Training in progress, step 250, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 912336848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6205cde9e35c6ed5e342280482fb19c3c14e67b145df66ba76a8ca5fdec00e23
|
| 3 |
size 912336848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 463916180
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69af6732fcfaa62dbe5fd22121a1bdbca2cb208e4e58004a9e5acd9783b1066b
|
| 3 |
size 463916180
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e057796212acd69adcf8f6a4f8b9390c11503374fb99c4ffc0cb516febcb3920
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d5e1448282b30a66cf7be83aef18a251fdb6205c0184b42e99ae724602144bf
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 50,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1089,6 +1089,722 @@
|
|
| 1089 |
"eval_samples_per_second": 8.878,
|
| 1090 |
"eval_steps_per_second": 2.22,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -1117,7 +1833,7 @@
|
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
-
"total_flos":
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.7680420875549316,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-250",
|
| 4 |
+
"epoch": 0.15852885225110971,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 250,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1089 |
"eval_samples_per_second": 8.878,
|
| 1090 |
"eval_steps_per_second": 2.22,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 0.09575142675967026,
|
| 1095 |
+
"grad_norm": 1.7044169902801514,
|
| 1096 |
+
"learning_rate": 7.107005538862646e-05,
|
| 1097 |
+
"loss": 4.1348,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 0.0963855421686747,
|
| 1102 |
+
"grad_norm": 1.5047065019607544,
|
| 1103 |
+
"learning_rate": 7.07041155014006e-05,
|
| 1104 |
+
"loss": 3.3954,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.09701965757767914,
|
| 1109 |
+
"grad_norm": 1.2422665357589722,
|
| 1110 |
+
"learning_rate": 7.033683215379002e-05,
|
| 1111 |
+
"loss": 3.3153,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.09765377298668358,
|
| 1116 |
+
"grad_norm": 5.329709053039551,
|
| 1117 |
+
"learning_rate": 6.996822917828477e-05,
|
| 1118 |
+
"loss": 3.3014,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.09828788839568801,
|
| 1123 |
+
"grad_norm": 0.9941044449806213,
|
| 1124 |
+
"learning_rate": 6.959833049300377e-05,
|
| 1125 |
+
"loss": 3.1952,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.09892200380469246,
|
| 1130 |
+
"grad_norm": 1.0377179384231567,
|
| 1131 |
+
"learning_rate": 6.922716010014255e-05,
|
| 1132 |
+
"loss": 3.3715,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 0.09955611921369689,
|
| 1137 |
+
"grad_norm": 1.0825616121292114,
|
| 1138 |
+
"learning_rate": 6.885474208441603e-05,
|
| 1139 |
+
"loss": 3.238,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.10019023462270134,
|
| 1144 |
+
"grad_norm": 1.0486564636230469,
|
| 1145 |
+
"learning_rate": 6.848110061149556e-05,
|
| 1146 |
+
"loss": 3.2671,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.10082435003170577,
|
| 1151 |
+
"grad_norm": 1.0950785875320435,
|
| 1152 |
+
"learning_rate": 6.810625992644085e-05,
|
| 1153 |
+
"loss": 3.387,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.10145846544071022,
|
| 1158 |
+
"grad_norm": 1.0026979446411133,
|
| 1159 |
+
"learning_rate": 6.773024435212678e-05,
|
| 1160 |
+
