Upload sigma_lambda_proto_v3/history.json with huggingface_hub
Browse files
sigma_lambda_proto_v3/history.json
CHANGED
|
@@ -5,7 +5,8 @@
|
|
| 5 |
3,
|
| 6 |
4,
|
| 7 |
5,
|
| 8 |
-
6
|
|
|
|
| 9 |
],
|
| 10 |
"gen_loss": [
|
| 11 |
18.370433433850607,
|
|
@@ -13,7 +14,8 @@
|
|
| 13 |
12.081629503279022,
|
| 14 |
11.843723264174027,
|
| 15 |
11.67701819376512,
|
| 16 |
-
11.95506536512664
|
|
|
|
| 17 |
],
|
| 18 |
"ret_loss": [
|
| 19 |
0.0,
|
|
@@ -21,6 +23,7 @@
|
|
| 21 |
0.0,
|
| 22 |
0.0,
|
| 23 |
0.0,
|
|
|
|
| 24 |
0.0
|
| 25 |
],
|
| 26 |
"ent_loss": [
|
|
@@ -29,7 +32,8 @@
|
|
| 29 |
0.6330166747714534,
|
| 30 |
0.6162352153026697,
|
| 31 |
0.6095350569667238,
|
| 32 |
-
0.6105213370467677
|
|
|
|
| 33 |
],
|
| 34 |
"ppl": [
|
| 35 |
40729.60576844635,
|
|
@@ -37,7 +41,8 @@
|
|
| 37 |
337.67548432577263,
|
| 38 |
123.54939191255174,
|
| 39 |
81.4022446898097,
|
| 40 |
-
73.53574793147658
|
|
|
|
| 41 |
],
|
| 42 |
"tf_acc": [
|
| 43 |
9.249682110189831e-06,
|
|
@@ -45,7 +50,8 @@
|
|
| 45 |
0.39405658955515016,
|
| 46 |
0.4409562865464462,
|
| 47 |
0.4514476024551921,
|
| 48 |
-
0.4529803291907453
|
|
|
|
| 49 |
],
|
| 50 |
"fr_acc": [
|
| 51 |
0.0,
|
|
@@ -53,7 +59,8 @@
|
|
| 53 |
0.11453107816089203,
|
| 54 |
0.09257571660973458,
|
| 55 |
0.1671656922337565,
|
| 56 |
-
0.16827387097913035
|
|
|
|
| 57 |
],
|
| 58 |
"def_rate": [
|
| 59 |
0.0,
|
|
@@ -61,6 +68,7 @@
|
|
| 61 |
0.0,
|
| 62 |
0.45,
|
| 63 |
1.0,
|
|
|
|
| 64 |
1.0
|
| 65 |
],
|
| 66 |
"mrr": [
|
|
@@ -69,7 +77,8 @@
|
|
| 69 |
0.004967095330357552,
|
| 70 |
0.004980043042451143,
|
| 71 |
0.004973841365426779,
|
| 72 |
-
0.004973897710442543
|
|
|
|
| 73 |
],
|
| 74 |
"r1": [
|
| 75 |
0.0006249999860301614,
|
|
@@ -77,6 +86,7 @@
|
|
| 77 |
0.0006249999860301614,
|
| 78 |
0.0006249999860301614,
|
| 79 |
0.0006249999860301614,
|
|
|
|
| 80 |
0.0006249999860301614
|
| 81 |
],
|
| 82 |
"ret_ok_pct": [
|
|
@@ -85,6 +95,7 @@
|
|
| 85 |
0.0,
|
| 86 |
0.0,
|
| 87 |
0.0,
|
|
|
|
| 88 |
0.0
|
| 89 |
],
|
| 90 |
"grad_norm": [
|
|
@@ -93,7 +104,8 @@
|
|
| 93 |
1329.5837538555174,
|
| 94 |
1217.5837738567411,
|
| 95 |
1152.3023670419057,
|
| 96 |
-
2138.4472111783607
|
|
|
|
| 97 |
],
|
| 98 |
"step_var": [
|
| 99 |
0.22542811930179596,
|
|
@@ -101,7 +113,8 @@
|
|
| 101 |
0.2280091941356659,
|
| 102 |
0.22952058911323547,
|
| 103 |
0.23098886013031006,
|
| 104 |
-
0.23129060864448547
|
|
|
|
| 105 |
],
|
| 106 |
"pos_slope": [
|
| 107 |
0.49990350008010864,
|
|
@@ -109,7 +122,8 @@
|
|
| 109 |
0.49926072359085083,
|
| 110 |
0.4988322854042053,
|
| 111 |
0.4983888268470764,
|
| 112 |
-
0.4982922673225403
|
|
|
|
| 113 |
],
|
| 114 |
"beta_mean": [
|
| 115 |
0.09998118132352829,
|
|
@@ -117,7 +131,8 @@
|
|
| 117 |
0.09985271841287613,
|
| 118 |
0.09976637363433838,
|
| 119 |
0.09967698156833649,
|
| 120 |
-
0.0996585488319397
|
|
|
|
| 121 |
],
|
| 122 |
"w_ret": [
|
| 123 |
0.0,
|
|
@@ -125,6 +140,7 @@
|
|
| 125 |
0.0,
|
| 126 |
0.0,
|
| 127 |
0.03,
|
| 128 |
-
