Training in progress, step 48000, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 140637408
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:55eab6c50cc6cba8adfeee5c05f44d22f3db36e57fea6c27663f868d2d205b27
|
| 3 |
size 140637408
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 281362507
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4614855000ef7710bdb6c8a716335ac27fd8b01da78602a185c5190e3fa081bb
|
| 3 |
size 281362507
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0591be5ab3c49004fb98678aba40d71597cb334b693aa1f9cf1270d955ae5462
|
| 3 |
size 14645
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c0a4f460e39e5e5c35812afe60a3df875ab624e766a314e245a1dd96c236d859
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -34113,6 +34113,365 @@
|
|
| 34113 |
"eval_samples_per_second": 33.469,
|
| 34114 |
"eval_steps_per_second": 0.536,
|
| 34115 |
"step": 47500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34116 |
}
|
| 34117 |
],
|
| 34118 |
"logging_steps": 10,
|
|
@@ -34132,7 +34491,7 @@
|
|
| 34132 |
"attributes": {}
|
| 34133 |
}
|
| 34134 |
},
|
| 34135 |
-
"total_flos": 8.
|
| 34136 |
"train_batch_size": 128,
|
| 34137 |
"trial_name": null,
|
| 34138 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.6291451490287572,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 48000,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 34113 |
"eval_samples_per_second": 33.469,
|
| 34114 |
"eval_steps_per_second": 0.536,
|
| 34115 |
"step": 47500
|
| 34116 |
+
},
|
| 34117 |
+
{
|
| 34118 |
+
"epoch": 0.622722625632422,
|
| 34119 |
+
"grad_norm": 0.6280195713043213,
|
| 34120 |
+
"learning_rate": 9.468467607524137e-05,
|
| 34121 |
+
"loss": 3.25,
|
| 34122 |
+
"step": 47510
|
| 34123 |
+
},
|
| 34124 |
+
{
|
| 34125 |
+
"epoch": 0.6228536975384696,
|
| 34126 |
+
"grad_norm": 0.6984219551086426,
|
| 34127 |
+
"learning_rate": 9.462688911437063e-05,
|
| 34128 |
+
"loss": 3.2415,
|
| 34129 |
+
"step": 47520
|
| 34130 |
+
},
|
| 34131 |
+
{
|
| 34132 |
+
"epoch": 0.6229847694445173,
|
| 34133 |
+
"grad_norm": 0.7601839303970337,
|
| 34134 |
+
"learning_rate": 9.456911166675204e-05,
|
| 34135 |
+
"loss": 3.3118,
|
| 34136 |
+
"step": 47530
|
| 34137 |
+
},
|
| 34138 |
+
{
|
| 34139 |
+
"epoch": 0.623115841350565,
|
| 34140 |
+
"grad_norm": 0.6929161548614502,
|
| 34141 |
+
"learning_rate": 9.451134374231191e-05,
|
| 34142 |
+
"loss": 3.2198,
|
| 34143 |
+
"step": 47540
|
| 34144 |
+
},
|
| 34145 |
+
{
|
| 34146 |
+
"epoch": 0.6232469132566125,
|
| 34147 |
+
"grad_norm": 0.6816897392272949,
|
| 34148 |
+
"learning_rate": 9.445358535097499e-05,
|
| 34149 |
+
"loss": 3.1733,
|
| 34150 |
+
"step": 47550
|
| 34151 |
+
},
|
| 34152 |
+
{
|
| 34153 |
+
"epoch": 0.6233779851626602,
|
| 34154 |
+
