,
|
429 |
+
"grad_norm": 0.7421719431877136,
|
430 |
+
"learning_rate": 1.0594353798403781e-06,
|
431 |
+
"loss": 0.6734,
|
432 |
+
"step": 270
|
433 |
+
},
|
434 |
+
{
|
435 |
+
"epoch": 5.73,
|
436 |
+
"grad_norm": 0.8555863499641418,
|
437 |
+
"learning_rate": 1.0342107279394168e-06,
|
438 |
+
"loss": 0.6399,
|
439 |
+
"step": 275
|
440 |
+
},
|
441 |
+
{
|
442 |
+
"epoch": 5.83,
|
443 |
+
"grad_norm": 0.8841156363487244,
|
444 |
+
"learning_rate": 1.0089860760384554e-06,
|
445 |
+
"loss": 0.6173,
|
446 |
+
"step": 280
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 5.94,
|
450 |
+
"grad_norm": 0.8565478324890137,
|
451 |
+
"learning_rate": 9.83761424137494e-07,
|
452 |
+
"loss": 0.5537,
|
453 |
+
"step": 285
|
454 |
+
},
|
455 |
+
{
|
456 |
+
"epoch": 6.0,
|
457 |
+
"eval_f1": 0.7326732673267327,
|
458 |
+
"eval_loss": 0.6864242553710938,
|
459 |
+
"eval_runtime": 1.3622,
|
460 |
+
"eval_samples_per_second": 46.983,
|
461 |
+
"eval_steps_per_second": 5.873,
|
462 |
+
"step": 288
|
463 |
+
},
|
464 |
+
{
|
465 |
+
"epoch": 6.04,
|
466 |
+
"grad_norm": 0.8750139474868774,
|
467 |
+
"learning_rate": 9.585367722365327e-07,
|
468 |
+
"loss": 0.5531,
|
469 |
+
"step": 290
|
470 |
+
},
|
471 |
+
{
|
472 |
+
"epoch": 6.15,
|
473 |
+
"grad_norm": 1.0445302724838257,
|
474 |
+
"learning_rate": 9.333121203355712e-07,
|
475 |
+
"loss": 0.638,
|
476 |
+
"step": 295
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 6.25,
|
480 |
+
"grad_norm": 0.7958914637565613,
|
481 |
+
"learning_rate": 9.080874684346099e-07,
|
482 |
+
"loss": 0.547,
|
483 |
+
"step": 300
|
484 |
+
},
|
485 |
+
{
|
486 |
+
"epoch": 6.35,
|
487 |
+
"grad_norm": 0.9992254376411438,
|
488 |
+
"learning_rate": 8.828628165336485e-07,
|
489 |
+
"loss": 0.6425,
|
490 |
+
"step": 305
|
491 |
+
},
|
492 |
+
{
|
493 |
+
"epoch": 6.46,
|
494 |
+
"grad_norm": 0.8400682806968689,
|
495 |
+
"learning_rate": 8.576381646326871e-07,
|
496 |
+
"loss": 0.6955,
|
497 |
+
"step": 310
|
498 |
+
},
|
499 |
+
{
|
500 |
+
"epoch": 6.56,
|
501 |
+
"grad_norm": 0.742438793182373,
|
502 |
+
"learning_rate": 8.324135127317256e-07,
|
503 |
+
"loss": 0.6473,
|
504 |
+
"step": 315
|
505 |
+
},
|
506 |
+
{
|
507 |
+
"epoch": 6.67,
|
508 |
+
"grad_norm": 0.6693254113197327,
|
509 |
+
"learning_rate": 8.071888608307642e-07,
|
510 |
+
"loss": 0.603,
|
511 |
+
"step": 320
|
512 |
+
},
|
513 |
+
{
|
514 |
+
"epoch": 6.77,
|
515 |
+
"grad_norm": 1.0816401243209839,
|
516 |
+
"learning_rate": 7.819642089298028e-07,
|
517 |
+
"loss": 0.6053,
|
518 |
+
"step": 325
|
519 |
+
},
|
520 |
+
{
|
521 |
+
"epoch": 6.88,
|
522 |
+
"grad_norm": 0.7275277376174927,
|
523 |
+
"learning_rate": 7.567395570288415e-07,
|
524 |
+
"loss": 0.612,
|
525 |
+
"step": 330
|
526 |
+
},
|
527 |
+
{
|
528 |
+
"epoch": 6.98,
|
529 |
+
"grad_norm": 0.7834873795509338,
|
530 |
+
"learning_rate": 7.315149051278801e-07,
|
531 |
+
"loss": 0.55,
|
532 |
+
"step": 335
|
533 |
+
},
|
534 |
+
{
|
535 |
+
"epoch": 7.0,
|
536 |
+
"eval_f1": 0.7326732673267327,
|
537 |
+
"eval_loss": 0.6889228820800781,
|
538 |
+
"eval_runtime": 1.3769,
|
539 |
+
"eval_samples_per_second": 46.483,
|
540 |
+
"eval_steps_per_second": 5.81,
|
541 |
+
"step": 336
|
542 |
+
},
|
543 |
+
{
|
544 |
+
"epoch": 7.08,
|
545 |
+
"grad_norm": 1.24147367477417,
|
546 |
+
"learning_rate": 7.062902532269188e-07,
|
547 |
+
"loss": 0.508,
|
548 |
+
"step": 340
|
