End of training
Browse files
all_results.json
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.9971844204598779,
|
3 |
+
"eval_accuracy": 0.7506270627062707,
|
4 |
+
"eval_loss": 0.9383056163787842,
|
5 |
+
"eval_runtime": 71.9354,
|
6 |
+
"eval_samples_per_second": 105.303,
|
7 |
+
"eval_steps_per_second": 3.295,
|
8 |
+
"total_flos": 3.394127568277555e+18,
|
9 |
+
"train_loss": 1.8402288484394103,
|
10 |
+
"train_runtime": 2568.7785,
|
11 |
+
"train_samples_per_second": 53.08,
|
12 |
+
"train_steps_per_second": 0.414
|
13 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.9971844204598779,
|
3 |
+
"eval_accuracy": 0.7506270627062707,
|
4 |
+
"eval_loss": 0.9383056163787842,
|
5 |
+
"eval_runtime": 71.9354,
|
6 |
+
"eval_samples_per_second": 105.303,
|
7 |
+
"eval_steps_per_second": 3.295
|
8 |
+
}
|
runs/May31_10-15-48_bfc6c9092c85/events.out.tfevents.1717153502.bfc6c9092c85.733.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:59288b408602209875dccf18875e11d2827c9f7ea3e9cc3c58baa1eef0293d8d
|
3 |
+
size 411
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.9971844204598779,
|
3 |
+
"total_flos": 3.394127568277555e+18,
|
4 |
+
"train_loss": 1.8402288484394103,
|
5 |
+
"train_runtime": 2568.7785,
|
6 |
+
"train_samples_per_second": 53.08,
|
7 |
+
"train_steps_per_second": 0.414
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,802 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.7506270627062707,
|
3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-1064",
|
4 |
+
"epoch": 1.9971844204598779,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 1064,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.018770530267480056,
|
13 |
+
"grad_norm": 3.569255828857422,
|
14 |
+
"learning_rate": 4.6728971962616825e-06,
|
15 |
+
"loss": 4.6843,
|
16 |
+
"step": 10
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.03754106053496011,
|
20 |
+
"grad_norm": 3.1433990001678467,
|
21 |
+
"learning_rate": 9.345794392523365e-06,
|
22 |
+
"loss": 4.6663,
|
23 |
+
"step": 20
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.05631159080244017,
|
27 |
+
"grad_norm": 3.3763270378112793,
|
28 |
+
"learning_rate": 1.4018691588785047e-05,
|
29 |
+
"loss": 4.6424,
|
30 |
+
"step": 30
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.07508212106992022,
|
34 |
+
"grad_norm": 3.4368538856506348,
|
35 |
+
"learning_rate": 1.869158878504673e-05,
|
36 |
+
"loss": 4.6077,
|
37 |
+
"step": 40
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.09385265133740028,
|
41 |
+
"grad_norm": 3.3751752376556396,
|
42 |
+
"learning_rate": 2.3364485981308414e-05,
|
43 |
+
"loss": 4.5724,
|
44 |
+
"step": 50
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.11262318160488034,
|
48 |
+
"grad_norm": 4.194823265075684,
|
49 |
+
"learning_rate": 2.8037383177570094e-05,
|
50 |
+
"loss": 4.5213,
|
51 |
+
"step": 60
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.1313937118723604,
|
55 |
+
"grad_norm": 4.447911262512207,
|
56 |
+
"learning_rate": 3.2710280373831774e-05,
|
57 |
+
"loss": 4.4357,
|
58 |
+
"step": 70
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.15016424213984045,
|
62 |
+
"grad_norm": 5.328530788421631,
