Training in progress, step 192, 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 639691872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c90bf8e9dd144d9601754a9b2cba7d26c60f5ab5da659a728e65f7b833bc1e3a
|
| 3 |
size 639691872
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 325339796
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:226d054da1d7ad0fe9cb6455e078f6180a8bf321d250443c9f36ca6d8a1c1ba0
|
| 3 |
size 325339796
|
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 sha256:9019d2347240393dd1a2c212e11f37d00de8c19095a9bc69b4bb6f2337ffe4b7
|
| 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 sha256:6dd7378526c26b25312f7867118a6b4fdf26dc64ecd157b60ae548c66129cd40
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 2.4987945556640625,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-100",
|
| 4 |
-
"epoch":
|
| 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 +1089,300 @@
|
|
| 1089 |
"eval_samples_per_second": 13.284,
|
| 1090 |
"eval_steps_per_second": 3.321,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -1112,12 +1406,12 @@
|
|
| 1112 |
"should_evaluate": false,
|
| 1113 |
"should_log": false,
|
| 1114 |
"should_save": true,
|
| 1115 |
-
"should_training_stop":
|
| 1116 |
},
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
-
"total_flos": 2.
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 2.4987945556640625,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-100",
|
| 4 |
+
"epoch": 3.011764705882353,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 192,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1089 |
"eval_samples_per_second": 13.284,
|
| 1090 |
"eval_steps_per_second": 3.321,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.368627450980392,
|
| 1095 |
+
"grad_norm": 0.6427918076515198,
|
| 1096 |
+
"learning_rate": 1.2007741992771065e-05,
|
| 1097 |
+
"loss": 2.0535,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 2.384313725490196,
|
| 1102 |
+
"grad_norm": 0.7129402160644531,
|
| 1103 |
+
"learning_rate": 1.145234126025102e-05,
|
| 1104 |
+
"loss": 2.108,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 2.4,
|
| 1109 |
+
"grad_norm": 0.7248530983924866,
|
| 1110 |
+
"learning_rate": 1.090842587659851e-05,
|
| 1111 |
+
"loss": 2.0189,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.4156862745098038,
|
| 1116 |
+
"grad_norm": 0.8135029673576355,
|
| 1117 |
+
"learning_rate": 1.0376157902445488e-05,
|
| 1118 |
+
"loss": 1.9979,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 2.431372549019608,
|
| 1123 |
+
"grad_norm": 0.8778344392776489,
|
| 1124 |
+
"learning_rate": 9.85569592805588e-06,
|
| 1125 |
+
"loss": 2.0551,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 2.447058823529412,
|
| 1130 |
+
"grad_norm": 0.9533079266548157,
|
| 1131 |
+
"learning_rate": 9.347195026073369e-06,
|
| 1132 |
+
"loss": 1.9452,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.462745098039216,
|
| 1137 |
+
"grad_norm": 1.3244249820709229,
|
| 1138 |
+
"learning_rate": 8.850806705317183e-06,
|
| 1139 |
+
"loss": 2.0356,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 2.4784313725490197,
|
| 1144 |
+
"grad_norm": 0.899782657623291,
|
| 1145 |
+
"learning_rate": 8.366678865639688e-06,
|
| 1146 |
+
"loss": 2.1336,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 2.4941176470588236,
|
| 1151 |
+
"grad_norm": 0.44793930649757385,
|
| 1152 |
+
"learning_rate": 7.894955753859413e-06,
|
| 1153 |
+
"loss": 1.9492,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.5098039215686274,
