Training in progress, step 200, 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 42487072
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:09802372ae6528ae60ef4239d9054890cac2425f1ea418a31b103f78903f12df
|
| 3 |
size 42487072
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 21735354
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d57935fa15c96a1643ffadf083c34db8a9d30de17b6da53e15140bebbff72098
|
| 3 |
size 21735354
|
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:f071ddc338c365927a962d7cc27b4008499647fb4bc152f33fb387fabacd7e5e
|
| 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:b60d6f1383abda4776549360effee800fe6cfe2c0604503e9e3fbaa79347f790
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 1.
|
| 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 +1089,364 @@
|
|
| 1089 |
"eval_samples_per_second": 155.797,
|
| 1090 |
"eval_steps_per_second": 38.96,
|
| 1091 |
"step": 150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
@@ -1112,12 +1470,12 @@
|
|
| 1112 |
"should_evaluate": false,
|
| 1113 |
"should_log": false,
|
| 1114 |
"should_save": true,
|
| 1115 |
-
"should_training_stop":
|
| 1116 |
},
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
-
"total_flos":
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.6883196830749512,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.04497161167013323,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 200,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1089 |
"eval_samples_per_second": 155.797,
|
| 1090 |
"eval_steps_per_second": 38.96,
|
| 1091 |
"step": 150
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 0.03395356681095059,
|
| 1095 |
+
"grad_norm": 3.9572837352752686,
|
| 1096 |
+
"learning_rate": 1.0872630680850196e-05,
|
| 1097 |
+
"loss": 6.7227,
|
| 1098 |
+
"step": 151
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 0.034178424869301255,
|
| 1102 |
+
"grad_norm": 3.420825719833374,
|
| 1103 |
+
"learning_rate": 1.0456712550462898e-05,
|
| 1104 |
+
"loss": 6.6782,
|
| 1105 |
+
"step": 152
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.03440328292765192,
|
| 1109 |
+
"grad_norm": 3.3071367740631104,
|
| 1110 |
+
"learning_rate": 1.0047504309801104e-05,
|
| 1111 |
+
"loss": 6.2229,
|
| 1112 |
+
"step": 153
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.03462814098600259,
|
| 1116 |
+
"grad_norm": 3.183055877685547,
|
| 1117 |
+
"learning_rate": 9.645117832311886e-06,
|
| 1118 |
+
"loss": 6.7772,
|
| 1119 |
+
"step": 154
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.03485299904435325,
|
| 1123 |
+
"grad_norm": 3.4305922985076904,
|
| 1124 |
+
"learning_rate": 9.249663126440394e-06,
|
| 1125 |
+
"loss": 6.0615,
|
| 1126 |
+
"step": 155
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.03507785710270392,
|
| 1130 |
+
"grad_norm": 2.9730641841888428,
|
| 1131 |
+
"learning_rate": 8.861248305554624e-06,
|
| 1132 |
+
"loss": 6.6021,
|
| 1133 |
+
"step": 156
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 0.035302715161054585,
|
| 1137 |
+
"grad_norm": 2.939992904663086,
|
| 1138 |
+
"learning_rate": 8.47997955838829e-06,
|
| 1139 |
+
"loss": 7.2011,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.03552757321940525,
|
| 1144 |
+
"grad_norm": 2.6817591190338135,
|
| 1145 |
+
"learning_rate": 8.10596112000994e-06,
|
| 1146 |
+
"loss": 7.2791,
|
| 1147 |
+
"step": 158