"loss": 3.2341,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.10209258084971465,
|
| 1165 |
+
"grad_norm": 1.0109604597091675,
|
| 1166 |
+
"learning_rate": 6.735307828766515e-05,
|
| 1167 |
+
"loss": 3.1852,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.10272669625871908,
|
| 1172 |
+
"grad_norm": 1.0047856569290161,
|
| 1173 |
+
"learning_rate": 6.697478620682137e-05,
|
| 1174 |
+
"loss": 3.316,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.10336081166772353,
|
| 1179 |
+
"grad_norm": 1.0846145153045654,
|
| 1180 |
+
"learning_rate": 6.659539265642643e-05,
|
| 1181 |
+
"loss": 3.2932,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 0.10399492707672796,
|
| 1186 |
+
"grad_norm": 1.1783442497253418,
|
| 1187 |
+
"learning_rate": 6.621492225478414e-05,
|
| 1188 |
+
"loss": 3.2566,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.1046290424857324,
|
| 1193 |
+
"grad_norm": 1.1924242973327637,
|
| 1194 |
+
"learning_rate": 6.583339969007363e-05,
|
| 1195 |
+
"loss": 3.2032,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.10526315789473684,
|
| 1200 |
+
"grad_norm": 1.0744683742523193,
|
| 1201 |
+
"learning_rate": 6.545084971874738e-05,
|
| 1202 |
+
"loss": 3.2983,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.10589727330374128,
|
| 1207 |
+
"grad_norm": 1.0177804231643677,
|
| 1208 |
+
"learning_rate": 6.506729716392481e-05,
|
| 1209 |
+
"loss": 3.1444,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.10653138871274571,
|
| 1214 |
+
"grad_norm": 1.000748872756958,
|
| 1215 |
+
"learning_rate": 6.468276691378155e-05,
|
| 1216 |
+
"loss": 3.1848,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 0.10716550412175016,
|
| 1221 |
+
"grad_norm": 0.9723776578903198,
|
| 1222 |
+
"learning_rate": 6.429728391993446e-05,
|
| 1223 |
+
"loss": 3.0663,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.10779961953075459,
|
| 1228 |
+
"grad_norm": 1.0768924951553345,
|
| 1229 |
+
"learning_rate": 6.391087319582264e-05,
|
| 1230 |
+
"loss": 3.4086,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.10843373493975904,
|
| 1235 |
+
"grad_norm": 1.0122095346450806,
|
| 1236 |
+
"learning_rate": 6.35235598150842e-05,
|
| 1237 |
+
"loss": 3.1676,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.10906785034876347,
|
| 1242 |
+
"grad_norm": 0.9996663331985474,
|
| 1243 |
+
"learning_rate": 6.313536890992935e-05,
|
| 1244 |
+
"loss": 3.226,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.10970196575776792,
|
| 1249 |
+
"grad_norm": 1.062509298324585,
|
| 1250 |
+
"learning_rate": 6.274632566950967e-05,
|
| 1251 |
+
"loss": 3.2257,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.11033608116677235,
|
| 1256 |
+
"grad_norm": 1.0289890766143799,
|
| 1257 |
+
"learning_rate": 6.235645533828349e-05,
|
| 1258 |
+
"loss": 3.2668,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 0.1109701965757768,
|
| 1263 |
+
"grad_norm": 1.0083286762237549,
|
| 1264 |
+
"learning_rate": 6.19657832143779e-05,
|
| 1265 |
+
"loss": 3.1725,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.11160431198478123,
|
| 1270 |
+
"grad_norm": 1.052514672279358,
|
| 1271 |
+
"learning_rate": 6.157433464794716e-05,
|
| 1272 |
+
"loss": 3.2222,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.11223842739378567,
|
| 1277 |
+
"grad_norm": 1.099542260169983,
|
| 1278 |
+
"learning_rate": 6.118213503952779e-05,
|
| 1279 |
+
