gustavomalkomes commited on
Commit
7210494
·
verified ·
1 Parent(s): 8b70c95

End of training

Browse files
README.md CHANGED
@@ -3,7 +3,11 @@ library_name: transformers
3
  license: apache-2.0
4
  base_model: google/vit-base-patch16-224-in21k
5
  tags:
 
 
6
  - generated_from_trainer
 
 
7
  model-index:
8
  - name: vit-base-patch16-224-in21k
9
  results: []
@@ -14,7 +18,13 @@ should probably proofread and complete it, then remove this comment. -->
14
 
15
  # vit-base-patch16-224-in21k
16
 
17
- This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
 
 
 
 
 
 
18
 
19
  ## Model description
20
 
 
3
  license: apache-2.0
4
  base_model: google/vit-base-patch16-224-in21k
5
  tags:
6
+ - image-classification
7
+ - vision
8
  - generated_from_trainer
9
+ metrics:
10
+ - accuracy
11
  model-index:
12
  - name: vit-base-patch16-224-in21k
13
  results: []
 
18
 
19
  # vit-base-patch16-224-in21k
20
 
21
+ This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the chainyo/rvl-cdip dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: nan
24
+ - Accuracy: 0.0630
25
+ - Memory Allocated (gb): 1.7
26
+ - Max Memory Allocated (gb): 2.31
27
+ - Total Memory Available (gb): 126.62
28
 
