Upload sigma_lambda_v5/history.json with huggingface_hub
Browse files- sigma_lambda_v5/history.json +21 -9
sigma_lambda_v5/history.json
CHANGED
|
@@ -6,7 +6,8 @@
|
|
| 6 |
4,
|
| 7 |
5,
|
| 8 |
6,
|
| 9 |
-
7
|
|
|
|
| 10 |
],
|
| 11 |
"gen_loss": [
|
| 12 |
16.586591267879488,
|
|
@@ -15,7 +16,8 @@
|
|
| 15 |
114.44467188566325,
|
| 16 |
205.98175486637822,
|
| 17 |
333.0307641626024,
|
| 18 |
-
465.32868966012995
|
|
|
|
| 19 |
],
|
| 20 |
"ret_loss": [
|
| 21 |
0.0,
|
|
@@ -24,6 +26,7 @@
|
|
| 24 |
0.0,
|
| 25 |
0.0,
|
| 26 |
0.0,
|
|
|
|
| 27 |
0.0
|
| 28 |
],
|
| 29 |
"ppl": [
|
|
@@ -33,7 +36,8 @@
|
|
| 33 |
485165195.4097903,
|
| 34 |
485165195.4097903,
|
| 35 |
485165195.4097903,
|
| 36 |
-
485165195.4097903
|
|
|
|
| 37 |
],
|
| 38 |
"tf_acc": [
|
| 39 |
0.46026214386231185,
|
|
@@ -42,7 +46,8 @@
|
|
| 42 |
0.4664326184178535,
|
| 43 |
0.4302312730702071,
|
| 44 |
0.4411152534131013,
|
| 45 |
-
0.4291537424388662
|
|
|
|
| 46 |
],
|
| 47 |
"fr_acc": [
|
| 48 |
0.12302286153389094,
|
|
@@ -51,7 +56,8 @@
|
|
| 51 |
0.2376122651857946,
|
| 52 |
0.13397800116550115,
|
| 53 |
0.1528195558526441,
|
| 54 |
-
0.04862503713606654
|
|
|
|
| 55 |
],
|
| 56 |
"mrr": [
|
| 57 |
0.006059600971639156,
|
|
@@ -60,7 +66,8 @@
|
|
| 60 |
0.006051086820662022,
|
| 61 |
0.006043311208486557,
|
| 62 |
0.0060408106073737144,
|
| 63 |
-
0.006041784770786762
|
|
|
|
| 64 |
],
|
| 65 |
"r1": [
|
| 66 |
0.0007812500116415322,
|
|
@@ -69,6 +76,7 @@
|
|
| 69 |
0.0007812500116415322,
|
| 70 |
0.0007812500116415322,
|
| 71 |
0.0007812500116415322,
|
|
|
|
| 72 |
0.0007812500116415322
|
| 73 |
],
|
| 74 |
"step_var": [
|
|
@@ -78,7 +86,8 @@
|
|
| 78 |
0.23860830068588257,
|
| 79 |
0.2775871157646179,
|
| 80 |
0.312656432390213,
|
| 81 |
-
0.34328973293304443
|
|
|
|
| 82 |
],
|
| 83 |
"pos_slope": [
|
| 84 |
0.4985616207122803,
|
|
@@ -87,7 +96,8 @@
|
|
| 87 |
0.48204389214515686,
|
| 88 |
0.47680673003196716,
|
| 89 |
0.47214341163635254,
|
| 90 |
-
0.46813833713531494
|
|
|
|
| 91 |
],
|
| 92 |
"grad_norm": [
|
| 93 |
NaN,
|
|
@@ -96,6 +106,7 @@
|
|
| 96 |
NaN,
|
| 97 |
NaN,
|
| 98 |
NaN,
|
|
|
|
| 99 |
NaN
|
| 100 |
],
|
| 101 |
"w_ret": [
|
|
@@ -105,6 +116,7 @@
|
|
| 105 |
0.01,
|
| 106 |
0.02,
|
| 107 |
