Hugging Face's logo

mehrshadk
/
electronics-DEBERTL-seed1759-lr3e-05

Safetensors
deberta
Model card Files Community
mehrshadk commited on
Commit
b0f8613
·
verified ·
1 Parent(s): 7b61729

Training in progress, epoch 4, checkpoint

Browse files
Files changed (4) hide show
  1. checkpoint-400/config.json +32 -0
  2. checkpoint-400/model.safetensors +3 -0
  3. checkpoint-400/trainer_state.json +218 -0
  4. checkpoint-400/training_args.bin +3 -0
checkpoint-400/config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "DebertaForSequenceClassification"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "hidden_act": "gelu",
7
+ "hidden_dropout_prob": 0.1,
8
+ "hidden_size": 1024,
9
+ "initializer_range": 0.02,
10
+ "intermediate_size": 4096,
11
+ "layer_norm_eps": 1e-07,
12
+ "legacy": true,
13
+ "max_position_embeddings": 512,
14
+ "max_relative_positions": -1,
15
+ "model_type": "deberta",
16
+ "num_attention_heads": 16,
17
+ "num_hidden_layers": 24,
18
+ "pad_token_id": 0,
19
+ "pooler_dropout": 0,
20
+ "pooler_hidden_act": "gelu",
21
+ "pooler_hidden_size": 1024,
22
+ "pos_att_type": [
23
+ "c2p",
24
+ "p2c"
25
+ ],
26
+ "position_biased_input": false,
27
+ "relative_attention": true,
28
+ "torch_dtype": "float32",
29
+ "transformers_version": "4.50.3",
30
+ "type_vocab_size": 0,
31
+ "vocab_size": 50265
32
+ }
checkpoint-400/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cdfe214b41badf0ded53e7fbeef9062d22d4e600859b5ba532187f8018a7e129
3
+ size 1624907048
checkpoint-400/trainer_state.json ADDED
@@ -0,0 +1,218 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 4.0,
6
+ "eval_steps": 50,
7
+ "global_step": 400,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.5,
14
+ "grad_norm": 2.202085256576538,
15
+ "learning_rate": 2.75e-05,
16
+ "loss": 0.7001,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.5,
21
+ "eval_electronics_AUC": 0.8274661508704062,
22
+ "eval_electronics_accuracy": 0.5875,
23
+ "eval_electronics_f1": 0.0,
24
+ "eval_electronics_loss": 0.7029502987861633,
25
+ "eval_electronics_macro_f1": 0.3700787401574803,
26
+ "eval_electronics_micro_f1": 0.5875,
27
+ "eval_electronics_precision": 0.0,
28
+ "eval_electronics_recall": 0.0,
29
+ "eval_electronics_runtime": 1.1511,
30
+ "eval_electronics_samples_per_second": 347.485,
31
+ "eval_electronics_specificity": 1.0,
32
+ "eval_electronics_steps_per_second": 0.869,
33
+ "step": 50
34
+ },
35
+ {
36
+ "epoch": 1.0,
37
+ "grad_norm": 5.9589738845825195,
38
+ "learning_rate": 2.5e-05,
39
+ "loss": 0.6259,
40
+ "step": 100
41
+ },
42
+ {
43
+ "epoch": 1.0,
44
+ "eval_electronics_AUC": 0.8812894906511928,
45
+ "eval_electronics_accuracy": 0.805,
46
+ "eval_electronics_f1": 0.7914438502673797,
47
+ "eval_electronics_loss": 0.44977375864982605,
48
+ "eval_electronics_macro_f1": 0.804172629359042,
49
+ "eval_electronics_micro_f1": 0.805,
50
+ "eval_electronics_precision": 0.7081339712918661,
51
+ "eval_electronics_recall": 0.896969696969697,
52
+ "eval_electronics_runtime": 1.1287,
53
+ "eval_electronics_samples_per_second": 354.379,
54
+ "eval_electronics_specificity": 0.7404255319148936,
55
+ "eval_electronics_steps_per_second": 0.886,
56
+ "step": 100
57
+ },
58
+ {
59
+ "epoch": 1.5,
60
+ "grad_norm": 15.954824447631836,
61
+ "learning_rate": 2.25e-05,
62
+ "loss": 0.5285,
63
+ "step": 150
64
+ },
65
+ {
66
+ "epoch": 1.5,
67
+ "eval_electronics_AUC": 0.8546228239845262,
68
+ "eval_electronics_accuracy": 0.75,
69
+ "eval_electronics_f1": 0.647887323943662,
70
+ "eval_electronics_loss": 0.5688791275024414,
71
+ "eval_electronics_macro_f1": 0.7270444371656295,
72
+ "eval_electronics_micro_f1": 0.75,
73
+ "eval_electronics_precision": 0.773109243697479,
74
+ "eval_electronics_recall": 0.5575757575757576,
75
+ "eval_electronics_runtime": 1.1417,
76
+ "eval_electronics_samples_per_second": 350.348,
77
+ "eval_electronics_specificity": 0.8851063829787233,
78
+ "eval_electronics_steps_per_second": 0.876,
79
+ "step": 150
80
+ },
81
+ {
82
+ "epoch": 2.0,
83
+ "grad_norm": 2.80739426612854,
84
+ "learning_rate": 1.9999999999999998e-05,
85
+ "loss": 0.492,
86
+ "step": 200
87
+ },
88
+ {
89
+ "epoch": 2.0,
90
+ "eval_electronics_AUC": 0.8878401031592521,
91
+ "eval_electronics_accuracy": 0.79,
92
+ "eval_electronics_f1": 0.7741935483870968,
93
+ "eval_electronics_loss": 0.5164185762405396,
94