29
  ## Model description
30
 
all_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "eval_accuracy": 0.06302083333333333,
4
+ "eval_loss": NaN,
5
+ "eval_runtime": 245.7518,
6
+ "eval_samples_per_second": 195.319,
7
+ "eval_steps_per_second": 3.052,
8
+ "max_memory_allocated (GB)": 2.31,
9
+ "memory_allocated (GB)": 1.7,
10
+ "total_flos": 6.324139790696448e+19,
11
+ "total_memory_available (GB)": 126.62,
12
+ "train_loss": NaN,
13
+ "train_runtime": 3737.7146,
14
+ "train_samples_per_second": 218.314,
15
+ "train_steps_per_second": 27.289
16
+ }
eval_results.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "eval_accuracy": 0.06302083333333333,
4
+ "eval_loss": NaN,
5
+ "eval_runtime": 245.7518,
6
+ "eval_samples_per_second": 195.319,
7
+ "eval_steps_per_second": 3.052,
8
+ "max_memory_allocated (GB)": 2.31,
9
+ "memory_allocated (GB)": 1.7,
10
+ "total_memory_available (GB)": 126.62
11
+ }
runs/Nov06_03-48-21_gtown-28NZK54/events.out.tfevents.1730869348.gtown-28NZK54.57.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d65270cc4bb9aa3f20bf81fe2af91f0090df4ede74e046f08c1f7e33fcaa9f44
3
+ size 580
train_results.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "max_memory_allocated (GB)": 2.31,
4
+ "memory_allocated (GB)": 1.67,
5
+ "total_flos": 6.324139790696448e+19,
6
+ "total_memory_available (GB)": 126.62,
7
+ "train_loss": NaN,
8
+ "train_runtime": 3737.7146,
9
+ "train_samples_per_second": 218.314,
10
+ "train_steps_per_second": 27.289
11
+ }
trainer_state.json ADDED
@@ -0,0 +1,2085 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "eval_steps": 500,
6
+ "global_step": 102000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.014705882352941176,
13
+ "grad_norm": NaN,
14
+ "learning_rate": 4.975490196078432e-05,
15
+ "loss": NaN,
16
+ "max_memory_allocated (GB)": 2.31,
17
+ "memory_allocated (GB)": 1.67,
18
+ "step": 500,
19
+ "total_memory_available (GB)": 126.62
20
+ },
21
+ {
22
+ "epoch": 0.029411764705882353,
23
+ "grad_norm": NaN,
24
+ "learning_rate": 4.9509803921568634e-05,
25
+ "loss": NaN,
26
+ "max_memory_allocated (GB)": 2.31,
27
+ "memory_allocated (GB)": 1.67,
28
+ "step": 1000,
29
+ "total_memory_available (GB)": 126.62
30
+ },
31
+ {
32
+ "epoch": 0.04411764705882353,
33
+ "grad_norm": NaN,
34
+ "learning_rate": 4.9264705882352944e-05,
35
+ "loss": NaN,
36
+ "max_memory_allocated (GB)": 2.31,
37
+ "memory_allocated (GB)": 1.67,
38
+ "step": 1500,
39
+ "total_memory_available (GB)": 126.62
40
+ },
41
+ {
42
+ "epoch": 0.058823529411764705,
43
+ "grad_norm": NaN,
44
+ "learning_rate": 4.901960784313725e-05,
45
+ "loss": NaN,
46
+ "max_memory_allocated (GB)": 2.31,
47
+ "memory_allocated (GB)": 1.67,
48
+ "step": 2000,
49
+ "total_memory_available (GB)": 126.62
50
+ },
51
+ {
52
+ "epoch": 0.07352941176470588,
53
+ "grad_norm": NaN,
54
+ "learning_rate": 4.877450980392157e-05,
55
+ "loss": NaN,
56
+ "max_memory_allocated (GB)": 2.31,
57
+ "memory_allocated (GB)": 1.67,
58
+ "step": 2500,
59
+ "total_memory_available (GB)": 126.62
60
+ },
61
+ {
62
+ "epoch": 0.08823529411764706,
63
+ "grad_norm": NaN,
64
+ "learning_rate": 4.8529411764705885e-05,
65
+ "loss": NaN,
66
+ "max_memory_allocated (GB)": 2.31,
67
+ "memory_allocated (GB)": 1.67,
68
+ "step": 3000,
69
+ "total_memory_available (GB)": 126.62
70
+ },
71
+ {
72
+ "epoch": 0.10294117647058823,
73
+ "grad_norm": NaN,
74
+ "learning_rate": 4.82843137254902e-05,
75
+ "loss": NaN,
76
+ "max_memory_allocated (GB)": 2.31,
77
+ "memory_allocated (GB)": 1.67,
78
+ "step": 3500,
79
+ "total_memory_available (GB)": 126.62
80
+ },
81
+ {
82
+ "epoch": 0.11764705882352941,
83
+ "grad_norm": NaN,
84
+ "learning_rate": 4.803921568627452e-05,
85
+ "loss": NaN,
86
+ "max_memory_allocated (GB)": 2.31,
87
+ "memory_allocated (GB)": 1.67,
88
+ "step": 4000,
89
+ "total_memory_available (GB)": 126.62
90
+ },
91
+ {
92
+ "epoch": 0.1323529411764706,
93
+ "grad_norm": NaN,
94
+ "learning_rate": 4.7794117647058826e-05,
95
+ "loss": NaN,
96
+ "max_memory_allocated (GB)": 2.31,
97
+ "memory_allocated (GB)": 1.67,
98
+ "step": 4500,
99
+ "total_memory_available (GB)": 126.62
100
+ },
101
+ {
102
+ "epoch": 0.14705882352941177,
103
+ "grad_norm": NaN,
104
+ "learning_rate": 4.7549019607843135e-05,
105
+ "loss": NaN,
106
+ "max_memory_allocated (GB)": 2.31,
107
+ "memory_allocated (GB)": 1.67,
108
+ "step": 5000,
109
+ "total_memory_available (GB)": 126.62
110
+ },
111
+ {
112
+ "epoch": 0.16176470588235295,
113
+ "grad_norm": NaN,
114
+ "learning_rate": 4.730392156862745e-05,
115
+ "loss": NaN,
116
+ "max_memory_allocated (GB)": 2.31,
117
+ "memory_allocated (GB)": 1.67,
118
+ "step": 5500,
119
+ "total_memory_available (GB)": 126.62
120
+ },
121
+ {
122
+ "epoch": 0.17647058823529413,
123
+ "grad_norm": NaN,
124
+ "learning_rate": 4.705882352941177e-05,
125
+ "loss": NaN,
126
+ "max_memory_allocated (GB)": 2.31,
127
+ "memory_allocated (GB)": 1.67,
128
+ "step": 6000,
129
+ "total_memory_available (GB)": 126.62
130
+ },
131
+ {
132
+ "epoch": 0.19117647058823528,
133
+ "grad_norm": NaN,
134
+ "learning_rate": 4.681372549019608e-05,
135
+ "loss": NaN,
136
+ "max_memory_allocated (GB)": 2.31,
137
+ "memory_allocated (GB)": 1.67,
138
+ "step": 6500,
139
+ "total_memory_available (GB)": 126.62
140
+ },
141
+ {
142
+ "epoch": 0.20588235294117646,
143
+ "grad_norm": NaN,
144
+ "learning_rate": 4.656862745098039e-05,
145
+ "loss": NaN,
146
+ "max_memory_allocated (GB)": 2.31,
147
+ "memory_allocated (GB)": 1.67,
148
+ "step": 7000,
149
+ "total_memory_available (GB)": 126.62
150
+ },
151
+ {
152
+ "epoch": 0.22058823529411764,
153
+ "grad_norm": NaN,
154
+ "learning_rate": 4.632352941176471e-05,
155
+ "loss": NaN,
156
+ "max_memory_allocated (GB)": 2.31,
157
+ "memory_allocated (GB)": 1.67,
158
+ "step": 7500,
159
+ "total_memory_available (GB)": 126.62
160
+ },
161
+ {
162
+ "epoch": 0.23529411764705882,
163
+ "grad_norm": NaN,
164
+ "learning_rate": 4.607843137254902e-05,
165
+ "loss": NaN,
166
+ "max_memory_allocated (GB)": 2.31,
167
+ "memory_allocated (GB)": 1.67,
168
+ "step": 8000,
169
+ "total_memory_available (GB)": 126.62
170
+ },
171
+ {
172
+ "epoch": 0.25,
173
+ "grad_norm": NaN,
174
+ "learning_rate": 4.5833333333333334e-05,
175
+ "loss": NaN,
176
+ "max_memory_allocated (GB)": 2.31,
177
+ "memory_allocated (GB)": 1.67,
178
+ "step": 8500,
179
+ "total_memory_available (GB)": 126.62
180
+ },
181
+ {
182
+ "epoch": 0.2647058823529412,
183
+ "grad_norm": NaN,
184
+ "learning_rate": 4.558823529411765e-05,
185
+ "loss": NaN,
186
+ "max_memory_allocated (GB)": 2.31,
187
+ "memory_allocated (GB)": 1.67,
188
+ "step": 9000,
189
+ "total_memory_available (GB)": 126.62
190
+ },
191
+ {
192
+ "epoch": 0.27941176470588236,
193
+ "grad_norm": NaN,
194
+ "learning_rate": 4.5343137254901966e-05,
195
+ "loss": NaN,
196
+ "max_memory_allocated (GB)": 2.31,
197
+ "memory_allocated (GB)": 1.67,
198
+ "step": 9500,
199
+ "total_memory_available (GB)": 126.62
200
+ },
201
+ {
202
+ "epoch": 0.29411764705882354,
203
+ "grad_norm": NaN,
204
+ "learning_rate": 4.5098039215686275e-05,
205
+ "loss": NaN,
206
+ "max_memory_allocated (GB)": 2.31,
207
+ "memory_allocated (GB)": 1.67,
208
+ "step": 10000,
209
+ "total_memory_available (GB)": 126.62
210
+ },
211
+ {
212
+ "epoch": 0.3088235294117647,
213
+ "grad_norm": NaN,
214
+ "learning_rate": 4.485294117647059e-05,
215
+ "loss": NaN,
216
+ "max_memory_allocated (GB)": 2.31,
217
+ "memory_allocated (GB)": 1.67,
218
+ "step": 10500,
219
+ "total_memory_available (GB)": 126.62
220
+ },
221
+ {
222
+ "epoch": 0.3235294117647059,
223
+ "grad_norm": NaN,
224
+ "learning_rate": 4.460784313725491e-05,
225
+ "loss": NaN,
226
+ "max_memory_allocated (GB)": 2.31,
227
+ "memory_allocated (GB)": 1.67,
228
+ "step": 11000,
229
+ "total_memory_available (GB)": 126.62
230
+ },
231
+ {
232
+ "epoch": 0.3382352941176471,
233
+ "grad_norm": NaN,
234
+ "learning_rate": 4.4362745098039216e-05,
235
+ "loss": NaN,
236
+ "max_memory_allocated (GB)": 2.31,
237
+ "memory_allocated (GB)": 1.67,
238
+ "step": 11500,
239
+ "total_memory_available (GB)": 126.62
240
+ },
241
+ {
242
+ "epoch": 0.35294117647058826,
243
+ "grad_norm": NaN,
244
+ "learning_rate": 4.411764705882353e-05,
245
+ "loss": NaN,
246
+ "max_memory_allocated (GB)": 2.31,
247
+ "memory_allocated (GB)": 1.67,
248
+ "step": 12000,
249
+ "total_memory_available (GB)": 126.62
250
+ },
251
+ {
252
+ "epoch": 0.36764705882352944,
253
+ "grad_norm": NaN,
254
+ "learning_rate": 4.387254901960784e-05,
255
+ "loss": NaN,
256
+ "max_memory_allocated (GB)": 2.31,
257
+ "memory_allocated (GB)": 1.67,
258
+ "step": 12500,
259
+ "total_memory_available (GB)": 126.62
260
+ },
261
+ {
262
+ "epoch": 0.38235294117647056,
263
+ "grad_norm": NaN,
264
+ "learning_rate": 4.362745098039216e-05,