"grad_norm": 0.6594375371932983,
|
| 34155 |
+
"learning_rate": 9.439583650266424e-05,
|
| 34156 |
+
"loss": 3.2467,
|
| 34157 |
+
"step": 47560
|
| 34158 |
+
},
|
| 34159 |
+
{
|
| 34160 |
+
"epoch": 0.6235090570687079,
|
| 34161 |
+
"grad_norm": 0.6885610818862915,
|
| 34162 |
+
"learning_rate": 9.433809720730105e-05,
|
| 34163 |
+
"loss": 3.2176,
|
| 34164 |
+
"step": 47570
|
| 34165 |
+
},
|
| 34166 |
+
{
|
| 34167 |
+
"epoch": 0.6236401289747555,
|
| 34168 |
+
"grad_norm": 0.927130937576294,
|
| 34169 |
+
"learning_rate": 9.428036747480529e-05,
|
| 34170 |
+
"loss": 3.2081,
|
| 34171 |
+
"step": 47580
|
| 34172 |
+
},
|
| 34173 |
+
{
|
| 34174 |
+
"epoch": 0.6237712008808032,
|
| 34175 |
+
"grad_norm": 0.6482477188110352,
|
| 34176 |
+
"learning_rate": 9.422264731509505e-05,
|
| 34177 |
+
"loss": 3.2077,
|
| 34178 |
+
"step": 47590
|
| 34179 |
+
},
|
| 34180 |
+
{
|
| 34181 |
+
"epoch": 0.6239022727868508,
|
| 34182 |
+
"grad_norm": 0.6166397929191589,
|
| 34183 |
+
"learning_rate": 9.416493673808673e-05,
|
| 34184 |
+
"loss": 3.2161,
|
| 34185 |
+
"step": 47600
|
| 34186 |
+
},
|
| 34187 |
+
{
|
| 34188 |
+
"epoch": 0.6240333446928985,
|
| 34189 |
+
"grad_norm": 0.7229884266853333,
|
| 34190 |
+
"learning_rate": 9.410723575369526e-05,
|
| 34191 |
+
"loss": 3.1857,
|
| 34192 |
+
"step": 47610
|
| 34193 |
+
},
|
| 34194 |
+
{
|
| 34195 |
+
"epoch": 0.6241644165989462,
|
| 34196 |
+
"grad_norm": 0.6214064955711365,
|
| 34197 |
+
"learning_rate": 9.404954437183386e-05,
|
| 34198 |
+
"loss": 3.2443,
|
| 34199 |
+
"step": 47620
|
| 34200 |
+
},
|
| 34201 |
+
{
|
| 34202 |
+
"epoch": 0.6242954885049938,
|
| 34203 |
+
"grad_norm": 0.6719332933425903,
|
| 34204 |
+
"learning_rate": 9.399186260241394e-05,
|
| 34205 |
+
"loss": 3.4056,
|
| 34206 |
+
"step": 47630
|
| 34207 |
+
},
|
| 34208 |
+
{
|
| 34209 |
+
"epoch": 0.6244265604110415,
|
| 34210 |
+
"grad_norm": 0.6824233531951904,
|
| 34211 |
+
"learning_rate": 9.393419045534546e-05,
|
| 34212 |
+
"loss": 3.2983,
|
| 34213 |
+
"step": 47640
|
| 34214 |
+
},
|
| 34215 |
+
{
|
| 34216 |
+
"epoch": 0.6245576323170892,
|
| 34217 |
+
"grad_norm": 0.6507546901702881,
|
| 34218 |
+
"learning_rate": 9.387652794053667e-05,
|
| 34219 |
+
"loss": 3.2472,
|
| 34220 |
+
"step": 47650
|
| 34221 |
+
},
|
| 34222 |
+
{
|
| 34223 |
+
"epoch": 0.6246887042231368,
|
| 34224 |
+
"grad_norm": 0.6942393183708191,
|
| 34225 |
+
"learning_rate": 9.381887506789415e-05,
|
| 34226 |
+
"loss": 3.2285,
|
| 34227 |
+
"step": 47660
|
| 34228 |
+
},
|
| 34229 |
+
{
|
| 34230 |
+
"epoch": 0.6248197761291845,
|
| 34231 |
+
"grad_norm": 0.648705005645752,
|
| 34232 |
+
"learning_rate": 9.376123184732274e-05,
|
| 34233 |
+
"loss": 3.2121,
|
| 34234 |
+
"step": 47670
|
| 34235 |
+
},
|
| 34236 |
+
{
|
| 34237 |