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.03575243127775592,
|
| 1151 |
+
"grad_norm": 2.6212384700775146,
|
| 1152 |
+
"learning_rate": 7.739295243326067e-06,
|
| 1153 |
+
"loss": 6.9312,
|
| 1154 |
+
"step": 159
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.03597728933610658,
|
| 1158 |
+
"grad_norm": 2.7194411754608154,
|
| 1159 |
+
"learning_rate": 7.380082171126228e-06,
|
| 1160 |
+
"loss": 6.8364,
|
| 1161 |
+
"step": 160
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.03620214739445725,
|
| 1165 |
+
"grad_norm": 2.6084587574005127,
|
| 1166 |
+
"learning_rate": 7.028420108677635e-06,
|
| 1167 |
+
"loss": 7.3086,
|
| 1168 |
+
"step": 161
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.036427005452807915,
|
| 1172 |
+
"grad_norm": 2.780031204223633,
|
| 1173 |
+
"learning_rate": 6.684405196876842e-06,
|
| 1174 |
+
"loss": 7.5853,
|
| 1175 |
+
"step": 162
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.03665186351115858,
|
| 1179 |
+
"grad_norm": 2.847130060195923,
|
| 1180 |
+
"learning_rate": 6.3481314859657675e-06,
|
| 1181 |
+
"loss": 7.0845,
|
| 1182 |
+
"step": 163
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 0.036876721569509247,
|
| 1186 |
+
"grad_norm": 2.741257429122925,
|
| 1187 |
+
"learning_rate": 6.019690909819298e-06,
|
| 1188 |
+
"loss": 7.2365,
|
| 1189 |
+
"step": 164
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.03710157962785991,
|
| 1193 |
+
"grad_norm": 2.826801300048828,
|
| 1194 |
+
"learning_rate": 5.6991732608115e-06,
|
| 1195 |
+
"loss": 7.1966,
|
| 1196 |
+
"step": 165
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.03732643768621058,
|
| 1200 |
+
"grad_norm": 2.8895692825317383,
|
| 1201 |
+
"learning_rate": 5.386666165267256e-06,
|
| 1202 |
+
"loss": 7.2495,
|
| 1203 |
+
"step": 166
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.037551295744561244,
|
| 1207 |
+
"grad_norm": 2.9744369983673096,
|
| 1208 |
+
"learning_rate": 5.08225505950613e-06,
|
| 1209 |
+
"loss": 7.0297,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.03777615380291191,
|
| 1214 |
+
"grad_norm": 2.860124349594116,
|
| 1215 |
+
"learning_rate": 4.786023166484913e-06,
|
| 1216 |
+
"loss": 7.2646,
|
| 1217 |
+
"step": 168
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 0.038001011861262576,
|
| 1221 |
+
"grad_norm": 2.9127304553985596,
|
| 1222 |
+
"learning_rate": 4.498051473045291e-06,
|
| 1223 |
+
"loss": 6.8972,
|
| 1224 |
+
"step": 169
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.03822586991961324,
|
| 1228 |
+
"grad_norm": 3.142061471939087,
|
| 1229 |
+
"learning_rate": 4.218418707772886e-06,
|
| 1230 |
+
"loss": 7.2819,
|
| 1231 |
+
"step": 170
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.03845072797796391,
|
| 1235 |
+
"grad_norm": 3.084571361541748,
|
| 1236 |
+
"learning_rate": 3.947201319473587e-06,
|
| 1237 |
+
"loss": 6.8547,
|
| 1238 |
+
"step": 171
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.038675586036314574,
|
| 1242 |
+
"grad_norm": 3.1179471015930176,
|
| 1243 |
+
"learning_rate": 3.684473456273278e-06,
|
| 1244 |
+
"loss": 7.1903,
|
| 1245 |
+
"step": 172
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.03890044409466524,
|
| 1249 |
+
"grad_norm": 3.02191162109375,
|
| 1250 |
+
"learning_rate": 3.4303069453464383e-06,
|
| 1251 |
+
"loss": 6.9666,
|
| 1252 |
+
"step": 173
