Upload sigma_lambda_proto_v3/history.json with huggingface_hub
Browse files
sigma_lambda_proto_v3/history.json
CHANGED
|
@@ -9,7 +9,8 @@
|
|
| 9 |
7,
|
| 10 |
8,
|
| 11 |
9,
|
| 12 |
-
10
|
|
|
|
| 13 |
],
|
| 14 |
"gen_loss": [
|
| 15 |
18.370433433850607,
|
|
@@ -21,7 +22,8 @@
|
|
| 21 |
11.651441534504746,
|
| 22 |
11.484277419465961,
|
| 23 |
11.311316739862615,
|
| 24 |
-
11.168713086735119
|
|
|
|
| 25 |
],
|
| 26 |
"ret_loss": [
|
| 27 |
0.0,
|
|
@@ -33,6 +35,7 @@
|
|
| 33 |
0.0,
|
| 34 |
0.0,
|
| 35 |
0.0,
|
|
|
|
| 36 |
0.0
|
| 37 |
],
|
| 38 |
"ent_loss": [
|
|
@@ -45,7 +48,8 @@
|
|
| 45 |
0.6074923021143133,
|
| 46 |
0.606140976638505,
|
| 47 |
0.6052777969114708,
|
| 48 |
-
0.6063350294575547
|
|
|
|
| 49 |
],
|
| 50 |
"ppl": [
|
| 51 |
40729.60576844635,
|
|
@@ -57,7 +61,8 @@
|
|
| 57 |
66.62547611610893,
|
| 58 |
61.79489383359564,
|
| 59 |
57.91288486607438,
|
| 60 |
-
54.61457791073815
|
|
|
|
| 61 |
],
|
| 62 |
"tf_acc": [
|
| 63 |
9.249682110189831e-06,
|
|
@@ -69,7 +74,8 @@
|
|
| 69 |
0.4559930050639289,
|
| 70 |
0.45953291281739334,
|
| 71 |
0.46299827901502855,
|
| 72 |
-
0.4662247857746418
|
|
|
|
| 73 |
],
|
| 74 |
"fr_acc": [
|
| 75 |
0.0,
|
|
@@ -81,7 +87,8 @@
|
|
| 81 |
0.1727579081837558,
|
| 82 |
0.18450658056400712,
|
| 83 |
0.19845598231894296,
|
| 84 |
-
0.19307914900769146
|
|
|
|
| 85 |
],
|
| 86 |
"def_rate": [
|
| 87 |
0.0,
|
|
@@ -93,6 +100,7 @@
|
|
| 93 |
1.0,
|
| 94 |
1.0,
|
| 95 |
1.0,
|
|
|
|
| 96 |
1.0
|
| 97 |
],
|
| 98 |
"mrr": [
|
|
@@ -105,7 +113,8 @@
|
|
| 105 |
0.004973262082785368,
|
| 106 |
0.004975350573658943,
|
| 107 |
0.004973086062818766,
|
| 108 |
-
0.0049743144772946835
|
|
|
|
| 109 |
],
|
| 110 |
"r1": [
|
| 111 |
0.0006249999860301614,
|
|
@@ -117,6 +126,7 @@
|
|
| 117 |
0.0006249999860301614,
|
| 118 |
0.0006249999860301614,
|
| 119 |
0.0006249999860301614,
|
|
|
|
| 120 |
0.0006249999860301614
|
| 121 |
],
|
| 122 |
"ret_ok_pct": [
|
|
@@ -129,6 +139,7 @@
|
|
| 129 |
0.0,
|
| 130 |
0.0,
|
| 131 |
0.0,
|
|
|
|
| 132 |
0.0
|
| 133 |
],
|
| 134 |
"grad_norm": [
|
|
@@ -141,7 +152,8 @@
|
|
| 141 |
1115.3582513533938,
|
| 142 |
1271.3913563283284,
|
| 143 |
954.6965318550485,
|
| 144 |
-
886.9761017929424
|
|
|
|
| 145 |
],
|
| 146 |
"step_var": [
|
| 147 |
0.22542811930179596,
|
|
@@ -153,7 +165,8 @@
|
|
| 153 |
0.23215171694755554,
|
| 154 |
0.23333927989006042,
|
| 155 |
0.23468747735023499,
|
| 156 |
-
0.23607377707958221
|
|
|
|
| 157 |
],
|
| 158 |
"pos_slope": [
|
| 159 |
0.49990350008010864,
|
|
@@ -165,7 +178,8 @@
|
|
| 165 |
0.49802443385124207,
|
| 166 |
0.49764952063560486,
|
| 167 |
0.49722111225128174,
|
| 168 |
-
0.49677765369415283
|
|
|
|
| 169 |
],
|
| 170 |
"beta_mean": [
|
| 171 |
0.09998118132352829,
|
|
@@ -177,7 +191,8 @@
|
|
| 177 |
0.09960417449474335,
|
| 178 |
0.09952875226736069,
|
| 179 |
0.0994425043463707,
|
| 180 |
-
0.09935332834720612
|
|
|
|
| 181 |
],
|
| 182 |
"w_ret": [
|
| 183 |
0.0,
|
|
@@ -189,6 +204,7 @@
|
|
| 189 |
0.09,
|
| 190 |
0.12,
|
| 191 |
0.15,
|
| 192 |