"loss": 3.2265,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.1128725428027901,
|
| 1284 |
+
"grad_norm": 1.0204391479492188,
|
| 1285 |
+
"learning_rate": 6.078920983839031e-05,
|
| 1286 |
+
"loss": 3.221,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.11350665821179455,
|
| 1291 |
+
"grad_norm": 1.131372094154358,
|
| 1292 |
+
"learning_rate": 6.0395584540887963e-05,
|
| 1293 |
+
"loss": 3.1018,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 0.11414077362079898,
|
| 1298 |
+
"grad_norm": 1.0413641929626465,
|
| 1299 |
+
"learning_rate": 6.0001284688802226e-05,
|
| 1300 |
+
"loss": 3.1977,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 0.11477488902980343,
|
| 1305 |
+
"grad_norm": 1.0400148630142212,
|
| 1306 |
+
"learning_rate": 5.960633586768543e-05,
|
| 1307 |
+
"loss": 3.2413,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.11540900443880786,
|
| 1312 |
+
"grad_norm": 1.1186673641204834,
|
| 1313 |
+
"learning_rate": 5.921076370520058e-05,
|
| 1314 |
+
"loss": 3.1271,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.1160431198478123,
|
| 1319 |
+
"grad_norm": 5.378909587860107,
|
| 1320 |
+
"learning_rate": 5.8814593869458455e-05,
|
| 1321 |
+
"loss": 3.017,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.11667723525681674,
|
| 1326 |
+
"grad_norm": 1.0467098951339722,
|
| 1327 |
+
"learning_rate": 5.841785206735192e-05,
|
| 1328 |
+
"loss": 3.1646,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.11731135066582118,
|
| 1333 |
+
"grad_norm": 1.2054715156555176,
|
| 1334 |
+
"learning_rate": 5.8020564042888015e-05,
|
| 1335 |
+
"loss": 3.1296,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.11794546607482562,
|
| 1340 |
+
"grad_norm": 1.1292742490768433,
|
| 1341 |
+
"learning_rate": 5.762275557551727e-05,
|
| 1342 |
+
"loss": 3.0876,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.11857958148383006,
|
| 1347 |
+
"grad_norm": 1.0788121223449707,
|
| 1348 |
+
"learning_rate": 5.7224452478461064e-05,
|
| 1349 |
+
"loss": 3.0394,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.1192136968928345,
|
| 1354 |
+
"grad_norm": 1.117724895477295,
|
| 1355 |
+
"learning_rate": 5.682568059703659e-05,
|
| 1356 |
+
"loss": 3.2083,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.11984781230183894,
|
| 1361 |
+
"grad_norm": 1.1697856187820435,
|
| 1362 |
+
"learning_rate": 5.642646580697973e-05,
|
| 1363 |
+
"loss": 3.1939,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.12048192771084337,
|
| 1368 |
+
"grad_norm": 1.1887933015823364,
|
| 1369 |
+
"learning_rate": 5.602683401276615e-05,
|
| 1370 |
+
"loss": 3.022,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.12111604311984782,
|
| 1375 |
+
"grad_norm": 1.121009111404419,
|
| 1376 |
+
"learning_rate": 5.562681114593028e-05,
|
| 1377 |
+
"loss": 2.9906,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 0.12175015852885225,
|
| 1382 |
+
"grad_norm": 1.1305354833602905,
|
| 1383 |
+
"learning_rate": 5.522642316338268e-05,
|
| 1384 |
+
"loss": 3.0668,
|
| 1385 |
+
"step": 192
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.1223842739378567,
|
| 1389 |
+
"grad_norm": 1.2224658727645874,
|
| 1390 |
+
"learning_rate": 5.482569604572576e-05,
|
| 1391 |
+
"loss": 3.1373,
|
| 1392 |
+
"step": 193
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.12301838934686113,
|
| 1396 |
+
"grad_norm": 1.1487756967544556,
|
| 1397 |
+
"learning_rate": 5.442465579556793e-05,
