Training in progress, step 1700, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +703 -3
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 159967880
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0930d6e64875820c22c7cffca09ca1acaf6cdcc1e7968d0c5a856968a87824e8
|
| 3 |
size 159967880
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 81735892
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b9d593632762fa95cd51142f1adb4c560e59f1f9d92ecb7de05d6485c887205
|
| 3 |
size 81735892
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4b283d316b0c499174401fc8457651f1fb183c6003c46a4d25e29dfecd151147
|
| 3 |
size 15024
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a6797f0f81e1d80bc4d2d6295ad3c421b4b433370ca9e0c209b11267f3ef64f
|
| 3 |
size 15024
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2bd17fa23f67ef7fbf6e377f7e0c23474bf385755bb96f63949a2752039f1de4
|
| 3 |
size 15024
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:212929e3bfad92319ef54b8b509922f96991c6c7d7791e9983b6f58b96c35aff
|
| 3 |
size 15024
|
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 sha256:671918de7ffc87d6187292033f79bb1cacaa6a7d5996a986d5989df4cdad43d1
|
| 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 +11207,706 @@
|
|
| 11207 |
"learning_rate": 6.759906759906761e-05,
|
| 11208 |
"loss": 0.0034,
|
| 11209 |
"step": 1600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11210 |
}
|
| 11211 |
],
|
| 11212 |
"logging_steps": 1,
|
|
@@ -11226,7 +11926,7 @@
|
|
| 11226 |
"attributes": {}
|
| 11227 |
}
|
| 11228 |
},
|
| 11229 |
-
"total_flos":
|
| 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": 0.9906759906759907,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1700,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11207 |
"learning_rate": 6.759906759906761e-05,
|
| 11208 |
"loss": 0.0034,
|
| 11209 |
"step": 1600
|
| 11210 |
+
},
|
| 11211 |
+
{
|
| 11212 |
+
"epoch": 0.932983682983683,
|
| 11213 |
+
"grad_norm": 0.0013091769069433212,
|
| 11214 |
+
"learning_rate": 6.701631701631703e-05,
|
| 11215 |
+
"loss": 0.0038,
|
| 11216 |
+
"step": 1601
|
| 11217 |
+
},
|
| 11218 |
+
{
|
| 11219 |
+
"epoch": 0.9335664335664335,
|
| 11220 |
+
"grad_norm": 0.0009162210044451058,
|
| 11221 |
+
"learning_rate": 6.643356643356644e-05,
|
| 11222 |
+
"loss": 0.0028,
|
| 11223 |
+
"step": 1602
|
| 11224 |
+
},
|
| 11225 |
+
{
|
| 11226 |
+
"epoch": 0.9341491841491841,
|
| 11227 |
+
"grad_norm": 0.0008717044838704169,
|
| 11228 |
+
"learning_rate": 6.585081585081586e-05,
|
| 11229 |
+
"loss": 0.003,
|
| 11230 |
+
"step": 1603
|
| 11231 |
+
},
|
| 11232 |
+
{
|
| 11233 |
+
"epoch": 0.9347319347319347,
|
| 11234 |
+
"grad_norm": 0.0010865787044167519,
|
| 11235 |
+
"learning_rate": 6.526806526806527e-05,
|
| 11236 |
+
"loss": 0.0035,
|
| 11237 |
+
"step": 1604
|
| 11238 |
+
},
|
| 11239 |
+
{
|
| 11240 |
+
"epoch": 0.9353146853146853,
|
| 11241 |
+
"grad_norm": 0.0007628489984199405,
|
| 11242 |
+
"learning_rate": 6.46853146853147e-05,
|
| 11243 |
+