265
+ "loss": NaN,
266
+ "max_memory_allocated (GB)": 2.31,
267
+ "memory_allocated (GB)": 1.67,
268
+ "step": 13000,
269
+ "total_memory_available (GB)": 126.62
270
+ },
271
+ {
272
+ "epoch": 0.39705882352941174,
273
+ "grad_norm": NaN,
274
+ "learning_rate": 4.3382352941176474e-05,
275
+ "loss": NaN,
276
+ "max_memory_allocated (GB)": 2.31,
277
+ "memory_allocated (GB)": 1.67,
278
+ "step": 13500,
279
+ "total_memory_available (GB)": 126.62
280
+ },
281
+ {
282
+ "epoch": 0.4117647058823529,
283
+ "grad_norm": NaN,
284
+ "learning_rate": 4.313725490196079e-05,
285
+ "loss": NaN,
286
+ "max_memory_allocated (GB)": 2.31,
287
+ "memory_allocated (GB)": 1.67,
288
+ "step": 14000,
289
+ "total_memory_available (GB)": 126.62
290
+ },
291
+ {
292
+ "epoch": 0.4264705882352941,
293
+ "grad_norm": NaN,
294
+ "learning_rate": 4.28921568627451e-05,
295
+ "loss": NaN,
296
+ "max_memory_allocated (GB)": 2.31,
297
+ "memory_allocated (GB)": 1.67,
298
+ "step": 14500,
299
+ "total_memory_available (GB)": 126.62
300
+ },
301
+ {
302
+ "epoch": 0.4411764705882353,
303
+ "grad_norm": NaN,
304
+ "learning_rate": 4.2647058823529415e-05,
305
+ "loss": NaN,
306
+ "max_memory_allocated (GB)": 2.31,
307
+ "memory_allocated (GB)": 1.67,
308
+ "step": 15000,
309
+ "total_memory_available (GB)": 126.62
310
+ },
311
+ {
312
+ "epoch": 0.45588235294117646,
313
+ "grad_norm": NaN,
314
+ "learning_rate": 4.2401960784313724e-05,
315
+ "loss": NaN,
316
+ "max_memory_allocated (GB)": 2.31,
317
+ "memory_allocated (GB)": 1.67,
318
+ "step": 15500,
319
+ "total_memory_available (GB)": 126.62
320
+ },
321
+ {
322
+ "epoch": 0.47058823529411764,
323
+ "grad_norm": NaN,
324
+ "learning_rate": 4.215686274509804e-05,
325
+ "loss": NaN,
326
+ "max_memory_allocated (GB)": 2.31,
327
+ "memory_allocated (GB)": 1.67,
328
+ "step": 16000,
329
+ "total_memory_available (GB)": 126.62
330
+ },
331
+ {
332
+ "epoch": 0.4852941176470588,
333
+ "grad_norm": NaN,
334
+ "learning_rate": 4.1911764705882356e-05,
335
+ "loss": NaN,
336
+ "max_memory_allocated (GB)": 2.31,
337
+ "memory_allocated (GB)": 1.67,
338
+ "step": 16500,
339
+ "total_memory_available (GB)": 126.62
340
+ },
341
+ {
342
+ "epoch": 0.5,
343
+ "grad_norm": NaN,
344
+ "learning_rate": 4.166666666666667e-05,
345
+ "loss": NaN,
346
+ "max_memory_allocated (GB)": 2.31,
347
+ "memory_allocated (GB)": 1.67,
348
+ "step": 17000,
349
+ "total_memory_available (GB)": 126.62
350
+ },
351
+ {
352
+ "epoch": 0.5147058823529411,
353
+ "grad_norm": NaN,
354
+ "learning_rate": 4.142156862745099e-05,
355
+ "loss": NaN,
356
+ "max_memory_allocated (GB)": 2.31,
357
+ "memory_allocated (GB)": 1.67,
358
+ "step": 17500,
359
+ "total_memory_available (GB)": 126.62
360
+ },
361
+ {
362
+ "epoch": 0.5294117647058824,
363
+ "grad_norm": NaN,
364
+ "learning_rate": 4.11764705882353e-05,
365
+ "loss": NaN,
366
+ "max_memory_allocated (GB)": 2.31,
367
+ "memory_allocated (GB)": 1.67,
368
+ "step": 18000,
369
+ "total_memory_available (GB)": 126.62
370
+ },
371
+ {
372
+ "epoch": 0.5441176470588235,
373
+ "grad_norm": NaN,
374
+ "learning_rate": 4.0931372549019607e-05,
375
+ "loss": NaN,
376
+ "max_memory_allocated (GB)": 2.31,
377
+ "memory_allocated (GB)": 1.67,
378
+ "step": 18500,
379
+ "total_memory_available (GB)": 126.62
380
+ },
381
+ {
382
+ "epoch": 0.5588235294117647,
383
+ "grad_norm": NaN,
384
+ "learning_rate": 4.068627450980392e-05,
385
+ "loss": NaN,
386
+ "max_memory_allocated (GB)": 2.31,
387
+ "memory_allocated (GB)": 1.67,
388
+ "step": 19000,
389
+ "total_memory_available (GB)": 126.62
390
+ },
391
+ {
392
+ "epoch": 0.5735294117647058,
393
+ "grad_norm": NaN,
394
+ "learning_rate": 4.044117647058824e-05,
395
+ "loss": NaN,
396
+ "max_memory_allocated (GB)": 2.31,
397
+ "memory_allocated (GB)": 1.67,
398
+ "step": 19500,
399
+ "total_memory_available (GB)": 126.62
400
+ },
401
+ {
402
+ "epoch": 0.5882352941176471,
403
+ "grad_norm": NaN,
404
+ "learning_rate": 4.0196078431372555e-05,
405
+ "loss": NaN,
406
+ "max_memory_allocated (GB)": 2.31,
407
+ "memory_allocated (GB)": 1.67,
408
+ "step": 20000,
409
+ "total_memory_available (GB)": 126.62
410
+ },
411
+ {
412
+ "epoch": 0.6029411764705882,
413
+ "grad_norm": NaN,
414
+ "learning_rate": 3.9950980392156864e-05,
415
+ "loss": NaN,
416
+ "max_memory_allocated (GB)": 2.31,
417
+ "memory_allocated (GB)": 1.67,
418
+ "step": 20500,
419
+ "total_memory_available (GB)": 126.62
420
+ },
421
+ {
422
+ "epoch": 0.6176470588235294,
423
+ "grad_norm": NaN,
424
+ "learning_rate": 3.970588235294117e-05,
425
+ "loss": NaN,
426
+ "max_memory_allocated (GB)": 2.31,
427
+ "memory_allocated (GB)": 1.67,
428
+ "step": 21000,
429
+ "total_memory_available (GB)": 126.62
430
+ },
431
+ {
432
+ "epoch": 0.6323529411764706,
433
+ "grad_norm": NaN,
434
+ "learning_rate": 3.946078431372549e-05,
435
+ "loss": NaN,
436
+ "max_memory_allocated (GB)": 2.31,
437
+ "memory_allocated (GB)": 1.67,
438
+ "step": 21500,
439
+ "total_memory_available (GB)": 126.62
440
+ },
441
+ {
442
+ "epoch": 0.6470588235294118,
443
+ "grad_norm": NaN,
444
+ "learning_rate": 3.9215686274509805e-05,
445
+ "loss": NaN,
446
+ "max_memory_allocated (GB)": 2.31,
447
+ "memory_allocated (GB)": 1.67,
448
+ "step": 22000,
449
+ "total_memory_available (GB)": 126.62
450
+ },
451
+ {
452
+ "epoch": 0.6617647058823529,
453
+ "grad_norm": NaN,
454
+ "learning_rate": 3.897058823529412e-05,
455
+ "loss": NaN,
456
+ "max_memory_allocated (GB)": 2.31,
457
+ "memory_allocated (GB)": 1.67,
458
+ "step": 22500,
459
+ "total_memory_available (GB)": 126.62
460
+ },
461
+ {
462
+ "epoch": 0.6764705882352942,
463
+ "grad_norm": NaN,
464
+ "learning_rate": 3.872549019607844e-05,
465
+ "loss": NaN,
466
+ "max_memory_allocated (GB)": 2.31,
467
+ "memory_allocated (GB)": 1.67,
468
+ "step": 23000,
469
+ "total_memory_available (GB)": 126.62
470
+ },
471
+ {
472
+ "epoch": 0.6911764705882353,
473
+ "grad_norm": NaN,
474
+ "learning_rate": 3.8480392156862746e-05,
475
+ "loss": NaN,
476
+ "max_memory_allocated (GB)": 2.31,
477
+ "memory_allocated (GB)": 1.67,
478
+ "step": 23500,
479
+ "total_memory_available (GB)": 126.62
480
+ },
481
+ {
482
+ "epoch": 0.7058823529411765,
483
+ "grad_norm": NaN,
484
+ "learning_rate": 3.8235294117647055e-05,
485
+ "loss": NaN,
486
+ "max_memory_allocated (GB)": 2.31,
487
+ "memory_allocated (GB)": 1.67,
488
+ "step": 24000,
489
+ "total_memory_available (GB)": 126.62
490
+ },
491
+ {
492
+ "epoch": 0.7205882352941176,
493
+ "grad_norm": NaN,
494
+ "learning_rate": 3.799019607843137e-05,
495
+ "loss": NaN,
496
+ "max_memory_allocated (GB)": 2.31,
497
+ "memory_allocated (GB)": 1.67,
498
+ "step": 24500,
499
+ "total_memory_available (GB)": 126.62
500
+ },
501
+ {
502
+ "epoch": 0.7352941176470589,
503
+ "grad_norm": NaN,
504
+ "learning_rate": 3.774509803921569e-05,
505
+ "loss": NaN,
506
+ "max_memory_allocated (GB)": 2.31,
507
+ "memory_allocated (GB)": 1.67,
508
+ "step": 25000,
509
+ "total_memory_available (GB)": 126.62
510
+ },
511
+ {
512
+ "epoch": 0.75,
513
+ "grad_norm": NaN,
514
+ "learning_rate": 3.7500000000000003e-05,
515
+ "loss": NaN,
516
+ "max_memory_allocated (GB)": 2.31,
517
+ "memory_allocated (GB)": 1.67,
518
+ "step": 25500,
519
+ "total_memory_available (GB)": 126.62
520
+ },
521
+ {
522
+ "epoch": 0.7647058823529411,
523
+ "grad_norm": NaN,
524
+ "learning_rate": 3.725490196078432e-05,
525
+ "loss": NaN,
526
+ "max_memory_allocated (GB)": 2.31,
527
+ "memory_allocated (GB)": 1.67,
528
+ "step": 26000,
529
+ "total_memory_available (GB)": 126.62
530
+ },
531
+ {
532
+ "epoch": 0.7794117647058824,
533
+ "grad_norm": NaN,
534
+ "learning_rate": 3.700980392156863e-05,
535
+ "loss": NaN,
536
+ "max_memory_allocated (GB)": 2.31,
537
+ "memory_allocated (GB)": 1.67,
538
+ "step": 26500,
539
+ "total_memory_available (GB)": 126.62
540
+ },
541
+ {
542
+ "epoch": 0.7941176470588235,
543
+ "grad_norm": NaN,
544
+ "learning_rate": 3.6764705882352945e-05,
545
+ "loss": NaN,
546
+ "max_memory_allocated (GB)": 2.31,
547
+ "memory_allocated (GB)": 1.67,
548
+ "step": 27000,
549
+ "total_memory_available (GB)": 126.62
550
+ },
551
+ {
552
+ "epoch": 0.8088235294117647,
553
+ "grad_norm": NaN,
554
+ "learning_rate": 3.6519607843137254e-05,
555
+ "loss": NaN,
556
+ "max_memory_allocated (GB)": 2.31,
557
+ "memory_allocated (GB)": 1.67,
558
+ "step": 27500,
559
+ "total_memory_available (GB)": 126.62
560
+ },
561
+ {
562
+ "epoch": 0.8235294117647058,
563
+ "grad_norm": NaN,
564
+ "learning_rate": 3.627450980392157e-05,
565
+ "loss": NaN,
566
+ "max_memory_allocated (GB)": 2.31,
567
+ "memory_allocated (GB)": 1.67,
568
+ "step": 28000,
569
+ "total_memory_available (GB)": 126.62
570
+ },
571
+ {
572
+ "epoch": 0.8382352941176471,
573
+ "grad_norm": NaN,
574