|
63 |
+
"learning_rate": 3.738317757009346e-05,
|
64 |
+
"loss": 4.2996,
|
65 |
+
"step": 80
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.1689347724073205,
|
69 |
+
"grad_norm": 7.00287389755249,
|
70 |
+
"learning_rate": 4.205607476635514e-05,
|
71 |
+
"loss": 4.0969,
|
72 |
+
"step": 90
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.18770530267480057,
|
76 |
+
"grad_norm": 8.038763999938965,
|
77 |
+
"learning_rate": 4.672897196261683e-05,
|
78 |
+
"loss": 3.7622,
|
79 |
+
"step": 100
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.2064758329422806,
|
83 |
+
"grad_norm": 10.13965129852295,
|
84 |
+
"learning_rate": 4.9843260188087774e-05,
|
85 |
+
"loss": 3.5079,
|
86 |
+
"step": 110
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.22524636320976069,
|
90 |
+
"grad_norm": 9.638181686401367,
|
91 |
+
"learning_rate": 4.932079414838036e-05,
|
92 |
+
"loss": 3.2635,
|
93 |
+
"step": 120
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.24401689347724073,
|
97 |
+
"grad_norm": 9.798678398132324,
|
98 |
+
"learning_rate": 4.879832810867294e-05,
|
99 |
+
"loss": 3.0213,
|
100 |
+
"step": 130
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.2627874237447208,
|
104 |
+
"grad_norm": 9.259151458740234,
|
105 |
+
"learning_rate": 4.827586206896552e-05,
|
106 |
+
"loss": 2.9042,
|
107 |
+
"step": 140
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.28155795401220085,
|
111 |
+
"grad_norm": 9.942826271057129,
|
112 |
+
"learning_rate": 4.77533960292581e-05,
|
113 |
+
"loss": 2.6591,
|
114 |
+
"step": 150
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.3003284842796809,
|
118 |
+
"grad_norm": 9.459359169006348,
|
119 |
+
"learning_rate": 4.723092998955068e-05,
|
120 |
+
"loss": 2.6108,
|
121 |
+
"step": 160
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.31909901454716094,
|
125 |
+
"grad_norm": 10.439133644104004,
|
126 |
+
"learning_rate": 4.670846394984326e-05,
|
127 |
+
"loss": 2.4054,
|
128 |
+
"step": 170
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.337869544814641,
|
132 |
+
"grad_norm": 9.522982597351074,
|
133 |
+
"learning_rate": 4.6185997910135844e-05,
|
134 |
+
"loss": 2.3187,
|
135 |
+
"step": 180
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.3566400750821211,
|
139 |
+
"grad_norm": 10.88189697265625,
|
140 |
+
"learning_rate": 4.566353187042842e-05,
|
141 |
+
"loss": 2.3281,
|
142 |
+
"step": 190
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.37541060534960113,
|
146 |
+
"grad_norm": 12.891951560974121,
|
147 |
+
"learning_rate": 4.514106583072101e-05,
|
148 |
+
"loss": 2.2129,
|
149 |
+
"step": 200
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.3941811356170812,
|
153 |
+
"grad_norm": 9.860595703125,
|
154 |
+
"learning_rate": 4.461859979101359e-05,
|
155 |
+
"loss": 2.1525,
|
156 |
+
"step": 210
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.4129516658845612,
|
160 |
+
"grad_norm": 9.62917709350586,
|
161 |
+
"learning_rate": 4.4096133751306166e-05,
|
162 |
+
"loss": 2.1014,
|
163 |
+
"step": 220
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.43172219615204127,
|
167 |
+
"grad_norm": 10.246869087219238,
