xzuyn commited on
Commit
15a6148
1 Parent(s): 614990f

Upload folder using huggingface_hub

Browse files
checkpoint-410/adapter_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
3
+ "bias": "none",
4
+ "fan_in_fan_out": false,
5
+ "inference_mode": true,
6
+ "init_lora_weights": true,
7
+ "layers_pattern": null,
8
+ "layers_to_transform": null,
9
+ "lora_alpha": 32,
10
+ "lora_dropout": 0.05,
11
+ "modules_to_save": null,
12
+ "peft_type": "LORA",
13
+ "r": 32,
14
+ "revision": null,
15
+ "target_modules": [
16
+ "gate_proj",
17
+ "down_proj",
18
+ "up_proj",
19
+ "q_proj",
20
+ "k_proj",
21
+ "v_proj",
22
+ "o_proj"
23
+ ],
24
+ "task_type": "CAUSAL_LM"
25
+ }
checkpoint-410/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fcfbf03fca21d2d80a967e91b1c805ec200ddb7af96238b308763b2eb358049
3
+ size 319977674
checkpoint-410/adapter_model/adapter_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
3
+ "bias": "none",
4
+ "fan_in_fan_out": false,
5
+ "inference_mode": true,
6
+ "init_lora_weights": true,
7
+ "layers_pattern": null,
8
+ "layers_to_transform": null,
9
+ "lora_alpha": 32,
10
+ "lora_dropout": 0.05,
11
+ "modules_to_save": null,
12
+ "peft_type": "LORA",
13
+ "r": 32,
14
+ "revision": null,
15
+ "target_modules": [
16
+ "gate_proj",
17
+ "down_proj",
18
+ "up_proj",
19
+ "q_proj",
20
+ "k_proj",
21
+ "v_proj",
22
+ "o_proj"
23
+ ],
24
+ "task_type": "CAUSAL_LM"
25
+ }
checkpoint-410/adapter_model/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fcfbf03fca21d2d80a967e91b1c805ec200ddb7af96238b308763b2eb358049
3
+ size 319977674
checkpoint-410/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ae50a75ddce530ef6e552bf1b1cb570d73afa3f2adfe8e9a51211635ad6f3d0
3
+ size 160747348
checkpoint-410/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1477dd3a59216ce5e60159a0d861391519048d4ffabf96470998d686d2c57275
3
+ size 14244
checkpoint-410/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:168ad8ce09c2b97c59b93b875893cb36af25784d53cfba426deac2507fbb65c5
3
+ size 1064
checkpoint-410/trainer_state.json ADDED
@@ -0,0 +1,2476 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 9.553398058252426,
5
+ "global_step": 410,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 2.631578947368421e-06,
13
+ "loss": 1.9044,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 5.263157894736842e-06,
19
+ "loss": 2.018,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 7.894736842105263e-06,
25
+ "loss": 2.0221,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.09,
30
+ "learning_rate": 1.0526315789473684e-05,
31
+ "loss": 1.8884,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 1.3157894736842106e-05,
37
+ "loss": 1.9326,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.14,
42
+ "learning_rate": 1.5789473684210526e-05,
43
+ "loss": 2.1193,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.16,
48
+ "learning_rate": 1.8421052631578947e-05,
49
+ "loss": 2.0827,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.19,
54
+ "learning_rate": 2.105263157894737e-05,
55
+ "loss": 1.8846,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.21,
60
+ "learning_rate": 2.368421052631579e-05,
61
+ "loss": 2.0943,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.23,
66
+ "learning_rate": 2.6315789473684212e-05,
67
+ "loss": 2.0881,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.26,
72
+ "learning_rate": 2.8947368421052634e-05,
73
+ "loss": 1.9275,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.28,
78
+ "learning_rate": 3.157894736842105e-05,
79
+ "loss": 1.6874,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.3,
84
+ "learning_rate": 3.421052631578947e-05,
85
+ "loss": 1.9874,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.33,
90
+ "learning_rate": 3.6842105263157895e-05,
91
+ "loss": 1.8901,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.35,
96
+ "learning_rate": 3.9473684210526316e-05,
97
+ "loss": 1.7683,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.37,
102
+ "learning_rate": 4.210526315789474e-05,
103
+ "loss": 2.0108,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.4,
108
+ "learning_rate": 4.473684210526316e-05,
109
+ "loss": 1.6579,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.42,
114
+ "learning_rate": 4.736842105263158e-05,
115
+ "loss": 1.9758,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.44,
120
+ "learning_rate": 5e-05,
121
+ "loss": 1.9118,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.47,
126
+ "learning_rate": 5.2631578947368424e-05,
127
+ "loss": 1.8964,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.49,
132
+ "learning_rate": 5.526315789473685e-05,
133
+ "loss": 1.9288,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.51,
138
+ "learning_rate": 5.789473684210527e-05,
139
+ "loss": 1.8706,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.54,
144
+ "learning_rate": 6.052631578947369e-05,
145
+ "loss": 1.8604,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.56,
150
+ "learning_rate": 6.31578947368421e-05,
151
+ "loss": 1.9001,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.58,
156
+ "learning_rate": 6.578947368421054e-05,
157
+ "loss": 1.8254,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.61,
162
+ "learning_rate": 6.842105263157895e-05,
163
+ "loss": 1.8111,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.63,
168
+ "learning_rate": 7.105263157894737e-05,
169
+ "loss": 1.7496,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.65,
174
+ "learning_rate": 7.368421052631579e-05,
175
+ "loss": 1.7604,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.68,
180
+ "learning_rate": 7.631578947368422e-05,
181
+ "loss": 1.6285,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.7,
186
+ "learning_rate": 7.894736842105263e-05,
187
+ "loss": 1.7226,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.72,
192
+ "learning_rate": 8.157894736842105e-05,
193
+ "loss": 1.9689,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.75,
198
+ "learning_rate": 8.421052631578948e-05,
199
+ "loss": 1.7737,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.77,
204
+ "learning_rate": 8.68421052631579e-05,
205
+ "loss": 1.8093,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.79,
210
+ "learning_rate": 8.947368421052632e-05,
211
+ "loss": 1.6292,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.82,
216
+ "learning_rate": 9.210526315789474e-05,
217
+ "loss": 1.8351,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.84,
222
+ "learning_rate": 9.473684210526316e-05,
223
+ "loss": 1.918,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.86,
228
+ "learning_rate": 9.736842105263158e-05,
229
+ "loss": 1.8161,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.89,
234
+ "learning_rate": 0.0001,
235
+ "loss": 1.8184,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.91,
240
