Upload sigma_lambda_proto_v1/history.json with huggingface_hub
Browse files
sigma_lambda_proto_v1/history.json
CHANGED
|
@@ -8,7 +8,8 @@
|
|
| 8 |
6,
|
| 9 |
7,
|
| 10 |
8,
|
| 11 |
-
9
|
|
|
|
| 12 |
],
|
| 13 |
"gen_loss": [
|
| 14 |
14.378914150180238,
|
|
@@ -19,7 +20,8 @@
|
|
| 19 |
9.424415427897916,
|
| 20 |
9.337661864071181,
|
| 21 |
9.261172692017121,
|
| 22 |
-
9.224323922919504
|
|
|
|
| 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 |
25.650242317335714,
|
| 42 |
25.075453392796494,
|
| 43 |
24.59513520962833,
|
| 44 |
-
24.370497339906883
|
|
|
|
| 45 |
],
|
| 46 |
"tf_acc": [
|
| 47 |
0.44998506448388675,
|
|
@@ -52,7 +56,8 @@
|
|
| 52 |
0.5145881089351385,
|
| 53 |
0.5173037067829007,
|
| 54 |
0.5190997774091205,
|
| 55 |
-
0.5202924423023874
|
|
|
|
| 56 |
],
|
| 57 |
"fr_acc": [
|
| 58 |
0.15843538324420675,
|
|
@@ -63,7 +68,8 @@
|
|
| 63 |
0.132664598473422,
|
| 64 |
0.2427502285296403,
|
| 65 |
0.14725610174139586,
|
| 66 |
-
0.16435602632661456
|
|
|
|
| 67 |
],
|
| 68 |
"mrr": [
|
| 69 |
0.01984136737883091,
|
|
@@ -74,7 +80,8 @@
|
|
| 74 |
0.028147408738732338,
|
| 75 |
0.031810443848371506,
|
| 76 |
0.036788471043109894,
|
| 77 |
-
0.03900999575853348
|
|
|
|
| 78 |
],
|
| 79 |
"r1": [
|
| 80 |
0.0031250000465661287,
|
|
@@ -85,7 +92,8 @@
|
|
| 85 |
0.0031250000465661287,
|
| 86 |
0.0031250000465661287,
|
| 87 |
0.0062500000931322575,
|
| 88 |
-
0.012500000186264515
|
|
|
|
| 89 |
],
|
| 90 |
"he_proxy": [
|
| 91 |
0.0,
|
|
@@ -96,7 +104,8 @@
|
|
| 96 |
0.067495606794824,
|
| 97 |
0.0,
|
| 98 |
0.0,
|
| 99 |
-
0.03818094680227914
|
|
|
|
| 100 |
],
|
| 101 |
"grad_norm": [
|
| 102 |
69556.54960830766,
|
|
@@ -107,7 +116,8 @@
|
|
| 107 |
169132.88870452382,
|
| 108 |
347530.32776330994,
|
| 109 |
319075.012837759,
|
| 110 |
-
164171.89088809182
|
|
|
|
| 111 |
],
|
| 112 |
"step_var": [
|
| 113 |
0.2371368706226349,
|
|
@@ -118,7 +128,8 @@
|
|
| 118 |
0.47667616605758667,
|
| 119 |
0.5235044360160828,
|
| 120 |
0.5645045638084412,
|
| 121 |
-
0.5979146957397461
|
|
|
|
| 122 |
],
|
| 123 |
"pos_slope": [
|
| 124 |
0.4982355535030365,
|
|
@@ -129,7 +140,8 @@
|
|
| 129 |
0.4636712968349457,
|
| 130 |
0.4582301080226898,
|
| 131 |
0.4536069333553314,
|
| 132 |
-
0.44983890652656555
|
|
|
|
| 133 |
],
|
| 134 |
"beta_mean": [
|
| 135 |
0.09964697062969208,
|
|
@@ -140,7 +152,8 @@
|
|
| 140 |
0.09273404628038406,
|
| 141 |
0.09163296222686768,
|
| 142 |
0.09070558100938797,
|
| 143 |
-
0.08995532989501953
|
|
|
|
| 144 |
],
|
| 145 |
"w_ret": [
|
| 146 |
0.0,
|
|
@@ -151,6 +164,7 @@
|
|
| 151 |
0.06,
|
| 152 |
0.09,
|
| 153 |
0.12,
|
|
|
|
| 154 |
0.15