0.06
|
|
|
|
| 129 |
]
|
| 130 |
}
|
|
|
|
| 5 |
3,
|
| 6 |
4,
|
| 7 |
5,
|
| 8 |
+
6,
|
| 9 |
+
7
|
| 10 |
],
|
| 11 |
"gen_loss": [
|
| 12 |
18.370433433850607,
|
|
|
|
| 14 |
12.081629503279022,
|
| 15 |
11.843723264174027,
|
| 16 |
11.67701819376512,
|
| 17 |
+
11.95506536512664,
|
| 18 |
+
11.651441534504746
|
| 19 |
],
|
| 20 |
"ret_loss": [
|
| 21 |
0.0,
|
|
|
|
| 23 |
0.0,
|
| 24 |
0.0,
|
| 25 |
0.0,
|
| 26 |
+
0.0,
|
| 27 |
0.0
|
| 28 |
],
|
| 29 |
"ent_loss": [
|
|
|
|
| 32 |
0.6330166747714534,
|
| 33 |
0.6162352153026697,
|
| 34 |
0.6095350569667238,
|
| 35 |
+
0.6105213370467677,
|
| 36 |
+
0.6074923021143133
|
| 37 |
],
|
| 38 |
"ppl": [
|
| 39 |
40729.60576844635,
|
|
|
|
| 41 |
337.67548432577263,
|
| 42 |
123.54939191255174,
|
| 43 |
81.4022446898097,
|
| 44 |
+
73.53574793147658,
|
| 45 |
+
66.62547611610893
|
| 46 |
],
|
| 47 |
"tf_acc": [
|
| 48 |
9.249682110189831e-06,
|
|
|
|
| 50 |
0.39405658955515016,
|
| 51 |
0.4409562865464462,
|
| 52 |
0.4514476024551921,
|
| 53 |
+
0.4529803291907453,
|
| 54 |
+
0.4559930050639289
|
| 55 |
],
|
| 56 |
"fr_acc": [
|
| 57 |
0.0,
|
|
|
|
| 59 |
0.11453107816089203,
|
| 60 |
0.09257571660973458,
|
| 61 |
0.1671656922337565,
|
| 62 |
+
0.16827387097913035,
|
| 63 |
+
0.1727579081837558
|
| 64 |
],
|
| 65 |
"def_rate": [
|
| 66 |
0.0,
|
|
|
|
| 68 |
0.0,
|
| 69 |
0.45,
|
| 70 |
1.0,
|
| 71 |
+
1.0,
|
| 72 |
1.0
|
| 73 |
],
|
| 74 |
"mrr": [
|
|
|
|
| 77 |
0.004967095330357552,
|
| 78 |
0.004980043042451143,
|
| 79 |
0.004973841365426779,
|
| 80 |
+
0.004973897710442543,
|
| 81 |
+
0.004973262082785368
|
| 82 |
],
|
| 83 |
"r1": [
|
| 84 |
0.0006249999860301614,
|
|
|
|
| 86 |
0.0006249999860301614,
|
| 87 |
0.0006249999860301614,
|
| 88 |
0.0006249999860301614,
|
| 89 |
+
0.0006249999860301614,
|
| 90 |
0.0006249999860301614
|
| 91 |
],
|
| 92 |
"ret_ok_pct": [
|
|
|
|
| 95 |
0.0,
|
| 96 |
0.0,
|
| 97 |
0.0,
|
| 98 |
+
0.0,
|
| 99 |
0.0
|
| 100 |
],
|
| 101 |
"grad_norm": [
|
|
|
|
| 104 |
1329.5837538555174,
|
| 105 |
1217.5837738567411,
|
| 106 |
1152.3023670419057,
|
| 107 |
+
2138.4472111783607,
|
| 108 |
+
1115.3582513533938
|
| 109 |
],
|
| 110 |
"step_var": [
|
| 111 |
0.22542811930179596,
|
|
|
|
| 113 |
0.2280091941356659,
|
| 114 |
0.22952058911323547,
|
| 115 |
0.23098886013031006,
|
| 116 |
+
0.23129060864448547,
|
| 117 |
+
0.23215171694755554
|
| 118 |
],
|
| 119 |
"pos_slope": [
|
| 120 |
0.49990350008010864,
|
|
|
|
| 122 |
0.49926072359085083,
|
| 123 |
0.4988322854042053,
|
| 124 |
0.4983888268470764,
|
| 125 |
+
0.4982922673225403,
|
| 126 |
+
0.49802443385124207
|
| 127 |
],
|
| 128 |
"beta_mean": [
|
| 129 |
0.09998118132352829,
|
|
|
|
| 131 |
0.09985271841287613,
|
| 132 |
0.09976637363433838,
|
| 133 |
0.09967698156833649,
|
| 134 |
+
0.0996585488319397,
|
| 135 |
+
0.09960417449474335
|
| 136 |
],
|
| 137 |
"w_ret": [
|
| 138 |
0.0,
|
|
|
|
| 140 |
0.0,
|
| 141 |
0.0,
|
| 142 |
0.03,
|
| 143 |
+
0.06,
|
| 144 |
+
0.09
|
| 145 |
]
|
| 146 |
}
|