|
| 1398 |
+
"loss": 2.9432,
|
| 1399 |
+
"step": 194
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.12365250475586556,
|
| 1403 |
+
"grad_norm": 1.124640703201294,
|
| 1404 |
+
"learning_rate": 5.402332843583631e-05,
|
| 1405 |
+
"loss": 2.8273,
|
| 1406 |
+
"step": 195
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.12428662016487001,
|
| 1410 |
+
"grad_norm": 1.1626677513122559,
|
| 1411 |
+
"learning_rate": 5.3621740008088126e-05,
|
| 1412 |
+
"loss": 2.7136,
|
| 1413 |
+
"step": 196
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.12492073557387444,
|
| 1417 |
+
"grad_norm": 1.2303059101104736,
|
| 1418 |
+
"learning_rate": 5.321991657082097e-05,
|
| 1419 |
+
"loss": 2.5001,
|
| 1420 |
+
"step": 197
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.12555485098287889,
|
| 1424 |
+
"grad_norm": 1.3254761695861816,
|
| 1425 |
+
"learning_rate": 5.281788419778187e-05,
|
| 1426 |
+
"loss": 2.637,
|
| 1427 |
+
"step": 198
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.12618896639188332,
|
| 1431 |
+
"grad_norm": 1.3234314918518066,
|
| 1432 |
+
"learning_rate": 5.2415668976275355e-05,
|
| 1433 |
+
"loss": 2.5926,
|
| 1434 |
+
"step": 199
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.12682308180088775,
|
| 1438 |
+
"grad_norm": 1.525715708732605,
|
| 1439 |
+
"learning_rate": 5.201329700547076e-05,
|
| 1440 |
+
"loss": 2.5566,
|
| 1441 |
+
"step": 200
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.12682308180088775,
|
| 1445 |
+
"eval_loss": 0.7819186449050903,
|
| 1446 |
+
"eval_runtime": 298.2576,
|
| 1447 |
+
"eval_samples_per_second": 8.905,
|
| 1448 |
+
"eval_steps_per_second": 2.226,
|
| 1449 |
+
"step": 200
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 0.1274571972098922,
|
| 1453 |
+
"grad_norm": 1.2543182373046875,
|
| 1454 |
+
"learning_rate": 5.161079439470866e-05,
|
| 1455 |
+
"loss": 3.6287,
|
| 1456 |
+
"step": 201
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 0.12809131261889664,
|
| 1460 |
+
"grad_norm": 1.2177551984786987,
|
| 1461 |
+
"learning_rate": 5.1208187261806615e-05,
|
| 1462 |
+
"loss": 3.3176,
|
| 1463 |
+
"step": 202
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 0.12872542802790107,
|
| 1467 |
+
"grad_norm": 1.089298129081726,
|
| 1468 |
+
"learning_rate": 5.080550173136457e-05,
|
| 1469 |
+
"loss": 3.3302,
|
| 1470 |
+
"step": 203
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 0.1293595434369055,
|
| 1474 |
+
"grad_norm": 0.9786179065704346,
|
| 1475 |
+
"learning_rate": 5.0402763933069496e-05,
|
| 1476 |
+
"loss": 3.2146,
|
| 1477 |
+
"step": 204
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 0.12999365884590997,
|
| 1481 |
+
"grad_norm": 0.9849365949630737,
|
| 1482 |
+
"learning_rate": 5e-05,
|
| 1483 |
+
"loss": 3.2046,
|
| 1484 |
+
"step": 205
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.1306277742549144,
|
| 1488 |
+
"grad_norm": 0.9649850130081177,
|
| 1489 |
+
"learning_rate": 4.9597236066930516e-05,
|
| 1490 |
+
"loss": 3.3047,
|
| 1491 |
+
"step": 206
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 0.13126188966391883,
|
| 1495 |
+
"grad_norm": 0.9902065992355347,
|
| 1496 |
+
"learning_rate": 4.919449826863544e-05,
|
| 1497 |
+
"loss": 3.2021,
|
| 1498 |
+
"step": 207
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 0.13189600507292326,
|
| 1502 |
+
"grad_norm": 0.9828966856002808,
|
