Upload sigma_lambda_v5/history.json with huggingface_hub
Browse files- sigma_lambda_v5/history.json +21 -9
sigma_lambda_v5/history.json
CHANGED
|
@@ -8,7 +8,8 @@
|
|
| 8 |
6,
|
| 9 |
7,
|
| 10 |
4,
|
| 11 |
-
5
|
|
|
|
| 12 |
],
|
| 13 |
"gen_loss": [
|
| 14 |
16.586591267879488,
|
|
@@ -19,7 +20,8 @@
|
|
| 19 |
333.0307641626024,
|
| 20 |
465.32868966012995,
|
| 21 |
48.11255316601957,
|
| 22 |
-
74.47382958961273
|
|
|
|
| 23 |
],
|
| 24 |
"ret_loss": [
|
| 25 |
0.0,
|
|
@@ -30,6 +32,7 @@
|
|
| 30 |
0.0,
|
| 31 |
0.0,
|
| 32 |
0.0,
|
|
|
|
| 33 |
0.0
|
| 34 |
],
|
| 35 |
"ppl": [
|
|
@@ -41,7 +44,8 @@
|
|
| 41 |
485165195.4097903,
|
| 42 |
485165195.4097903,
|
| 43 |
15048.086253037915,
|
| 44 |
-
1624266.5983328498
|
|
|
|
| 45 |
],
|
| 46 |
"tf_acc": [
|
| 47 |
0.46026214386231185,
|
|
@@ -52,7 +56,8 @@
|
|
| 52 |
0.4411152534131013,
|
| 53 |
0.4291537424388662,
|
| 54 |
0.4800979171424993,
|
| 55 |
-
0.4732017211019178
|
|
|
|
| 56 |
],
|
| 57 |
"fr_acc": [
|
| 58 |
0.12302286153389094,
|
|
@@ -63,7 +68,8 @@
|
|
| 63 |
0.1528195558526441,
|
| 64 |
0.04862503713606654,
|
| 65 |
0.2002858762969057,
|
| 66 |
-
0.24578219982631747
|
|
|
|
| 67 |
],
|
| 68 |
"mrr": [
|
| 69 |
0.006059600971639156,
|
|
@@ -74,7 +80,8 @@
|
|
| 74 |
0.0060408106073737144,
|
| 75 |
0.006041784770786762,
|
| 76 |
0.0060499003157019615,
|
| 77 |
-
0.006042419467121363
|
|
|
|
| 78 |
],
|
| 79 |
"r1": [
|
| 80 |
0.0007812500116415322,
|
|
@@ -85,6 +92,7 @@
|
|
| 85 |
0.0007812500116415322,
|
| 86 |
0.0007812500116415322,
|
| 87 |
0.0007812500116415322,
|
|
|
|
| 88 |
0.0007812500116415322
|
| 89 |
],
|
| 90 |
"step_var": [
|
|
@@ -96,7 +104,8 @@
|
|
| 96 |
0.312656432390213,
|
| 97 |
0.34328973293304443,
|
| 98 |
0.20889317989349365,
|
| 99 |
-
0.22339501976966858
|
|
|
|
| 100 |
],
|
| 101 |
"pos_slope": [
|
| 102 |
0.4985616207122803,
|
|
@@ -107,7 +116,8 @@
|
|
| 107 |
0.47214341163635254,
|
| 108 |
0.46813833713531494,
|
| 109 |
0.48697665333747864,
|
| 110 |
-
0.48436233401298523
|
|
|
|
| 111 |
],
|
| 112 |
"grad_norm": [
|
| 113 |
NaN,
|
|
@@ -118,6 +128,7 @@
|
|
| 118 |
NaN,
|
| 119 |
NaN,
|
| 120 |
NaN,
|
|
|
|
| 121 |
NaN
|
| 122 |
],
|
| 123 |
"w_ret": [
|
|
@@ -129,6 +140,7 @@
|
|
| 129 |
0.03,
|
| 130 |
0.04,
|
| 131 |