|
168 |
+
"learning_rate": 4.357366771159875e-05,
|
169 |
+
"loss": 2.0448,
|
170 |
+
"step": 230
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.45049272641952137,
|
174 |
+
"grad_norm": 10.559744834899902,
|
175 |
+
"learning_rate": 4.305120167189133e-05,
|
176 |
+
"loss": 1.9428,
|
177 |
+
"step": 240
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.4692632566870014,
|
181 |
+
"grad_norm": 10.051835060119629,
|
182 |
+
"learning_rate": 4.252873563218391e-05,
|
183 |
+
"loss": 1.9952,
|
184 |
+
"step": 250
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.48803378695448146,
|
188 |
+
"grad_norm": 10.530678749084473,
|
189 |
+
"learning_rate": 4.2006269592476494e-05,
|
190 |
+
"loss": 1.8988,
|
191 |
+
"step": 260
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.5068043172219615,
|
195 |
+
"grad_norm": 14.675883293151855,
|
196 |
+
"learning_rate": 4.148380355276907e-05,
|
197 |
+
"loss": 1.8214,
|
198 |
+
"step": 270
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.5255748474894416,
|
202 |
+
"grad_norm": 10.792617797851562,
|
203 |
+
"learning_rate": 4.096133751306165e-05,
|
204 |
+
"loss": 1.8403,
|
205 |
+
"step": 280
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.5443453777569216,
|
209 |
+
"grad_norm": 11.018345832824707,
|
210 |
+
"learning_rate": 4.0438871473354236e-05,
|
211 |
+
"loss": 1.8173,
|
212 |
+
"step": 290
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.5631159080244017,
|
216 |
+
"grad_norm": 10.860880851745605,
|
217 |
+
"learning_rate": 3.9916405433646815e-05,
|
218 |
+
"loss": 1.8106,
|
219 |
+
"step": 300
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.5818864382918817,
|
223 |
+
"grad_norm": 12.167359352111816,
|
224 |
+
"learning_rate": 3.939393939393939e-05,
|
225 |
+
"loss": 1.785,
|
226 |
+
"step": 310
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.6006569685593618,
|
230 |
+
"grad_norm": 10.483503341674805,
|
231 |
+
"learning_rate": 3.887147335423197e-05,
|
232 |
+
"loss": 1.7222,
|
233 |
+
"step": 320
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.6194274988268419,
|
237 |
+
"grad_norm": 9.229653358459473,
|
238 |
+
"learning_rate": 3.834900731452456e-05,
|
239 |
+
"loss": 1.6953,
|
240 |
+
"step": 330
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.6381980290943219,
|
244 |
+
"grad_norm": 9.605615615844727,
|
245 |
+
"learning_rate": 3.782654127481714e-05,
|
246 |
+
"loss": 1.7395,
|
247 |
+
"step": 340
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.656968559361802,
|
251 |
+
"grad_norm": 9.905487060546875,
|
252 |
+
"learning_rate": 3.730407523510972e-05,
|
253 |
+
"loss": 1.6551,
|
254 |
+
"step": 350
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.675739089629282,
|
258 |
+
"grad_norm": 13.063451766967773,
|
259 |
+
"learning_rate": 3.67816091954023e-05,
|
260 |
+
"loss": 1.6907,
|
261 |
+
"step": 360
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.6945096198967621,
|
265 |
+
"grad_norm": 11.43543815612793,
|
266 |
+
"learning_rate": 3.625914315569488e-05,
|
267 |
+
"loss": 1.6405,
|
268 |
+
"step": 370
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 0.7132801501642422,
|
272 |