"loss": 0.0025,
|
| 11244 |
+
"step": 1605
|
| 11245 |
+
},
|
| 11246 |
+
{
|
| 11247 |
+
"epoch": 0.9358974358974359,
|
| 11248 |
+
"grad_norm": 0.0009772854391485453,
|
| 11249 |
+
"learning_rate": 6.41025641025641e-05,
|
| 11250 |
+
"loss": 0.0041,
|
| 11251 |
+
"step": 1606
|
| 11252 |
+
},
|
| 11253 |
+
{
|
| 11254 |
+
"epoch": 0.9364801864801865,
|
| 11255 |
+
"grad_norm": 0.0008730532717891037,
|
| 11256 |
+
"learning_rate": 6.351981351981353e-05,
|
| 11257 |
+
"loss": 0.0028,
|
| 11258 |
+
"step": 1607
|
| 11259 |
+
},
|
| 11260 |
+
{
|
| 11261 |
+
"epoch": 0.9370629370629371,
|
| 11262 |
+
"grad_norm": 0.0007411614060401917,
|
| 11263 |
+
"learning_rate": 6.293706293706295e-05,
|
| 11264 |
+
"loss": 0.0032,
|
| 11265 |
+
"step": 1608
|
| 11266 |
+
},
|
| 11267 |
+
{
|
| 11268 |
+
"epoch": 0.9376456876456877,
|
| 11269 |
+
"grad_norm": 0.0009279727819375694,
|
| 11270 |
+
"learning_rate": 6.235431235431236e-05,
|
| 11271 |
+
"loss": 0.0022,
|
| 11272 |
+
"step": 1609
|
| 11273 |
+
},
|
| 11274 |
+
{
|
| 11275 |
+
"epoch": 0.9382284382284383,
|
| 11276 |
+
"grad_norm": 0.0010616903891786933,
|
| 11277 |
+
"learning_rate": 6.177156177156177e-05,
|
| 11278 |
+
"loss": 0.0044,
|
| 11279 |
+
"step": 1610
|
| 11280 |
+
},
|
| 11281 |
+
{
|
| 11282 |
+
"epoch": 0.9388111888111889,
|
| 11283 |
+
"grad_norm": 0.0009532080148346722,
|
| 11284 |
+
"learning_rate": 6.118881118881119e-05,
|
| 11285 |
+
"loss": 0.0026,
|
| 11286 |
+
"step": 1611
|
| 11287 |
+
},
|
| 11288 |
+
{
|
| 11289 |
+
"epoch": 0.9393939393939394,
|
| 11290 |
+
"grad_norm": 0.0007426452939398587,
|
| 11291 |
+
"learning_rate": 6.060606060606061e-05,
|
| 11292 |
+
"loss": 0.0027,
|
| 11293 |
+
"step": 1612
|
| 11294 |
+
},
|
| 11295 |
+
{
|
| 11296 |
+
"epoch": 0.9399766899766899,
|
| 11297 |
+
"grad_norm": 0.0010714689269661903,
|
| 11298 |
+
"learning_rate": 6.002331002331003e-05,
|
| 11299 |
+
"loss": 0.0041,
|
| 11300 |
+
"step": 1613
|
| 11301 |
+
},
|
| 11302 |
+
{
|
| 11303 |
+
"epoch": 0.9405594405594405,
|
| 11304 |
+
"grad_norm": 0.0008739576442167163,
|
| 11305 |
+
"learning_rate": 5.944055944055944e-05,
|
| 11306 |
+
"loss": 0.0035,
|
| 11307 |
+
"step": 1614
|
| 11308 |
+
},
|
| 11309 |
+
{
|
| 11310 |
+
"epoch": 0.9411421911421911,
|
| 11311 |
+
"grad_norm": 0.0018457169644534588,
|
| 11312 |
+
"learning_rate": 5.885780885780886e-05,
|
| 11313 |
+
"loss": 0.0048,
|
| 11314 |
+
"step": 1615
|
| 11315 |
+
},
|
| 11316 |
+
{
|
| 11317 |
+
"epoch": 0.9417249417249417,
|
| 11318 |
+
"grad_norm": 0.0008747098036110401,
|
| 11319 |
+
"learning_rate": 5.8275058275058275e-05,
|
| 11320 |
+
"loss": 0.003,
|
| 11321 |
+
"step": 1616
|
| 11322 |
+
},
|
| 11323 |
+
{
|
| 11324 |
+
"epoch": 0.9423076923076923,
|
| 11325 |
+
"grad_norm": 0.001420872751623392,
|
| 11326 |
+
"learning_rate": 5.76923076923077e-05,
|
| 11327 |
+
"loss": 0.0048,
|
| 11328 |
+
"step": 1617
|
| 11329 |
+
},
|
| 11330 |
+
{
|
| 11331 |
+
"epoch": 0.9428904428904429,
|
| 11332 |
+
"grad_norm": 0.0008954692748375237,
|
| 11333 |