+ "learning_rate": 9.983108108108109e-05,
241
+ "loss": 1.5839,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.93,
246
+ "learning_rate": 9.966216216216217e-05,
247
+ "loss": 1.8736,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.96,
252
+ "learning_rate": 9.949324324324325e-05,
253
+ "loss": 1.77,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.98,
258
+ "learning_rate": 9.932432432432433e-05,
259
+ "loss": 1.7912,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 1.0,
264
+ "learning_rate": 9.915540540540541e-05,
265
+ "loss": 1.6989,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 1.03,
270
+ "learning_rate": 9.89864864864865e-05,
271
+ "loss": 1.7327,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 1.05,
276
+ "learning_rate": 9.881756756756757e-05,
277
+ "loss": 1.7799,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 1.07,
282
+ "learning_rate": 9.864864864864865e-05,
283
+ "loss": 1.8298,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 1.1,
288
+ "learning_rate": 9.847972972972973e-05,
289
+ "loss": 1.8616,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 1.12,
294
+ "learning_rate": 9.831081081081081e-05,
295
+ "loss": 1.7984,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 1.14,
300
+ "learning_rate": 9.81418918918919e-05,
301
+ "loss": 1.7626,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 1.17,
306
+ "learning_rate": 9.797297297297297e-05,
307
+ "loss": 1.8424,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 1.19,
312
+ "learning_rate": 9.780405405405407e-05,
313
+ "loss": 1.6436,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 1.21,
318
+ "learning_rate": 9.763513513513513e-05,
319
+ "loss": 1.6303,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 1.23,
324
+ "learning_rate": 9.746621621621623e-05,
325
+ "loss": 1.7199,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 1.26,
330
+ "learning_rate": 9.729729729729731e-05,
331
+ "loss": 1.6789,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 1.28,
336
+ "learning_rate": 9.712837837837838e-05,
337
+ "loss": 1.8519,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "learning_rate": 9.695945945945947e-05,
343
+ "loss": 1.7961,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 1.33,
348
+ "learning_rate": 9.679054054054054e-05,
349
+ "loss": 1.6725,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 1.35,
354
+ "learning_rate": 9.662162162162163e-05,
355
+ "loss": 1.7249,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 1.37,
360
+ "learning_rate": 9.64527027027027e-05,
361
+ "loss": 1.5996,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 1.4,
366
+ "learning_rate": 9.628378378378379e-05,
367
+ "loss": 1.7033,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 1.42,
372
+ "learning_rate": 9.611486486486487e-05,
373
+ "loss": 1.7387,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 1.44,
378
+ "learning_rate": 9.594594594594595e-05,
379
+ "loss": 1.7804,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 1.47,
384
+ "learning_rate": 9.577702702702703e-05,
385
+ "loss": 1.7484,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 1.49,
390
+ "learning_rate": 9.56081081081081e-05,
391
+ "loss": 1.7526,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 1.51,
396
+ "learning_rate": 9.54391891891892e-05,
397
+ "loss": 1.7247,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 1.54,
402
+ "learning_rate": 9.527027027027028e-05,
403
+ "loss": 1.6276,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 1.56,
408
+ "learning_rate": 9.510135135135136e-05,
409
+ "loss": 1.8279,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 1.58,
414
+ "learning_rate": 9.493243243243244e-05,
415
+ "loss": 1.7067,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 1.61,
420
+ "learning_rate": 9.476351351351352e-05,
421
+ "loss": 1.7865,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 1.63,
426
+ "learning_rate": 9.45945945945946e-05,
427
+ "loss": 1.6693,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 1.65,
432
+ "learning_rate": 9.442567567567568e-05,
433
+ "loss": 1.6939,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 1.68,
438
+ "learning_rate": 9.425675675675676e-05,
439
+ "loss": 1.8667,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 1.7,
444
+ "learning_rate": 9.408783783783784e-05,
445
+ "loss": 1.746,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 1.72,
450
+ "learning_rate": 9.391891891891892e-05,
451
+ "loss": 1.6951,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 1.75,
456
+ "learning_rate": 9.375e-05,
457
+ "loss": 1.6749,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 1.77,
462
+ "learning_rate": 9.35810810810811e-05,
463
+ "loss": 1.8773,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 1.79,
468
+ "learning_rate": 9.341216216216216e-05,
469
+ "loss": 1.8064,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 1.82,
474
+ "learning_rate": 9.324324324324324e-05,
475
+ "loss": 1.6596,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 1.84,
480
+ "learning_rate": 9.307432432432432e-05,
481
+ "loss": 1.7321,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 1.86,
486
+ "learning_rate": 9.29054054054054e-05,
487
+ "loss": 1.7029,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 1.89,
492
+ "learning_rate": 9.27364864864865e-05,
493
+ "loss": 1.7392,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 1.91,
498
+ "learning_rate": 9.256756756756757e-05,
499
+ "loss": 1.7904,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 1.93,
504
+ "learning_rate": 9.239864864864866e-05,
505
+ "loss": 1.6794,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 1.96,
510
+ "learning_rate": 9.222972972972973e-05,
511
+ "loss": 1.8827,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 1.98,
516
+ "learning_rate": 9.206081081081082e-05,
517
+ "loss": 1.7051,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 2.0,
522
+ "learning_rate": 9.18918918918919e-05,
523
+ "loss": 1.7069,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 2.03,
528
+ "learning_rate": 9.172297297297297e-05,
529
+ "loss": 1.8051,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 2.05,
534
+ "learning_rate": 9.155405405405406e-05,
535
+ "loss": 1.6665,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 2.07,
540
+ "learning_rate": 9.138513513513513e-05,
541
+ "loss": 1.6218,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 2.1,
546
+ "learning_rate": 9.121621621621623e-05,
547
+ "loss": 1.5848,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 2.12,
552
+ "learning_rate": 9.10472972972973e-05,
553
+ "loss": 1.7544,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 2.14,
558
+ "learning_rate": 9.087837837837839e-05,
559
+ "loss": 1.6454,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 2.17,
564
+ "learning_rate": 9.070945945945947e-05,