|
| 1158 |
+
"grad_norm": 0.5112513899803162,
|
| 1159 |
+
"learning_rate": 7.435777920782444e-06,
|
| 1160 |
+
"loss": 2.1618,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 2.5254901960784313,
|
| 1165 |
+
"grad_norm": 0.5470635890960693,
|
| 1166 |
+
"learning_rate": 6.989282179324963e-06,
|
| 1167 |
+
"loss": 2.1548,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 2.541176470588235,
|
| 1172 |
+
"grad_norm": 0.5480771064758301,
|
| 1173 |
+
"learning_rate": 6.555601563749675e-06,
|
| 1174 |
+
"loss": 2.1317,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.556862745098039,
|
| 1179 |
+
"grad_norm": 0.6045755743980408,
|
| 1180 |
+
"learning_rate": 6.1348652900279025e-06,
|
| 1181 |
+
"loss": 2.273,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 2.572549019607843,
|
| 1186 |
+
"grad_norm": 0.6052933931350708,
|
| 1187 |
+
"learning_rate": 5.727198717339511e-06,
|
| 1188 |
+
"loss": 2.1034,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 2.588235294117647,
|
| 1193 |
+
"grad_norm": 0.6706366539001465,
|
| 1194 |
+
"learning_rate": 5.332723310721854e-06,
|
| 1195 |
+
"loss": 2.1542,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.603921568627451,
|
| 1200 |
+
"grad_norm": 0.7189547419548035,
|
| 1201 |
+
"learning_rate": 4.951556604879048e-06,
|
| 1202 |
+
"loss": 2.1109,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 2.619607843137255,
|
| 1207 |
+
"grad_norm": 0.7314002513885498,
|
| 1208 |
+
"learning_rate": 4.5838121691623e-06,
|
| 1209 |
+
"loss": 2.0928,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 2.635294117647059,
|
| 1214 |
+
"grad_norm": 0.8130326867103577,
|
| 1215 |
+
"learning_rate": 4.229599573731685e-06,
|
| 1216 |
+
"loss": 2.184,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.6509803921568627,
|
| 1221 |
+
"grad_norm": 0.8824214339256287,
|
| 1222 |
+
"learning_rate": 3.8890243569094874e-06,
|
| 1223 |
+
"loss": 2.28,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 2.6666666666666665,
|
| 1228 |
+
"grad_norm": 0.9210173487663269,
|
| 1229 |
+
"learning_rate": 3.5621879937348836e-06,
|
| 1230 |
+
"loss": 2.0636,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 2.682352941176471,
|
| 1235 |
+
"grad_norm": 0.9852874875068665,
|
| 1236 |
+
"learning_rate": 3.249187865729264e-06,
|
| 1237 |
+
"loss": 1.9793,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.6980392156862747,
|
| 1242 |
+
"grad_norm": 1.3538250923156738,
|
| 1243 |
+
"learning_rate": 2.950117231881183e-06,
|
| 1244 |
+
"loss": 2.0158,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 2.7137254901960786,
|
| 1249 |
+
"grad_norm": 1.0066053867340088,
|
| 1250 |
+
"learning_rate": 2.6650652008597068e-06,
|
| 1251 |
+
"loss": 2.3757,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 2.7294117647058824,
|
| 1256 |
+
"grad_norm": 0.45928213000297546,
|
| 1257 |
+
"learning_rate": 2.3941167044642944e-06,
|
| 1258 |
+
"loss": 1.8917,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.7450980392156863,
|
| 1263 |
+
"grad_norm": 0.5406450629234314,
|
| 1264 |
+
"learning_rate": 2.137352472319215e-06,
|
| 1265 |
+
"loss": 2.2532,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 2.76078431372549,
|
| 1270 |
+
"grad_norm": 0.5478349328041077,
|
| 1271 |
+
"learning_rate": 1.8948490078199764e-06,
|
| 1272 |
+
"loss": 2.2074,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 2.776470588235294,
|
| 1277 |
+
"grad_norm": 0.5485714077949524,
|
| 1278 |
+