+ "learning_rate": 3.6029411764705886e-05,
575
+ "loss": NaN,
576
+ "max_memory_allocated (GB)": 2.31,
577
+ "memory_allocated (GB)": 1.67,
578
+ "step": 28500,
579
+ "total_memory_available (GB)": 126.62
580
+ },
581
+ {
582
+ "epoch": 0.8529411764705882,
583
+ "grad_norm": NaN,
584
+ "learning_rate": 3.5784313725490195e-05,
585
+ "loss": NaN,
586
+ "max_memory_allocated (GB)": 2.31,
587
+ "memory_allocated (GB)": 1.67,
588
+ "step": 29000,
589
+ "total_memory_available (GB)": 126.62
590
+ },
591
+ {
592
+ "epoch": 0.8676470588235294,
593
+ "grad_norm": NaN,
594
+ "learning_rate": 3.553921568627451e-05,
595
+ "loss": NaN,
596
+ "max_memory_allocated (GB)": 2.31,
597
+ "memory_allocated (GB)": 1.67,
598
+ "step": 29500,
599
+ "total_memory_available (GB)": 126.62
600
+ },
601
+ {
602
+ "epoch": 0.8823529411764706,
603
+ "grad_norm": NaN,
604
+ "learning_rate": 3.529411764705883e-05,
605
+ "loss": NaN,
606
+ "max_memory_allocated (GB)": 2.31,
607
+ "memory_allocated (GB)": 1.67,
608
+ "step": 30000,
609
+ "total_memory_available (GB)": 126.62
610
+ },
611
+ {
612
+ "epoch": 0.8970588235294118,
613
+ "grad_norm": NaN,
614
+ "learning_rate": 3.5049019607843136e-05,
615
+ "loss": NaN,
616
+ "max_memory_allocated (GB)": 2.31,
617
+ "memory_allocated (GB)": 1.67,
618
+ "step": 30500,
619
+ "total_memory_available (GB)": 126.62
620
+ },
621
+ {
622
+ "epoch": 0.9117647058823529,
623
+ "grad_norm": NaN,
624
+ "learning_rate": 3.480392156862745e-05,
625
+ "loss": NaN,
626
+ "max_memory_allocated (GB)": 2.31,
627
+ "memory_allocated (GB)": 1.67,
628
+ "step": 31000,
629
+ "total_memory_available (GB)": 126.62
630
+ },
631
+ {
632
+ "epoch": 0.9264705882352942,
633
+ "grad_norm": NaN,
634
+ "learning_rate": 3.455882352941177e-05,
635
+ "loss": NaN,
636
+ "max_memory_allocated (GB)": 2.31,
637
+ "memory_allocated (GB)": 1.67,
638
+ "step": 31500,
639
+ "total_memory_available (GB)": 126.62
640
+ },
641
+ {
642
+ "epoch": 0.9411764705882353,
643
+ "grad_norm": NaN,
644
+ "learning_rate": 3.431372549019608e-05,
645
+ "loss": NaN,
646
+ "max_memory_allocated (GB)": 2.31,
647
+ "memory_allocated (GB)": 1.67,
648
+ "step": 32000,
649
+ "total_memory_available (GB)": 126.62
650
+ },
651
+ {
652
+ "epoch": 0.9558823529411765,
653
+ "grad_norm": NaN,
654
+ "learning_rate": 3.4068627450980394e-05,
655
+ "loss": NaN,
656
+ "max_memory_allocated (GB)": 2.31,
657
+ "memory_allocated (GB)": 1.67,
658
+ "step": 32500,
659
+ "total_memory_available (GB)": 126.62
660
+ },
661
+ {
662
+ "epoch": 0.9705882352941176,
663
+ "grad_norm": NaN,
664
+ "learning_rate": 3.382352941176471e-05,
665
+ "loss": NaN,
666
+ "max_memory_allocated (GB)": 2.31,
667
+ "memory_allocated (GB)": 1.67,
668
+ "step": 33000,
669
+ "total_memory_available (GB)": 126.62
670
+ },
671
+ {
672
+ "epoch": 0.9852941176470589,
673
+ "grad_norm": NaN,
674
+ "learning_rate": 3.357843137254902e-05,
675
+ "loss": NaN,
676
+ "max_memory_allocated (GB)": 2.31,
677
+ "memory_allocated (GB)": 1.67,
678
+ "step": 33500,
679
+ "total_memory_available (GB)": 126.62
680
+ },
681
+ {
682
+ "epoch": 1.0,
683
+ "grad_norm": NaN,
684
+ "learning_rate": 3.3333333333333335e-05,
685
+ "loss": NaN,
686
+ "max_memory_allocated (GB)": 2.31,
687
+ "memory_allocated (GB)": 1.67,
688
+ "step": 34000,
689
+ "total_memory_available (GB)": 126.62
690
+ },
691
+ {
692
+ "epoch": 1.0147058823529411,
693
+ "grad_norm": NaN,
694
+ "learning_rate": 3.308823529411765e-05,
695
+ "loss": NaN,
696
+ "max_memory_allocated (GB)": 2.31,
697
+ "memory_allocated (GB)": 1.67,
698
+ "step": 34500,
699
+ "total_memory_available (GB)": 126.62
700
+ },
701
+ {
702
+ "epoch": 1.0294117647058822,
703
+ "grad_norm": NaN,
704
+ "learning_rate": 3.284313725490196e-05,
705
+ "loss": NaN,
706
+ "max_memory_allocated (GB)": 2.31,
707
+ "memory_allocated (GB)": 1.67,
708
+ "step": 35000,
709
+ "total_memory_available (GB)": 126.62
710
+ },
711
+ {
712
+ "epoch": 1.0441176470588236,
713
+ "grad_norm": NaN,
714
+ "learning_rate": 3.2598039215686276e-05,
715
+ "loss": NaN,
716
+ "max_memory_allocated (GB)": 2.31,
717
+ "memory_allocated (GB)": 1.67,
718
+ "step": 35500,
719
+ "total_memory_available (GB)": 126.62
720
+ },
721
+ {
722
+ "epoch": 1.0588235294117647,
723
+ "grad_norm": NaN,
724
+ "learning_rate": 3.235294117647059e-05,
725
+ "loss": NaN,
726
+ "max_memory_allocated (GB)": 2.31,
727
+ "memory_allocated (GB)": 1.67,
728
+ "step": 36000,
729
+ "total_memory_available (GB)": 126.62
730
+ },
731
+ {
732
+ "epoch": 1.0735294117647058,
733
+ "grad_norm": NaN,
734
+ "learning_rate": 3.210784313725491e-05,
735
+ "loss": NaN,
736
+ "max_memory_allocated (GB)": 2.31,
737
+ "memory_allocated (GB)": 1.67,
738
+ "step": 36500,
739
+ "total_memory_available (GB)": 126.62
740
+ },
741
+ {
742
+ "epoch": 1.088235294117647,
743
+ "grad_norm": NaN,
744
+ "learning_rate": 3.186274509803922e-05,
745
+ "loss": NaN,
746
+ "max_memory_allocated (GB)": 2.31,
747
+ "memory_allocated (GB)": 1.67,
748
+ "step": 37000,
749
+ "total_memory_available (GB)": 126.62
750
+ },
751
+ {
752
+ "epoch": 1.1029411764705883,
753
+ "grad_norm": NaN,
754
+ "learning_rate": 3.161764705882353e-05,
755
+ "loss": NaN,
756
+ "max_memory_allocated (GB)": 2.31,
757
+ "memory_allocated (GB)": 1.67,
758
+ "step": 37500,
759
+ "total_memory_available (GB)": 126.62
760
+ },
761
+ {
762
+ "epoch": 1.1176470588235294,
763
+ "grad_norm": NaN,
764
+ "learning_rate": 3.137254901960784e-05,
765
+ "loss": NaN,
766
+ "max_memory_allocated (GB)": 2.31,
767
+ "memory_allocated (GB)": 1.67,
768
+ "step": 38000,
769
+ "total_memory_available (GB)": 126.62
770
+ },
771
+ {
772
+ "epoch": 1.1323529411764706,
773
+ "grad_norm": NaN,
774
+ "learning_rate": 3.112745098039216e-05,
775
+ "loss": NaN,
776
+ "max_memory_allocated (GB)": 2.31,
777
+ "memory_allocated (GB)": 1.67,
778
+ "step": 38500,
779
+ "total_memory_available (GB)": 126.62
780
+ },
781
+ {
782
+ "epoch": 1.1470588235294117,
783
+ "grad_norm": NaN,
784
+ "learning_rate": 3.0882352941176475e-05,
785
+ "loss": NaN,
786
+ "max_memory_allocated (GB)": 2.31,
787
+ "memory_allocated (GB)": 1.67,
788
+ "step": 39000,
789
+ "total_memory_available (GB)": 126.62
790
+ },
791
+ {
792
+ "epoch": 1.161764705882353,
793
+ "grad_norm": NaN,
794
+ "learning_rate": 3.063725490196079e-05,
795
+ "loss": NaN,
796
+ "max_memory_allocated (GB)": 2.31,
797
+ "memory_allocated (GB)": 1.67,
798
+ "step": 39500,
799
+ "total_memory_available (GB)": 126.62
800
+ },
801
+ {
802
+ "epoch": 1.1764705882352942,
803
+ "grad_norm": NaN,
804
+ "learning_rate": 3.0392156862745097e-05,
805
+ "loss": NaN,
806
+ "max_memory_allocated (GB)": 2.31,
807
+ "memory_allocated (GB)": 1.67,
808
+ "step": 40000,
809
+ "total_memory_available (GB)": 126.62
810
+ },
811
+ {
812
+ "epoch": 1.1911764705882353,
813
+ "grad_norm": NaN,
814
+ "learning_rate": 3.0147058823529413e-05,
815
+ "loss": NaN,
816
+ "max_memory_allocated (GB)": 2.31,
817
+ "memory_allocated (GB)": 1.67,
818
+ "step": 40500,
819
+ "total_memory_available (GB)": 126.62
820
+ },
821
+ {
822
+ "epoch": 1.2058823529411764,
823
+ "grad_norm": NaN,
824
+ "learning_rate": 2.9901960784313725e-05,
825
+ "loss": NaN,
826
+ "max_memory_allocated (GB)": 2.31,
827
+ "memory_allocated (GB)": 1.67,
828
+ "step": 41000,
829
+ "total_memory_available (GB)": 126.62
830
+ },
831
+ {
832
+ "epoch": 1.2205882352941178,
833
+ "grad_norm": NaN,
834
+ "learning_rate": 2.965686274509804e-05,
835
+ "loss": NaN,
836
+ "max_memory_allocated (GB)": 2.31,
837
+ "memory_allocated (GB)": 1.67,
838
+ "step": 41500,
839
+ "total_memory_available (GB)": 126.62
840
+ },
841
+ {
842
+ "epoch": 1.2352941176470589,
843
+ "grad_norm": NaN,
844
+ "learning_rate": 2.9411764705882354e-05,
845
+ "loss": NaN,
846
+ "max_memory_allocated (GB)": 2.31,
847
+ "memory_allocated (GB)": 1.67,
848
+ "step": 42000,
849
+ "total_memory_available (GB)": 126.62
850
+ },
851
+ {
852
+ "epoch": 1.25,
853
+ "grad_norm": NaN,
854
+ "learning_rate": 2.916666666666667e-05,
855
+ "loss": NaN,
856
+ "max_memory_allocated (GB)": 2.31,
857
+ "memory_allocated (GB)": 1.67,
858
+ "step": 42500,
859
+ "total_memory_available (GB)": 126.62
860
+ },
861
+ {
862
+ "epoch": 1.2647058823529411,
863
+ "grad_norm": NaN,
864
+ "learning_rate": 2.8921568627450986e-05,
865
+ "loss": NaN,
866
+ "max_memory_allocated (GB)": 2.31,
867
+ "memory_allocated (GB)": 1.67,
868
+ "step": 43000,
869
+ "total_memory_available (GB)": 126.62
870
+ },
871
+ {
872
+ "epoch": 1.2794117647058822,
873
+ "grad_norm": NaN,
874
+ "learning_rate": 2.8676470588235295e-05,
875
+ "loss": NaN,
876
+ "max_memory_allocated (GB)": 2.31,
877
+ "memory_allocated (GB)": 1.67,
878
+ "step": 43500,
879
+ "total_memory_available (GB)": 126.62
880
+ },
881
+ {
882
+ "epoch": 1.2941176470588236,
883
+ "grad_norm": NaN,