|
| 155 |
]
|
| 156 |
}
|
|
|
|
| 8 |
6,
|
| 9 |
7,
|
| 10 |
8,
|
| 11 |
+
9,
|
| 12 |
+
10
|
| 13 |
],
|
| 14 |
"gen_loss": [
|
| 15 |
14.378914150180238,
|
|
|
|
| 20 |
9.424415427897916,
|
| 21 |
9.337661864071181,
|
| 22 |
9.261172692017121,
|
| 23 |
+
9.224323922919504,
|
| 24 |
+
9.187259725458695
|
| 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 |
25.650242317335714,
|
| 45 |
25.075453392796494,
|
| 46 |
24.59513520962833,
|
| 47 |
+
24.370497339906883,
|
| 48 |
+
24.162149260824183
|
| 49 |
],
|
| 50 |
"tf_acc": [
|
| 51 |
0.44998506448388675,
|
|
|
|
| 56 |
0.5145881089351385,
|
| 57 |
0.5173037067829007,
|
| 58 |
0.5190997774091205,
|
| 59 |
+
0.5202924423023874,
|
| 60 |
+
0.5209306703679905
|
| 61 |
],
|
| 62 |
"fr_acc": [
|
| 63 |
0.15843538324420675,
|
|
|
|
| 68 |
0.132664598473422,
|
| 69 |
0.2427502285296403,
|
| 70 |
0.14725610174139586,
|
| 71 |
+
0.16435602632661456,
|
| 72 |
+
0.1527788861465332
|
| 73 |
],
|
| 74 |
"mrr": [
|
| 75 |
0.01984136737883091,
|
|
|
|
| 80 |
0.028147408738732338,
|
| 81 |
0.031810443848371506,
|
| 82 |
0.036788471043109894,
|
| 83 |
+
0.03900999575853348,
|
| 84 |
+
0.033143915235996246
|
| 85 |
],
|
| 86 |
"r1": [
|
| 87 |
0.0031250000465661287,
|
|
|
|
| 92 |
0.0031250000465661287,
|
| 93 |
0.0031250000465661287,
|
| 94 |
0.0062500000931322575,
|
| 95 |
+
0.012500000186264515,
|
| 96 |
+
0.0031250000465661287
|
| 97 |
],
|
| 98 |
"he_proxy": [
|
| 99 |
0.0,
|
|
|
|
| 104 |
0.067495606794824,
|
| 105 |
0.0,
|
| 106 |
0.0,
|
| 107 |
+
0.03818094680227914,
|
| 108 |
+
0.0
|
| 109 |
],
|
| 110 |
"grad_norm": [
|
| 111 |
69556.54960830766,
|
|
|
|
| 116 |
169132.88870452382,
|
| 117 |
347530.32776330994,
|
| 118 |
319075.012837759,
|
| 119 |
+
164171.89088809182,
|
| 120 |
+
823332.3090508484
|
| 121 |
],
|
| 122 |
"step_var": [
|
| 123 |
0.2371368706226349,
|
|
|
|
| 128 |
0.47667616605758667,
|
| 129 |
0.5235044360160828,
|
| 130 |
0.5645045638084412,
|
| 131 |
+
0.5979146957397461,
|
| 132 |
+
0.6259930729866028
|
| 133 |
],
|
| 134 |
"pos_slope": [
|
| 135 |
0.4982355535030365,
|
|
|
|
| 140 |
0.4636712968349457,
|
| 141 |
0.4582301080226898,
|
| 142 |
0.4536069333553314,
|
| 143 |
+
0.44983890652656555,
|
| 144 |
+
0.4469045102596283
|
| 145 |
],
|
| 146 |
"beta_mean": [
|
| 147 |
0.09964697062969208,
|
|
|
|
| 152 |
0.09273404628038406,
|
| 153 |
0.09163296222686768,
|
| 154 |
0.09070558100938797,
|
| 155 |
+
0.08995532989501953,
|
| 156 |
+
0.0893704891204834
|
| 157 |
],
|
| 158 |
"w_ret": [
|
| 159 |
0.0,
|
|
|
|
| 164 |
0.06,
|
| 165 |
0.09,
|
| 166 |
0.12,
|
| 167 |
+
0.15,
|
| 168 |
0.15
|
| 169 |
]
|
| 170 |
}
|