+
"grad_norm": 8.96221923828125,
|
273 |
+
"learning_rate": 3.573667711598746e-05,
|
274 |
+
"loss": 1.6317,
|
275 |
+
"step": 380
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 0.7320506804317222,
|
279 |
+
"grad_norm": 9.918800354003906,
|
280 |
+
"learning_rate": 3.521421107628005e-05,
|
281 |
+
"loss": 1.6596,
|
282 |
+
"step": 390
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 0.7508212106992023,
|
286 |
+
"grad_norm": 9.336323738098145,
|
287 |
+
"learning_rate": 3.469174503657263e-05,
|
288 |
+
"loss": 1.6412,
|
289 |
+
"step": 400
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 0.7695917409666823,
|
293 |
+
"grad_norm": 9.826384544372559,
|
294 |
+
"learning_rate": 3.4169278996865206e-05,
|
295 |
+
"loss": 1.5947,
|
296 |
+
"step": 410
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.7883622712341624,
|
300 |
+
"grad_norm": 11.160934448242188,
|
301 |
+
"learning_rate": 3.3646812957157785e-05,
|
302 |
+
"loss": 1.5278,
|
303 |
+
"step": 420
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 0.8071328015016425,
|
307 |
+
"grad_norm": 13.808747291564941,
|
308 |
+
"learning_rate": 3.3124346917450364e-05,
|
309 |
+
"loss": 1.5359,
|
310 |
+
"step": 430
|
311 |
+
},
|
312 |
+
{
|
313 |
+
"epoch": 0.8259033317691225,
|
314 |
+
"grad_norm": 9.443258285522461,
|
315 |
+
"learning_rate": 3.260188087774295e-05,
|
316 |
+
"loss": 1.544,
|
317 |
+
"step": 440
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 0.8446738620366026,
|
321 |
+
"grad_norm": 9.449996948242188,
|
322 |
+
"learning_rate": 3.2079414838035534e-05,
|
323 |
+
"loss": 1.5372,
|
324 |
+
"step": 450
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 0.8634443923040825,
|
328 |
+
"grad_norm": 10.876261711120605,
|
329 |
+
"learning_rate": 3.155694879832811e-05,
|
330 |
+
"loss": 1.5632,
|
331 |
+
"step": 460
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 0.8822149225715626,
|
335 |
+
"grad_norm": 9.270318031311035,
|
336 |
+
"learning_rate": 3.103448275862069e-05,
|
337 |
+
"loss": 1.4651,
|
338 |
+
"step": 470
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.9009854528390427,
|
342 |
+
"grad_norm": 10.690321922302246,
|
343 |
+
"learning_rate": 3.0512016718913274e-05,
|
344 |
+
"loss": 1.5552,
|
345 |
+
"step": 480
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 0.9197559831065227,
|
349 |
+
"grad_norm": 8.963105201721191,
|
350 |
+
"learning_rate": 2.9989550679205852e-05,
|
351 |
+
"loss": 1.5143,
|
352 |
+
"step": 490
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 0.9385265133740028,
|
356 |
+
"grad_norm": 9.028443336486816,
|
357 |
+
"learning_rate": 2.946708463949843e-05,
|
358 |
+
"loss": 1.4811,
|
359 |
+
"step": 500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 0.9572970436414828,
|
363 |
+
"grad_norm": 9.890482902526855,
|
364 |
+
"learning_rate": 2.8944618599791013e-05,
|
365 |
+
"loss": 1.4827,
|
366 |
+
"step": 510
|
367 |
+
},
|
368 |
+
{
|
369 |
+
"epoch": 0.9760675739089629,
|
370 |
+
"grad_norm": 11.68128490447998,
|
371 |
+
"learning_rate": 2.8422152560083598e-05,
|
372 |
+
"loss": 1.3682,
|
373 |
+
"step": 520
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 0.994838104176443,