884
+ "learning_rate": 2.8431372549019608e-05,
885
+ "loss": NaN,
886
+ "max_memory_allocated (GB)": 2.31,
887
+ "memory_allocated (GB)": 1.67,
888
+ "step": 44000,
889
+ "total_memory_available (GB)": 126.62
890
+ },
891
+ {
892
+ "epoch": 1.3088235294117647,
893
+ "grad_norm": NaN,
894
+ "learning_rate": 2.8186274509803924e-05,
895
+ "loss": NaN,
896
+ "max_memory_allocated (GB)": 2.31,
897
+ "memory_allocated (GB)": 1.67,
898
+ "step": 44500,
899
+ "total_memory_available (GB)": 126.62
900
+ },
901
+ {
902
+ "epoch": 1.3235294117647058,
903
+ "grad_norm": NaN,
904
+ "learning_rate": 2.7941176470588236e-05,
905
+ "loss": NaN,
906
+ "max_memory_allocated (GB)": 2.31,
907
+ "memory_allocated (GB)": 1.67,
908
+ "step": 45000,
909
+ "total_memory_available (GB)": 126.62
910
+ },
911
+ {
912
+ "epoch": 1.3382352941176472,
913
+ "grad_norm": NaN,
914
+ "learning_rate": 2.7696078431372552e-05,
915
+ "loss": NaN,
916
+ "max_memory_allocated (GB)": 2.31,
917
+ "memory_allocated (GB)": 1.67,
918
+ "step": 45500,
919
+ "total_memory_available (GB)": 126.62
920
+ },
921
+ {
922
+ "epoch": 1.3529411764705883,
923
+ "grad_norm": NaN,
924
+ "learning_rate": 2.7450980392156865e-05,
925
+ "loss": NaN,
926
+ "max_memory_allocated (GB)": 2.31,
927
+ "memory_allocated (GB)": 1.67,
928
+ "step": 46000,
929
+ "total_memory_available (GB)": 126.62
930
+ },
931
+ {
932
+ "epoch": 1.3676470588235294,
933
+ "grad_norm": NaN,
934
+ "learning_rate": 2.7205882352941174e-05,
935
+ "loss": NaN,
936
+ "max_memory_allocated (GB)": 2.31,
937
+ "memory_allocated (GB)": 1.67,
938
+ "step": 46500,
939
+ "total_memory_available (GB)": 126.62
940
+ },
941
+ {
942
+ "epoch": 1.3823529411764706,
943
+ "grad_norm": NaN,
944
+ "learning_rate": 2.696078431372549e-05,
945
+ "loss": NaN,
946
+ "max_memory_allocated (GB)": 2.31,
947
+ "memory_allocated (GB)": 1.67,
948
+ "step": 47000,
949
+ "total_memory_available (GB)": 126.62
950
+ },
951
+ {
952
+ "epoch": 1.3970588235294117,
953
+ "grad_norm": NaN,
954
+ "learning_rate": 2.6715686274509806e-05,
955
+ "loss": NaN,
956
+ "max_memory_allocated (GB)": 2.31,
957
+ "memory_allocated (GB)": 1.67,
958
+ "step": 47500,
959
+ "total_memory_available (GB)": 126.62
960
+ },
961
+ {
962
+ "epoch": 1.4117647058823528,
963
+ "grad_norm": NaN,
964
+ "learning_rate": 2.647058823529412e-05,
965
+ "loss": NaN,
966
+ "max_memory_allocated (GB)": 2.31,
967
+ "memory_allocated (GB)": 1.67,
968
+ "step": 48000,
969
+ "total_memory_available (GB)": 126.62
970
+ },
971
+ {
972
+ "epoch": 1.4264705882352942,
973
+ "grad_norm": NaN,
974
+ "learning_rate": 2.6225490196078435e-05,
975
+ "loss": NaN,
976
+ "max_memory_allocated (GB)": 2.31,
977
+ "memory_allocated (GB)": 1.67,
978
+ "step": 48500,
979
+ "total_memory_available (GB)": 126.62
980
+ },
981
+ {
982
+ "epoch": 1.4411764705882353,
983
+ "grad_norm": NaN,
984
+ "learning_rate": 2.5980392156862747e-05,
985
+ "loss": NaN,
986
+ "max_memory_allocated (GB)": 2.31,
987
+ "memory_allocated (GB)": 1.67,
988
+ "step": 49000,
989
+ "total_memory_available (GB)": 126.62
990
+ },
991
+ {
992
+ "epoch": 1.4558823529411764,
993
+ "grad_norm": NaN,
994
+ "learning_rate": 2.5735294117647057e-05,
995
+ "loss": NaN,
996
+ "max_memory_allocated (GB)": 2.31,
997
+ "memory_allocated (GB)": 1.67,
998
+ "step": 49500,
999
+ "total_memory_available (GB)": 126.62
1000
+ },
1001
+ {
1002
+ "epoch": 1.4705882352941178,
1003
+ "grad_norm": NaN,
1004
+ "learning_rate": 2.5490196078431373e-05,
1005
+ "loss": NaN,
1006
+ "max_memory_allocated (GB)": 2.31,
1007
+ "memory_allocated (GB)": 1.67,
1008
+ "step": 50000,
1009
+ "total_memory_available (GB)": 126.62
1010
+ },
1011
+ {
1012
+ "epoch": 1.4852941176470589,
1013
+ "grad_norm": NaN,
1014
+ "learning_rate": 2.5245098039215685e-05,
1015
+ "loss": NaN,
1016
+ "max_memory_allocated (GB)": 2.31,
1017
+ "memory_allocated (GB)": 1.67,
1018
+ "step": 50500,
1019
+ "total_memory_available (GB)": 126.62
1020
+ },
1021
+ {
1022
+ "epoch": 1.5,
1023
+ "grad_norm": NaN,
1024
+ "learning_rate": 2.5e-05,
1025
+ "loss": NaN,
1026
+ "max_memory_allocated (GB)": 2.31,
1027
+ "memory_allocated (GB)": 1.67,
1028
+ "step": 51000,
1029
+ "total_memory_available (GB)": 126.62
1030
+ },
1031
+ {
1032
+ "epoch": 1.5147058823529411,
1033
+ "grad_norm": NaN,
1034
+ "learning_rate": 2.4754901960784317e-05,
1035
+ "loss": NaN,
1036
+ "max_memory_allocated (GB)": 2.31,
1037
+ "memory_allocated (GB)": 1.67,
1038
+ "step": 51500,
1039
+ "total_memory_available (GB)": 126.62
1040
+ },
1041
+ {
1042
+ "epoch": 1.5294117647058822,
1043
+ "grad_norm": NaN,
1044
+ "learning_rate": 2.4509803921568626e-05,
1045
+ "loss": NaN,
1046
+ "max_memory_allocated (GB)": 2.31,
1047
+ "memory_allocated (GB)": 1.67,
1048
+ "step": 52000,
1049
+ "total_memory_available (GB)": 126.62
1050
+ },
1051
+ {
1052
+ "epoch": 1.5441176470588234,
1053
+ "grad_norm": NaN,
1054
+ "learning_rate": 2.4264705882352942e-05,
1055
+ "loss": NaN,
1056
+ "max_memory_allocated (GB)": 2.31,
1057
+ "memory_allocated (GB)": 1.67,
1058
+ "step": 52500,
1059
+ "total_memory_available (GB)": 126.62
1060
+ },
1061
+ {
1062
+ "epoch": 1.5588235294117647,
1063
+ "grad_norm": NaN,
1064
+ "learning_rate": 2.401960784313726e-05,
1065
+ "loss": NaN,
1066
+ "max_memory_allocated (GB)": 2.31,
1067
+ "memory_allocated (GB)": 1.67,
1068
+ "step": 53000,
1069
+ "total_memory_available (GB)": 126.62
1070
+ },
1071
+ {
1072
+ "epoch": 1.5735294117647058,
1073
+ "grad_norm": NaN,
1074
+ "learning_rate": 2.3774509803921568e-05,
1075
+ "loss": NaN,
1076
+ "max_memory_allocated (GB)": 2.31,
1077
+ "memory_allocated (GB)": 1.67,
1078
+ "step": 53500,
1079
+ "total_memory_available (GB)": 126.62
1080
+ },
1081
+ {
1082
+ "epoch": 1.5882352941176472,
1083
+ "grad_norm": NaN,
1084
+ "learning_rate": 2.3529411764705884e-05,
1085
+ "loss": NaN,
1086
+ "max_memory_allocated (GB)": 2.31,
1087
+ "memory_allocated (GB)": 1.67,
1088
+ "step": 54000,
1089
+ "total_memory_available (GB)": 126.62
1090
+ },
1091
+ {
1092
+ "epoch": 1.6029411764705883,
1093
+ "grad_norm": NaN,
1094
+ "learning_rate": 2.3284313725490196e-05,
1095
+ "loss": NaN,
1096
+ "max_memory_allocated (GB)": 2.31,
1097
+ "memory_allocated (GB)": 1.67,
1098
+ "step": 54500,
1099
+ "total_memory_available (GB)": 126.62
1100
+ },
1101
+ {
1102
+ "epoch": 1.6176470588235294,
1103
+ "grad_norm": NaN,
1104
+ "learning_rate": 2.303921568627451e-05,
1105
+ "loss": NaN,
1106
+ "max_memory_allocated (GB)": 2.31,
1107
+ "memory_allocated (GB)": 1.67,
1108
+ "step": 55000,
1109
+ "total_memory_available (GB)": 126.62
1110
+ },
1111
+ {
1112
+ "epoch": 1.6323529411764706,
1113
+ "grad_norm": NaN,
1114
+ "learning_rate": 2.2794117647058825e-05,
1115
+ "loss": NaN,
1116
+ "max_memory_allocated (GB)": 2.31,
1117
+ "memory_allocated (GB)": 1.67,
1118
+ "step": 55500,
1119
+ "total_memory_available (GB)": 126.62
1120
+ },
1121
+ {
1122
+ "epoch": 1.6470588235294117,
1123
+ "grad_norm": NaN,
1124
+ "learning_rate": 2.2549019607843138e-05,
1125
+ "loss": NaN,
1126
+ "max_memory_allocated (GB)": 2.31,
1127
+ "memory_allocated (GB)": 1.67,
1128
+ "step": 56000,
1129
+ "total_memory_available (GB)": 126.62
1130
+ },
1131
+ {
1132
+ "epoch": 1.6617647058823528,
1133
+ "grad_norm": NaN,
1134
+ "learning_rate": 2.2303921568627454e-05,
1135
+ "loss": NaN,
1136
+ "max_memory_allocated (GB)": 2.31,
1137
+ "memory_allocated (GB)": 1.67,
1138
+ "step": 56500,
1139
+ "total_memory_available (GB)": 126.62
1140
+ },
1141
+ {
1142
+ "epoch": 1.6764705882352942,
1143
+ "grad_norm": NaN,
1144
+ "learning_rate": 2.2058823529411766e-05,
1145
+ "loss": NaN,
1146
+ "max_memory_allocated (GB)": 2.31,
1147
+ "memory_allocated (GB)": 1.67,
1148
+ "step": 57000,
1149
+ "total_memory_available (GB)": 126.62
1150
+ },
1151
+ {
1152
+ "epoch": 1.6911764705882353,
1153
+ "grad_norm": NaN,
1154
+ "learning_rate": 2.181372549019608e-05,
1155
+ "loss": NaN,
1156
+ "max_memory_allocated (GB)": 2.31,
1157
+ "memory_allocated (GB)": 1.67,
1158
+ "step": 57500,
1159
+ "total_memory_available (GB)": 126.62
1160
+ },
1161
+ {
1162
+ "epoch": 1.7058823529411766,
1163
+ "grad_norm": NaN,
1164
+ "learning_rate": 2.1568627450980395e-05,
1165
+ "loss": NaN,
1166
+ "max_memory_allocated (GB)": 2.31,
1167
+ "memory_allocated (GB)": 1.67,
1168
+ "step": 58000,
1169
+ "total_memory_available (GB)": 126.62
1170
+ },
1171
+ {
1172
+ "epoch": 1.7205882352941178,
1173
+ "grad_norm": NaN,
1174
+ "learning_rate": 2.1323529411764707e-05,
1175
+ "loss": NaN,
1176
+ "max_memory_allocated (GB)": 2.31,
1177
+ "memory_allocated (GB)": 1.67,
1178
+ "step": 58500,
1179
+ "total_memory_available (GB)": 126.62
1180
+ },
1181
+ {
1182
+ "epoch": 1.7352941176470589,
1183
+ "grad_norm": NaN,
1184
+ "learning_rate": 2.107843137254902e-05,
1185
+ "loss": NaN,