+
"learning_rate": 5.7109557109557114e-05,
|
| 11334 |
+
"loss": 0.0026,
|
| 11335 |
+
"step": 1618
|
| 11336 |
+
},
|
| 11337 |
+
{
|
| 11338 |
+
"epoch": 0.9434731934731935,
|
| 11339 |
+
"grad_norm": 0.000952814007177949,
|
| 11340 |
+
"learning_rate": 5.652680652680653e-05,
|
| 11341 |
+
"loss": 0.004,
|
| 11342 |
+
"step": 1619
|
| 11343 |
+
},
|
| 11344 |
+
{
|
| 11345 |
+
"epoch": 0.9440559440559441,
|
| 11346 |
+
"grad_norm": 0.0011160552967339754,
|
| 11347 |
+
"learning_rate": 5.5944055944055945e-05,
|
| 11348 |
+
"loss": 0.0034,
|
| 11349 |
+
"step": 1620
|
| 11350 |
+
},
|
| 11351 |
+
{
|
| 11352 |
+
"epoch": 0.9446386946386947,
|
| 11353 |
+
"grad_norm": 0.0009834656957536936,
|
| 11354 |
+
"learning_rate": 5.536130536130536e-05,
|
| 11355 |
+
"loss": 0.0042,
|
| 11356 |
+
"step": 1621
|
| 11357 |
+
},
|
| 11358 |
+
{
|
| 11359 |
+
"epoch": 0.9452214452214452,
|
| 11360 |
+
"grad_norm": 0.0007495367899537086,
|
| 11361 |
+
"learning_rate": 5.477855477855478e-05,
|
| 11362 |
+
"loss": 0.0028,
|
| 11363 |
+
"step": 1622
|
| 11364 |
+
},
|
| 11365 |
+
{
|
| 11366 |
+
"epoch": 0.9458041958041958,
|
| 11367 |
+
"grad_norm": 0.0011641675373539329,
|
| 11368 |
+
"learning_rate": 5.419580419580419e-05,
|
| 11369 |
+
"loss": 0.004,
|
| 11370 |
+
"step": 1623
|
| 11371 |
+
},
|
| 11372 |
+
{
|
| 11373 |
+
"epoch": 0.9463869463869464,
|
| 11374 |
+
"grad_norm": 0.0013844856293871999,
|
| 11375 |
+
"learning_rate": 5.3613053613053616e-05,
|
| 11376 |
+
"loss": 0.014,
|
| 11377 |
+
"step": 1624
|
| 11378 |
+
},
|
| 11379 |
+
{
|
| 11380 |
+
"epoch": 0.946969696969697,
|
| 11381 |
+
"grad_norm": 0.0008486348669975996,
|
| 11382 |
+
"learning_rate": 5.303030303030303e-05,
|
| 11383 |
+
"loss": 0.0029,
|
| 11384 |
+
"step": 1625
|
| 11385 |
+
},
|
| 11386 |
+
{
|
| 11387 |
+
"epoch": 0.9475524475524476,
|
| 11388 |
+
"grad_norm": 0.001107304822653532,
|
| 11389 |
+
"learning_rate": 5.244755244755245e-05,
|
| 11390 |
+
"loss": 0.0028,
|
| 11391 |
+
"step": 1626
|
| 11392 |
+
},
|
| 11393 |
+
{
|
| 11394 |
+
"epoch": 0.9481351981351981,
|
| 11395 |
+
"grad_norm": 0.001052669482305646,
|
| 11396 |
+
"learning_rate": 5.1864801864801863e-05,
|
| 11397 |
+
"loss": 0.0035,
|
| 11398 |
+
"step": 1627
|
| 11399 |
+
},
|
| 11400 |
+
{
|
| 11401 |
+
"epoch": 0.9487179487179487,
|
| 11402 |
+
"grad_norm": 0.001127295778132975,
|
| 11403 |
+
"learning_rate": 5.128205128205128e-05,
|
| 11404 |
+
"loss": 0.0033,
|
| 11405 |
+
"step": 1628
|
| 11406 |
+
},
|
| 11407 |
+
{
|
| 11408 |
+
"epoch": 0.9493006993006993,
|
| 11409 |
+
"grad_norm": 0.00099327159114182,
|
| 11410 |
+
"learning_rate": 5.0699300699300695e-05,
|
| 11411 |
+
"loss": 0.0029,
|
| 11412 |
+
"step": 1629
|
| 11413 |
+
},
|
| 11414 |
+
{
|
| 11415 |
+
"epoch": 0.9498834498834499,
|
| 11416 |
+
"grad_norm": 0.0008510001935064793,
|
| 11417 |
+
"learning_rate": 5.011655011655012e-05,
|
| 11418 |
+
"loss": 0.0031,
|
| 11419 |
+
"step": 1630
|
| 11420 |
+
},
|
| 11421 |
+
{
|
| 11422 |
+
"epoch": 0.9504662004662005,
|