|
377 |
+
"grad_norm": 10.292035102844238,
|
378 |
+
"learning_rate": 2.7899686520376177e-05,
|
379 |
+
"loss": 1.3766,
|
380 |
+
"step": 530
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.998592210229939,
|
384 |
+
"eval_accuracy": 0.7071947194719472,
|
385 |
+
"eval_loss": 1.12041175365448,
|
386 |
+
"eval_runtime": 73.0098,
|
387 |
+
"eval_samples_per_second": 103.753,
|
388 |
+
"eval_steps_per_second": 3.246,
|
389 |
+
"step": 532
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 1.013608634443923,
|
393 |
+
"grad_norm": 10.038961410522461,
|
394 |
+
"learning_rate": 2.737722048066876e-05,
|
395 |
+
"loss": 1.3601,
|
396 |
+
"step": 540
|
397 |
+
},
|
398 |
+
{
|
399 |
+
"epoch": 1.0323791647114031,
|
400 |
+
"grad_norm": 15.009783744812012,
|
401 |
+
"learning_rate": 2.6854754440961337e-05,
|
402 |
+
"loss": 1.4014,
|
403 |
+
"step": 550
|
404 |
+
},
|
405 |
+
{
|
406 |
+
"epoch": 1.0511496949788832,
|
407 |
+
"grad_norm": 9.596123695373535,
|
408 |
+
"learning_rate": 2.633228840125392e-05,
|
409 |
+
"loss": 1.3601,
|
410 |
+
"step": 560
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 1.069920225246363,
|
414 |
+
"grad_norm": 11.943960189819336,
|
415 |
+
"learning_rate": 2.5809822361546498e-05,
|
416 |
+
"loss": 1.3876,
|
417 |
+
"step": 570
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 1.0886907555138432,
|
421 |
+
"grad_norm": 10.823417663574219,
|
422 |
+
"learning_rate": 2.5287356321839083e-05,
|
423 |
+
"loss": 1.3681,
|
424 |
+
"step": 580
|
425 |
+
},
|
426 |
+
{
|
427 |
+
"epoch": 1.1074612857813233,
|
428 |
+
"grad_norm": 8.811203002929688,
|
429 |
+
"learning_rate": 2.4764890282131662e-05,
|
430 |
+
"loss": 1.3331,
|
431 |
+
"step": 590
|
432 |
+
},
|
433 |
+
{
|
434 |
+
"epoch": 1.1262318160488034,
|
435 |
+
"grad_norm": 10.159242630004883,
|
436 |
+
"learning_rate": 2.4242424242424244e-05,
|
437 |
+
"loss": 1.4147,
|
438 |
+
"step": 600
|
439 |
+
},
|
440 |
+
{
|
441 |
+
"epoch": 1.1450023463162835,
|
442 |
+
"grad_norm": 10.464547157287598,
|
443 |
+
"learning_rate": 2.3719958202716826e-05,
|
444 |
+
"loss": 1.3975,
|
445 |
+
"step": 610
|
446 |
+
},
|
447 |
+
{
|
448 |
+
"epoch": 1.1637728765837636,
|
449 |
+
"grad_norm": 11.621941566467285,
|
450 |
+
"learning_rate": 2.3197492163009404e-05,
|
451 |
+
"loss": 1.3553,
|
452 |
+
"step": 620
|
453 |
+
},
|
454 |
+
{
|
455 |
+
"epoch": 1.1825434068512435,
|
456 |
+
"grad_norm": 10.114766120910645,
|
457 |
+
"learning_rate": 2.2675026123301986e-05,
|
458 |
+
"loss": 1.3037,
|
459 |
+
"step": 630
|
460 |
+
},
|
461 |
+
{
|
462 |
+
"epoch": 1.2013139371187236,
|
463 |
+
"grad_norm": 13.734535217285156,
|
464 |
+
"learning_rate": 2.215256008359457e-05,
|
465 |
+
"loss": 1.3696,
|
466 |
+
"step": 640
|
467 |
+
},
|
468 |
+
{
|
469 |
+
"epoch": 1.2200844673862037,
|
470 |
+
"grad_norm": 10.053481101989746,
|
471 |
+
"learning_rate": 2.1630094043887147e-05,
|
472 |
+
"loss": 1.3007,
|
473 |
+
"step": 650
|
474 |
+
},
|
475 |
+
{
|
476 |
+
"epoch": 1.2388549976536838,
|
477 |
+
"grad_norm": 10.231536865234375,
|
478 |
+
"learning_rate": 2.110762800417973e-05,
|
479 |
+
"loss": 1.3774,
|
480 |
+
"step": 660
|
