Training in progress, step 1650, 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 131146352
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6029a505f307e3098b30acc19cdd7ba452e55709d8c353bc4a3f4f8ba146e277
|
| 3 |
size 131146352
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 67210516
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e748695332f398e0372a0342f533eda6dda257cbd0c6ff0c31662fd1d9df830
|
| 3 |
size 67210516
|
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:c2eb5cad9f3cbd36a844058d2bce505f26319b38c69d84d8607ffff4425c91e1
|
| 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 sha256:03bf9dc9befb01615f74b39d2b43ebf93f55dc1a1259dddadf80e9de69443c5a
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -11207,6 +11207,356 @@
|
|
| 11207 |
"learning_rate": 9.99999984307167e-05,
|
| 11208 |
"loss": 3.5878,
|
| 11209 |
"step": 1600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11210 |
}
|
| 11211 |
],
|
| 11212 |
"logging_steps": 1,
|
|
@@ -11226,7 +11576,7 @@
|
|
| 11226 |
"attributes": {}
|
| 11227 |
}
|
| 11228 |
},
|
| 11229 |
-
"total_flos": 2.
|
| 11230 |
"train_batch_size": 4,
|
| 11231 |
"trial_name": null,
|
| 11232 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.04432147845707532,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1650,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11207 |
"learning_rate": 9.99999984307167e-05,
|
| 11208 |
"loss": 3.5878,
|
| 11209 |
"step": 1600
|
| 11210 |
+
},
|
| 11211 |
+
{
|
| 11212 |
+
"epoch": 0.04300526485441066,
|
| 11213 |
+
"grad_norm": 1.3168871402740479,
|
| 11214 |
+
"learning_rate": 9.999999842874834e-05,
|
| 11215 |
+
"loss": 2.8627,
|
| 11216 |
+
"step": 1601
|
| 11217 |
+
},
|
| 11218 |
+
{
|
| 11219 |
+
"epoch": 0.043032126356505856,
|
| 11220 |
+
"grad_norm": 1.3844327926635742,
|
| 11221 |
+
"learning_rate": 9.999999842677874e-05,
|
| 11222 |
+
"loss": 2.8701,
|
| 11223 |
+
"step": 1602
|
| 11224 |
+
},
|
| 11225 |
+
{
|
| 11226 |
+
"epoch": 0.04305898785860105,
|
| 11227 |
+
"grad_norm": 1.3635506629943848,
|
| 11228 |
+
"learning_rate": 9.999999842480789e-05,
|
| 11229 |
+
"loss": 2.6733,
|
| 11230 |
+
"step": 1603
|
| 11231 |
+
},
|
| 11232 |
+
{
|
| 11233 |
+
"epoch": 0.04308584936069625,
|
| 11234 |
+
"grad_norm": 1.3588570356369019,
|
| 11235 |
+
"learning_rate": 9.999999842283582e-05,
|
| 11236 |
+
"loss": 2.8806,
|
| 11237 |
+
"step": 1604
|
| 11238 |
+
},
|
| 11239 |
+
{
|
| 11240 |
+
"epoch": 0.04311271086279145,
|
| 11241 |
+
"grad_norm": 1.3737713098526,
|
| 11242 |
+
"learning_rate": 9.999999842086252e-05,
|
| 11243 |
+
"loss": 2.9025,
|
| 11244 |
+
"step": 1605
|
| 11245 |
+
},
|
| 11246 |
+
{
|
| 11247 |
+