+
"epoch": 0.6249508480352322,
|
| 34238 |
+
"grad_norm": 0.5787950754165649,
|
| 34239 |
+
"learning_rate": 9.370359828872578e-05,
|
| 34240 |
+
"loss": 3.2125,
|
| 34241 |
+
"step": 47680
|
| 34242 |
+
},
|
| 34243 |
+
{
|
| 34244 |
+
"epoch": 0.6250819199412798,
|
| 34245 |
+
"grad_norm": 0.753471314907074,
|
| 34246 |
+
"learning_rate": 9.364597440200482e-05,
|
| 34247 |
+
"loss": 3.2391,
|
| 34248 |
+
"step": 47690
|
| 34249 |
+
},
|
| 34250 |
+
{
|
| 34251 |
+
"epoch": 0.6252129918473275,
|
| 34252 |
+
"grad_norm": 0.6960231065750122,
|
| 34253 |
+
"learning_rate": 9.358836019705991e-05,
|
| 34254 |
+
"loss": 3.2272,
|
| 34255 |
+
"step": 47700
|
| 34256 |
+
},
|
| 34257 |
+
{
|
| 34258 |
+
"epoch": 0.6253440637533751,
|
| 34259 |
+
"grad_norm": 0.6731109023094177,
|
| 34260 |
+
"learning_rate": 9.353075568378918e-05,
|
| 34261 |
+
"loss": 3.2203,
|
| 34262 |
+
"step": 47710
|
| 34263 |
+
},
|
| 34264 |
+
{
|
| 34265 |
+
"epoch": 0.6254751356594228,
|
| 34266 |
+
"grad_norm": 0.7097241282463074,
|
| 34267 |
+
"learning_rate": 9.347316087208932e-05,
|
| 34268 |
+
"loss": 3.2546,
|
| 34269 |
+
"step": 47720
|
| 34270 |
+
},
|
| 34271 |
+
{
|
| 34272 |
+
"epoch": 0.6256062075654704,
|
| 34273 |
+
"grad_norm": 0.6814481019973755,
|
| 34274 |
+
"learning_rate": 9.341557577185528e-05,
|
| 34275 |
+
"loss": 3.1959,
|
| 34276 |
+
"step": 47730
|
| 34277 |
+
},
|
| 34278 |
+
{
|
| 34279 |
+
"epoch": 0.625737279471518,
|
| 34280 |
+
"grad_norm": 0.6176202893257141,
|
| 34281 |
+
"learning_rate": 9.335800039298038e-05,
|
| 34282 |
+
"loss": 3.2576,
|
| 34283 |
+
"step": 47740
|
| 34284 |
+
},
|
| 34285 |
+
{
|
| 34286 |
+
"epoch": 0.6258683513775657,
|
| 34287 |
+
"grad_norm": 0.6454567313194275,
|
| 34288 |
+
"learning_rate": 9.330043474535613e-05,
|
| 34289 |
+
"loss": 3.1762,
|
| 34290 |
+
"step": 47750
|
| 34291 |
+
},
|
| 34292 |
+
{
|
| 34293 |
+
"epoch": 0.6259994232836134,
|
| 34294 |
+
"grad_norm": 0.6443287134170532,
|
| 34295 |
+
"learning_rate": 9.32428788388725e-05,
|
| 34296 |
+
"loss": 3.2684,
|
| 34297 |
+
"step": 47760
|
| 34298 |
+
},
|
| 34299 |
+
{
|
| 34300 |
+
"epoch": 0.626130495189661,
|
| 34301 |
+
"grad_norm": 0.6694857478141785,
|
| 34302 |
+
"learning_rate": 9.318533268341784e-05,
|
| 34303 |
+
"loss": 3.2302,
|
| 34304 |
+
"step": 47770
|
| 34305 |
+
},
|
| 34306 |
+
{
|
| 34307 |
+
"epoch": 0.6262615670957087,
|
| 34308 |
+
"grad_norm": 0.6364569067955017,
|
| 34309 |
+
"learning_rate": 9.312779628887859e-05,
|
| 34310 |
+
"loss": 3.2618,
|
| 34311 |
+
"step": 47780
|
| 34312 |
+
},
|
| 34313 |
+
{
|
| 34314 |
+
"epoch": 0.6263926390017563,
|
| 34315 |
+
"grad_norm": 0.646737813949585,
|
| 34316 |
+
"learning_rate": 9.307026966513973e-05,
|
| 34317 |
+
"loss": 3.224,
|
| 34318 |
+