565
+ "loss": 1.8591,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 2.19,
570
+ "learning_rate": 9.054054054054055e-05,
571
+ "loss": 1.7025,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 2.21,
576
+ "learning_rate": 9.037162162162163e-05,
577
+ "loss": 1.7824,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 2.24,
582
+ "learning_rate": 9.02027027027027e-05,
583
+ "loss": 1.6049,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 2.26,
588
+ "learning_rate": 9.003378378378379e-05,
589
+ "loss": 1.6961,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 2.28,
594
+ "learning_rate": 8.986486486486487e-05,
595
+ "loss": 1.8705,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 2.31,
600
+ "learning_rate": 8.969594594594595e-05,
601
+ "loss": 1.4498,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 2.33,
606
+ "learning_rate": 8.952702702702703e-05,
607
+ "loss": 1.6967,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 2.35,
612
+ "learning_rate": 8.935810810810811e-05,
613
+ "loss": 1.6345,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 2.38,
618
+ "learning_rate": 8.918918918918919e-05,
619
+ "loss": 1.8037,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 2.4,
624
+ "learning_rate": 8.902027027027027e-05,
625
+ "loss": 1.7896,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 2.42,
630
+ "learning_rate": 8.885135135135135e-05,
631
+ "loss": 1.751,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 2.45,
636
+ "learning_rate": 8.868243243243243e-05,
637
+ "loss": 1.482,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 2.47,
642
+ "learning_rate": 8.851351351351352e-05,
643
+ "loss": 1.6049,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 2.49,
648
+ "learning_rate": 8.83445945945946e-05,
649
+ "loss": 1.6029,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 2.52,
654
+ "learning_rate": 8.817567567567569e-05,
655
+ "loss": 1.875,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 2.54,
660
+ "learning_rate": 8.800675675675676e-05,
661
+ "loss": 1.5798,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 2.56,
666
+ "learning_rate": 8.783783783783784e-05,
667
+ "loss": 1.4883,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 2.59,
672
+ "learning_rate": 8.766891891891892e-05,
673
+ "loss": 1.7249,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 2.61,
678
+ "learning_rate": 8.75e-05,
679
+ "loss": 1.5086,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 2.63,
684
+ "learning_rate": 8.73310810810811e-05,
685
+ "loss": 1.39,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 2.66,
690
+ "learning_rate": 8.716216216216216e-05,
691
+ "loss": 1.7481,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 2.68,
696
+ "learning_rate": 8.699324324324325e-05,
697
+ "loss": 1.6514,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 2.7,
702
+ "learning_rate": 8.682432432432432e-05,
703
+ "loss": 1.8731,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 2.73,
708
+ "learning_rate": 8.665540540540542e-05,
709
+ "loss": 1.6415,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 2.75,
714
+ "learning_rate": 8.64864864864865e-05,
715
+ "loss": 1.7659,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 2.77,
720
+ "learning_rate": 8.631756756756756e-05,
721
+ "loss": 1.5691,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 2.8,
726
+ "learning_rate": 8.614864864864866e-05,
727
+ "loss": 1.6315,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 2.82,
732
+ "learning_rate": 8.597972972972972e-05,
733
+ "loss": 1.367,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 2.84,
738
+ "learning_rate": 8.581081081081082e-05,
739
+ "loss": 1.6088,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 2.87,
744
+ "learning_rate": 8.56418918918919e-05,
745
+ "loss": 1.7598,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 2.89,
750
+ "learning_rate": 8.547297297297298e-05,
751
+ "loss": 1.6255,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 2.91,
756
+ "learning_rate": 8.530405405405406e-05,
757
+ "loss": 1.5739,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 2.94,
762
+ "learning_rate": 8.513513513513514e-05,
763
+ "loss": 1.5542,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 2.96,
768
+ "learning_rate": 8.496621621621622e-05,
769
+ "loss": 1.5758,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 2.98,
774
+ "learning_rate": 8.47972972972973e-05,
775
+ "loss": 1.668,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 3.01,
780
+ "learning_rate": 8.462837837837838e-05,
781
+ "loss": 1.6283,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 3.03,
786
+ "learning_rate": 8.445945945945946e-05,
787
+ "loss": 1.6111,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 3.05,
792
+ "learning_rate": 8.429054054054054e-05,
793
+ "loss": 1.8061,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 3.08,
798
+ "learning_rate": 8.412162162162163e-05,
799
+ "loss": 1.4882,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 3.1,
804
+ "learning_rate": 8.39527027027027e-05,
805
+ "loss": 1.592,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 3.12,
810
+ "learning_rate": 8.378378378378379e-05,
811
+ "loss": 1.638,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 3.15,
816
+ "learning_rate": 8.361486486486487e-05,
817
+ "loss": 1.5036,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 3.17,
822
+ "learning_rate": 8.344594594594595e-05,
823
+ "loss": 1.6721,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 3.19,
828
+ "learning_rate": 8.327702702702703e-05,
829
+ "loss": 1.7742,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 3.22,
834
+ "learning_rate": 8.310810810810811e-05,
835
+ "loss": 1.484,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 3.24,
840
+ "learning_rate": 8.293918918918919e-05,
841
+ "loss": 1.3698,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 3.26,
846
+ "learning_rate": 8.277027027027028e-05,
847
+ "loss": 1.5644,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 3.29,
852
+ "learning_rate": 8.260135135135135e-05,
853
+ "loss": 1.6195,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 3.31,
858
+ "learning_rate": 8.243243243243243e-05,
859
+ "loss": 1.5031,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 3.33,
864
+ "learning_rate": 8.226351351351351e-05,
865
+ "loss": 1.5403,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 3.36,
870
+ "learning_rate": 8.209459459459459e-05,
871
+ "loss": 1.57,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 3.38,
876
+ "learning_rate": 8.192567567567569e-05,
877
+ "loss": 1.6454,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 3.4,
882
+ "learning_rate": 8.175675675675675e-05,
883
+ "loss": 1.5069,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 3.43,
888
+ "learning_rate": 8.158783783783785e-05,
889