"epoch": 0.04313957236488664,
|
| 11248 |
+
"grad_norm": 1.40369713306427,
|
| 11249 |
+
"learning_rate": 9.999999841888799e-05,
|
| 11250 |
+
"loss": 2.8484,
|
| 11251 |
+
"step": 1606
|
| 11252 |
+
},
|
| 11253 |
+
{
|
| 11254 |
+
"epoch": 0.043166433866981845,
|
| 11255 |
+
"grad_norm": 1.2884548902511597,
|
| 11256 |
+
"learning_rate": 9.99999984169122e-05,
|
| 11257 |
+
"loss": 2.7603,
|
| 11258 |
+
"step": 1607
|
| 11259 |
+
},
|
| 11260 |
+
{
|
| 11261 |
+
"epoch": 0.04319329536907704,
|
| 11262 |
+
"grad_norm": 1.351904273033142,
|
| 11263 |
+
"learning_rate": 9.99999984149352e-05,
|
| 11264 |
+
"loss": 3.0526,
|
| 11265 |
+
"step": 1608
|
| 11266 |
+
},
|
| 11267 |
+
{
|
| 11268 |
+
"epoch": 0.043220156871172234,
|
| 11269 |
+
"grad_norm": 1.4556187391281128,
|
| 11270 |
+
"learning_rate": 9.999999841295697e-05,
|
| 11271 |
+
"loss": 3.0934,
|
| 11272 |
+
"step": 1609
|
| 11273 |
+
},
|
| 11274 |
+
{
|
| 11275 |
+
"epoch": 0.043247018373267436,
|
| 11276 |
+
"grad_norm": 1.3308874368667603,
|
| 11277 |
+
"learning_rate": 9.999999841097748e-05,
|
| 11278 |
+
"loss": 2.7632,
|
| 11279 |
+
"step": 1610
|
| 11280 |
+
},
|
| 11281 |
+
{
|
| 11282 |
+
"epoch": 0.04327387987536263,
|
| 11283 |
+
"grad_norm": 1.3838894367218018,
|
| 11284 |
+
"learning_rate": 9.999999840899678e-05,
|
| 11285 |
+
"loss": 3.0159,
|
| 11286 |
+
"step": 1611
|
| 11287 |
+
},
|
| 11288 |
+
{
|
| 11289 |
+
"epoch": 0.043300741377457826,
|
| 11290 |
+
"grad_norm": 1.504431962966919,
|
| 11291 |
+
"learning_rate": 9.999999840701483e-05,
|
| 11292 |
+
"loss": 3.0511,
|
| 11293 |
+
"step": 1612
|
| 11294 |
+
},
|
| 11295 |
+
{
|
| 11296 |
+
"epoch": 0.04332760287955303,
|
| 11297 |
+
"grad_norm": 1.2878671884536743,
|
| 11298 |
+
"learning_rate": 9.999999840503167e-05,
|
| 11299 |
+
"loss": 2.8596,
|
| 11300 |
+
"step": 1613
|
| 11301 |
+
},
|
| 11302 |
+
{
|
| 11303 |
+
"epoch": 0.04335446438164822,
|
| 11304 |
+
"grad_norm": 1.3833988904953003,
|
| 11305 |
+
"learning_rate": 9.999999840304725e-05,
|
| 11306 |
+
"loss": 2.7807,
|
| 11307 |
+
"step": 1614
|
| 11308 |
+
},
|
| 11309 |
+
{
|
| 11310 |
+
"epoch": 0.04338132588374342,
|
| 11311 |
+
"grad_norm": 1.3519421815872192,
|
| 11312 |
+
"learning_rate": 9.999999840106163e-05,
|
| 11313 |
+
"loss": 2.7777,
|
| 11314 |
+
"step": 1615
|
| 11315 |
+
},
|
| 11316 |
+
{
|
| 11317 |
+
"epoch": 0.04340818738583862,
|
| 11318 |
+
"grad_norm": 1.4081404209136963,
|
| 11319 |
+
"learning_rate": 9.999999839907475e-05,
|
| 11320 |
+
"loss": 2.7198,
|
| 11321 |
+
"step": 1616
|
| 11322 |
+
},
|
| 11323 |
+
{
|
| 11324 |
+
"epoch": 0.043435048887933814,
|
| 11325 |
+
"grad_norm": 1.3711954355239868,
|
| 11326 |
+
"learning_rate": 9.999999839708662e-05,