"step": 47790
|
| 34319 |
+
},
|
| 34320 |
+
{
|
| 34321 |
+
"epoch": 0.626523710907804,
|
| 34322 |
+
"grad_norm": 0.6261652112007141,
|
| 34323 |
+
"learning_rate": 9.30127528220845e-05,
|
| 34324 |
+
"loss": 3.2228,
|
| 34325 |
+
"step": 47800
|
| 34326 |
+
},
|
| 34327 |
+
{
|
| 34328 |
+
"epoch": 0.6266547828138517,
|
| 34329 |
+
"grad_norm": 0.7065747976303101,
|
| 34330 |
+
"learning_rate": 9.29552457695945e-05,
|
| 34331 |
+
"loss": 3.2767,
|
| 34332 |
+
"step": 47810
|
| 34333 |
+
},
|
| 34334 |
+
{
|
| 34335 |
+
"epoch": 0.6267858547198993,
|
| 34336 |
+
"grad_norm": 0.617780864238739,
|
| 34337 |
+
"learning_rate": 9.289774851754949e-05,
|
| 34338 |
+
"loss": 3.2327,
|
| 34339 |
+
"step": 47820
|
| 34340 |
+
},
|
| 34341 |
+
{
|
| 34342 |
+
"epoch": 0.626916926625947,
|
| 34343 |
+
"grad_norm": 0.615584671497345,
|
| 34344 |
+
"learning_rate": 9.284026107582774e-05,
|
| 34345 |
+
"loss": 3.3459,
|
| 34346 |
+
"step": 47830
|
| 34347 |
+
},
|
| 34348 |
+
{
|
| 34349 |
+
"epoch": 0.6270479985319947,
|
| 34350 |
+
"grad_norm": 0.6487901210784912,
|
| 34351 |
+
"learning_rate": 9.27827834543057e-05,
|
| 34352 |
+
"loss": 3.2493,
|
| 34353 |
+
"step": 47840
|
| 34354 |
+
},
|
| 34355 |
+
{
|
| 34356 |
+
"epoch": 0.6271790704380423,
|
| 34357 |
+
"grad_norm": 0.711365818977356,
|
| 34358 |
+
"learning_rate": 9.272531566285824e-05,
|
| 34359 |
+
"loss": 3.245,
|
| 34360 |
+
"step": 47850
|
| 34361 |
+
},
|
| 34362 |
+
{
|
| 34363 |
+
"epoch": 0.62731014234409,
|
| 34364 |
+
"grad_norm": 0.7267693877220154,
|
| 34365 |
+
"learning_rate": 9.26678577113584e-05,
|
| 34366 |
+
"loss": 3.1991,
|
| 34367 |
+
"step": 47860
|
| 34368 |
+
},
|
| 34369 |
+
{
|
| 34370 |
+
"epoch": 0.6274412142501377,
|
| 34371 |
+
"grad_norm": 0.610774040222168,
|
| 34372 |
+
"learning_rate": 9.261040960967766e-05,
|
| 34373 |
+
"loss": 3.2446,
|
| 34374 |
+
"step": 47870
|
| 34375 |
+
},
|
| 34376 |
+
{
|
| 34377 |
+
"epoch": 0.6275722861561853,
|
| 34378 |
+
"grad_norm": 0.600073516368866,
|
| 34379 |
+
"learning_rate": 9.255297136768582e-05,
|
| 34380 |
+
"loss": 3.2238,
|
| 34381 |
+
"step": 47880
|
| 34382 |
+
},
|
| 34383 |
+
{
|
| 34384 |
+
"epoch": 0.627703358062233,
|
| 34385 |
+
"grad_norm": 0.6462284326553345,
|
| 34386 |
+
"learning_rate": 9.249554299525081e-05,
|
| 34387 |
+
"loss": 3.2325,
|
| 34388 |
+
"step": 47890
|
| 34389 |
+
},
|
| 34390 |
+
{
|
| 34391 |
+
"epoch": 0.6278344299682806,
|
| 34392 |
+
"grad_norm": 0.6960959434509277,
|
| 34393 |
+
"learning_rate": 9.243812450223904e-05,
|
| 34394 |
+
"loss": 3.2073,
|
| 34395 |
+
"step": 47900
|
| 34396 |
+
},
|
| 34397 |
+
{
|
| 34398 |
+
"epoch": 0.6279655018743282,
|
| 34399 |
+
"grad_norm": 0.6119141578674316,
|
| 34400 |
+
"learning_rate": 9.23807158985152e-05,
|
| 34401 |