+ "loss": 1.3782,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 3.45,
894
+ "learning_rate": 8.141891891891892e-05,
895
+ "loss": 1.4296,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 3.47,
900
+ "learning_rate": 8.125000000000001e-05,
901
+ "loss": 1.4303,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 3.5,
906
+ "learning_rate": 8.108108108108109e-05,
907
+ "loss": 1.6368,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 3.52,
912
+ "learning_rate": 8.091216216216216e-05,
913
+ "loss": 1.5951,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 3.54,
918
+ "learning_rate": 8.074324324324325e-05,
919
+ "loss": 1.5769,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 3.57,
924
+ "learning_rate": 8.057432432432432e-05,
925
+ "loss": 1.482,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 3.59,
930
+ "learning_rate": 8.040540540540541e-05,
931
+ "loss": 1.5415,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 3.61,
936
+ "learning_rate": 8.02364864864865e-05,
937
+ "loss": 1.6084,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 3.63,
942
+ "learning_rate": 8.006756756756757e-05,
943
+ "loss": 1.5483,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 3.66,
948
+ "learning_rate": 7.989864864864865e-05,
949
+ "loss": 1.4329,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 3.68,
954
+ "learning_rate": 7.972972972972974e-05,
955
+ "loss": 1.6068,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 3.7,
960
+ "learning_rate": 7.956081081081082e-05,
961
+ "loss": 1.5519,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 3.73,
966
+ "learning_rate": 7.93918918918919e-05,
967
+ "loss": 1.7387,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 3.75,
972
+ "learning_rate": 7.922297297297298e-05,
973
+ "loss": 1.5583,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 3.77,
978
+ "learning_rate": 7.905405405405406e-05,
979
+ "loss": 1.5003,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 3.8,
984
+ "learning_rate": 7.888513513513514e-05,
985
+ "loss": 1.5025,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 3.82,
990
+ "learning_rate": 7.871621621621622e-05,
991
+ "loss": 1.5322,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 3.84,
996
+ "learning_rate": 7.85472972972973e-05,
997
+ "loss": 1.6915,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 3.87,
1002
+ "learning_rate": 7.837837837837838e-05,
1003
+ "loss": 1.4906,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 3.89,
1008
+ "learning_rate": 7.820945945945946e-05,
1009
+ "loss": 1.4796,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 3.91,
1014
+ "learning_rate": 7.804054054054054e-05,
1015
+ "loss": 1.5712,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 3.94,
1020
+ "learning_rate": 7.787162162162162e-05,
1021
+ "loss": 1.459,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 3.96,
1026
+ "learning_rate": 7.77027027027027e-05,
1027
+ "loss": 1.4917,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 3.98,
1032
+ "learning_rate": 7.753378378378378e-05,
1033
+ "loss": 1.5149,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 4.01,
1038
+ "learning_rate": 7.736486486486488e-05,
1039
+ "loss": 1.4344,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 4.03,
1044
+ "learning_rate": 7.719594594594595e-05,
1045
+ "loss": 1.4152,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 4.05,
1050
+ "learning_rate": 7.702702702702703e-05,
1051
+ "loss": 1.3729,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 4.08,
1056
+ "learning_rate": 7.68581081081081e-05,
1057
+ "loss": 1.4462,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 4.1,
1062
+ "learning_rate": 7.668918918918919e-05,
1063
+ "loss": 1.5449,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 4.12,
1068
+ "learning_rate": 7.652027027027028e-05,
1069
+ "loss": 1.2145,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 4.15,
1074
+ "learning_rate": 7.635135135135135e-05,
1075
+ "loss": 1.4754,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 4.17,
1080
+ "learning_rate": 7.618243243243244e-05,
1081
+ "loss": 1.3256,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 4.19,
1086
+ "learning_rate": 7.601351351351351e-05,
1087
+ "loss": 1.3754,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 4.22,
1092
+ "learning_rate": 7.58445945945946e-05,
1093
+ "loss": 1.4267,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 4.24,
1098
+ "learning_rate": 7.567567567567568e-05,
1099
+ "loss": 1.4389,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 4.26,
1104
+ "learning_rate": 7.550675675675675e-05,
1105
+ "loss": 1.4266,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 4.29,
1110
+ "learning_rate": 7.533783783783785e-05,
1111
+ "loss": 1.4098,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 4.31,
1116
+ "learning_rate": 7.516891891891891e-05,
1117
+ "loss": 1.0715,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 4.33,
1122
+ "learning_rate": 7.500000000000001e-05,
1123
+ "loss": 1.4547,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 4.36,
1128
+ "learning_rate": 7.483108108108109e-05,
1129
+ "loss": 1.5758,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 4.38,
1134
+ "learning_rate": 7.466216216216217e-05,
1135
+ "loss": 1.3765,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 4.4,
1140
+ "learning_rate": 7.449324324324325e-05,
1141
+ "loss": 1.4207,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 4.43,
1146
+ "learning_rate": 7.432432432432433e-05,
1147
+ "loss": 1.4264,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 4.45,
1152
+ "learning_rate": 7.415540540540541e-05,
1153
+ "loss": 1.2975,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 4.47,
1158
+ "learning_rate": 7.398648648648649e-05,
1159
+ "loss": 1.5207,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 4.5,
1164
+ "learning_rate": 7.381756756756757e-05,
1165
+ "loss": 1.5805,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 4.52,
1170
+ "learning_rate": 7.364864864864865e-05,
1171
+ "loss": 1.3546,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 4.54,
1176
+ "learning_rate": 7.347972972972973e-05,
1177
+ "loss": 1.6047,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 4.57,
1182
+ "learning_rate": 7.331081081081081e-05,
1183
+ "loss": 1.4313,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 4.59,
1188
+ "learning_rate": 7.31418918918919e-05,
1189
+ "loss": 1.371,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 4.61,
1194
+ "learning_rate": 7.297297297297297e-05,
1195
+ "loss": 1.4083,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 4.64,
1200
+ "learning_rate": 7.280405405405406e-05,
1201
+ "loss": 1.6423,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 4.66,
1206
+ "learning_rate": 7.263513513513514e-05,
1207
+ "loss": 1.4229,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 4.68,