|
| 11327 |
+
"loss": 2.717,
|
| 11328 |
+
"step": 1617
|
| 11329 |
+
},
|
| 11330 |
+
{
|
| 11331 |
+
"epoch": 0.04346191039002901,
|
| 11332 |
+
"grad_norm": 1.3935779333114624,
|
| 11333 |
+
"learning_rate": 9.999999839509728e-05,
|
| 11334 |
+
"loss": 3.0255,
|
| 11335 |
+
"step": 1618
|
| 11336 |
+
},
|
| 11337 |
+
{
|
| 11338 |
+
"epoch": 0.04348877189212421,
|
| 11339 |
+
"grad_norm": 1.4566105604171753,
|
| 11340 |
+
"learning_rate": 9.99999983931067e-05,
|
| 11341 |
+
"loss": 3.1218,
|
| 11342 |
+
"step": 1619
|
| 11343 |
+
},
|
| 11344 |
+
{
|
| 11345 |
+
"epoch": 0.043515633394219405,
|
| 11346 |
+
"grad_norm": 1.4367201328277588,
|
| 11347 |
+
"learning_rate": 9.99999983911149e-05,
|
| 11348 |
+
"loss": 3.0931,
|
| 11349 |
+
"step": 1620
|
| 11350 |
+
},
|
| 11351 |
+
{
|
| 11352 |
+
"epoch": 0.0435424948963146,
|
| 11353 |
+
"grad_norm": 1.390453815460205,
|
| 11354 |
+
"learning_rate": 9.999999838912187e-05,
|
| 11355 |
+
"loss": 2.903,
|
| 11356 |
+
"step": 1621
|
| 11357 |
+
},
|
| 11358 |
+
{
|
| 11359 |
+
"epoch": 0.0435693563984098,
|
| 11360 |
+
"grad_norm": 1.2955039739608765,
|
| 11361 |
+
"learning_rate": 9.999999838712759e-05,
|
| 11362 |
+
"loss": 2.9835,
|
| 11363 |
+
"step": 1622
|
| 11364 |
+
},
|
| 11365 |
+
{
|
| 11366 |
+
"epoch": 0.043596217900505,
|
| 11367 |
+
"grad_norm": 1.5156861543655396,
|
| 11368 |
+
"learning_rate": 9.999999838513207e-05,
|
| 11369 |
+
"loss": 3.1388,
|
| 11370 |
+
"step": 1623
|
| 11371 |
+
},
|
| 11372 |
+
{
|
| 11373 |
+
"epoch": 0.04362307940260019,
|
| 11374 |
+
"grad_norm": 1.4825620651245117,
|
| 11375 |
+
"learning_rate": 9.999999838313532e-05,
|
| 11376 |
+
"loss": 3.1583,
|
| 11377 |
+
"step": 1624
|
| 11378 |
+
},
|
| 11379 |
+
{
|
| 11380 |
+
"epoch": 0.04364994090469539,
|
| 11381 |
+
"grad_norm": 1.369698166847229,
|
| 11382 |
+
"learning_rate": 9.999999838113734e-05,
|
| 11383 |
+
"loss": 2.9362,
|
| 11384 |
+
"step": 1625
|
| 11385 |
+
},
|
| 11386 |
+
{
|
| 11387 |
+
"epoch": 0.04367680240679059,
|
| 11388 |
+
"grad_norm": 1.3391295671463013,
|
| 11389 |
+
"learning_rate": 9.999999837913813e-05,
|
| 11390 |
+
"loss": 2.8266,
|
| 11391 |
+
"step": 1626
|
| 11392 |
+
},
|
| 11393 |
+
{
|
| 11394 |
+
"epoch": 0.04370366390888578,
|
| 11395 |
+
"grad_norm": 1.4127203226089478,
|
| 11396 |
+
"learning_rate": 9.999999837713768e-05,
|
| 11397 |
+
"loss": 2.827,
|
| 11398 |
+
"step": 1627
|
| 11399 |
+
},
|
| 11400 |
+
{
|
| 11401 |
+
"epoch": 0.043730525410980985,
|
| 11402 |
+
"grad_norm": 1.542043685913086,
|
| 11403 |
+
"learning_rate": 9.999999837513601e-05,
|
| 11404 |
+
"loss": 3.0887,
|
| 11405 |
+
"step": 1628
|
| 11406 |
+
},
|
| 11407 |
+
{
|
| 11408 |
+