-
0.15
|
|
|
|
| 193 |
]
|
| 194 |
}
|
|
|
|
| 9 |
7,
|
| 10 |
8,
|
| 11 |
9,
|
| 12 |
+
10,
|
| 13 |
+
11
|
| 14 |
],
|
| 15 |
"gen_loss": [
|
| 16 |
18.370433433850607,
|
|
|
|
| 22 |
11.651441534504746,
|
| 23 |
11.484277419465961,
|
| 24 |
11.311316739862615,
|
| 25 |
+
11.168713086735119,
|
| 26 |
+
11.069438566150087
|
| 27 |
],
|
| 28 |
"ret_loss": [
|
| 29 |
0.0,
|
|
|
|
| 35 |
0.0,
|
| 36 |
0.0,
|
| 37 |
0.0,
|
| 38 |
+
0.0,
|
| 39 |
0.0
|
| 40 |
],
|
| 41 |
"ent_loss": [
|
|
|
|
| 48 |
0.6074923021143133,
|
| 49 |
0.606140976638505,
|
| 50 |
0.6052777969114708,
|
| 51 |
+
0.6063350294575547,
|
| 52 |
+
0.601400239648241
|
| 53 |
],
|
| 54 |
"ppl": [
|
| 55 |
40729.60576844635,
|
|
|
|
| 61 |
66.62547611610893,
|
| 62 |
61.79489383359564,
|
| 63 |
57.91288486607438,
|
| 64 |
+
54.61457791073815,
|
| 65 |
+
51.96517694532289
|
| 66 |
],
|
| 67 |
"tf_acc": [
|
| 68 |
9.249682110189831e-06,
|
|
|
|
| 74 |
0.4559930050639289,
|
| 75 |
0.45953291281739334,
|
| 76 |
0.46299827901502855,
|
| 77 |
+
0.4662247857746418,
|
| 78 |
+
0.4692156976899191
|
| 79 |
],
|
| 80 |
"fr_acc": [
|
| 81 |
0.0,
|
|
|
|
| 87 |
0.1727579081837558,
|
| 88 |
0.18450658056400712,
|
| 89 |
0.19845598231894296,
|
| 90 |
+
0.19307914900769146,
|
| 91 |
+
0.16651888731103695
|
| 92 |
],
|
| 93 |
"def_rate": [
|
| 94 |
0.0,
|
|
|
|
| 100 |
1.0,
|
| 101 |
1.0,
|
| 102 |
1.0,
|
| 103 |
+
1.0,
|
| 104 |
1.0
|
| 105 |
],
|
| 106 |
"mrr": [
|
|
|
|
| 113 |
0.004973262082785368,
|
| 114 |
0.004975350573658943,
|
| 115 |
0.004973086062818766,
|
| 116 |
+
0.0049743144772946835,
|
| 117 |
+
0.004973770119249821
|
| 118 |
],
|
| 119 |
"r1": [
|
| 120 |
0.0006249999860301614,
|
|
|
|
| 126 |
0.0006249999860301614,
|
| 127 |
0.0006249999860301614,
|
| 128 |
0.0006249999860301614,
|
| 129 |
+
0.0006249999860301614,
|
| 130 |
0.0006249999860301614
|
| 131 |
],
|
| 132 |
"ret_ok_pct": [
|
|
|
|
| 139 |
0.0,
|
| 140 |
0.0,
|
| 141 |
0.0,
|
| 142 |
+
0.0,
|
| 143 |
0.0
|
| 144 |
],
|
| 145 |
"grad_norm": [
|
|
|
|
| 152 |
1115.3582513533938,
|
| 153 |
1271.3913563283284,
|
| 154 |
954.6965318550485,
|
| 155 |
+
886.9761017929424,
|
| 156 |
+
1321.5633731934518
|
| 157 |
],
|
| 158 |
"step_var": [
|
| 159 |
0.22542811930179596,
|
|
|
|
| 165 |
0.23215171694755554,
|
| 166 |
0.23333927989006042,
|
| 167 |
0.23468747735023499,
|
| 168 |
+
0.23607377707958221,
|
| 169 |
+
0.23739510774612427
|
| 170 |
],
|
| 171 |
"pos_slope": [
|
| 172 |
0.49990350008010864,
|
|
|
|
| 178 |
0.49802443385124207,
|
| 179 |
0.49764952063560486,
|
| 180 |
0.49722111225128174,
|
| 181 |
+
0.49677765369415283,
|
| 182 |
+
0.4963497519493103
|
| 183 |
],
|
| 184 |
"beta_mean": [
|
| 185 |
0.09998118132352829,
|
|
|
|
| 191 |
0.09960417449474335,
|
| 192 |
0.09952875226736069,
|
| 193 |
0.0994425043463707,
|
| 194 |
+
0.09935332834720612,
|
| 195 |
+
0.09926743805408478
|
| 196 |
],
|
| 197 |
"w_ret": [
|
| 198 |
0.0,
|
|
|
|
| 204 |
0.09,
|
| 205 |
0.12,
|
| 206 |
0.15,
|
| 207 |
+
0.15,
|
| 208 |
+
0.2
|
| 209 |
]
|
| 210 |
}
|