1212
+ "learning_rate": 7.246621621621622e-05,
1213
+ "loss": 1.4711,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 4.71,
1218
+ "learning_rate": 7.229729729729731e-05,
1219
+ "loss": 1.3806,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 4.73,
1224
+ "learning_rate": 7.212837837837838e-05,
1225
+ "loss": 1.4641,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 4.75,
1230
+ "learning_rate": 7.195945945945947e-05,
1231
+ "loss": 1.6534,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 4.78,
1236
+ "learning_rate": 7.179054054054054e-05,
1237
+ "loss": 1.3054,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 4.8,
1242
+ "learning_rate": 7.162162162162162e-05,
1243
+ "loss": 1.4058,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 4.82,
1248
+ "learning_rate": 7.14527027027027e-05,
1249
+ "loss": 1.5076,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 4.85,
1254
+ "learning_rate": 7.128378378378378e-05,
1255
+ "loss": 1.5715,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 4.87,
1260
+ "learning_rate": 7.111486486486488e-05,
1261
+ "loss": 1.4723,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 4.89,
1266
+ "learning_rate": 7.094594594594594e-05,
1267
+ "loss": 1.3515,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 4.92,
1272
+ "learning_rate": 7.077702702702704e-05,
1273
+ "loss": 1.2815,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 4.94,
1278
+ "learning_rate": 7.06081081081081e-05,
1279
+ "loss": 1.2335,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 4.96,
1284
+ "learning_rate": 7.04391891891892e-05,
1285
+ "loss": 1.2843,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 4.99,
1290
+ "learning_rate": 7.027027027027028e-05,
1291
+ "loss": 1.518,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 5.01,
1296
+ "learning_rate": 7.010135135135135e-05,
1297
+ "loss": 1.3281,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 5.03,
1302
+ "learning_rate": 6.993243243243244e-05,
1303
+ "loss": 1.3571,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 5.06,
1308
+ "learning_rate": 6.97635135135135e-05,
1309
+ "loss": 1.2656,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 5.08,
1314
+ "learning_rate": 6.95945945945946e-05,
1315
+ "loss": 1.3251,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 5.1,
1320
+ "learning_rate": 6.942567567567568e-05,
1321
+ "loss": 1.2875,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 5.13,
1326
+ "learning_rate": 6.925675675675676e-05,
1327
+ "loss": 1.4429,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 5.15,
1332
+ "learning_rate": 6.908783783783784e-05,
1333
+ "loss": 1.1838,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 5.17,
1338
+ "learning_rate": 6.891891891891892e-05,
1339
+ "loss": 1.1694,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 5.2,
1344
+ "learning_rate": 6.875e-05,
1345
+ "loss": 1.2747,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 5.22,
1350
+ "learning_rate": 6.858108108108108e-05,
1351
+ "loss": 1.3806,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 5.24,
1356
+ "learning_rate": 6.841216216216217e-05,
1357
+ "loss": 1.1208,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 5.27,
1362
+ "learning_rate": 6.824324324324325e-05,
1363
+ "loss": 1.4053,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 5.29,
1368
+ "learning_rate": 6.807432432432433e-05,
1369
+ "loss": 1.3146,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 5.31,
1374
+ "learning_rate": 6.790540540540541e-05,
1375
+ "loss": 1.3767,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 5.34,
1380
+ "learning_rate": 6.773648648648649e-05,
1381
+ "loss": 1.2501,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 5.36,
1386
+ "learning_rate": 6.756756756756757e-05,
1387
+ "loss": 1.3812,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 5.38,
1392
+ "learning_rate": 6.739864864864865e-05,
1393
+ "loss": 1.1772,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 5.41,
1398
+ "learning_rate": 6.722972972972973e-05,
1399
+ "loss": 1.3066,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 5.43,
1404
+ "learning_rate": 6.706081081081081e-05,
1405
+ "loss": 1.3026,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 5.45,
1410
+ "learning_rate": 6.68918918918919e-05,
1411
+ "loss": 1.2601,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 5.48,
1416
+ "learning_rate": 6.672297297297297e-05,
1417
+ "loss": 1.1287,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 5.5,
1422
+ "learning_rate": 6.655405405405407e-05,
1423
+ "loss": 1.4575,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 5.52,
1428
+ "learning_rate": 6.638513513513513e-05,
1429
+ "loss": 1.2495,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 5.55,
1434
+ "learning_rate": 6.621621621621621e-05,
1435
+ "loss": 1.1763,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 5.57,
1440
+ "learning_rate": 6.604729729729731e-05,
1441
+ "loss": 1.2768,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 5.59,
1446
+ "learning_rate": 6.587837837837837e-05,
1447
+ "loss": 1.3268,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 5.62,
1452
+ "learning_rate": 6.570945945945947e-05,
1453
+ "loss": 1.2082,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 5.64,
1458
+ "learning_rate": 6.554054054054054e-05,
1459
+ "loss": 1.3312,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 5.66,
1464
+ "learning_rate": 6.537162162162163e-05,
1465
+ "loss": 1.3582,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 5.69,
1470
+ "learning_rate": 6.52027027027027e-05,
1471
+ "loss": 1.4009,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 5.71,
1476
+ "learning_rate": 6.503378378378379e-05,
1477
+ "loss": 1.3129,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 5.73,
1482
+ "learning_rate": 6.486486486486487e-05,
1483
+ "loss": 1.0997,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 5.76,
1488
+ "learning_rate": 6.469594594594594e-05,
1489
+ "loss": 1.3952,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 5.78,
1494
+ "learning_rate": 6.452702702702703e-05,
1495
+ "loss": 1.328,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 5.8,
1500
+ "learning_rate": 6.43581081081081e-05,
1501
+ "loss": 1.3402,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 5.83,
1506
+ "learning_rate": 6.41891891891892e-05,
1507
+ "loss": 1.6361,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 5.85,
1512
+ "learning_rate": 6.402027027027028e-05,
1513
+ "loss": 1.2595,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 5.87,
1518
+ "learning_rate": 6.385135135135136e-05,
1519
+ "loss": 1.2653,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 5.9,
1524
+ "learning_rate": 6.368243243243244e-05,
1525
+ "loss": 1.2824,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 5.92,
1530