+
"loss": 3.2906,
|
| 34402 |
+
"step": 47910
|
| 34403 |
+
},
|
| 34404 |
+
{
|
| 34405 |
+
"epoch": 0.6280965737803759,
|
| 34406 |
+
"grad_norm": 0.6240313649177551,
|
| 34407 |
+
"learning_rate": 9.23233171939422e-05,
|
| 34408 |
+
"loss": 3.23,
|
| 34409 |
+
"step": 47920
|
| 34410 |
+
},
|
| 34411 |
+
{
|
| 34412 |
+
"epoch": 0.6282276456864235,
|
| 34413 |
+
"grad_norm": 0.6524478197097778,
|
| 34414 |
+
"learning_rate": 9.226592839838131e-05,
|
| 34415 |
+
"loss": 3.2276,
|
| 34416 |
+
"step": 47930
|
| 34417 |
+
},
|
| 34418 |
+
{
|
| 34419 |
+
"epoch": 0.6283587175924712,
|
| 34420 |
+
"grad_norm": 0.64365553855896,
|
| 34421 |
+
"learning_rate": 9.220854952169205e-05,
|
| 34422 |
+
"loss": 3.243,
|
| 34423 |
+
"step": 47940
|
| 34424 |
+
},
|
| 34425 |
+
{
|
| 34426 |
+
"epoch": 0.6284897894985189,
|
| 34427 |
+
"grad_norm": 0.6872357130050659,
|
| 34428 |
+
"learning_rate": 9.21511805737323e-05,
|
| 34429 |
+
"loss": 3.202,
|
| 34430 |
+
"step": 47950
|
| 34431 |
+
},
|
| 34432 |
+
{
|
| 34433 |
+
"epoch": 0.6286208614045665,
|
| 34434 |
+
"grad_norm": 0.6357784867286682,
|
| 34435 |
+
"learning_rate": 9.209382156435824e-05,
|
| 34436 |
+
"loss": 3.237,
|
| 34437 |
+
"step": 47960
|
| 34438 |
+
},
|
| 34439 |
+
{
|
| 34440 |
+
"epoch": 0.6287519333106142,
|
| 34441 |
+
"grad_norm": 0.7168740034103394,
|
| 34442 |
+
"learning_rate": 9.203647250342423e-05,
|
| 34443 |
+
"loss": 3.2618,
|
| 34444 |
+
"step": 47970
|
| 34445 |
+
},
|
| 34446 |
+
{
|
| 34447 |
+
"epoch": 0.6288830052166619,
|
| 34448 |
+
"grad_norm": 0.6022654175758362,
|
| 34449 |
+
"learning_rate": 9.1979133400783e-05,
|
| 34450 |
+
"loss": 3.2212,
|
| 34451 |
+
"step": 47980
|
| 34452 |
+
},
|
| 34453 |
+
{
|
| 34454 |
+
"epoch": 0.6290140771227095,
|
| 34455 |
+
"grad_norm": 0.6093465089797974,
|
| 34456 |
+
"learning_rate": 9.192180426628567e-05,
|
| 34457 |
+
"loss": 3.2531,
|
| 34458 |
+
"step": 47990
|
| 34459 |
+
},
|
| 34460 |
+
{
|
| 34461 |
+
"epoch": 0.6291451490287572,
|
| 34462 |
+
"grad_norm": 0.6737045645713806,
|
| 34463 |
+
"learning_rate": 9.18644851097814e-05,
|
| 34464 |
+
"loss": 3.2229,
|
| 34465 |
+
"step": 48000
|
| 34466 |
+
},
|
| 34467 |
+
{
|
| 34468 |
+
"epoch": 0.6291451490287572,
|
| 34469 |
+
"eval_accuracy": 0.39996041055718473,
|
| 34470 |
+
"eval_loss": 3.2245070934295654,
|
| 34471 |
+
"eval_runtime": 62.1503,
|
| 34472 |
+
"eval_samples_per_second": 32.18,
|
| 34473 |
+
"eval_steps_per_second": 0.515,
|
| 34474 |
+
"step": 48000
|
| 34475 |
}
|
| 34476 |
],
|
| 34477 |
"logging_steps": 10,
|
|
|
|
| 34491 |
"attributes": {}
|
| 34492 |
}
|
| 34493 |
},
|
| 34494 |
+
"total_flos": 8.63222391373824e+17,
|
| 34495 |
"train_batch_size": 128,
|
| 34496 |
"trial_name": null,
|
| 34497 |
"trial_params": null
|