1186
+ "max_memory_allocated (GB)": 2.31,
1187
+ "memory_allocated (GB)": 1.67,
1188
+ "step": 59000,
1189
+ "total_memory_available (GB)": 126.62
1190
+ },
1191
+ {
1192
+ "epoch": 1.75,
1193
+ "grad_norm": NaN,
1194
+ "learning_rate": 2.0833333333333336e-05,
1195
+ "loss": NaN,
1196
+ "max_memory_allocated (GB)": 2.31,
1197
+ "memory_allocated (GB)": 1.67,
1198
+ "step": 59500,
1199
+ "total_memory_available (GB)": 126.62
1200
+ },
1201
+ {
1202
+ "epoch": 1.7647058823529411,
1203
+ "grad_norm": NaN,
1204
+ "learning_rate": 2.058823529411765e-05,
1205
+ "loss": NaN,
1206
+ "max_memory_allocated (GB)": 2.31,
1207
+ "memory_allocated (GB)": 1.67,
1208
+ "step": 60000,
1209
+ "total_memory_available (GB)": 126.62
1210
+ },
1211
+ {
1212
+ "epoch": 1.7794117647058822,
1213
+ "grad_norm": NaN,
1214
+ "learning_rate": 2.034313725490196e-05,
1215
+ "loss": NaN,
1216
+ "max_memory_allocated (GB)": 2.31,
1217
+ "memory_allocated (GB)": 1.67,
1218
+ "step": 60500,
1219
+ "total_memory_available (GB)": 126.62
1220
+ },
1221
+ {
1222
+ "epoch": 1.7941176470588234,
1223
+ "grad_norm": NaN,
1224
+ "learning_rate": 2.0098039215686277e-05,
1225
+ "loss": NaN,
1226
+ "max_memory_allocated (GB)": 2.31,
1227
+ "memory_allocated (GB)": 1.67,
1228
+ "step": 61000,
1229
+ "total_memory_available (GB)": 126.62
1230
+ },
1231
+ {
1232
+ "epoch": 1.8088235294117647,
1233
+ "grad_norm": NaN,
1234
+ "learning_rate": 1.9852941176470586e-05,
1235
+ "loss": NaN,
1236
+ "max_memory_allocated (GB)": 2.31,
1237
+ "memory_allocated (GB)": 1.67,
1238
+ "step": 61500,
1239
+ "total_memory_available (GB)": 126.62
1240
+ },
1241
+ {
1242
+ "epoch": 1.8235294117647058,
1243
+ "grad_norm": NaN,
1244
+ "learning_rate": 1.9607843137254903e-05,
1245
+ "loss": NaN,
1246
+ "max_memory_allocated (GB)": 2.31,
1247
+ "memory_allocated (GB)": 1.67,
1248
+ "step": 62000,
1249
+ "total_memory_available (GB)": 126.62
1250
+ },
1251
+ {
1252
+ "epoch": 1.8382352941176472,
1253
+ "grad_norm": NaN,
1254
+ "learning_rate": 1.936274509803922e-05,
1255
+ "loss": NaN,
1256
+ "max_memory_allocated (GB)": 2.31,
1257
+ "memory_allocated (GB)": 1.67,
1258
+ "step": 62500,
1259
+ "total_memory_available (GB)": 126.62
1260
+ },
1261
+ {
1262
+ "epoch": 1.8529411764705883,
1263
+ "grad_norm": NaN,
1264
+ "learning_rate": 1.9117647058823528e-05,
1265
+ "loss": NaN,
1266
+ "max_memory_allocated (GB)": 2.31,
1267
+ "memory_allocated (GB)": 1.67,
1268
+ "step": 63000,
1269
+ "total_memory_available (GB)": 126.62
1270
+ },
1271
+ {
1272
+ "epoch": 1.8676470588235294,
1273
+ "grad_norm": NaN,
1274
+ "learning_rate": 1.8872549019607844e-05,
1275
+ "loss": NaN,
1276
+ "max_memory_allocated (GB)": 2.31,
1277
+ "memory_allocated (GB)": 1.67,
1278
+ "step": 63500,
1279
+ "total_memory_available (GB)": 126.62
1280
+ },
1281
+ {
1282
+ "epoch": 1.8823529411764706,
1283
+ "grad_norm": NaN,
1284
+ "learning_rate": 1.862745098039216e-05,
1285
+ "loss": NaN,
1286
+ "max_memory_allocated (GB)": 2.31,
1287
+ "memory_allocated (GB)": 1.67,
1288
+ "step": 64000,
1289
+ "total_memory_available (GB)": 126.62
1290
+ },
1291
+ {
1292
+ "epoch": 1.8970588235294117,
1293
+ "grad_norm": NaN,
1294
+ "learning_rate": 1.8382352941176472e-05,
1295
+ "loss": NaN,
1296
+ "max_memory_allocated (GB)": 2.31,
1297
+ "memory_allocated (GB)": 1.67,
1298
+ "step": 64500,
1299
+ "total_memory_available (GB)": 126.62
1300
+ },
1301
+ {
1302
+ "epoch": 1.9117647058823528,
1303
+ "grad_norm": NaN,
1304
+ "learning_rate": 1.8137254901960785e-05,
1305
+ "loss": NaN,
1306
+ "max_memory_allocated (GB)": 2.31,
1307
+ "memory_allocated (GB)": 1.67,
1308
+ "step": 65000,
1309
+ "total_memory_available (GB)": 126.62
1310
+ },
1311
+ {
1312
+ "epoch": 1.9264705882352942,
1313
+ "grad_norm": NaN,
1314
+ "learning_rate": 1.7892156862745098e-05,
1315
+ "loss": NaN,
1316
+ "max_memory_allocated (GB)": 2.31,
1317
+ "memory_allocated (GB)": 1.67,
1318
+ "step": 65500,
1319
+ "total_memory_available (GB)": 126.62
1320
+ },
1321
+ {
1322
+ "epoch": 1.9411764705882353,
1323
+ "grad_norm": NaN,
1324
+ "learning_rate": 1.7647058823529414e-05,
1325
+ "loss": NaN,
1326
+ "max_memory_allocated (GB)": 2.31,
1327
+ "memory_allocated (GB)": 1.67,
1328
+ "step": 66000,
1329
+ "total_memory_available (GB)": 126.62
1330
+ },
1331
+ {
1332
+ "epoch": 1.9558823529411766,
1333
+ "grad_norm": NaN,
1334
+ "learning_rate": 1.7401960784313726e-05,
1335
+ "loss": NaN,
1336
+ "max_memory_allocated (GB)": 2.31,
1337
+ "memory_allocated (GB)": 1.67,
1338
+ "step": 66500,
1339
+ "total_memory_available (GB)": 126.62
1340
+ },
1341
+ {
1342
+ "epoch": 1.9705882352941178,
1343
+ "grad_norm": NaN,
1344
+ "learning_rate": 1.715686274509804e-05,
1345
+ "loss": NaN,
1346
+ "max_memory_allocated (GB)": 2.31,
1347
+ "memory_allocated (GB)": 1.67,
1348
+ "step": 67000,
1349
+ "total_memory_available (GB)": 126.62
1350
+ },
1351
+ {
1352
+ "epoch": 1.9852941176470589,
1353
+ "grad_norm": NaN,
1354
+ "learning_rate": 1.6911764705882355e-05,
1355
+ "loss": NaN,
1356
+ "max_memory_allocated (GB)": 2.31,
1357
+ "memory_allocated (GB)": 1.67,
1358
+ "step": 67500,
1359
+ "total_memory_available (GB)": 126.62
1360
+ },
1361
+ {
1362
+ "epoch": 2.0,
1363
+ "grad_norm": NaN,
1364
+ "learning_rate": 1.6666666666666667e-05,
1365
+ "loss": NaN,
1366
+ "max_memory_allocated (GB)": 2.31,
1367
+ "memory_allocated (GB)": 1.67,
1368
+ "step": 68000,
1369
+ "total_memory_available (GB)": 126.62
1370
+ },
1371
+ {
1372
+ "epoch": 2.014705882352941,
1373
+ "grad_norm": NaN,
1374
+ "learning_rate": 1.642156862745098e-05,
1375
+ "loss": NaN,
1376
+ "max_memory_allocated (GB)": 2.31,
1377
+ "memory_allocated (GB)": 1.67,
1378
+ "step": 68500,
1379
+ "total_memory_available (GB)": 126.62
1380
+ },
1381
+ {
1382
+ "epoch": 2.0294117647058822,
1383
+ "grad_norm": NaN,
1384
+ "learning_rate": 1.6176470588235296e-05,
1385
+ "loss": NaN,
1386
+ "max_memory_allocated (GB)": 2.31,
1387
+ "memory_allocated (GB)": 1.67,
1388
+ "step": 69000,
1389
+ "total_memory_available (GB)": 126.62
1390
+ },
1391
+ {
1392
+ "epoch": 2.0441176470588234,
1393
+ "grad_norm": NaN,
1394
+ "learning_rate": 1.593137254901961e-05,
1395
+ "loss": NaN,
1396
+ "max_memory_allocated (GB)": 2.31,
1397
+ "memory_allocated (GB)": 1.67,
1398
+ "step": 69500,
1399
+ "total_memory_available (GB)": 126.62
1400
+ },
1401
+ {
1402
+ "epoch": 2.0588235294117645,
1403
+ "grad_norm": NaN,
1404
+ "learning_rate": 1.568627450980392e-05,
1405
+ "loss": NaN,
1406
+ "max_memory_allocated (GB)": 2.31,
1407
+ "memory_allocated (GB)": 1.67,
1408
+ "step": 70000,
1409
+ "total_memory_available (GB)": 126.62
1410
+ },
1411
+ {
1412
+ "epoch": 2.073529411764706,
1413
+ "grad_norm": NaN,
1414
+ "learning_rate": 1.5441176470588237e-05,
1415
+ "loss": NaN,
1416
+ "max_memory_allocated (GB)": 2.31,
1417
+ "memory_allocated (GB)": 1.67,
1418
+ "step": 70500,
1419
+ "total_memory_available (GB)": 126.62
1420
+ },
1421
+ {
1422
+ "epoch": 2.088235294117647,
1423
+ "grad_norm": NaN,
1424
+ "learning_rate": 1.5196078431372548e-05,
1425
+ "loss": NaN,
1426
+ "max_memory_allocated (GB)": 2.31,
1427
+ "memory_allocated (GB)": 1.67,
1428
+ "step": 71000,
1429
+ "total_memory_available (GB)": 126.62
1430
+ },
1431
+ {
1432
+ "epoch": 2.1029411764705883,
1433
+ "grad_norm": NaN,
1434
+ "learning_rate": 1.4950980392156863e-05,
1435
+ "loss": NaN,
1436
+ "max_memory_allocated (GB)": 2.31,
1437
+ "memory_allocated (GB)": 1.67,
1438
+ "step": 71500,
1439
+ "total_memory_available (GB)": 126.62
1440
+ },
1441
+ {
1442
+ "epoch": 2.1176470588235294,
1443
+ "grad_norm": NaN,
1444
+ "learning_rate": 1.4705882352941177e-05,
1445
+ "loss": NaN,
1446
+ "max_memory_allocated (GB)": 2.31,
1447
+ "memory_allocated (GB)": 1.67,
1448
+ "step": 72000,
1449
+ "total_memory_available (GB)": 126.62
1450
+ },
1451
+ {
1452
+ "epoch": 2.1323529411764706,
1453
+ "grad_norm": NaN,
1454
+ "learning_rate": 1.4460784313725493e-05,
1455
+ "loss": NaN,
1456
+ "max_memory_allocated (GB)": 2.31,
1457
+ "memory_allocated (GB)": 1.67,
1458
+ "step": 72500,
1459
+ "total_memory_available (GB)": 126.62
1460
+ },
1461
+ {
1462
+ "epoch": 2.1470588235294117,
1463
+ "grad_norm": NaN,
1464
+ "learning_rate": 1.4215686274509804e-05,
1465
+ "loss": NaN,
1466
+ "max_memory_allocated (GB)": 2.31,
1467
+ "memory_allocated (GB)": 1.67,
1468
+ "step": 73000,
1469
+ "total_memory_available (GB)": 126.62
1470
+ },
1471
+ {
1472
+ "epoch": 2.161764705882353,
1473
+ "grad_norm": NaN,
1474
+ "learning_rate": 1.3970588235294118e-05,
1475
+ "loss": NaN,
1476
+ "max_memory_allocated (GB)": 2.31,
1477
+ "memory_allocated (GB)": 1.67,
1478
+ "step": 73500,
1479
+ "total_memory_available (GB)": 126.62
1480
+ },
1481
+ {
1482
+ "epoch": 2.176470588235294,
1483
+ "grad_norm": NaN,
1484
+ "learning_rate": 1.3725490196078432e-05,