"learning_rate": 1.6666785653390249e-06,
|
| 1279 |
+
"loss": 2.1245,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.792156862745098,
|
| 1284 |
+
"grad_norm": 0.6040245294570923,
|
| 1285 |
+
"learning_rate": 1.4529091286973995e-06,
|
| 1286 |
+
"loss": 2.2487,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 2.8078431372549018,
|
| 1291 |
+
"grad_norm": 0.6607996821403503,
|
| 1292 |
+
"learning_rate": 1.2536043909088191e-06,
|
| 1293 |
+
"loss": 2.2545,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 2.8235294117647056,
|
| 1298 |
+
"grad_norm": 0.6602091789245605,
|
| 1299 |
+
"learning_rate": 1.0688237352022345e-06,
|
| 1300 |
+
"loss": 2.234,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.83921568627451,
|
| 1305 |
+
"grad_norm": 0.687446653842926,
|
| 1306 |
+
"learning_rate": 8.986222173284875e-07,
|
| 1307 |
+
"loss": 2.1189,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 2.854901960784314,
|
| 1312 |
+
"grad_norm": 0.7298436760902405,
|
| 1313 |
+
"learning_rate": 7.4305054915631e-07,
|
| 1314 |
+
"loss": 2.1273,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 2.8705882352941177,
|
| 1319 |
+
"grad_norm": 0.7987807989120483,
|
| 1320 |
+
"learning_rate": 6.021550835626777e-07,
|
| 1321 |
+
"loss": 2.1664,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.8862745098039215,
|
| 1326 |
+
"grad_norm": 0.860362708568573,
|
| 1327 |
+
"learning_rate": 4.7597780062184073e-07,
|
| 1328 |
+
"loss": 2.1681,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 2.9019607843137254,
|
| 1333 |
+
"grad_norm": 0.9093377590179443,
|
| 1334 |
+
"learning_rate": 3.6455629509730136e-07,
|
| 1335 |
+
"loss": 1.9198,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.9176470588235293,
|
| 1340 |
+
"grad_norm": 0.9680060744285583,
|
| 1341 |
+
"learning_rate": 2.6792376524036877e-07,
|
| 1342 |
+
"loss": 1.827,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.9333333333333336,
|
| 1347 |
+
"grad_norm": 1.2015458345413208,
|
| 1348 |
+
"learning_rate": 1.8610900289867673e-07,
|
| 1349 |
+
"loss": 2.0031,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.9490196078431374,
|
| 1354 |
+
"grad_norm": 0.9453166127204895,
|
| 1355 |
+
"learning_rate": 1.191363849376237e-07,
|
| 1356 |
+
"loss": 2.2456,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.9647058823529413,
|
| 1361 |
+
"grad_norm": 0.5282714366912842,
|
| 1362 |
+
"learning_rate": 6.702586597719385e-08,
|
| 1363 |
+
"loss": 1.8368,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.980392156862745,
|
| 1368 |
+
"grad_norm": 0.7178200483322144,
|
| 1369 |
+
"learning_rate": 2.9792972446479605e-08,
|
| 1370 |
+
"loss": 2.2156,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 2.996078431372549,
|
| 1375 |
+
"grad_norm": 1.3029696941375732,
|
| 1376 |
+
"learning_rate": 7.448797957526621e-09,
|
| 1377 |
+
"loss": 2.4417,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 3.011764705882353,
|
| 1382 |
+
"grad_norm": 0.6315656304359436,
|
| 1383 |
+
"learning_rate": 0.0,
|
| 1384 |
+
"loss": 2.2333,
|
| 1385 |
+
"step": 192
|
| 1386 |
}
|
| 1387 |
],
|
| 1388 |
"logging_steps": 1,
|
|
|
|
| 1406 |
"should_evaluate": false,
|
| 1407 |
"should_log": false,
|
| 1408 |
"should_save": true,
|
| 1409 |
+
"should_training_stop": true
|
| 1410 |
},
|
| 1411 |
"attributes": {}
|
| 1412 |
}
|
| 1413 |
},
|
| 1414 |
+
"total_flos": 2.62298517110784e+17,
|
| 1415 |
"train_batch_size": 8,
|
| 1416 |
"trial_name": null,
|
| 1417 |
"trial_params": null
|