0.01,
|
| 132 |
-
0.02
|
|
|
|
| 133 |
]
|
| 134 |
}
|
|
|
|
| 8 |
6,
|
| 9 |
7,
|
| 10 |
4,
|
| 11 |
+
5,
|
| 12 |
+
6
|
| 13 |
],
|
| 14 |
"gen_loss": [
|
| 15 |
16.586591267879488,
|
|
|
|
| 20 |
333.0307641626024,
|
| 21 |
465.32868966012995,
|
| 22 |
48.11255316601957,
|
| 23 |
+
74.47382958961273,
|
| 24 |
+
116.53298210471324
|
| 25 |
],
|
| 26 |
"ret_loss": [
|
| 27 |
0.0,
|
|
|
|
| 32 |
0.0,
|
| 33 |
0.0,
|
| 34 |
0.0,
|
| 35 |
+
0.0,
|
| 36 |
0.0
|
| 37 |
],
|
| 38 |
"ppl": [
|
|
|
|
| 44 |
485165195.4097903,
|
| 45 |
485165195.4097903,
|
| 46 |
15048.086253037915,
|
| 47 |
+
1624266.5983328498,
|
| 48 |
+
485165195.4097903
|
| 49 |
],
|
| 50 |
"tf_acc": [
|
| 51 |
0.46026214386231185,
|
|
|
|
| 56 |
0.4411152534131013,
|
| 57 |
0.4291537424388662,
|
| 58 |
0.4800979171424993,
|
| 59 |
+
0.4732017211019178,
|
| 60 |
+
0.46131388743580864
|
| 61 |
],
|
| 62 |
"fr_acc": [
|
| 63 |
0.12302286153389094,
|
|
|
|
| 68 |
0.1528195558526441,
|
| 69 |
0.04862503713606654,
|
| 70 |
0.2002858762969057,
|
| 71 |
+
0.24578219982631747,
|
| 72 |
+
0.27369709538827186
|
| 73 |
],
|
| 74 |
"mrr": [
|
| 75 |
0.006059600971639156,
|
|
|
|
| 80 |
0.0060408106073737144,
|
| 81 |
0.006041784770786762,
|
| 82 |
0.0060499003157019615,
|
| 83 |
+
0.006042419467121363,
|
| 84 |
+
0.006052091252058744
|
| 85 |
],
|
| 86 |
"r1": [
|
| 87 |
0.0007812500116415322,
|
|
|
|
| 92 |
0.0007812500116415322,
|
| 93 |
0.0007812500116415322,
|
| 94 |
0.0007812500116415322,
|
| 95 |
+
0.0007812500116415322,
|
| 96 |
0.0007812500116415322
|
| 97 |
],
|
| 98 |
"step_var": [
|
|
|
|
| 104 |
0.312656432390213,
|
| 105 |
0.34328973293304443,
|
| 106 |
0.20889317989349365,
|
| 107 |
+
0.22339501976966858,
|
| 108 |
+
0.24468941986560822
|
| 109 |
],
|
| 110 |
"pos_slope": [
|
| 111 |
0.4985616207122803,
|
|
|
|
| 116 |
0.47214341163635254,
|
| 117 |
0.46813833713531494,
|
| 118 |
0.48697665333747864,
|
| 119 |
+
0.48436233401298523,
|
| 120 |
+
0.4812909960746765
|
| 121 |
],
|
| 122 |
"grad_norm": [
|
| 123 |
NaN,
|
|
|
|
| 128 |
NaN,
|
| 129 |
NaN,
|
| 130 |
NaN,
|
| 131 |
+
NaN,
|
| 132 |
NaN
|
| 133 |
],
|
| 134 |
"w_ret": [
|
|
|
|
| 140 |
0.03,
|
| 141 |
0.04,
|
| 142 |
0.01,
|
| 143 |
+
0.02,
|
| 144 |
+
0.03
|
| 145 |
]
|
| 146 |
}
|