| 11423 |
+
"grad_norm": 0.0006990230758674443,
|
| 11424 |
+
"learning_rate": 4.9533799533799534e-05,
|
| 11425 |
+
"loss": 0.0028,
|
| 11426 |
+
"step": 1631
|
| 11427 |
+
},
|
| 11428 |
+
{
|
| 11429 |
+
"epoch": 0.951048951048951,
|
| 11430 |
+
"grad_norm": 0.0008159316494129598,
|
| 11431 |
+
"learning_rate": 4.895104895104895e-05,
|
| 11432 |
+
"loss": 0.0028,
|
| 11433 |
+
"step": 1632
|
| 11434 |
+
},
|
| 11435 |
+
{
|
| 11436 |
+
"epoch": 0.9516317016317016,
|
| 11437 |
+
"grad_norm": 0.0008230661042034626,
|
| 11438 |
+
"learning_rate": 4.836829836829837e-05,
|
| 11439 |
+
"loss": 0.0033,
|
| 11440 |
+
"step": 1633
|
| 11441 |
+
},
|
| 11442 |
+
{
|
| 11443 |
+
"epoch": 0.9522144522144522,
|
| 11444 |
+
"grad_norm": 0.0009854782838374376,
|
| 11445 |
+
"learning_rate": 4.778554778554779e-05,
|
| 11446 |
+
"loss": 0.0034,
|
| 11447 |
+
"step": 1634
|
| 11448 |
+
},
|
| 11449 |
+
{
|
| 11450 |
+
"epoch": 0.9527972027972028,
|
| 11451 |
+
"grad_norm": 0.0010709573980420828,
|
| 11452 |
+
"learning_rate": 4.7202797202797204e-05,
|
| 11453 |
+
"loss": 0.0029,
|
| 11454 |
+
"step": 1635
|
| 11455 |
+
},
|
| 11456 |
+
{
|
| 11457 |
+
"epoch": 0.9533799533799534,
|
| 11458 |
+
"grad_norm": 0.0010027334792539477,
|
| 11459 |
+
"learning_rate": 4.662004662004663e-05,
|
| 11460 |
+
"loss": 0.0033,
|
| 11461 |
+
"step": 1636
|
| 11462 |
+
},
|
| 11463 |
+
{
|
| 11464 |
+
"epoch": 0.953962703962704,
|
| 11465 |
+
"grad_norm": 0.0009396614041179419,
|
| 11466 |
+
"learning_rate": 4.603729603729604e-05,
|
| 11467 |
+
"loss": 0.0088,
|
| 11468 |
+
"step": 1637
|
| 11469 |
+
},
|
| 11470 |
+
{
|
| 11471 |
+
"epoch": 0.9545454545454546,
|
| 11472 |
+
"grad_norm": 0.0007274977397173643,
|
| 11473 |
+
"learning_rate": 4.545454545454546e-05,
|
| 11474 |
+
"loss": 0.0032,
|
| 11475 |
+
"step": 1638
|
| 11476 |
+
},
|
| 11477 |
+
{
|
| 11478 |
+
"epoch": 0.9551282051282052,
|
| 11479 |
+
"grad_norm": 0.0010892600985243917,
|
| 11480 |
+
"learning_rate": 4.4871794871794874e-05,
|
| 11481 |
+
"loss": 0.0042,
|
| 11482 |
+
"step": 1639
|
| 11483 |
+
},
|
| 11484 |
+
{
|
| 11485 |
+
"epoch": 0.9557109557109557,
|
| 11486 |
+
"grad_norm": 0.0007885160739533603,
|
| 11487 |
+
"learning_rate": 4.428904428904429e-05,
|
| 11488 |
+
"loss": 0.0031,
|
| 11489 |
+
"step": 1640
|
| 11490 |
+
},
|
| 11491 |
+
{
|
| 11492 |
+
"epoch": 0.9562937062937062,
|
| 11493 |
+
"grad_norm": 0.0012906527845188975,
|
| 11494 |
+
"learning_rate": 4.3706293706293706e-05,
|
| 11495 |
+
"loss": 0.0044,
|
| 11496 |
+
"step": 1641
|
| 11497 |
+
},
|
| 11498 |
+
{
|
| 11499 |
+
"epoch": 0.9568764568764568,
|
| 11500 |
+
"grad_norm": 0.0012814976507797837,
|
| 11501 |
+
"learning_rate": 4.312354312354312e-05,
|
| 11502 |
+
"loss": 0.0038,
|
| 11503 |
+
"step": 1642
|
| 11504 |
+
},
|
| 11505 |
+
{
|
| 11506 |
+
"epoch": 0.9574592074592074,
|
| 11507 |
+
"grad_norm": 0.0011575610842555761,
|
| 11508 |
+
"learning_rate": 4.2540792540792545e-05,
|
| 11509 |
+
"loss": 0.004,
|
| 11510 |
+
"step": 1643
|
| 11511 |
