Upload sigma_lambda_v5/history.json with huggingface_hub
Browse files- sigma_lambda_v5/history.json +20 -8
sigma_lambda_v5/history.json
CHANGED
|
@@ -5,7 +5,8 @@
|
|
| 5 |
3,
|
| 6 |
4,
|
| 7 |
5,
|
| 8 |
-
6
|
|
|
|
| 9 |
],
|
| 10 |
"gen_loss": [
|
| 11 |
16.586591267879488,
|
|
@@ -13,7 +14,8 @@
|
|
| 13 |
50.61838954157162,
|
| 14 |
114.44467188566325,
|
| 15 |
205.98175486637822,
|
| 16 |
-
333.0307641626024
|
|
|
|
| 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 |
"ppl": [
|
|
@@ -29,6 +32,7 @@
|
|
| 29 |
15470.654177938472,
|
| 30 |
485165195.4097903,
|
| 31 |
485165195.4097903,
|
|
|
|
| 32 |
485165195.4097903
|
| 33 |
],
|
| 34 |
"tf_acc": [
|
|
@@ -37,7 +41,8 @@
|
|
| 37 |
0.47478209936121774,
|
| 38 |
0.4664326184178535,
|
| 39 |
0.4302312730702071,
|
| 40 |
-
0.4411152534131013
|
|
|
|
| 41 |
],
|
| 42 |
"fr_acc": [
|
| 43 |
0.12302286153389094,
|
|
@@ -45,7 +50,8 @@
|
|
| 45 |
0.24578219982631747,
|
| 46 |
0.2376122651857946,
|
| 47 |
0.13397800116550115,
|
| 48 |
-
0.1528195558526441
|
|
|
|
| 49 |
],
|
| 50 |
"mrr": [
|
| 51 |
0.006059600971639156,
|
|
@@ -53,7 +59,8 @@
|
|
| 53 |
0.006044529378414154,
|
| 54 |
0.006051086820662022,
|
| 55 |
0.006043311208486557,
|
| 56 |
-
0.0060408106073737144
|
|
|
|
| 57 |
],
|
| 58 |
"r1": [
|
| 59 |
0.0007812500116415322,
|
|
@@ -61,6 +68,7 @@
|
|
| 61 |
0.0007812500116415322,
|
| 62 |
0.0007812500116415322,
|
| 63 |
0.0007812500116415322,
|
|
|
|
| 64 |
0.0007812500116415322
|
| 65 |
],
|
| 66 |
"step_var": [
|
|
@@ -69,7 +77,8 @@
|
|
| 69 |
0.20552654564380646,
|
| 70 |
0.23860830068588257,
|
| 71 |
0.2775871157646179,
|
| 72 |
-
0.312656432390213
|
|
|
|
| 73 |
],
|
| 74 |
"pos_slope": [
|
| 75 |
0.4985616207122803,
|
|
@@ -77,7 +86,8 @@
|
|
| 77 |
0.48772382736206055,
|
| 78 |
0.48204389214515686,
|
| 79 |
0.47680673003196716,
|
| 80 |
-
0.47214341163635254
|
|
|
|
| 81 |
],
|
| 82 |
"grad_norm": [
|
| 83 |
NaN,
|
|
@@ -85,6 +95,7 @@
|
|
| 85 |
NaN,
|
| 86 |
NaN,
|
| 87 |
NaN,
|
|
|
|
| 88 |
NaN
|
| 89 |
],
|
| 90 |
"w_ret": [
|
|
@@ -93,6 +104,7 @@
|
|
| 93 |
0.0,
|
| 94 |
0.01,
|
| 95 |
0.02,
|
| 96 |
-
0.03
|
|
|
|
| 97 |
]
|
| 98 |
}
|
|
|
|
| 5 |
3,
|
| 6 |
4,
|
| 7 |
5,
|
| 8 |
+
6,
|
| 9 |
+
7
|
| 10 |
],
|
| 11 |
"gen_loss": [
|
| 12 |
16.586591267879488,
|
|
|
|
| 14 |
50.61838954157162,
|
| 15 |
114.44467188566325,
|
| 16 |
205.98175486637822,
|
| 17 |
+
333.0307641626024,
|
| 18 |
+
465.32868966012995
|
| 19 |
],
|
| 20 |
"ret_loss": [
|
| 21 |
0.0,
|
|
|
|
| 23 |
0.0,
|
| 24 |
0.0,
|
| 25 |
0.0,
|
| 26 |
+
0.0,
|
| 27 |
0.0
|
| 28 |
],
|
| 29 |
"ppl": [
|
|
|
|
| 32 |
15470.654177938472,
|
| 33 |
485165195.4097903,
|
| 34 |
485165195.4097903,
|
| 35 |
+
485165195.4097903,
|
| 36 |
485165195.4097903
|
| 37 |
],
|
| 38 |
"tf_acc": [
|
|
|
|
| 41 |
0.47478209936121774,
|
| 42 |
0.4664326184178535,
|
| 43 |
0.4302312730702071,
|
| 44 |
+
0.4411152534131013,
|
| 45 |
+
0.4291537424388662
|
| 46 |
],
|
| 47 |
"fr_acc": [
|
| 48 |
0.12302286153389094,
|
|
|
|
| 50 |
0.24578219982631747,
|
| 51 |
0.2376122651857946,
|
| 52 |
0.13397800116550115,
|
| 53 |
+
0.1528195558526441,
|
| 54 |
+
0.04862503713606654
|
| 55 |
],
|
| 56 |
"mrr": [
|
| 57 |
0.006059600971639156,
|
|
|
|
| 59 |
0.006044529378414154,
|
| 60 |
0.006051086820662022,
|
| 61 |
0.006043311208486557,
|
| 62 |
+
0.0060408106073737144,
|
| 63 |
+
0.006041784770786762
|
| 64 |
],
|
| 65 |
"r1": [
|
| 66 |
0.0007812500116415322,
|
|
|
|
| 68 |
0.0007812500116415322,
|
| 69 |
0.0007812500116415322,
|
| 70 |
0.0007812500116415322,
|
| 71 |
+
0.0007812500116415322,
|
| 72 |
0.0007812500116415322
|
| 73 |
],
|
| 74 |
"step_var": [
|
|
|
|
| 77 |
0.20552654564380646,
|
| 78 |
0.23860830068588257,
|
| 79 |
0.2775871157646179,
|
| 80 |
+
0.312656432390213,
|
| 81 |
+
0.34328973293304443
|
| 82 |
],
|
| 83 |
"pos_slope": [
|
| 84 |
0.4985616207122803,
|
|
|
|
| 86 |
0.48772382736206055,
|
| 87 |
0.48204389214515686,
|
| 88 |
0.47680673003196716,
|
| 89 |
+
0.47214341163635254,
|
| 90 |
+
0.46813833713531494
|
| 91 |
],
|
| 92 |
"grad_norm": [
|
| 93 |
NaN,
|
|
|
|
| 95 |
NaN,
|
| 96 |
NaN,
|
| 97 |
NaN,
|
| 98 |
+
NaN,
|
| 99 |
NaN
|
| 100 |
],
|
| 101 |
"w_ret": [
|
|
|
|
| 104 |
0.0,
|
| 105 |
0.01,
|
| 106 |
0.02,
|
| 107 |
+
0.03,
|
| 108 |
+
0.04
|
| 109 |
]
|
| 110 |
}
|