| 1503 |
+
"learning_rate": 4.87918127381934e-05,
|
| 1504 |
+
"loss": 3.2409,
|
| 1505 |
+
"step": 208
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 0.13253012048192772,
|
| 1509 |
+
"grad_norm": 1.0174707174301147,
|
| 1510 |
+
"learning_rate": 4.8389205605291365e-05,
|
| 1511 |
+
"loss": 3.3236,
|
| 1512 |
+
"step": 209
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 0.13316423589093215,
|
| 1516 |
+
"grad_norm": 0.9697160124778748,
|
| 1517 |
+
"learning_rate": 4.798670299452926e-05,
|
| 1518 |
+
"loss": 3.2682,
|
| 1519 |
+
"step": 210
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 0.13379835129993659,
|
| 1523 |
+
"grad_norm": 1.0232049226760864,
|
| 1524 |
+
"learning_rate": 4.758433102372466e-05,
|
| 1525 |
+
"loss": 3.2173,
|
| 1526 |
+
"step": 211
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.13443246670894102,
|
| 1530 |
+
"grad_norm": 0.9926899671554565,
|
| 1531 |
+
"learning_rate": 4.7182115802218126e-05,
|
| 1532 |
+
"loss": 3.2237,
|
| 1533 |
+
"step": 212
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 0.13506658211794548,
|
| 1537 |
+
"grad_norm": 1.0102065801620483,
|
| 1538 |
+
"learning_rate": 4.678008342917903e-05,
|
| 1539 |
+
"loss": 3.1939,
|
| 1540 |
+
"step": 213
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 0.1357006975269499,
|
| 1544 |
+
"grad_norm": 1.0068438053131104,
|
| 1545 |
+
"learning_rate": 4.6378259991911886e-05,
|
| 1546 |
+
"loss": 3.2868,
|
| 1547 |
+
"step": 214
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 0.13633481293595434,
|
| 1551 |
+
"grad_norm": 0.9734184741973877,
|
| 1552 |
+
"learning_rate": 4.597667156416371e-05,
|
| 1553 |
+
"loss": 3.1584,
|
| 1554 |
+
"step": 215
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 0.13696892834495877,
|
| 1558 |
+
"grad_norm": 15.009803771972656,
|
| 1559 |
+
"learning_rate": 4.5575344204432084e-05,
|
| 1560 |
+
"loss": 3.2934,
|
| 1561 |
+
"step": 216
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 0.13760304375396323,
|
| 1565 |
+
"grad_norm": 0.9599727988243103,
|
| 1566 |
+
"learning_rate": 4.5174303954274244e-05,
|
| 1567 |
+
"loss": 3.1804,
|
| 1568 |
+
"step": 217
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.13823715916296767,
|
| 1572 |
+
"grad_norm": 1.007860779762268,
|
| 1573 |
+
"learning_rate": 4.477357683661734e-05,
|
| 1574 |
+
"loss": 3.2044,
|
| 1575 |
+
"step": 218
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 0.1388712745719721,
|
| 1579 |
+
"grad_norm": 0.9852535128593445,
|
| 1580 |
+
"learning_rate": 4.437318885406973e-05,
|
| 1581 |
+
"loss": 3.0663,
|
| 1582 |
+
"step": 219
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 0.13950538998097653,
|
| 1586 |
+
"grad_norm": 0.9651182293891907,
|
| 1587 |
+
"learning_rate": 4.397316598723385e-05,
|
| 1588 |
+
"loss": 3.2841,
|
| 1589 |
+
"step": 220
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 0.140139505389981,
|
| 1593 |
+
"grad_norm": 0.9738379120826721,
|
| 1594 |
+
"learning_rate": 4.3573534193020274e-05,
|
| 1595 |
+
"loss": 3.1556,
|
| 1596 |
+
"step": 221
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 0.14077362079898542,
|
| 1600 |
+
"grad_norm": 1.0059845447540283,
|
| 1601 |
+
"learning_rate": 4.317431940296343e-05,
|
| 1602 |
+
"loss": 3.1792,
|
| 1603 |
+
"step": 222
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 0.14140773620798985,
|
| 1607 |
+
"grad_norm": 1.044323444366455,