549 |
+
},
|
550 |
+
{
|
551 |
+
"epoch": 7.19,
|
552 |
+
"grad_norm": 1.8932181596755981,
|
553 |
+
"learning_rate": 6.810656013259573e-07,
|
554 |
+
"loss": 0.6358,
|
555 |
+
"step": 345
|
556 |
+
},
|
557 |
+
{
|
558 |
+
"epoch": 7.29,
|
559 |
+
"grad_norm": 1.861436128616333,
|
560 |
+
"learning_rate": 6.55840949424996e-07,
|
561 |
+
"loss": 0.5741,
|
562 |
+
"step": 350
|
563 |
+
},
|
564 |
+
{
|
565 |
+
"epoch": 7.4,
|
566 |
+
"grad_norm": 0.8429200053215027,
|
567 |
+
"learning_rate": 6.306162975240346e-07,
|
568 |
+
"loss": 0.5717,
|
569 |
+
"step": 355
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 7.5,
|
573 |
+
"grad_norm": 1.8665741682052612,
|
574 |
+
"learning_rate": 6.053916456230732e-07,
|
575 |
+
"loss": 0.6992,
|
576 |
+
"step": 360
|
577 |
+
},
|
578 |
+
{
|
579 |
+
"epoch": 7.6,
|
580 |
+
"grad_norm": 2.312748908996582,
|
581 |
+
"learning_rate": 5.801669937221119e-07,
|
582 |
+
"loss": 0.6151,
|
583 |
+
"step": 365
|
584 |
+
},
|
585 |
+
{
|
586 |
+
"epoch": 7.71,
|
587 |
+
"grad_norm": 1.1628329753875732,
|
588 |
+
"learning_rate": 5.549423418211505e-07,
|
589 |
+
"loss": 0.5354,
|
590 |
+
"step": 370
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 7.81,
|
594 |
+
"grad_norm": 1.8674992322921753,
|
595 |
+
"learning_rate": 5.297176899201891e-07,
|
596 |
+
"loss": 0.6411,
|
597 |
+
"step": 375
|
598 |
+
},
|
599 |
+
{
|
600 |
+
"epoch": 7.92,
|
601 |
+
"grad_norm": 0.7112137675285339,
|
602 |
+
"learning_rate": 5.044930380192277e-07,
|
603 |
+
"loss": 0.6063,
|
604 |
+
"step": 380
|
605 |
+
},
|
606 |
+
{
|
607 |
+
"epoch": 8.0,
|
608 |
+
"eval_f1": 0.7326732673267327,
|
609 |
+
"eval_loss": 0.6909217834472656,
|
610 |
+
"eval_runtime": 1.3913,
|
611 |
+
"eval_samples_per_second": 45.999,
|
612 |
+
"eval_steps_per_second": 5.75,
|
613 |
+
"step": 384
|
614 |
+
}
|
615 |
+
],
|
616 |
+
"logging_steps": 5,
|
617 |
+
"max_steps": 480,
|
618 |
+
"num_input_tokens_seen": 0,
|
619 |
+
"num_train_epochs": 10,
|
620 |
+
"save_steps": 500,
|
621 |
+
"total_flos": 5808783041309760.0,
|
622 |
+
"train_batch_size": 4,
|
623 |
+
"trial_name": null,
|
624 |
+
"trial_params": {
|
625 |
+
"learning_rate": 2.1794099242430636e-06,
|
626 |
+
"per_device_train_batch_size": 4
|
627 |
+
}
|
628 |
+
}
|
run-4/checkpoint-384/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6eadd8415a707c8e4de7440233358b89448cac114a2ada36aef1d9186553c0c3
|
3 |
+
size 4920
|
run-4/checkpoint-432/config.json
ADDED
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "NOT_WORD",
|
53 |
+
"1": "WORD"
|
54 |
+
},
|
55 |
+
"initializer_range": 0.02,
|
56 |
+
"intermediate_size": 3072,
|
57 |
+
"label2id": {
|
58 |
+
"NOT_WORD": "0",
|
59 |
+
"WORD": "1"
|
60 |
+
},
|
61 |
+
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
+
"mask_feature_length": 10,
|
64 |
+
"mask_feature_min_masks": 0,
|
65 |
+
"mask_feature_prob": 0.0,
|
66 |
+
"mask_time_length": 10,
|
67 |
+
"mask_time_min_masks": 2,
|
68 |
+
"mask_time_prob": 0.05,
|
69 |
+
"model_type": "hubert",
|
70 |
+
"num_attention_heads": 12,
|
71 |
+
"num_conv_pos_embedding_groups": 16,
|
72 |
+
"num_conv_pos_embeddings": 128,
|
73 |
+
"num_feat_extract_layers": 7,
|
74 |
+
"num_hidden_layers": 2,
|
75 |
+
"pad_token_id": 0,
|
76 |
+
"torch_dtype": "float32",
|
77 |
+
"transformers_version": "4.38.1",
|
78 |
+
"use_weighted_layer_sum": false,
|
79 |
+
"vocab_size": 32
|
80 |
+
}
|