"epoch": 0.04375738691307618,
|
| 11409 |
+
"grad_norm": 1.513839602470398,
|
| 11410 |
+
"learning_rate": 9.99999983731331e-05,
|
| 11411 |
+
"loss": 3.0853,
|
| 11412 |
+
"step": 1629
|
| 11413 |
+
},
|
| 11414 |
+
{
|
| 11415 |
+
"epoch": 0.043784248415171374,
|
| 11416 |
+
"grad_norm": 1.4729801416397095,
|
| 11417 |
+
"learning_rate": 9.999999837112895e-05,
|
| 11418 |
+
"loss": 2.98,
|
| 11419 |
+
"step": 1630
|
| 11420 |
+
},
|
| 11421 |
+
{
|
| 11422 |
+
"epoch": 0.043811109917266576,
|
| 11423 |
+
"grad_norm": 1.509283185005188,
|
| 11424 |
+
"learning_rate": 9.999999836912355e-05,
|
| 11425 |
+
"loss": 3.2404,
|
| 11426 |
+
"step": 1631
|
| 11427 |
+
},
|
| 11428 |
+
{
|
| 11429 |
+
"epoch": 0.04383797141936177,
|
| 11430 |
+
"grad_norm": 1.543927550315857,
|
| 11431 |
+
"learning_rate": 9.999999836711694e-05,
|
| 11432 |
+
"loss": 3.119,
|
| 11433 |
+
"step": 1632
|
| 11434 |
+
},
|
| 11435 |
+
{
|
| 11436 |
+
"epoch": 0.043864832921456966,
|
| 11437 |
+
"grad_norm": 1.5025025606155396,
|
| 11438 |
+
"learning_rate": 9.999999836510909e-05,
|
| 11439 |
+
"loss": 3.1039,
|
| 11440 |
+
"step": 1633
|
| 11441 |
+
},
|
| 11442 |
+
{
|
| 11443 |
+
"epoch": 0.04389169442355217,
|
| 11444 |
+
"grad_norm": 1.3757987022399902,
|
| 11445 |
+
"learning_rate": 9.999999836310001e-05,
|
| 11446 |
+
"loss": 2.8637,
|
| 11447 |
+
"step": 1634
|
| 11448 |
+
},
|
| 11449 |
+
{
|
| 11450 |
+
"epoch": 0.04391855592564736,
|
| 11451 |
+
"grad_norm": 1.5544795989990234,
|
| 11452 |
+
"learning_rate": 9.99999983610897e-05,
|
| 11453 |
+
"loss": 3.0113,
|
| 11454 |
+
"step": 1635
|
| 11455 |
+
},
|
| 11456 |
+
{
|
| 11457 |
+
"epoch": 0.04394541742774256,
|
| 11458 |
+
"grad_norm": 1.4375749826431274,
|
| 11459 |
+
"learning_rate": 9.999999835907815e-05,
|
| 11460 |
+
"loss": 2.9539,
|
| 11461 |
+
"step": 1636
|
| 11462 |
+
},
|
| 11463 |
+
{
|
| 11464 |
+
"epoch": 0.04397227892983776,
|
| 11465 |
+
"grad_norm": 1.557188868522644,
|
| 11466 |
+
"learning_rate": 9.999999835706537e-05,
|
| 11467 |
+
"loss": 3.1977,
|
| 11468 |
+
"step": 1637
|
| 11469 |
+
},
|
| 11470 |
+
{
|
| 11471 |
+
"epoch": 0.043999140431932954,
|
| 11472 |
+
"grad_norm": 1.6400185823440552,
|
| 11473 |
+
"learning_rate": 9.999999835505136e-05,
|
| 11474 |
+
"loss": 3.1995,
|
| 11475 |
+
"step": 1638
|
| 11476 |
+
},
|
| 11477 |
+
{
|
| 11478 |
+
"epoch": 0.04402600193402815,
|
| 11479 |
+
"grad_norm": 1.4655009508132935,
|
| 11480 |
+
"learning_rate": 9.99999983530361e-05,
|
| 11481 |
+
"loss": 2.9985,
|
| 11482 |
+
"step": 1639
|
| 11483 |
+
},
|
| 11484 |
+
{
|
| 11485 |
+
"epoch": 0.04405286343612335,
|
| 11486 |
+
"grad_norm": 1.4965097904205322,
|
| 11487 |
+
"learning_rate": 9.999999835101961e-05,