+ "learning_rate": 6.351351351351352e-05,
1531
+ "loss": 1.3233,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 5.94,
1536
+ "learning_rate": 6.33445945945946e-05,
1537
+ "loss": 1.1506,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 5.97,
1542
+ "learning_rate": 6.317567567567568e-05,
1543
+ "loss": 1.1696,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 5.99,
1548
+ "learning_rate": 6.300675675675676e-05,
1549
+ "loss": 1.3595,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 6.01,
1554
+ "learning_rate": 6.283783783783784e-05,
1555
+ "loss": 1.2825,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 6.03,
1560
+ "learning_rate": 6.266891891891892e-05,
1561
+ "loss": 1.4051,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 6.06,
1566
+ "learning_rate": 6.25e-05,
1567
+ "loss": 1.2085,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 6.08,
1572
+ "learning_rate": 6.233108108108108e-05,
1573
+ "loss": 1.2621,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 6.1,
1578
+ "learning_rate": 6.216216216216216e-05,
1579
+ "loss": 1.293,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 6.13,
1584
+ "learning_rate": 6.199324324324324e-05,
1585
+ "loss": 1.2971,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 6.15,
1590
+ "learning_rate": 6.182432432432432e-05,
1591
+ "loss": 1.0583,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 6.17,
1596
+ "learning_rate": 6.16554054054054e-05,
1597
+ "loss": 1.225,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 6.2,
1602
+ "learning_rate": 6.14864864864865e-05,
1603
+ "loss": 1.0673,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 6.22,
1608
+ "learning_rate": 6.131756756756757e-05,
1609
+ "loss": 1.1158,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 6.24,
1614
+ "learning_rate": 6.114864864864866e-05,
1615
+ "loss": 1.1983,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 6.27,
1620
+ "learning_rate": 6.097972972972973e-05,
1621
+ "loss": 1.1234,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 6.29,
1626
+ "learning_rate": 6.0810810810810814e-05,
1627
+ "loss": 1.2147,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 6.31,
1632
+ "learning_rate": 6.06418918918919e-05,
1633
+ "loss": 1.0961,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 6.34,
1638
+ "learning_rate": 6.0472972972972976e-05,
1639
+ "loss": 1.2271,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 6.36,
1644
+ "learning_rate": 6.030405405405406e-05,
1645
+ "loss": 0.8272,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 6.38,
1650
+ "learning_rate": 6.013513513513514e-05,
1651
+ "loss": 1.0314,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 6.41,
1656
+ "learning_rate": 5.996621621621622e-05,
1657
+ "loss": 1.172,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 6.43,
1662
+ "learning_rate": 5.9797297297297305e-05,
1663
+ "loss": 1.1037,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 6.45,
1668
+ "learning_rate": 5.962837837837838e-05,
1669
+ "loss": 1.0629,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 6.48,
1674
+ "learning_rate": 5.9459459459459466e-05,
1675
+ "loss": 1.1261,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 6.5,
1680
+ "learning_rate": 5.929054054054054e-05,
1681
+ "loss": 1.2574,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 6.52,
1686
+ "learning_rate": 5.912162162162163e-05,
1687
+ "loss": 1.3313,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 6.55,
1692
+ "learning_rate": 5.89527027027027e-05,
1693
+ "loss": 0.924,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 6.57,
1698
+ "learning_rate": 5.878378378378379e-05,
1699
+ "loss": 1.1797,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 6.59,
1704
+ "learning_rate": 5.861486486486487e-05,
1705
+ "loss": 1.1554,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 6.62,
1710
+ "learning_rate": 5.8445945945945943e-05,
1711
+ "loss": 1.2582,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 6.64,
1716
+ "learning_rate": 5.827702702702703e-05,
1717
+ "loss": 1.2177,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 6.66,
1722
+ "learning_rate": 5.8108108108108105e-05,
1723
+ "loss": 1.1448,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 6.69,
1728
+ "learning_rate": 5.793918918918919e-05,
1729
+ "loss": 1.1025,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 6.71,
1734
+ "learning_rate": 5.777027027027028e-05,
1735
+ "loss": 1.1173,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 6.73,
1740
+ "learning_rate": 5.760135135135135e-05,
1741
+ "loss": 1.1789,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 6.76,
1746
+ "learning_rate": 5.7432432432432434e-05,
1747
+ "loss": 1.3601,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 6.78,
1752
+ "learning_rate": 5.7263513513513515e-05,
1753
+ "loss": 1.2766,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 6.8,
1758
+ "learning_rate": 5.7094594594594595e-05,
1759
+ "loss": 1.2541,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 6.83,
1764
+ "learning_rate": 5.692567567567568e-05,
1765
+ "loss": 1.2084,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 6.85,
1770
+ "learning_rate": 5.6756756756756757e-05,
1771
+ "loss": 1.3129,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 6.87,
1776
+ "learning_rate": 5.6587837837837844e-05,
1777
+ "loss": 1.1758,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 6.9,
1782
+ "learning_rate": 5.641891891891892e-05,
1783
+ "loss": 1.341,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 6.92,
1788
+ "learning_rate": 5.6250000000000005e-05,
1789
+ "loss": 1.3618,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 6.94,
1794
+ "learning_rate": 5.6081081081081086e-05,
1795
+ "loss": 1.0368,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 6.97,
1800
+ "learning_rate": 5.591216216216216e-05,
1801
+ "loss": 1.1718,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 6.99,
1806
+ "learning_rate": 5.574324324324325e-05,
1807
+ "loss": 1.1364,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 7.01,
1812
+ "learning_rate": 5.557432432432432e-05,
1813
+ "loss": 1.1075,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 7.04,
1818
+ "learning_rate": 5.540540540540541e-05,
1819
+ "loss": 1.2325,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 7.06,
1824
+ "learning_rate": 5.5236486486486496e-05,
1825
+ "loss": 0.8635,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 7.08,
1830
+ "learning_rate": 5.506756756756757e-05,
1831
+ "loss": 0.7712,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 7.11,
1836
+ "learning_rate": 5.489864864864866e-05,
1837
+ "loss": 0.9811,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 7.13,
1842
+ "learning_rate": 5.472972972972973e-05,
1843
+ "loss": 1.1597,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 7.15,
1848