+
},
|
| 11512 |
+
{
|
| 11513 |
+
"epoch": 0.958041958041958,
|
| 11514 |
+
"grad_norm": 0.0012061079032719135,
|
| 11515 |
+
"learning_rate": 4.195804195804196e-05,
|
| 11516 |
+
"loss": 0.0043,
|
| 11517 |
+
"step": 1644
|
| 11518 |
+
},
|
| 11519 |
+
{
|
| 11520 |
+
"epoch": 0.9586247086247086,
|
| 11521 |
+
"grad_norm": 0.0008972581708803773,
|
| 11522 |
+
"learning_rate": 4.1375291375291377e-05,
|
| 11523 |
+
"loss": 0.0038,
|
| 11524 |
+
"step": 1645
|
| 11525 |
+
},
|
| 11526 |
+
{
|
| 11527 |
+
"epoch": 0.9592074592074592,
|
| 11528 |
+
"grad_norm": 0.0008104901062324643,
|
| 11529 |
+
"learning_rate": 4.079254079254079e-05,
|
| 11530 |
+
"loss": 0.0028,
|
| 11531 |
+
"step": 1646
|
| 11532 |
+
},
|
| 11533 |
+
{
|
| 11534 |
+
"epoch": 0.9597902097902098,
|
| 11535 |
+
"grad_norm": 0.0007863112259656191,
|
| 11536 |
+
"learning_rate": 4.020979020979021e-05,
|
| 11537 |
+
"loss": 0.003,
|
| 11538 |
+
"step": 1647
|
| 11539 |
+
},
|
| 11540 |
+
{
|
| 11541 |
+
"epoch": 0.9603729603729604,
|
| 11542 |
+
"grad_norm": 0.0006893305107951164,
|
| 11543 |
+
"learning_rate": 3.9627039627039624e-05,
|
| 11544 |
+
"loss": 0.0026,
|
| 11545 |
+
"step": 1648
|
| 11546 |
+
},
|
| 11547 |
+
{
|
| 11548 |
+
"epoch": 0.960955710955711,
|
| 11549 |
+
"grad_norm": 0.0009396909736096859,
|
| 11550 |
+
"learning_rate": 3.904428904428905e-05,
|
| 11551 |
+
"loss": 0.0029,
|
| 11552 |
+
"step": 1649
|
| 11553 |
+
},
|
| 11554 |
+
{
|
| 11555 |
+
"epoch": 0.9615384615384616,
|
| 11556 |
+
"grad_norm": 0.0007228578324429691,
|
| 11557 |
+
"learning_rate": 3.846153846153846e-05,
|
| 11558 |
+
"loss": 0.0023,
|
| 11559 |
+
"step": 1650
|
| 11560 |
+
},
|
| 11561 |
+
{
|
| 11562 |
+
"epoch": 0.9621212121212122,
|
| 11563 |
+
"grad_norm": 0.0011825780384242535,
|
| 11564 |
+
"learning_rate": 3.787878787878788e-05,
|
| 11565 |
+
"loss": 0.0044,
|
| 11566 |
+
"step": 1651
|
| 11567 |
+
},
|
| 11568 |
+
{
|
| 11569 |
+
"epoch": 0.9627039627039627,
|
| 11570 |
+
"grad_norm": 0.0007647788152098656,
|
| 11571 |
+
"learning_rate": 3.7296037296037295e-05,
|
| 11572 |
+
"loss": 0.003,
|
| 11573 |
+
"step": 1652
|
| 11574 |
+
},
|
| 11575 |
+
{
|
| 11576 |
+
"epoch": 0.9632867132867133,
|
| 11577 |
+
"grad_norm": 0.0010042464127764106,
|
| 11578 |
+
"learning_rate": 3.671328671328671e-05,
|
| 11579 |
+
"loss": 0.0041,
|
| 11580 |
+
"step": 1653
|
| 11581 |
+
},
|
| 11582 |
+
{
|
| 11583 |
+
"epoch": 0.9638694638694638,
|
| 11584 |
+
"grad_norm": 0.0009029952925629914,
|
| 11585 |
+
"learning_rate": 3.6130536130536126e-05,
|
| 11586 |
+
"loss": 0.0032,
|
| 11587 |
+
"step": 1654
|
| 11588 |
+
},
|
| 11589 |
+
{
|
| 11590 |
+
"epoch": 0.9644522144522144,
|
| 11591 |
+
"grad_norm": 0.0010356158018112183,
|
| 11592 |
+
"learning_rate": 3.554778554778554e-05,
|
| 11593 |
+
"loss": 0.0031,
|
| 11594 |
+
"step": 1655
|
| 11595 |
+
},
|
| 11596 |
+
{
|
| 11597 |
+
"epoch": 0.965034965034965,
|
| 11598 |
+
"grad_norm": 0.0010094497120007873,
|
| 11599 |
+
"learning_rate": 3.4965034965034965e-05,
|