0.03,
|
| 108 |
-
0.04
|
|
|
|
| 109 |
]
|
| 110 |
}
|
|
|
|
| 6 |
4,
|
| 7 |
5,
|
| 8 |
6,
|
| 9 |
+
7,
|
| 10 |
+
4
|
| 11 |
],
|
| 12 |
"gen_loss": [
|
| 13 |
16.586591267879488,
|
|
|
|
| 16 |
114.44467188566325,
|
| 17 |
205.98175486637822,
|
| 18 |
333.0307641626024,
|
| 19 |
+
465.32868966012995,
|
| 20 |
+
48.11255316601957
|
| 21 |
],
|
| 22 |
"ret_loss": [
|
| 23 |
0.0,
|
|
|
|
| 26 |
0.0,
|
| 27 |
0.0,
|
| 28 |
0.0,
|
| 29 |
+
0.0,
|
| 30 |
0.0
|
| 31 |
],
|
| 32 |
"ppl": [
|
|
|
|
| 36 |
485165195.4097903,
|
| 37 |
485165195.4097903,
|
| 38 |
485165195.4097903,
|
| 39 |
+
485165195.4097903,
|
| 40 |
+
15048.086253037915
|
| 41 |
],
|
| 42 |
"tf_acc": [
|
| 43 |
0.46026214386231185,
|
|
|
|
| 46 |
0.4664326184178535,
|
| 47 |
0.4302312730702071,
|
| 48 |
0.4411152534131013,
|
| 49 |
+
0.4291537424388662,
|
| 50 |
+
0.4800979171424993
|
| 51 |
],
|
| 52 |
"fr_acc": [
|
| 53 |
0.12302286153389094,
|
|
|
|
| 56 |
0.2376122651857946,
|
| 57 |
0.13397800116550115,
|
| 58 |
0.1528195558526441,
|
| 59 |
+
0.04862503713606654,
|
| 60 |
+
0.2002858762969057
|
| 61 |
],
|
| 62 |
"mrr": [
|
| 63 |
0.006059600971639156,
|
|
|
|
| 66 |
0.006051086820662022,
|
| 67 |
0.006043311208486557,
|
| 68 |
0.0060408106073737144,
|
| 69 |
+
0.006041784770786762,
|
| 70 |
+
0.0060499003157019615
|
| 71 |
],
|
| 72 |
"r1": [
|
| 73 |
0.0007812500116415322,
|
|
|
|
| 76 |
0.0007812500116415322,
|
| 77 |
0.0007812500116415322,
|
| 78 |
0.0007812500116415322,
|
| 79 |
+
0.0007812500116415322,
|
| 80 |
0.0007812500116415322
|
| 81 |
],
|
| 82 |
"step_var": [
|
|
|
|
| 86 |
0.23860830068588257,
|
| 87 |
0.2775871157646179,
|
| 88 |
0.312656432390213,
|
| 89 |
+
0.34328973293304443,
|
| 90 |
+
0.20889317989349365
|
| 91 |
],
|
| 92 |
"pos_slope": [
|
| 93 |
0.4985616207122803,
|
|
|
|
| 96 |
0.48204389214515686,
|
| 97 |
0.47680673003196716,
|
| 98 |
0.47214341163635254,
|
| 99 |
+
0.46813833713531494,
|
| 100 |
+
0.48697665333747864
|
| 101 |
],
|
| 102 |
"grad_norm": [
|
| 103 |
NaN,
|
|
|
|
| 106 |
NaN,
|
| 107 |
NaN,
|
| 108 |
NaN,
|
| 109 |
+
NaN,
|
| 110 |
NaN
|
| 111 |
],
|
| 112 |
"w_ret": [
|
|
|
|
| 116 |
0.01,
|
| 117 |
0.02,
|
| 118 |
0.03,
|
| 119 |
+
0.04,
|
| 120 |
+
0.01
|
| 121 |
]
|
| 122 |
}
|