|
| 1608 |
+
"learning_rate": 4.277554752153895e-05,
|
| 1609 |
+
"loss": 3.029,
|
| 1610 |
+
"step": 223
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.14204185161699429,
|
| 1614 |
+
"grad_norm": 1.0486642122268677,
|
| 1615 |
+
"learning_rate": 4.237724442448273e-05,
|
| 1616 |
+
"loss": 3.1771,
|
| 1617 |
+
"step": 224
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 0.14267596702599875,
|
| 1621 |
+
"grad_norm": 1.0943423509597778,
|
| 1622 |
+
"learning_rate": 4.197943595711198e-05,
|
| 1623 |
+
"loss": 3.0797,
|
| 1624 |
+
"step": 225
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 0.14331008243500318,
|
| 1628 |
+
"grad_norm": 1.032805323600769,
|
| 1629 |
+
"learning_rate": 4.1582147932648074e-05,
|
| 1630 |
+
"loss": 3.2204,
|
| 1631 |
+
"step": 226
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 0.1439441978440076,
|
| 1635 |
+
"grad_norm": 1.0199164152145386,
|
| 1636 |
+
"learning_rate": 4.118540613054156e-05,
|
| 1637 |
+
"loss": 3.1531,
|
| 1638 |
+
"step": 227
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 0.14457831325301204,
|
| 1642 |
+
"grad_norm": 1.031343698501587,
|
| 1643 |
+
"learning_rate": 4.078923629479943e-05,
|
| 1644 |
+
"loss": 3.1324,
|
| 1645 |
+
"step": 228
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 0.14521242866201647,
|
| 1649 |
+
"grad_norm": 1.0214942693710327,
|
| 1650 |
+
"learning_rate": 4.039366413231458e-05,
|
| 1651 |
+
"loss": 3.0392,
|
| 1652 |
+
"step": 229
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.14584654407102093,
|
| 1656 |
+
"grad_norm": 1.0384721755981445,
|
| 1657 |
+
"learning_rate": 3.9998715311197785e-05,
|
| 1658 |
+
"loss": 3.1602,
|
| 1659 |
+
"step": 230
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 0.14648065948002537,
|
| 1663 |
+
"grad_norm": 1.062154769897461,
|
| 1664 |
+
"learning_rate": 3.960441545911204e-05,
|
| 1665 |
+
"loss": 3.1477,
|
| 1666 |
+
"step": 231
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 0.1471147748890298,
|
| 1670 |
+
"grad_norm": 1.1326991319656372,
|
| 1671 |
+
"learning_rate": 3.92107901616097e-05,
|
| 1672 |
+
"loss": 3.0429,
|
| 1673 |
+
"step": 232
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 0.14774889029803423,
|
| 1677 |
+
"grad_norm": 1.0269993543624878,
|
| 1678 |
+
"learning_rate": 3.8817864960472236e-05,
|
| 1679 |
+
"loss": 3.038,
|
| 1680 |
+
"step": 233
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 0.1483830057070387,
|
| 1684 |
+
"grad_norm": 1.0221335887908936,
|
| 1685 |
+
"learning_rate": 3.842566535205286e-05,
|
| 1686 |
+
"loss": 3.0781,
|
| 1687 |
+
"step": 234
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 0.14901712111604312,
|
| 1691 |
+
"grad_norm": 1.4944452047348022,
|
| 1692 |
+
"learning_rate": 3.803421678562213e-05,
|
| 1693 |
+
"loss": 3.1389,
|
| 1694 |
+
"step": 235
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.14965123652504755,
|
| 1698 |
+
"grad_norm": 1.0359796285629272,
|
| 1699 |
+
"learning_rate": 3.764354466171652e-05,
|
| 1700 |
+
"loss": 3.0269,
|
| 1701 |
+
"step": 236
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 0.15028535193405199,
|
| 1705 |
+
"grad_norm": 1.1374757289886475,
|
| 1706 |
+
"learning_rate": 3.725367433049033e-05,
|
| 1707 |
+
"loss": 3.0241,
|
| 1708 |
+
"step": 237
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 0.15091946734305645,
|
| 1712 |
+