| 11600 |
+
"loss": 0.0043,
|
| 11601 |
+
"step": 1656
|
| 11602 |
+
},
|
| 11603 |
+
{
|
| 11604 |
+
"epoch": 0.9656177156177156,
|
| 11605 |
+
"grad_norm": 0.0008370497962459922,
|
| 11606 |
+
"learning_rate": 3.438228438228439e-05,
|
| 11607 |
+
"loss": 0.0035,
|
| 11608 |
+
"step": 1657
|
| 11609 |
+
},
|
| 11610 |
+
{
|
| 11611 |
+
"epoch": 0.9662004662004662,
|
| 11612 |
+
"grad_norm": 0.000803111121058464,
|
| 11613 |
+
"learning_rate": 3.3799533799533804e-05,
|
| 11614 |
+
"loss": 0.0029,
|
| 11615 |
+
"step": 1658
|
| 11616 |
+
},
|
| 11617 |
+
{
|
| 11618 |
+
"epoch": 0.9667832167832168,
|
| 11619 |
+
"grad_norm": 0.0011143162846565247,
|
| 11620 |
+
"learning_rate": 3.321678321678322e-05,
|
| 11621 |
+
"loss": 0.0035,
|
| 11622 |
+
"step": 1659
|
| 11623 |
+
},
|
| 11624 |
+
{
|
| 11625 |
+
"epoch": 0.9673659673659674,
|
| 11626 |
+
"grad_norm": 0.0008031773613765836,
|
| 11627 |
+
"learning_rate": 3.2634032634032635e-05,
|
| 11628 |
+
"loss": 0.0031,
|
| 11629 |
+
"step": 1660
|
| 11630 |
+
},
|
| 11631 |
+
{
|
| 11632 |
+
"epoch": 0.967948717948718,
|
| 11633 |
+
"grad_norm": 0.0011737227905541658,
|
| 11634 |
+
"learning_rate": 3.205128205128205e-05,
|
| 11635 |
+
"loss": 0.004,
|
| 11636 |
+
"step": 1661
|
| 11637 |
+
},
|
| 11638 |
+
{
|
| 11639 |
+
"epoch": 0.9685314685314685,
|
| 11640 |
+
"grad_norm": 0.001068048644810915,
|
| 11641 |
+
"learning_rate": 3.1468531468531474e-05,
|
| 11642 |
+
"loss": 0.0038,
|
| 11643 |
+
"step": 1662
|
| 11644 |
+
},
|
| 11645 |
+
{
|
| 11646 |
+
"epoch": 0.9691142191142191,
|
| 11647 |
+
"grad_norm": 0.0008520625997334719,
|
| 11648 |
+
"learning_rate": 3.088578088578088e-05,
|
| 11649 |
+
"loss": 0.0032,
|
| 11650 |
+
"step": 1663
|
| 11651 |
+
},
|
| 11652 |
+
{
|
| 11653 |
+
"epoch": 0.9696969696969697,
|
| 11654 |
+
"grad_norm": 0.0006241014925763011,
|
| 11655 |
+
"learning_rate": 3.0303030303030306e-05,
|
| 11656 |
+
"loss": 0.0023,
|
| 11657 |
+
"step": 1664
|
| 11658 |
+
},
|
| 11659 |
+
{
|
| 11660 |
+
"epoch": 0.9702797202797203,
|
| 11661 |
+
"grad_norm": 0.0013019571779295802,
|
| 11662 |
+
"learning_rate": 2.972027972027972e-05,
|
| 11663 |
+
"loss": 0.0048,
|
| 11664 |
+
"step": 1665
|
| 11665 |
+
},
|
| 11666 |
+
{
|
| 11667 |
+
"epoch": 0.9708624708624709,
|
| 11668 |
+
"grad_norm": 0.0011066205333918333,
|
| 11669 |
+
"learning_rate": 2.9137529137529138e-05,
|
| 11670 |
+
"loss": 0.0031,
|
| 11671 |
+
"step": 1666
|
| 11672 |
+
},
|
| 11673 |
+
{
|
| 11674 |
+
"epoch": 0.9714452214452215,
|
| 11675 |
+
"grad_norm": 0.0010444342624396086,
|
| 11676 |
+
"learning_rate": 2.8554778554778557e-05,
|
| 11677 |
+
"loss": 0.0033,
|
| 11678 |
+
"step": 1667
|
| 11679 |
+
},
|
| 11680 |
+
{
|
| 11681 |
+
"epoch": 0.972027972027972,
|
| 11682 |
+
"grad_norm": 0.0007983744144439697,
|
| 11683 |
+
"learning_rate": 2.7972027972027973e-05,
|
| 11684 |
+
"loss": 0.0032,
|
| 11685 |
+
"step": 1668
|
| 11686 |
+
},
|
| 11687 |
+
{
|
| 11688 |
+
"epoch": 0.9726107226107226,
|
| 11689 |
+
"grad_norm": 0.0008884937269613147,