1485
+ "loss": NaN,
1486
+ "max_memory_allocated (GB)": 2.31,
1487
+ "memory_allocated (GB)": 1.67,
1488
+ "step": 74000,
1489
+ "total_memory_available (GB)": 126.62
1490
+ },
1491
+ {
1492
+ "epoch": 2.1911764705882355,
1493
+ "grad_norm": NaN,
1494
+ "learning_rate": 1.3480392156862745e-05,
1495
+ "loss": NaN,
1496
+ "max_memory_allocated (GB)": 2.31,
1497
+ "memory_allocated (GB)": 1.67,
1498
+ "step": 74500,
1499
+ "total_memory_available (GB)": 126.62
1500
+ },
1501
+ {
1502
+ "epoch": 2.2058823529411766,
1503
+ "grad_norm": NaN,
1504
+ "learning_rate": 1.323529411764706e-05,
1505
+ "loss": NaN,
1506
+ "max_memory_allocated (GB)": 2.31,
1507
+ "memory_allocated (GB)": 1.67,
1508
+ "step": 75000,
1509
+ "total_memory_available (GB)": 126.62
1510
+ },
1511
+ {
1512
+ "epoch": 2.2205882352941178,
1513
+ "grad_norm": NaN,
1514
+ "learning_rate": 1.2990196078431374e-05,
1515
+ "loss": NaN,
1516
+ "max_memory_allocated (GB)": 2.31,
1517
+ "memory_allocated (GB)": 1.67,
1518
+ "step": 75500,
1519
+ "total_memory_available (GB)": 126.62
1520
+ },
1521
+ {
1522
+ "epoch": 2.235294117647059,
1523
+ "grad_norm": NaN,
1524
+ "learning_rate": 1.2745098039215686e-05,
1525
+ "loss": NaN,
1526
+ "max_memory_allocated (GB)": 2.31,
1527
+ "memory_allocated (GB)": 1.67,
1528
+ "step": 76000,
1529
+ "total_memory_available (GB)": 126.62
1530
+ },
1531
+ {
1532
+ "epoch": 2.25,
1533
+ "grad_norm": NaN,
1534
+ "learning_rate": 1.25e-05,
1535
+ "loss": NaN,
1536
+ "max_memory_allocated (GB)": 2.31,
1537
+ "memory_allocated (GB)": 1.67,
1538
+ "step": 76500,
1539
+ "total_memory_available (GB)": 126.62
1540
+ },
1541
+ {
1542
+ "epoch": 2.264705882352941,
1543
+ "grad_norm": NaN,
1544
+ "learning_rate": 1.2254901960784313e-05,
1545
+ "loss": NaN,
1546
+ "max_memory_allocated (GB)": 2.31,
1547
+ "memory_allocated (GB)": 1.67,
1548
+ "step": 77000,
1549
+ "total_memory_available (GB)": 126.62
1550
+ },
1551
+ {
1552
+ "epoch": 2.2794117647058822,
1553
+ "grad_norm": NaN,
1554
+ "learning_rate": 1.200980392156863e-05,
1555
+ "loss": NaN,
1556
+ "max_memory_allocated (GB)": 2.31,
1557
+ "memory_allocated (GB)": 1.67,
1558
+ "step": 77500,
1559
+ "total_memory_available (GB)": 126.62
1560
+ },
1561
+ {
1562
+ "epoch": 2.2941176470588234,
1563
+ "grad_norm": NaN,
1564
+ "learning_rate": 1.1764705882352942e-05,
1565
+ "loss": NaN,
1566
+ "max_memory_allocated (GB)": 2.31,
1567
+ "memory_allocated (GB)": 1.67,
1568
+ "step": 78000,
1569
+ "total_memory_available (GB)": 126.62
1570
+ },
1571
+ {
1572
+ "epoch": 2.3088235294117645,
1573
+ "grad_norm": NaN,
1574
+ "learning_rate": 1.1519607843137254e-05,
1575
+ "loss": NaN,
1576
+ "max_memory_allocated (GB)": 2.31,
1577
+ "memory_allocated (GB)": 1.67,
1578
+ "step": 78500,
1579
+ "total_memory_available (GB)": 126.62
1580
+ },
1581
+ {
1582
+ "epoch": 2.323529411764706,
1583
+ "grad_norm": NaN,
1584
+ "learning_rate": 1.1274509803921569e-05,
1585
+ "loss": NaN,
1586
+ "max_memory_allocated (GB)": 2.31,
1587
+ "memory_allocated (GB)": 1.67,
1588
+ "step": 79000,
1589
+ "total_memory_available (GB)": 126.62
1590
+ },
1591
+ {
1592
+ "epoch": 2.338235294117647,
1593
+ "grad_norm": NaN,
1594
+ "learning_rate": 1.1029411764705883e-05,
1595
+ "loss": NaN,
1596
+ "max_memory_allocated (GB)": 2.31,
1597
+ "memory_allocated (GB)": 1.67,
1598
+ "step": 79500,
1599
+ "total_memory_available (GB)": 126.62
1600
+ },
1601
+ {
1602
+ "epoch": 2.3529411764705883,
1603
+ "grad_norm": NaN,
1604
+ "learning_rate": 1.0784313725490197e-05,
1605
+ "loss": NaN,
1606
+ "max_memory_allocated (GB)": 2.31,
1607
+ "memory_allocated (GB)": 1.67,
1608
+ "step": 80000,
1609
+ "total_memory_available (GB)": 126.62
1610
+ },
1611
+ {
1612
+ "epoch": 2.3676470588235294,
1613
+ "grad_norm": NaN,
1614
+ "learning_rate": 1.053921568627451e-05,
1615
+ "loss": NaN,
1616
+ "max_memory_allocated (GB)": 2.31,
1617
+ "memory_allocated (GB)": 1.67,
1618
+ "step": 80500,
1619
+ "total_memory_available (GB)": 126.62
1620
+ },
1621
+ {
1622
+ "epoch": 2.3823529411764706,
1623
+ "grad_norm": NaN,
1624
+ "learning_rate": 1.0294117647058824e-05,
1625
+ "loss": NaN,
1626
+ "max_memory_allocated (GB)": 2.31,
1627
+ "memory_allocated (GB)": 1.67,
1628
+ "step": 81000,
1629
+ "total_memory_available (GB)": 126.62
1630
+ },
1631
+ {
1632
+ "epoch": 2.3970588235294117,
1633
+ "grad_norm": NaN,
1634
+ "learning_rate": 1.0049019607843139e-05,
1635
+ "loss": NaN,
1636
+ "max_memory_allocated (GB)": 2.31,
1637
+ "memory_allocated (GB)": 1.67,
1638
+ "step": 81500,
1639
+ "total_memory_available (GB)": 126.62
1640
+ },
1641
+ {
1642
+ "epoch": 2.411764705882353,
1643
+ "grad_norm": NaN,
1644
+ "learning_rate": 9.803921568627451e-06,
1645
+ "loss": NaN,
1646
+ "max_memory_allocated (GB)": 2.31,
1647
+ "memory_allocated (GB)": 1.67,
1648
+ "step": 82000,
1649
+ "total_memory_available (GB)": 126.62
1650
+ },
1651
+ {
1652
+ "epoch": 2.426470588235294,
1653
+ "grad_norm": NaN,
1654
+ "learning_rate": 9.558823529411764e-06,
1655
+ "loss": NaN,
1656
+ "max_memory_allocated (GB)": 2.31,
1657
+ "memory_allocated (GB)": 1.67,
1658
+ "step": 82500,
1659
+ "total_memory_available (GB)": 126.62
1660
+ },
1661
+ {
1662
+ "epoch": 2.4411764705882355,
1663
+ "grad_norm": NaN,
1664
+ "learning_rate": 9.31372549019608e-06,
1665
+ "loss": NaN,
1666
+ "max_memory_allocated (GB)": 2.31,
1667
+ "memory_allocated (GB)": 1.67,
1668
+ "step": 83000,
1669
+ "total_memory_available (GB)": 126.62
1670
+ },
1671
+ {
1672
+ "epoch": 2.4558823529411766,
1673
+ "grad_norm": NaN,
1674
+ "learning_rate": 9.068627450980392e-06,
1675
+ "loss": NaN,
1676
+ "max_memory_allocated (GB)": 2.31,
1677
+ "memory_allocated (GB)": 1.67,
1678
+ "step": 83500,
1679
+ "total_memory_available (GB)": 126.62
1680
+ },
1681
+ {
1682
+ "epoch": 2.4705882352941178,
1683
+ "grad_norm": NaN,
1684
+ "learning_rate": 8.823529411764707e-06,
1685
+ "loss": NaN,
1686
+ "max_memory_allocated (GB)": 2.31,
1687
+ "memory_allocated (GB)": 1.67,
1688
+ "step": 84000,
1689
+ "total_memory_available (GB)": 126.62
1690
+ },
1691
+ {
1692
+ "epoch": 2.485294117647059,
1693
+ "grad_norm": NaN,
1694
+ "learning_rate": 8.57843137254902e-06,
1695
+ "loss": NaN,
1696
+ "max_memory_allocated (GB)": 2.31,
1697
+ "memory_allocated (GB)": 1.67,
1698
+ "step": 84500,
1699
+ "total_memory_available (GB)": 126.62
1700
+ },
1701
+ {
1702
+ "epoch": 2.5,
1703
+ "grad_norm": NaN,
1704
+ "learning_rate": 8.333333333333334e-06,
1705
+ "loss": NaN,
1706
+ "max_memory_allocated (GB)": 2.31,
1707
+ "memory_allocated (GB)": 1.67,
1708
+ "step": 85000,
1709
+ "total_memory_available (GB)": 126.62
1710
+ },
1711
+ {
1712
+ "epoch": 2.514705882352941,
1713
+ "grad_norm": NaN,
1714
+ "learning_rate": 8.088235294117648e-06,
1715
+ "loss": NaN,
1716
+ "max_memory_allocated (GB)": 2.31,
1717
+ "memory_allocated (GB)": 1.67,
1718
+ "step": 85500,
1719
+ "total_memory_available (GB)": 126.62
1720
+ },
1721
+ {
1722
+ "epoch": 2.5294117647058822,
1723
+ "grad_norm": NaN,
1724
+ "learning_rate": 7.84313725490196e-06,
1725
+ "loss": NaN,
1726
+ "max_memory_allocated (GB)": 2.31,
1727
+ "memory_allocated (GB)": 1.67,
1728
+ "step": 86000,
1729
+ "total_memory_available (GB)": 126.62
1730
+ },
1731
+ {
1732
+ "epoch": 2.5441176470588234,
1733
+ "grad_norm": NaN,
1734
+ "learning_rate": 7.598039215686274e-06,
1735
+ "loss": NaN,
1736
+ "max_memory_allocated (GB)": 2.31,
1737
+ "memory_allocated (GB)": 1.67,
1738
+ "step": 86500,
1739
+ "total_memory_available (GB)": 126.62
1740
+ },
1741
+ {
1742
+ "epoch": 2.5588235294117645,
1743
+ "grad_norm": NaN,
1744
+ "learning_rate": 7.3529411764705884e-06,
1745
+ "loss": NaN,
1746
+ "max_memory_allocated (GB)": 2.31,
1747
+ "memory_allocated (GB)": 1.67,
1748
+ "step": 87000,
1749
+ "total_memory_available (GB)": 126.62
1750
+ },
1751
+ {
1752
+ "epoch": 2.5735294117647056,
1753
+ "grad_norm": NaN,
1754
+ "learning_rate": 7.107843137254902e-06,
1755
+ "loss": NaN,
1756
+ "max_memory_allocated (GB)": 2.31,
1757
+ "memory_allocated (GB)": 1.67,
1758
+ "step": 87500,
1759
+ "total_memory_available (GB)": 126.62
1760
+ },
1761
+ {
1762
+ "epoch": 2.588235294117647,
1763
+ "grad_norm": NaN,
1764
+ "learning_rate": 6.862745098039216e-06,
1765
+ "loss": NaN,
1766
+ "max_memory_allocated (GB)": 2.31,
1767
+ "memory_allocated (GB)": 1.67,
1768
+ "step": 88000,
1769
+ "total_memory_available (GB)": 126.62
1770
+ },
1771
+ {
1772
+ "epoch": 2.6029411764705883,
1773
+ "grad_norm": NaN,
1774
+ "learning_rate": 6.61764705882353e-06,
1775
+ "loss": NaN,
1776
+ "max_memory_allocated (GB)": 2.31,
1777
+ "memory_allocated (GB)": 1.67,
1778
+ "step": 88500,
1779
+ "total_memory_available (GB)": 126.62
1780
+ },
1781
+ {
1782
+ "epoch": 2.6176470588235294,
1783
+ "grad_norm": NaN,
1784