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.039125302153015906,
|
| 1256 |
+
"grad_norm": 3.0188581943511963,
|
| 1257 |
+
"learning_rate": 3.184771273279312e-06,
|
| 1258 |
+
"loss": 6.6101,
|
| 1259 |
+
"step": 174
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 0.03935016021136657,
|
| 1263 |
+
"grad_norm": 3.0796701908111572,
|
| 1264 |
+
"learning_rate": 2.947933567072987e-06,
|
| 1265 |
+
"loss": 7.0151,
|
| 1266 |
+
"step": 175
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.03957501826971724,
|
| 1270 |
+
"grad_norm": 3.1712894439697266,
|
| 1271 |
+
"learning_rate": 2.719858575791534e-06,
|
| 1272 |
+
"loss": 6.8814,
|
| 1273 |
+
"step": 176
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.039799876328067904,
|
| 1277 |
+
"grad_norm": 3.122129201889038,
|
| 1278 |
+
"learning_rate": 2.500608652860256e-06,
|
| 1279 |
+
"loss": 7.0466,
|
| 1280 |
+
"step": 177
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.04002473438641857,
|
| 1284 |
+
"grad_norm": 3.104573965072632,
|
| 1285 |
+
"learning_rate": 2.2902437390188737e-06,
|
| 1286 |
+
"loss": 6.9377,
|
| 1287 |
+
"step": 178
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.040249592444769236,
|
| 1291 |
+
"grad_norm": 3.207752227783203,
|
| 1292 |
+
"learning_rate": 2.0888213459343587e-06,
|
| 1293 |
+
"loss": 7.2979,
|
| 1294 |
+
"step": 179
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 0.04047445050311991,
|
| 1298 |
+
"grad_norm": 3.4283640384674072,
|
| 1299 |
+
"learning_rate": 1.8963965404777875e-06,
|
| 1300 |
+
"loss": 7.3337,
|
| 1301 |
+
"step": 180
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 0.040699308561470575,
|
| 1305 |
+
"grad_norm": 3.580007791519165,
|
| 1306 |
+
"learning_rate": 1.7130219296696263e-06,
|
| 1307 |
+
"loss": 7.1523,
|
| 1308 |
+
"step": 181
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.04092416661982124,
|
| 1312 |
+
"grad_norm": 3.8134102821350098,
|
| 1313 |
+
"learning_rate": 1.5387476462974824e-06,
|
| 1314 |
+
"loss": 6.9493,
|
| 1315 |
+
"step": 182
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.04114902467817191,
|
| 1319 |
+
"grad_norm": 3.512928009033203,
|
| 1320 |
+
"learning_rate": 1.3736213352103147e-06,
|
| 1321 |
+
"loss": 7.1148,
|
| 1322 |
+
"step": 183
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.04137388273652257,
|
| 1326 |
+
"grad_norm": 3.850146770477295,
|
| 1327 |
+
"learning_rate": 1.2176881402928002e-06,
|
| 1328 |
+
"loss": 7.2365,
|
| 1329 |
+
"step": 184
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.04159874079487324,
|
| 1333 |
+
"grad_norm": 3.849551200866699,
|
| 1334 |
+
"learning_rate": 1.0709906921234367e-06,
|
| 1335 |
+
"loss": 7.0453,
|
| 1336 |
+
"step": 185
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.041823598853223905,
|
| 1340 |
+
"grad_norm": 3.682288408279419,
|
| 1341 |
+
"learning_rate": 9.33569096319799e-07,
|
| 1342 |
+
"loss": 7.0223,
|
| 1343 |
+
"step": 186
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.04204845691157457,
|
| 1347 |
+
"grad_norm": 3.688978433609009,
|
| 1348 |
+
"learning_rate": 8.054609225740255e-07,
|
| 1349 |
+
"loss": 6.9328,
|
| 1350 |
+
"step": 187
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.04227331496992524,
|
| 1354 |
+
"grad_norm": 4.038215160369873,
|
| 1355 |
+
"learning_rate": 6.867011943816724e-07,
|
| 1356 |
+
"loss": 7.2454,
|
| 1357 |
+
"step": 188