|
| 11690 |
+
"learning_rate": 2.738927738927739e-05,
|
| 11691 |
+
"loss": 0.0037,
|
| 11692 |
+
"step": 1669
|
| 11693 |
+
},
|
| 11694 |
+
{
|
| 11695 |
+
"epoch": 0.9731934731934732,
|
| 11696 |
+
"grad_norm": 0.0012803805293515325,
|
| 11697 |
+
"learning_rate": 2.6806526806526808e-05,
|
| 11698 |
+
"loss": 0.0034,
|
| 11699 |
+
"step": 1670
|
| 11700 |
+
},
|
| 11701 |
+
{
|
| 11702 |
+
"epoch": 0.9737762237762237,
|
| 11703 |
+
"grad_norm": 0.0011096763191744685,
|
| 11704 |
+
"learning_rate": 2.6223776223776224e-05,
|
| 11705 |
+
"loss": 0.0035,
|
| 11706 |
+
"step": 1671
|
| 11707 |
+
},
|
| 11708 |
+
{
|
| 11709 |
+
"epoch": 0.9743589743589743,
|
| 11710 |
+
"grad_norm": 0.0011874607298523188,
|
| 11711 |
+
"learning_rate": 2.564102564102564e-05,
|
| 11712 |
+
"loss": 0.0043,
|
| 11713 |
+
"step": 1672
|
| 11714 |
+
},
|
| 11715 |
+
{
|
| 11716 |
+
"epoch": 0.9749417249417249,
|
| 11717 |
+
"grad_norm": 0.0009078698931261897,
|
| 11718 |
+
"learning_rate": 2.505827505827506e-05,
|
| 11719 |
+
"loss": 0.0027,
|
| 11720 |
+
"step": 1673
|
| 11721 |
+
},
|
| 11722 |
+
{
|
| 11723 |
+
"epoch": 0.9755244755244755,
|
| 11724 |
+
"grad_norm": 0.0009188731200993061,
|
| 11725 |
+
"learning_rate": 2.4475524475524475e-05,
|
| 11726 |
+
"loss": 0.0046,
|
| 11727 |
+
"step": 1674
|
| 11728 |
+
},
|
| 11729 |
+
{
|
| 11730 |
+
"epoch": 0.9761072261072261,
|
| 11731 |
+
"grad_norm": 0.0009996923618018627,
|
| 11732 |
+
"learning_rate": 2.3892773892773894e-05,
|
| 11733 |
+
"loss": 0.006,
|
| 11734 |
+
"step": 1675
|
| 11735 |
+
},
|
| 11736 |
+
{
|
| 11737 |
+
"epoch": 0.9766899766899767,
|
| 11738 |
+
"grad_norm": 0.0008249058737419546,
|
| 11739 |
+
"learning_rate": 2.3310023310023313e-05,
|
| 11740 |
+
"loss": 0.0033,
|
| 11741 |
+
"step": 1676
|
| 11742 |
+
},
|
| 11743 |
+
{
|
| 11744 |
+
"epoch": 0.9772727272727273,
|
| 11745 |
+
"grad_norm": 0.0008003967232070863,
|
| 11746 |
+
"learning_rate": 2.272727272727273e-05,
|
| 11747 |
+
"loss": 0.003,
|
| 11748 |
+
"step": 1677
|
| 11749 |
+
},
|
| 11750 |
+
{
|
| 11751 |
+
"epoch": 0.9778554778554779,
|
| 11752 |
+
"grad_norm": 0.0014063924318179488,
|
| 11753 |
+
"learning_rate": 2.2144522144522145e-05,
|
| 11754 |
+
"loss": 0.0036,
|
| 11755 |
+
"step": 1678
|
| 11756 |
+
},
|
| 11757 |
+
{
|
| 11758 |
+
"epoch": 0.9784382284382285,
|
| 11759 |
+
"grad_norm": 0.0008004964329302311,
|
| 11760 |
+
"learning_rate": 2.156177156177156e-05,
|
| 11761 |
+
"loss": 0.0028,
|
| 11762 |
+
"step": 1679
|
| 11763 |
+
},
|
| 11764 |
+
{
|
| 11765 |
+
"epoch": 0.9790209790209791,
|
| 11766 |
+
"grad_norm": 0.0009626666433177888,
|
| 11767 |
+
"learning_rate": 2.097902097902098e-05,
|
| 11768 |
+
"loss": 0.0027,
|
| 11769 |
+
"step": 1680
|
| 11770 |
+
},
|
| 11771 |
+
{
|
| 11772 |
+
"epoch": 0.9796037296037297,
|
| 11773 |
+
"grad_norm": 0.0008628361392766237,
|
| 11774 |
+
"learning_rate": 2.0396270396270396e-05,
|
| 11775 |
+
"loss": 0.0027,
|
| 11776 |
+
"step": 1681
|
| 11777 |
+
},
|
| 11778 |
+
{
|
| 11779 |
+