+ "learning_rate": 6.372549019607843e-06,
1785
+ "loss": NaN,
1786
+ "max_memory_allocated (GB)": 2.31,
1787
+ "memory_allocated (GB)": 1.67,
1788
+ "step": 89000,
1789
+ "total_memory_available (GB)": 126.62
1790
+ },
1791
+ {
1792
+ "epoch": 2.6323529411764706,
1793
+ "grad_norm": NaN,
1794
+ "learning_rate": 6.127450980392157e-06,
1795
+ "loss": NaN,
1796
+ "max_memory_allocated (GB)": 2.31,
1797
+ "memory_allocated (GB)": 1.67,
1798
+ "step": 89500,
1799
+ "total_memory_available (GB)": 126.62
1800
+ },
1801
+ {
1802
+ "epoch": 2.6470588235294117,
1803
+ "grad_norm": NaN,
1804
+ "learning_rate": 5.882352941176471e-06,
1805
+ "loss": NaN,
1806
+ "max_memory_allocated (GB)": 2.31,
1807
+ "memory_allocated (GB)": 1.67,
1808
+ "step": 90000,
1809
+ "total_memory_available (GB)": 126.62
1810
+ },
1811
+ {
1812
+ "epoch": 2.661764705882353,
1813
+ "grad_norm": NaN,
1814
+ "learning_rate": 5.637254901960784e-06,
1815
+ "loss": NaN,
1816
+ "max_memory_allocated (GB)": 2.31,
1817
+ "memory_allocated (GB)": 1.67,
1818
+ "step": 90500,
1819
+ "total_memory_available (GB)": 126.62
1820
+ },
1821
+ {
1822
+ "epoch": 2.6764705882352944,
1823
+ "grad_norm": NaN,
1824
+ "learning_rate": 5.392156862745099e-06,
1825
+ "loss": NaN,
1826
+ "max_memory_allocated (GB)": 2.31,
1827
+ "memory_allocated (GB)": 1.67,
1828
+ "step": 91000,
1829
+ "total_memory_available (GB)": 126.62
1830
+ },
1831
+ {
1832
+ "epoch": 2.6911764705882355,
1833
+ "grad_norm": NaN,
1834
+ "learning_rate": 5.147058823529412e-06,
1835
+ "loss": NaN,
1836
+ "max_memory_allocated (GB)": 2.31,
1837
+ "memory_allocated (GB)": 1.67,
1838
+ "step": 91500,
1839
+ "total_memory_available (GB)": 126.62
1840
+ },
1841
+ {
1842
+ "epoch": 2.7058823529411766,
1843
+ "grad_norm": NaN,
1844
+ "learning_rate": 4.901960784313726e-06,
1845
+ "loss": NaN,
1846
+ "max_memory_allocated (GB)": 2.31,
1847
+ "memory_allocated (GB)": 1.67,
1848
+ "step": 92000,
1849
+ "total_memory_available (GB)": 126.62
1850
+ },
1851
+ {
1852
+ "epoch": 2.7205882352941178,
1853
+ "grad_norm": NaN,
1854
+ "learning_rate": 4.65686274509804e-06,
1855
+ "loss": NaN,
1856
+ "max_memory_allocated (GB)": 2.31,
1857
+ "memory_allocated (GB)": 1.67,
1858
+ "step": 92500,
1859
+ "total_memory_available (GB)": 126.62
1860
+ },
1861
+ {
1862
+ "epoch": 2.735294117647059,
1863
+ "grad_norm": NaN,
1864
+ "learning_rate": 4.411764705882353e-06,
1865
+ "loss": NaN,
1866
+ "max_memory_allocated (GB)": 2.31,
1867
+ "memory_allocated (GB)": 1.67,
1868
+ "step": 93000,
1869
+ "total_memory_available (GB)": 126.62
1870
+ },
1871
+ {
1872
+ "epoch": 2.75,
1873
+ "grad_norm": NaN,
1874
+ "learning_rate": 4.166666666666667e-06,
1875
+ "loss": NaN,
1876
+ "max_memory_allocated (GB)": 2.31,
1877
+ "memory_allocated (GB)": 1.67,
1878
+ "step": 93500,
1879
+ "total_memory_available (GB)": 126.62
1880
+ },
1881
+ {
1882
+ "epoch": 2.764705882352941,
1883
+ "grad_norm": NaN,
1884
+ "learning_rate": 3.92156862745098e-06,
1885
+ "loss": NaN,
1886
+ "max_memory_allocated (GB)": 2.31,
1887
+ "memory_allocated (GB)": 1.67,
1888
+ "step": 94000,
1889
+ "total_memory_available (GB)": 126.62
1890
+ },
1891
+ {
1892
+ "epoch": 2.7794117647058822,
1893
+ "grad_norm": NaN,
1894
+ "learning_rate": 3.6764705882352942e-06,
1895
+ "loss": NaN,
1896
+ "max_memory_allocated (GB)": 2.31,
1897
+ "memory_allocated (GB)": 1.67,
1898
+ "step": 94500,
1899
+ "total_memory_available (GB)": 126.62
1900
+ },
1901
+ {
1902
+ "epoch": 2.7941176470588234,
1903
+ "grad_norm": NaN,
1904
+ "learning_rate": 3.431372549019608e-06,
1905
+ "loss": NaN,
1906
+ "max_memory_allocated (GB)": 2.31,
1907
+ "memory_allocated (GB)": 1.67,
1908
+ "step": 95000,
1909
+ "total_memory_available (GB)": 126.62
1910
+ },
1911
+ {
1912
+ "epoch": 2.8088235294117645,
1913
+ "grad_norm": NaN,
1914
+ "learning_rate": 3.1862745098039216e-06,
1915
+ "loss": NaN,
1916
+ "max_memory_allocated (GB)": 2.31,
1917
+ "memory_allocated (GB)": 1.67,
1918
+ "step": 95500,
1919
+ "total_memory_available (GB)": 126.62
1920
+ },
1921
+ {
1922
+ "epoch": 2.8235294117647056,
1923
+ "grad_norm": NaN,
1924
+ "learning_rate": 2.9411764705882355e-06,
1925
+ "loss": NaN,
1926
+ "max_memory_allocated (GB)": 2.31,
1927
+ "memory_allocated (GB)": 1.67,
1928
+ "step": 96000,
1929
+ "total_memory_available (GB)": 126.62
1930
+ },
1931
+ {
1932
+ "epoch": 2.838235294117647,
1933
+ "grad_norm": NaN,
1934
+ "learning_rate": 2.6960784313725493e-06,
1935
+ "loss": NaN,
1936
+ "max_memory_allocated (GB)": 2.31,
1937
+ "memory_allocated (GB)": 1.67,
1938
+ "step": 96500,
1939
+ "total_memory_available (GB)": 126.62
1940
+ },
1941
+ {
1942
+ "epoch": 2.8529411764705883,
1943
+ "grad_norm": NaN,
1944
+ "learning_rate": 2.450980392156863e-06,
1945
+ "loss": NaN,
1946
+ "max_memory_allocated (GB)": 2.31,
1947
+ "memory_allocated (GB)": 1.67,
1948
+ "step": 97000,
1949
+ "total_memory_available (GB)": 126.62
1950
+ },
1951
+ {
1952
+ "epoch": 2.8676470588235294,
1953
+ "grad_norm": NaN,
1954
+ "learning_rate": 2.2058823529411767e-06,
1955
+ "loss": NaN,
1956
+ "max_memory_allocated (GB)": 2.31,
1957
+ "memory_allocated (GB)": 1.67,
1958
+ "step": 97500,
1959
+ "total_memory_available (GB)": 126.62
1960
+ },
1961
+ {
1962
+ "epoch": 2.8823529411764706,
1963
+ "grad_norm": NaN,
1964
+ "learning_rate": 1.96078431372549e-06,
1965
+ "loss": NaN,
1966
+ "max_memory_allocated (GB)": 2.31,
1967
+ "memory_allocated (GB)": 1.67,
1968
+ "step": 98000,
1969
+ "total_memory_available (GB)": 126.62
1970
+ },
1971
+ {
1972
+ "epoch": 2.8970588235294117,
1973
+ "grad_norm": NaN,
1974
+ "learning_rate": 1.715686274509804e-06,
1975
+ "loss": NaN,
1976
+ "max_memory_allocated (GB)": 2.31,
1977
+ "memory_allocated (GB)": 1.67,
1978
+ "step": 98500,
1979
+ "total_memory_available (GB)": 126.62
1980
+ },
1981
+ {
1982
+ "epoch": 2.911764705882353,
1983
+ "grad_norm": NaN,
1984
+ "learning_rate": 1.4705882352941177e-06,
1985
+ "loss": NaN,
1986
+ "max_memory_allocated (GB)": 2.31,
1987
+ "memory_allocated (GB)": 1.67,
1988
+ "step": 99000,
1989
+ "total_memory_available (GB)": 126.62
1990
+ },
1991
+ {
1992
+ "epoch": 2.9264705882352944,
1993
+ "grad_norm": NaN,
1994
+ "learning_rate": 1.2254901960784314e-06,
1995
+ "loss": NaN,
1996
+ "max_memory_allocated (GB)": 2.31,
1997
+ "memory_allocated (GB)": 1.67,
1998
+ "step": 99500,
1999
+ "total_memory_available (GB)": 126.62
2000
+ },
2001
+ {
2002
+ "epoch": 2.9411764705882355,
2003
+ "grad_norm": NaN,
2004
+ "learning_rate": 9.80392156862745e-07,
2005
+ "loss": NaN,
2006
+ "max_memory_allocated (GB)": 2.31,
2007
+ "memory_allocated (GB)": 1.67,
2008
+ "step": 100000,
2009
+ "total_memory_available (GB)": 126.62
2010
+ },
2011
+ {
2012
+ "epoch": 2.9558823529411766,
2013
+ "grad_norm": NaN,
2014
+ "learning_rate": 7.352941176470589e-07,
2015
+ "loss": NaN,
2016
+ "max_memory_allocated (GB)": 2.31,
2017
+ "memory_allocated (GB)": 1.67,
2018
+ "step": 100500,
2019
+ "total_memory_available (GB)": 126.62
2020
+ },
2021
+ {
2022
+ "epoch": 2.9705882352941178,
2023
+ "grad_norm": NaN,
2024
+ "learning_rate": 4.901960784313725e-07,
2025
+ "loss": NaN,
2026
+ "max_memory_allocated (GB)": 2.31,
2027
+ "memory_allocated (GB)": 1.67,
2028
+ "step": 101000,
2029
+ "total_memory_available (GB)": 126.62
2030
+ },
2031
+ {
2032
+ "epoch": 2.985294117647059,
2033
+ "grad_norm": NaN,
2034
+ "learning_rate": 2.4509803921568627e-07,
2035
+ "loss": NaN,
2036
+ "max_memory_allocated (GB)": 2.31,
2037
+ "memory_allocated (GB)": 1.67,
2038
+ "step": 101500,
2039
+ "total_memory_available (GB)": 126.62
2040
+ },
2041
+ {
2042
+ "epoch": 3.0,
2043
+ "grad_norm": NaN,
2044
+ "learning_rate": 0.0,
2045
+ "loss": NaN,
2046
+ "max_memory_allocated (GB)": 2.31,
2047
+ "memory_allocated (GB)": 1.67,
2048
+ "step": 102000,
2049
+ "total_memory_available (GB)": 126.62
2050
+ },
2051
+ {
2052
+ "epoch": 3.0,
2053
+ "max_memory_allocated (GB)": 2.31,
2054
+ "memory_allocated (GB)": 1.67,
2055
+ "step": 102000,
2056
+ "total_flos": 6.324139790696448e+19,
2057
+ "total_memory_available (GB)": 126.62,
2058
+ "train_loss": NaN,
2059
+ "train_runtime": 3737.7146,
2060
+ "train_samples_per_second": 218.314,
2061
+ "train_steps_per_second": 27.289
2062
+ }
2063
+ ],
2064
+ "logging_steps": 500,
2065
+ "max_steps": 102000,
2066
+ "num_input_tokens_seen": 0,
2067
+ "num_train_epochs": 3,
2068
+ "save_steps": 500,
2069
+ "stateful_callbacks": {
2070
+ "TrainerControl": {
2071
+ "args": {
2072
+ "should_epoch_stop": false,
2073
+ "should_evaluate": false,
2074
+ "should_log": false,
2075
+ "should_save": true,
2076
+ "should_training_stop": true
2077
+ },
2078
+ "attributes": {}
2079
+ }
2080
+ },
2081
+ "total_flos": 6.324139790696448e+19,
2082
+ "train_batch_size": 8,
2083
+ "trial_name": null,
2084
+ "trial_params": null
2085
+ }