481 |
+
},
|
482 |
+
{
|
483 |
+
"epoch": 1.2576255279211637,
|
484 |
+
"grad_norm": 12.41734504699707,
|
485 |
+
"learning_rate": 2.058516196447231e-05,
|
486 |
+
"loss": 1.3729,
|
487 |
+
"step": 670
|
488 |
+
},
|
489 |
+
{
|
490 |
+
"epoch": 1.2763960581886438,
|
491 |
+
"grad_norm": 10.437618255615234,
|
492 |
+
"learning_rate": 2.006269592476489e-05,
|
493 |
+
"loss": 1.3175,
|
494 |
+
"step": 680
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 1.2951665884561239,
|
498 |
+
"grad_norm": 10.459583282470703,
|
499 |
+
"learning_rate": 1.9540229885057475e-05,
|
500 |
+
"loss": 1.3438,
|
501 |
+
"step": 690
|
502 |
+
},
|
503 |
+
{
|
504 |
+
"epoch": 1.313937118723604,
|
505 |
+
"grad_norm": 9.832696914672852,
|
506 |
+
"learning_rate": 1.9017763845350054e-05,
|
507 |
+
"loss": 1.2885,
|
508 |
+
"step": 700
|
509 |
+
},
|
510 |
+
{
|
511 |
+
"epoch": 1.332707648991084,
|
512 |
+
"grad_norm": 12.73790454864502,
|
513 |
+
"learning_rate": 1.8495297805642632e-05,
|
514 |
+
"loss": 1.3343,
|
515 |
+
"step": 710
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 1.3514781792585642,
|
519 |
+
"grad_norm": 12.657564163208008,
|
520 |
+
"learning_rate": 1.7972831765935218e-05,
|
521 |
+
"loss": 1.2856,
|
522 |
+
"step": 720
|
523 |
+
},
|
524 |
+
{
|
525 |
+
"epoch": 1.370248709526044,
|
526 |
+
"grad_norm": 10.494688034057617,
|
527 |
+
"learning_rate": 1.7450365726227796e-05,
|
528 |
+
"loss": 1.2428,
|
529 |
+
"step": 730
|
530 |
+
},
|
531 |
+
{
|
532 |
+
"epoch": 1.3890192397935242,
|
533 |
+
"grad_norm": 9.145332336425781,
|
534 |
+
"learning_rate": 1.6927899686520378e-05,
|
535 |
+
"loss": 1.2834,
|
536 |
+
"step": 740
|
537 |
+
},
|
538 |
+
{
|
539 |
+
"epoch": 1.4077897700610043,
|
540 |
+
"grad_norm": 9.796390533447266,
|
541 |
+
"learning_rate": 1.6405433646812957e-05,
|
542 |
+
"loss": 1.3583,
|
543 |
+
"step": 750
|
544 |
+
},
|
545 |
+
{
|
546 |
+
"epoch": 1.4265603003284844,
|
547 |
+
"grad_norm": 9.282154083251953,
|
548 |
+
"learning_rate": 1.588296760710554e-05,
|
549 |
+
"loss": 1.2639,
|
550 |
+
"step": 760
|
551 |
+
},
|
552 |
+
{
|
553 |
+
"epoch": 1.4453308305959642,
|
554 |
+
"grad_norm": 11.123170852661133,
|
555 |
+
"learning_rate": 1.536050156739812e-05,
|
556 |
+
"loss": 1.3345,
|
557 |
+
"step": 770
|
558 |
+
},
|
559 |
+
{
|
560 |
+
"epoch": 1.4641013608634443,
|
561 |
+
"grad_norm": 10.234476089477539,
|
562 |
+
"learning_rate": 1.48380355276907e-05,
|
563 |
+
"loss": 1.3293,
|
564 |
+
"step": 780
|
565 |
+
},
|
566 |
+
{
|
567 |
+
"epoch": 1.4828718911309244,
|
568 |
+
"grad_norm": 8.585359573364258,
|
569 |
+
"learning_rate": 1.4315569487983283e-05,
|
570 |
+
"loss": 1.2923,
|
571 |
+
"step": 790
|
572 |
+
},
|
573 |
+
{
|
574 |
+
"epoch": 1.5016424213984045,
|
575 |
+
"grad_norm": 9.973495483398438,
|
576 |
+
"learning_rate": 1.3793103448275863e-05,
|
577 |
+
"loss": 1.2716,
|
578 |
+
"step": 800
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 1.5204129516658846,
|
582 |
+
"grad_norm": 10.079455375671387,
|
583 |
+
"learning_rate": 1.3270637408568442e-05,
|
584 |
+
"loss": 1.2429,
|
585 |
+
"step": 810
|
586 |
+
},
|
587 |
+
{
|
588 |
+
"epoch": 1.5391834819333647,
|