+ "learning_rate": 5.456081081081081e-05,
1849
+ "loss": 0.9311,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 7.18,
1854
+ "learning_rate": 5.43918918918919e-05,
1855
+ "loss": 1.1239,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 7.2,
1860
+ "learning_rate": 5.422297297297297e-05,
1861
+ "loss": 1.2107,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 7.22,
1866
+ "learning_rate": 5.405405405405406e-05,
1867
+ "loss": 1.2178,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 7.25,
1872
+ "learning_rate": 5.3885135135135134e-05,
1873
+ "loss": 1.1905,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 7.27,
1878
+ "learning_rate": 5.371621621621622e-05,
1879
+ "loss": 1.2219,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 7.29,
1884
+ "learning_rate": 5.35472972972973e-05,
1885
+ "loss": 1.0516,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 7.32,
1890
+ "learning_rate": 5.337837837837838e-05,
1891
+ "loss": 1.0192,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 7.34,
1896
+ "learning_rate": 5.3209459459459463e-05,
1897
+ "loss": 0.9752,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 7.36,
1902
+ "learning_rate": 5.304054054054054e-05,
1903
+ "loss": 0.9826,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 7.39,
1908
+ "learning_rate": 5.2871621621621625e-05,
1909
+ "loss": 1.3142,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 7.41,
1914
+ "learning_rate": 5.27027027027027e-05,
1915
+ "loss": 1.1342,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 7.43,
1920
+ "learning_rate": 5.2533783783783786e-05,
1921
+ "loss": 1.1025,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 7.46,
1926
+ "learning_rate": 5.2364864864864873e-05,
1927
+ "loss": 1.2068,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 7.48,
1932
+ "learning_rate": 5.219594594594595e-05,
1933
+ "loss": 1.0683,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 7.5,
1938
+ "learning_rate": 5.202702702702703e-05,
1939
+ "loss": 0.859,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 7.53,
1944
+ "learning_rate": 5.185810810810811e-05,
1945
+ "loss": 1.0902,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 7.55,
1950
+ "learning_rate": 5.168918918918919e-05,
1951
+ "loss": 1.0506,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 7.57,
1956
+ "learning_rate": 5.152027027027028e-05,
1957
+ "loss": 1.019,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 7.6,
1962
+ "learning_rate": 5.135135135135135e-05,
1963
+ "loss": 0.6924,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 7.62,
1968
+ "learning_rate": 5.118243243243244e-05,
1969
+ "loss": 1.1495,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 7.64,
1974
+ "learning_rate": 5.101351351351351e-05,
1975
+ "loss": 1.0561,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 7.67,
1980
+ "learning_rate": 5.08445945945946e-05,
1981
+ "loss": 1.1057,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 7.69,
1986
+ "learning_rate": 5.067567567567568e-05,
1987
+ "loss": 1.0621,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 7.71,
1992
+ "learning_rate": 5.0506756756756754e-05,
1993
+ "loss": 1.0147,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 7.74,
1998
+ "learning_rate": 5.033783783783784e-05,
1999
+ "loss": 1.0904,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 7.76,
2004
+ "learning_rate": 5.0168918918918915e-05,
2005
+ "loss": 1.1281,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 7.78,
2010
+ "learning_rate": 5e-05,
2011
+ "loss": 1.0421,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 7.81,
2016
+ "learning_rate": 4.983108108108108e-05,
2017
+ "loss": 1.1218,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 7.83,
2022
+ "learning_rate": 4.9662162162162164e-05,
2023
+ "loss": 1.0854,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 7.85,
2028
+ "learning_rate": 4.949324324324325e-05,
2029
+ "loss": 1.1577,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 7.88,
2034
+ "learning_rate": 4.9324324324324325e-05,
2035
+ "loss": 1.078,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 7.9,
2040
+ "learning_rate": 4.9155405405405406e-05,
2041
+ "loss": 1.1829,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 7.92,
2046
+ "learning_rate": 4.8986486486486486e-05,
2047
+ "loss": 1.09,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 7.95,
2052
+ "learning_rate": 4.881756756756757e-05,
2053
+ "loss": 0.9816,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 7.97,
2058
+ "learning_rate": 4.8648648648648654e-05,
2059
+ "loss": 0.9653,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 7.99,
2064
+ "learning_rate": 4.8479729729729735e-05,
2065
+ "loss": 1.1243,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 8.02,
2070
+ "learning_rate": 4.8310810810810816e-05,
2071
+ "loss": 0.9883,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 8.04,
2076
+ "learning_rate": 4.8141891891891896e-05,
2077
+ "loss": 0.8322,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 8.06,
2082
+ "learning_rate": 4.797297297297298e-05,
2083
+ "loss": 0.9729,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 8.09,
2088
+ "learning_rate": 4.780405405405405e-05,
2089
+ "loss": 1.0465,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 8.11,
2094
+ "learning_rate": 4.763513513513514e-05,
2095
+ "loss": 0.8629,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 8.13,
2100
+ "learning_rate": 4.746621621621622e-05,
2101
+ "loss": 1.15,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 8.16,
2106
+ "learning_rate": 4.72972972972973e-05,
2107
+ "loss": 0.9984,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 8.18,
2112
+ "learning_rate": 4.712837837837838e-05,
2113
+ "loss": 0.939,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 8.2,
2118
+ "learning_rate": 4.695945945945946e-05,
2119
+ "loss": 0.9231,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 8.23,
2124
+ "learning_rate": 4.679054054054055e-05,
2125
+ "loss": 1.1219,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 8.25,
2130
+ "learning_rate": 4.662162162162162e-05,
2131
+ "loss": 1.1236,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 8.27,
2136
+ "learning_rate": 4.64527027027027e-05,
2137
+ "loss": 0.9729,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 8.3,
2142
+ "learning_rate": 4.628378378378378e-05,
2143
+ "loss": 1.0206,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 8.32,
2148
+ "learning_rate": 4.6114864864864864e-05,
2149
+ "loss": 1.1789,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 8.34,
2154
+ "learning_rate": 4.594594594594595e-05,
2155
+ "loss": 0.9932,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 8.37,
2160
+ "learning_rate": 4.577702702702703e-05,
2161
+ "loss": 1.0049,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 8.39,
2166
+ "learning_rate": 4.560810810810811e-05,