"grad_norm": 1.0923086404800415,
|
| 1713 |
+
"learning_rate": 3.6864631090070655e-05,
|
| 1714 |
+
"loss": 2.9826,
|
| 1715 |
+
"step": 238
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 0.15155358275206088,
|
| 1719 |
+
"grad_norm": 1.085671305656433,
|
| 1720 |
+
"learning_rate": 3.6476440184915815e-05,
|
| 1721 |
+
"loss": 3.0977,
|
| 1722 |
+
"step": 239
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 0.1521876981610653,
|
| 1726 |
+
"grad_norm": 1.0791702270507812,
|
| 1727 |
+
"learning_rate": 3.608912680417737e-05,
|
| 1728 |
+
"loss": 2.9639,
|
| 1729 |
+
"step": 240
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 0.15282181357006974,
|
| 1733 |
+
"grad_norm": 1.1274312734603882,
|
| 1734 |
+
"learning_rate": 3.570271608006555e-05,
|
| 1735 |
+
"loss": 2.8279,
|
| 1736 |
+
"step": 241
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.1534559289790742,
|
| 1740 |
+
"grad_norm": 1.0965379476547241,
|
| 1741 |
+
"learning_rate": 3.531723308621847e-05,
|
| 1742 |
+
"loss": 2.9191,
|
| 1743 |
+
"step": 242
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 0.15409004438807863,
|
| 1747 |
+
"grad_norm": 1.2408086061477661,
|
| 1748 |
+
"learning_rate": 3.493270283607522e-05,
|
| 1749 |
+
"loss": 2.8455,
|
| 1750 |
+
"step": 243
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 0.15472415979708307,
|
| 1754 |
+
"grad_norm": 1.1654704809188843,
|
| 1755 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 1756 |
+
"loss": 2.8552,
|
| 1757 |
+
"step": 244
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 0.1553582752060875,
|
| 1761 |
+
"grad_norm": 1.2121531963348389,
|
| 1762 |
+
"learning_rate": 3.4166600309926387e-05,
|
| 1763 |
+
"loss": 2.8707,
|
| 1764 |
+
"step": 245
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 0.15599239061509196,
|
| 1768 |
+
"grad_norm": 1.23041832447052,
|
| 1769 |
+
"learning_rate": 3.3785077745215873e-05,
|
| 1770 |
+
"loss": 2.8064,
|
| 1771 |
+
"step": 246
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 0.1566265060240964,
|
| 1775 |
+
"grad_norm": 1.237444281578064,
|
| 1776 |
+
"learning_rate": 3.340460734357359e-05,
|
| 1777 |
+
"loss": 2.5598,
|
| 1778 |
+
"step": 247
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.15726062143310082,
|
| 1782 |
+
"grad_norm": 1.2903790473937988,
|
| 1783 |
+
"learning_rate": 3.3025213793178646e-05,
|
| 1784 |
+
"loss": 2.6563,
|
| 1785 |
+
"step": 248
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 0.15789473684210525,
|
| 1789 |
+
"grad_norm": 1.252323865890503,
|
| 1790 |
+
"learning_rate": 3.264692171233485e-05,
|
| 1791 |
+
"loss": 2.4976,
|
| 1792 |
+
"step": 249
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 0.15852885225110971,
|
| 1796 |
+
"grad_norm": 1.4426056146621704,
|
| 1797 |
+
"learning_rate": 3.226975564787322e-05,
|
| 1798 |
+
"loss": 2.5239,
|
| 1799 |
+
"step": 250
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 0.15852885225110971,
|
| 1803 |
+
"eval_loss": 0.7680420875549316,
|
| 1804 |
+
"eval_runtime": 299.1281,
|
| 1805 |
+
"eval_samples_per_second": 8.879,
|
| 1806 |
+
"eval_steps_per_second": 2.22,
|
| 1807 |
+
"step": 250
|
| 1808 |
}
|
| 1809 |
],
|
| 1810 |
"logging_steps": 1,
|
|
|
|
| 1833 |
"attributes": {}
|
| 1834 |
}
|
| 1835 |
},
|
| 1836 |
+
"total_flos": 5.854495639810867e+17,
|
| 1837 |
"train_batch_size": 8,
|
| 1838 |
"trial_name": null,
|
| 1839 |
"trial_params": null
|