589 |
+
"grad_norm": 11.303532600402832,
|
590 |
+
"learning_rate": 1.2748171368861026e-05,
|
591 |
+
"loss": 1.3177,
|
592 |
+
"step": 820
|
593 |
+
},
|
594 |
+
{
|
595 |
+
"epoch": 1.5579540122008448,
|
596 |
+
"grad_norm": 11.871292114257812,
|
597 |
+
"learning_rate": 1.2225705329153606e-05,
|
598 |
+
"loss": 1.3239,
|
599 |
+
"step": 830
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 1.5767245424683247,
|
603 |
+
"grad_norm": 8.814146041870117,
|
604 |
+
"learning_rate": 1.1703239289446186e-05,
|
605 |
+
"loss": 1.1561,
|
606 |
+
"step": 840
|
607 |
+
},
|
608 |
+
{
|
609 |
+
"epoch": 1.5954950727358048,
|
610 |
+
"grad_norm": 10.472289085388184,
|
611 |
+
"learning_rate": 1.1180773249738766e-05,
|
612 |
+
"loss": 1.3309,
|
613 |
+
"step": 850
|
614 |
+
},
|
615 |
+
{
|
616 |
+
"epoch": 1.6142656030032847,
|
617 |
+
"grad_norm": 10.017809867858887,
|
618 |
+
"learning_rate": 1.0658307210031348e-05,
|
619 |
+
"loss": 1.2212,
|
620 |
+
"step": 860
|
621 |
+
},
|
622 |
+
{
|
623 |
+
"epoch": 1.6330361332707648,
|
624 |
+
"grad_norm": 9.293713569641113,
|
625 |
+
"learning_rate": 1.013584117032393e-05,
|
626 |
+
"loss": 1.2172,
|
627 |
+
"step": 870
|
628 |
+
},
|
629 |
+
{
|
630 |
+
"epoch": 1.651806663538245,
|
631 |
+
"grad_norm": 10.923977851867676,
|
632 |
+
"learning_rate": 9.61337513061651e-06,
|
633 |
+
"loss": 1.2031,
|
634 |
+
"step": 880
|
635 |
+
},
|
636 |
+
{
|
637 |
+
"epoch": 1.670577193805725,
|
638 |
+
"grad_norm": 8.05986213684082,
|
639 |
+
"learning_rate": 9.090909090909091e-06,
|
640 |
+
"loss": 1.2337,
|
641 |
+
"step": 890
|
642 |
+
},
|
643 |
+
{
|
644 |
+
"epoch": 1.689347724073205,
|
645 |
+
"grad_norm": 9.424606323242188,
|
646 |
+
"learning_rate": 8.568443051201671e-06,
|
647 |
+
"loss": 1.2331,
|
648 |
+
"step": 900
|
649 |
+
},
|
650 |
+
{
|
651 |
+
"epoch": 1.7081182543406852,
|
652 |
+
"grad_norm": 9.5239896774292,
|
653 |
+
"learning_rate": 8.045977011494253e-06,
|
654 |
+
"loss": 1.1844,
|
655 |
+
"step": 910
|
656 |
+
},
|
657 |
+
{
|
658 |
+
"epoch": 1.7268887846081653,
|
659 |
+
"grad_norm": 9.687056541442871,
|
660 |
+
"learning_rate": 7.5235109717868345e-06,
|
661 |
+
"loss": 1.2396,
|
662 |
+
"step": 920
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 1.7456593148756452,
|
666 |
+
"grad_norm": 11.813615798950195,
|
667 |
+
"learning_rate": 7.001044932079415e-06,
|
668 |
+
"loss": 1.2617,
|
669 |
+
"step": 930
|
670 |
+
},
|
671 |
+
{
|
672 |
+
"epoch": 1.7644298451431253,
|
673 |
+
"grad_norm": 9.889193534851074,
|
674 |
+
"learning_rate": 6.478578892371996e-06,
|
675 |
+
"loss": 1.2525,
|
676 |
+
"step": 940
|
677 |
+
},
|
678 |
+
{
|
679 |
+
"epoch": 1.7832003754106054,
|
680 |
+
"grad_norm": 9.544778823852539,
|
681 |
+
"learning_rate": 5.956112852664577e-06,
|
682 |
+
"loss": 1.2417,
|
683 |
+
"step": 950
|
684 |
+
},
|
685 |
+
{
|
686 |
+
"epoch": 1.8019709056780853,
|
687 |
+
"grad_norm": 10.579482078552246,
|
688 |
+
"learning_rate": 5.433646812957157e-06,
|
689 |
+
"loss": 1.1964,
|
690 |
+
"step": 960
|
691 |
+
},
|
692 |
+
{
|
693 |
+
"epoch": 1.8207414359455654,
|
694 |
+
"grad_norm": 8.800308227539062,
|
695 |
+
"learning_rate": 4.911180773249739e-06,