|
| 11488 |
+
"loss": 3.1393,
|
| 11489 |
+
"step": 1640
|
| 11490 |
+
},
|
| 11491 |
+
{
|
| 11492 |
+
"epoch": 0.044079724938218545,
|
| 11493 |
+
"grad_norm": 1.5797890424728394,
|
| 11494 |
+
"learning_rate": 9.999999834900189e-05,
|
| 11495 |
+
"loss": 3.0014,
|
| 11496 |
+
"step": 1641
|
| 11497 |
+
},
|
| 11498 |
+
{
|
| 11499 |
+
"epoch": 0.04410658644031374,
|
| 11500 |
+
"grad_norm": 1.567730188369751,
|
| 11501 |
+
"learning_rate": 9.999999834698295e-05,
|
| 11502 |
+
"loss": 2.9215,
|
| 11503 |
+
"step": 1642
|
| 11504 |
+
},
|
| 11505 |
+
{
|
| 11506 |
+
"epoch": 0.04413344794240894,
|
| 11507 |
+
"grad_norm": 1.6687631607055664,
|
| 11508 |
+
"learning_rate": 9.999999834496276e-05,
|
| 11509 |
+
"loss": 3.2531,
|
| 11510 |
+
"step": 1643
|
| 11511 |
+
},
|
| 11512 |
+
{
|
| 11513 |
+
"epoch": 0.04416030944450414,
|
| 11514 |
+
"grad_norm": 1.6272687911987305,
|
| 11515 |
+
"learning_rate": 9.999999834294133e-05,
|
| 11516 |
+
"loss": 3.3399,
|
| 11517 |
+
"step": 1644
|
| 11518 |
+
},
|
| 11519 |
+
{
|
| 11520 |
+
"epoch": 0.04418717094659933,
|
| 11521 |
+
"grad_norm": 1.6043704748153687,
|
| 11522 |
+
"learning_rate": 9.999999834091869e-05,
|
| 11523 |
+
"loss": 3.0772,
|
| 11524 |
+
"step": 1645
|
| 11525 |
+
},
|
| 11526 |
+
{
|
| 11527 |
+
"epoch": 0.04421403244869453,
|
| 11528 |
+
"grad_norm": 1.6957578659057617,
|
| 11529 |
+
"learning_rate": 9.999999833889479e-05,
|
| 11530 |
+
"loss": 3.3066,
|
| 11531 |
+
"step": 1646
|
| 11532 |
+
},
|
| 11533 |
+
{
|
| 11534 |
+
"epoch": 0.04424089395078973,
|
| 11535 |
+
"grad_norm": 1.7268688678741455,
|
| 11536 |
+
"learning_rate": 9.999999833686968e-05,
|
| 11537 |
+
"loss": 3.3125,
|
| 11538 |
+
"step": 1647
|
| 11539 |
+
},
|
| 11540 |
+
{
|
| 11541 |
+
"epoch": 0.04426775545288492,
|
| 11542 |
+
"grad_norm": 1.7406071424484253,
|
| 11543 |
+
"learning_rate": 9.999999833484333e-05,
|
| 11544 |
+
"loss": 3.2264,
|
| 11545 |
+
"step": 1648
|
| 11546 |
+
},
|
| 11547 |
+
{
|
| 11548 |
+
"epoch": 0.044294616954980125,
|
| 11549 |
+
"grad_norm": 1.8623542785644531,
|
| 11550 |
+
"learning_rate": 9.999999833281574e-05,
|
| 11551 |
+
"loss": 3.3522,
|
| 11552 |
+
"step": 1649
|
| 11553 |
+
},
|
| 11554 |
+
{
|
| 11555 |
+
"epoch": 0.04432147845707532,
|
| 11556 |
+
"grad_norm": 1.8407955169677734,
|
| 11557 |
+
"learning_rate": 9.999999833078691e-05,
|
| 11558 |
+
"loss": 3.3308,
|
| 11559 |
+
"step": 1650
|
| 11560 |
}
|
| 11561 |
],
|
| 11562 |
"logging_steps": 1,
|
|
|
|
| 11576 |
"attributes": {}
|
| 11577 |
}
|
| 11578 |
},
|
| 11579 |
+
"total_flos": 2.2518960104669184e+18,
|
| 11580 |
"train_batch_size": 4,
|
| 11581 |
"trial_name": null,
|
| 11582 |
"trial_params": null
|