+ "eval_electronics_macro_f1": 0.7889659330720531,
95
+ "eval_electronics_micro_f1": 0.79,
96
+ "eval_electronics_precision": 0.6956521739130435,
97
+ "eval_electronics_recall": 0.8727272727272727,
98
+ "eval_electronics_runtime": 1.1374,
99
+ "eval_electronics_samples_per_second": 351.668,
100
+ "eval_electronics_specificity": 0.7319148936170212,
101
+ "eval_electronics_steps_per_second": 0.879,
102
+ "step": 200
103
+ },
104
+ {
105
+ "epoch": 2.5,
106
+ "grad_norm": 48.75403594970703,
107
+ "learning_rate": 1.7500000000000002e-05,
108
+ "loss": 0.4042,
109
+ "step": 250
110
+ },
111
+ {
112
+ "epoch": 2.5,
113
+ "eval_electronics_AUC": 0.8810831721470018,
114
+ "eval_electronics_accuracy": 0.7525,
115
+ "eval_electronics_f1": 0.7579462102689487,
116
+ "eval_electronics_loss": 0.6417726874351501,
117
+ "eval_electronics_macro_f1": 0.7523746396613286,
118
+ "eval_electronics_micro_f1": 0.7525,
119
+ "eval_electronics_precision": 0.6352459016393442,
120
+ "eval_electronics_recall": 0.9393939393939394,
121
+ "eval_electronics_runtime": 1.176,
122
+ "eval_electronics_samples_per_second": 340.145,
123
+ "eval_electronics_specificity": 0.6212765957446809,
124
+ "eval_electronics_steps_per_second": 0.85,
125
+ "step": 250
126
+ },
127
+ {
128
+ "epoch": 3.0,
129
+ "grad_norm": 17.9029541015625,
130
+ "learning_rate": 1.5e-05,
131
+ "loss": 0.3676,
132
+ "step": 300
133
+ },
134
+ {
135
+ "epoch": 3.0,
136
+ "eval_electronics_AUC": 0.8895422308188267,
137
+ "eval_electronics_accuracy": 0.81,
138
+ "eval_electronics_f1": 0.7432432432432432,
139
+ "eval_electronics_loss": 0.5099908709526062,
140
+ "eval_electronics_macro_f1": 0.7962247962247961,
141
+ "eval_electronics_micro_f1": 0.81,
142
+ "eval_electronics_precision": 0.8396946564885496,
143
+ "eval_electronics_recall": 0.6666666666666666,
144
+ "eval_electronics_runtime": 1.1391,
145
+ "eval_electronics_samples_per_second": 351.16,
146
+ "eval_electronics_specificity": 0.9106382978723404,
147
+ "eval_electronics_steps_per_second": 0.878,
148
+ "step": 300
149
+ },
150
+ {
151
+ "epoch": 3.5,
152
+ "grad_norm": 0.9845223426818848,
153
+ "learning_rate": 1.25e-05,
154
+ "loss": 0.2304,
155
+ "step": 350
156
+ },
157
+ {
158
+ "epoch": 3.5,
159
+ "eval_electronics_AUC": 0.885880077369439,
160
+ "eval_electronics_accuracy": 0.8275,
161
+ "eval_electronics_f1": 0.7964601769911505,
162
+ "eval_electronics_loss": 0.5810662508010864,
163
+ "eval_electronics_macro_f1": 0.8233927783003474,
164
+ "eval_electronics_micro_f1": 0.8275,
165
+ "eval_electronics_precision": 0.7758620689655172,
166
+ "eval_electronics_recall": 0.8181818181818182,
167
+ "eval_electronics_runtime": 1.1575,
168
+ "eval_electronics_samples_per_second": 345.569,
169
+ "eval_electronics_specificity": 0.8340425531914893,
170
+ "eval_electronics_steps_per_second": 0.864,
171
+ "step": 350
172
+ },
173
+ {
174
+ "epoch": 4.0,
175
+ "grad_norm": 0.663503110408783,
176
+ "learning_rate": 9.999999999999999e-06,
177
+ "loss": 0.2429,
178
+ "step": 400
179
+ },
180
+ {
181
+ "epoch": 4.0,
182
+ "eval_electronics_AUC": 0.8947001934235976,
183
+ "eval_electronics_accuracy": 0.825,
184
+ "eval_electronics_f1": 0.7988505747126436,
185
+ "eval_electronics_loss": 0.4967746436595917,
186
+ "eval_electronics_macro_f1": 0.8219916590377377,
187
+ "eval_electronics_micro_f1": 0.825,
188
+ "eval_electronics_precision": 0.7595628415300546,
189
+ "eval_electronics_recall": 0.8424242424242424,
190
+ "eval_electronics_runtime": 1.1523,
191
+ "eval_electronics_samples_per_second": 347.119,
192
+ "eval_electronics_specificity": 0.8127659574468085,
193
+ "eval_electronics_steps_per_second": 0.868,
194
+ "step": 400
195
+ }
196
+ ],
197
+ "logging_steps": 50,
198
+ "max_steps": 600,
199
+ "num_input_tokens_seen": 0,
200
+ "num_train_epochs": 6,
201
+ "save_steps": 500,
202
+ "stateful_callbacks": {
203
+ "TrainerControl": {
204
+ "args": {
205
+ "should_epoch_stop": false,
206
+ "should_evaluate": false,
207
+ "should_log": false,
208
+ "should_save": true,
209
+ "should_training_stop": false
210
+ },
211
+ "attributes": {}
212
+ }
213
+ },
214
+ "total_flos": 543275092992000.0,
215
+ "train_batch_size": 16,
216
+ "trial_name": null,
217
+ "trial_params": null
218
+ }
checkpoint-400/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47871193157119808f8d00b7d97c1b2cdfffd44e2a841575fc1c03ac79100709
3
+ size 5432