2167
+ "loss": 1.0889,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 8.41,
2172
+ "learning_rate": 4.543918918918919e-05,
2173
+ "loss": 0.8475,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 8.43,
2178
+ "learning_rate": 4.5270270270270274e-05,
2179
+ "loss": 1.0319,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 8.46,
2184
+ "learning_rate": 4.510135135135135e-05,
2185
+ "loss": 0.9336,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 8.48,
2190
+ "learning_rate": 4.4932432432432435e-05,
2191
+ "loss": 0.9678,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 8.5,
2196
+ "learning_rate": 4.4763513513513516e-05,
2197
+ "loss": 0.8758,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 8.53,
2202
+ "learning_rate": 4.4594594594594596e-05,
2203
+ "loss": 0.9342,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 8.55,
2208
+ "learning_rate": 4.442567567567568e-05,
2209
+ "loss": 1.1054,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 8.57,
2214
+ "learning_rate": 4.425675675675676e-05,
2215
+ "loss": 0.8595,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 8.6,
2220
+ "learning_rate": 4.4087837837837845e-05,
2221
+ "loss": 0.9209,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 8.62,
2226
+ "learning_rate": 4.391891891891892e-05,
2227
+ "loss": 0.9508,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 8.64,
2232
+ "learning_rate": 4.375e-05,
2233
+ "loss": 0.9705,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 8.67,
2238
+ "learning_rate": 4.358108108108108e-05,
2239
+ "loss": 0.9421,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 8.69,
2244
+ "learning_rate": 4.341216216216216e-05,
2245
+ "loss": 1.0525,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 8.71,
2250
+ "learning_rate": 4.324324324324325e-05,
2251
+ "loss": 0.9324,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 8.74,
2256
+ "learning_rate": 4.307432432432433e-05,
2257
+ "loss": 0.9463,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 8.76,
2262
+ "learning_rate": 4.290540540540541e-05,
2263
+ "loss": 0.7919,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 8.78,
2268
+ "learning_rate": 4.273648648648649e-05,
2269
+ "loss": 1.0147,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 8.81,
2274
+ "learning_rate": 4.256756756756757e-05,
2275
+ "loss": 0.9019,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 8.83,
2280
+ "learning_rate": 4.239864864864865e-05,
2281
+ "loss": 0.8662,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 8.85,
2286
+ "learning_rate": 4.222972972972973e-05,
2287
+ "loss": 0.9442,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 8.88,
2292
+ "learning_rate": 4.206081081081081e-05,
2293
+ "loss": 1.1238,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 8.9,
2298
+ "learning_rate": 4.189189189189189e-05,
2299
+ "loss": 1.0176,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 8.92,
2304
+ "learning_rate": 4.1722972972972974e-05,
2305
+ "loss": 1.1425,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 8.95,
2310
+ "learning_rate": 4.1554054054054055e-05,
2311
+ "loss": 0.7894,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 8.97,
2316
+ "learning_rate": 4.138513513513514e-05,
2317
+ "loss": 1.0243,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 8.99,
2322
+ "learning_rate": 4.1216216216216216e-05,
2323
+ "loss": 1.1354,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 9.02,
2328
+ "learning_rate": 4.1047297297297297e-05,
2329
+ "loss": 0.8477,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 9.04,
2334
+ "learning_rate": 4.087837837837838e-05,
2335
+ "loss": 0.9073,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 9.06,
2340
+ "learning_rate": 4.070945945945946e-05,
2341
+ "loss": 1.0169,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 9.09,
2346
+ "learning_rate": 4.0540540540540545e-05,
2347
+ "loss": 0.7684,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 9.11,
2352
+ "learning_rate": 4.0371621621621626e-05,
2353
+ "loss": 1.1133,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 9.13,
2358
+ "learning_rate": 4.0202702702702707e-05,
2359
+ "loss": 0.9712,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 9.16,
2364
+ "learning_rate": 4.003378378378379e-05,
2365
+ "loss": 0.9081,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 9.18,
2370
+ "learning_rate": 3.986486486486487e-05,
2371
+ "loss": 0.788,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 9.2,
2376
+ "learning_rate": 3.969594594594595e-05,
2377
+ "loss": 1.0385,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 9.23,
2382
+ "learning_rate": 3.952702702702703e-05,
2383
+ "loss": 1.1285,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 9.25,
2388
+ "learning_rate": 3.935810810810811e-05,
2389
+ "loss": 0.7035,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 9.27,
2394
+ "learning_rate": 3.918918918918919e-05,
2395
+ "loss": 0.9234,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 9.3,
2400
+ "learning_rate": 3.902027027027027e-05,
2401
+ "loss": 1.0529,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 9.32,
2406
+ "learning_rate": 3.885135135135135e-05,
2407
+ "loss": 0.7203,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 9.34,
2412
+ "learning_rate": 3.868243243243244e-05,
2413
+ "loss": 0.7964,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 9.37,
2418
+ "learning_rate": 3.851351351351351e-05,
2419
+ "loss": 0.8433,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 9.39,
2424
+ "learning_rate": 3.8344594594594594e-05,
2425
+ "loss": 0.9627,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 9.41,
2430
+ "learning_rate": 3.8175675675675674e-05,
2431
+ "loss": 0.7189,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 9.44,
2436
+ "learning_rate": 3.8006756756756755e-05,
2437
+ "loss": 0.9945,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 9.46,
2442
+ "learning_rate": 3.783783783783784e-05,
2443
+ "loss": 0.9255,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 9.48,
2448
+ "learning_rate": 3.766891891891892e-05,
2449
+ "loss": 0.8539,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 9.51,
2454
+ "learning_rate": 3.7500000000000003e-05,
2455
+ "loss": 0.8397,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 9.53,
2460
+ "learning_rate": 3.7331081081081084e-05,
2461
+ "loss": 0.7326,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 9.55,
2466
+ "learning_rate": 3.7162162162162165e-05,
2467
+ "loss": 0.8779,
2468
+ "step": 410
2469
+ }
2470
+ ],
2471
+ "max_steps": 630,
2472
+ "num_train_epochs": 15,
2473
+ "total_flos": 1.4824693973424538e+17,
2474
+ "trial_name": null,
2475
+ "trial_params": null
2476
+ }
checkpoint-410/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba2fe107dab361c51487de4fc8c9ce45e8134a0096f693cda683706e603b37d9
3
+ size 4344