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.0424981730282759,
|
| 1361 |
+
"grad_norm": 4.048027515411377,
|
| 1362 |
+
"learning_rate": 5.77322379466617e-07,
|
| 1363 |
+
"loss": 6.4791,
|
| 1364 |
+
"step": 189
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.04272303108662657,
|
| 1368 |
+
"grad_norm": 4.304159641265869,
|
| 1369 |
+
"learning_rate": 4.773543809047186e-07,
|
| 1370 |
+
"loss": 7.1091,
|
| 1371 |
+
"step": 190
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.042947889144977235,
|
| 1375 |
+
"grad_norm": 4.445484161376953,
|
| 1376 |
+
"learning_rate": 3.868245289486027e-07,
|
| 1377 |
+
"loss": 6.5871,
|
| 1378 |
+
"step": 191
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 0.0431727472033279,
|
| 1382 |
+
"grad_norm": 4.866727352142334,
|
| 1383 |
+
"learning_rate": 3.0575757355586817e-07,
|
| 1384 |
+
"loss": 7.2476,
|
| 1385 |
+
"step": 192
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.043397605261678567,
|
| 1389 |
+
"grad_norm": 5.467154502868652,
|
| 1390 |
+
"learning_rate": 2.3417567762266497e-07,
|
| 1391 |
+
"loss": 7.0307,
|
| 1392 |
+
"step": 193
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.04362246332002923,
|
| 1396 |
+
"grad_norm": 5.4806013107299805,
|
| 1397 |
+
"learning_rate": 1.7209841092460043e-07,
|
| 1398 |
+
"loss": 7.371,
|
| 1399 |
+
"step": 194
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.0438473213783799,
|
| 1403 |
+
"grad_norm": 6.067953109741211,
|
| 1404 |
+
"learning_rate": 1.1954274476655534e-07,
|
| 1405 |
+
"loss": 7.1333,
|
| 1406 |
+
"step": 195
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.044072179436730564,
|
| 1410 |
+
"grad_norm": 6.512303352355957,
|
| 1411 |
+
"learning_rate": 7.652304734289127e-08,
|
| 1412 |
+
"loss": 7.4444,
|
| 1413 |
+
"step": 196
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.04429703749508123,
|
| 1417 |
+
"grad_norm": 7.601706504821777,
|
| 1418 |
+
"learning_rate": 4.30510798093342e-08,
|
| 1419 |
+
"loss": 7.0462,
|
| 1420 |
+
"step": 197
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.044521895553431896,
|
| 1424 |
+
"grad_norm": 8.405488967895508,
|
| 1425 |
+
"learning_rate": 1.9135993067588284e-08,
|
| 1426 |
+
"loss": 7.2042,
|
| 1427 |
+
"step": 198
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.04474675361178256,
|
| 1431 |
+
"grad_norm": 8.733073234558105,
|
| 1432 |
+
"learning_rate": 4.784325263584854e-09,
|
| 1433 |
+
"loss": 7.4388,
|
| 1434 |
+
"step": 199
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.04497161167013323,
|
| 1438 |
+
"grad_norm": 14.473525047302246,
|
| 1439 |
+
"learning_rate": 0.0,
|
| 1440 |
+
"loss": 7.8313,
|
| 1441 |
+
"step": 200
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.04497161167013323,
|
| 1445 |
+
"eval_loss": 1.6883196830749512,
|
| 1446 |
+
"eval_runtime": 47.9579,
|
| 1447 |
+
"eval_samples_per_second": 156.179,
|
| 1448 |
+
"eval_steps_per_second": 39.055,
|
| 1449 |
+
"step": 200
|
| 1450 |
}
|
| 1451 |
],
|
| 1452 |
"logging_steps": 1,
|
|
|
|
| 1470 |
"should_evaluate": false,
|
| 1471 |
"should_log": false,
|
| 1472 |
"should_save": true,
|
| 1473 |
+
"should_training_stop": true
|
| 1474 |
},
|
| 1475 |
"attributes": {}
|
| 1476 |
}
|
| 1477 |
},
|
| 1478 |
+
"total_flos": 4227557423579136.0,
|
| 1479 |
"train_batch_size": 8,
|
| 1480 |
"trial_name": null,
|
| 1481 |
"trial_params": null
|