|
696 |
+
"loss": 1.2285,
|
697 |
+
"step": 970
|
698 |
+
},
|
699 |
+
{
|
700 |
+
"epoch": 1.8395119662130455,
|
701 |
+
"grad_norm": 7.766645431518555,
|
702 |
+
"learning_rate": 4.38871473354232e-06,
|
703 |
+
"loss": 1.204,
|
704 |
+
"step": 980
|
705 |
+
},
|
706 |
+
{
|
707 |
+
"epoch": 1.8582824964805256,
|
708 |
+
"grad_norm": 9.686829566955566,
|
709 |
+
"learning_rate": 3.866248693834901e-06,
|
710 |
+
"loss": 1.2163,
|
711 |
+
"step": 990
|
712 |
+
},
|
713 |
+
{
|
714 |
+
"epoch": 1.8770530267480057,
|
715 |
+
"grad_norm": 9.186079025268555,
|
716 |
+
"learning_rate": 3.3437826541274823e-06,
|
717 |
+
"loss": 1.1354,
|
718 |
+
"step": 1000
|
719 |
+
},
|
720 |
+
{
|
721 |
+
"epoch": 1.8958235570154858,
|
722 |
+
"grad_norm": 9.512069702148438,
|
723 |
+
"learning_rate": 2.821316614420063e-06,
|
724 |
+
"loss": 1.218,
|
725 |
+
"step": 1010
|
726 |
+
},
|
727 |
+
{
|
728 |
+
"epoch": 1.9145940872829659,
|
729 |
+
"grad_norm": 11.176709175109863,
|
730 |
+
"learning_rate": 2.2988505747126437e-06,
|
731 |
+
"loss": 1.2341,
|
732 |
+
"step": 1020
|
733 |
+
},
|
734 |
+
{
|
735 |
+
"epoch": 1.9333646175504458,
|
736 |
+
"grad_norm": 9.307233810424805,
|
737 |
+
"learning_rate": 1.7763845350052249e-06,
|
738 |
+
"loss": 1.1894,
|
739 |
+
"step": 1030
|
740 |
+
},
|
741 |
+
{
|
742 |
+
"epoch": 1.9521351478179259,
|
743 |
+
"grad_norm": 11.547785758972168,
|
744 |
+
"learning_rate": 1.2539184952978056e-06,
|
745 |
+
"loss": 1.2157,
|
746 |
+
"step": 1040
|
747 |
+
},
|
748 |
+
{
|
749 |
+
"epoch": 1.970905678085406,
|
750 |
+
"grad_norm": 10.38868236541748,
|
751 |
+
"learning_rate": 7.314524555903866e-07,
|
752 |
+
"loss": 1.1476,
|
753 |
+
"step": 1050
|
754 |
+
},
|
755 |
+
{
|
756 |
+
"epoch": 1.9896762083528858,
|
757 |
+
"grad_norm": 15.987852096557617,
|
758 |
+
"learning_rate": 2.0898641588296764e-07,
|
759 |
+
"loss": 1.2542,
|
760 |
+
"step": 1060
|
761 |
+
},
|
762 |
+
{
|
763 |
+
"epoch": 1.9971844204598779,
|
764 |
+
"eval_accuracy": 0.7506270627062707,
|
765 |
+
"eval_loss": 0.9383056163787842,
|
766 |
+
"eval_runtime": 72.7566,
|
767 |
+
"eval_samples_per_second": 104.114,
|
768 |
+
"eval_steps_per_second": 3.257,
|
769 |
+
"step": 1064
|
770 |
+
},
|
771 |
+
{
|
772 |
+
"epoch": 1.9971844204598779,
|
773 |
+
"step": 1064,
|
774 |
+
"total_flos": 3.394127568277555e+18,
|
775 |
+
"train_loss": 1.8402288484394103,
|
776 |
+
"train_runtime": 2568.7785,
|
777 |
+
"train_samples_per_second": 53.08,
|
778 |
+
"train_steps_per_second": 0.414
|
779 |
+
}
|
780 |
+
],
|
781 |
+
"logging_steps": 10,
|
782 |
+
"max_steps": 1064,
|
783 |
+
"num_input_tokens_seen": 0,
|
784 |
+
"num_train_epochs": 2,
|
785 |
+
"save_steps": 500,
|
786 |
+
"stateful_callbacks": {
|
787 |
+
"TrainerControl": {
|
788 |
+
"args": {
|
789 |
+
"should_epoch_stop": false,
|
790 |
+
"should_evaluate": false,
|
791 |
+
"should_log": false,
|
792 |
+
"should_save": true,
|
793 |
+
"should_training_stop": true
|
794 |
+
},
|
795 |
+
"attributes": {}
|
796 |
+
}
|
797 |
+
},
|
798 |
+
"total_flos": 3.394127568277555e+18,
|
799 |
+
"train_batch_size": 32,
|
800 |
+
"trial_name": null,
|
801 |
+
"trial_params": null
|
802 |
+
}
|