"epoch": 0.9801864801864801,
|
| 11780 |
+
"grad_norm": 0.0008874722989276052,
|
| 11781 |
+
"learning_rate": 1.9813519813519812e-05,
|
| 11782 |
+
"loss": 0.0042,
|
| 11783 |
+
"step": 1682
|
| 11784 |
+
},
|
| 11785 |
+
{
|
| 11786 |
+
"epoch": 0.9807692307692307,
|
| 11787 |
+
"grad_norm": 0.0015157037414610386,
|
| 11788 |
+
"learning_rate": 1.923076923076923e-05,
|
| 11789 |
+
"loss": 0.0052,
|
| 11790 |
+
"step": 1683
|
| 11791 |
+
},
|
| 11792 |
+
{
|
| 11793 |
+
"epoch": 0.9813519813519813,
|
| 11794 |
+
"grad_norm": 0.0011951492633670568,
|
| 11795 |
+
"learning_rate": 1.8648018648018647e-05,
|
| 11796 |
+
"loss": 0.0049,
|
| 11797 |
+
"step": 1684
|
| 11798 |
+
},
|
| 11799 |
+
{
|
| 11800 |
+
"epoch": 0.9819347319347319,
|
| 11801 |
+
"grad_norm": 0.0009483549511060119,
|
| 11802 |
+
"learning_rate": 1.8065268065268063e-05,
|
| 11803 |
+
"loss": 0.0026,
|
| 11804 |
+
"step": 1685
|
| 11805 |
+
},
|
| 11806 |
+
{
|
| 11807 |
+
"epoch": 0.9825174825174825,
|
| 11808 |
+
"grad_norm": 0.000737398280762136,
|
| 11809 |
+
"learning_rate": 1.7482517482517483e-05,
|
| 11810 |
+
"loss": 0.0027,
|
| 11811 |
+
"step": 1686
|
| 11812 |
+
},
|
| 11813 |
+
{
|
| 11814 |
+
"epoch": 0.9831002331002331,
|
| 11815 |
+
"grad_norm": 0.001031695050187409,
|
| 11816 |
+
"learning_rate": 1.6899766899766902e-05,
|
| 11817 |
+
"loss": 0.0025,
|
| 11818 |
+
"step": 1687
|
| 11819 |
+
},
|
| 11820 |
+
{
|
| 11821 |
+
"epoch": 0.9836829836829837,
|
| 11822 |
+
"grad_norm": 0.0009539015591144562,
|
| 11823 |
+
"learning_rate": 1.6317016317016318e-05,
|
| 11824 |
+
"loss": 0.0027,
|
| 11825 |
+
"step": 1688
|
| 11826 |
+
},
|
| 11827 |
+
{
|
| 11828 |
+
"epoch": 0.9842657342657343,
|
| 11829 |
+
"grad_norm": 0.0007205713191069663,
|
| 11830 |
+
"learning_rate": 1.5734265734265737e-05,
|
| 11831 |
+
"loss": 0.0026,
|
| 11832 |
+
"step": 1689
|
| 11833 |
+
},
|
| 11834 |
+
{
|
| 11835 |
+
"epoch": 0.9848484848484849,
|
| 11836 |
+
"grad_norm": 0.0009316445211879909,
|
| 11837 |
+
"learning_rate": 1.5151515151515153e-05,
|
| 11838 |
+
"loss": 0.0032,
|
| 11839 |
+
"step": 1690
|
| 11840 |
+
},
|
| 11841 |
+
{
|
| 11842 |
+
"epoch": 0.9854312354312355,
|
| 11843 |
+
"grad_norm": 0.0010426414664834738,
|
| 11844 |
+
"learning_rate": 1.4568764568764569e-05,
|
| 11845 |
+
"loss": 0.0041,
|
| 11846 |
+
"step": 1691
|
| 11847 |
+
},
|
| 11848 |
+
{
|
| 11849 |
+
"epoch": 0.986013986013986,
|
| 11850 |
+
"grad_norm": 0.0006400300771929324,
|
| 11851 |
+
"learning_rate": 1.3986013986013986e-05,
|
| 11852 |
+
"loss": 0.0024,
|
| 11853 |
+
"step": 1692
|
| 11854 |
+
},
|
| 11855 |
+
{
|
| 11856 |
+
"epoch": 0.9865967365967366,
|
| 11857 |
+
"grad_norm": 0.0009383864235132933,
|
| 11858 |
+
"learning_rate": 1.3403263403263404e-05,
|
| 11859 |
+
"loss": 0.0029,
|
| 11860 |
+
"step": 1693
|
| 11861 |
+
},
|
| 11862 |
+
{
|
| 11863 |
+
"epoch": 0.9871794871794872,
|
| 11864 |
+
"grad_norm": 0.0008284033392556012,
|
| 11865 |
+
"learning_rate": 1.282051282051282e-05,
|
| 11866 |
+
"loss": 0.0029,
|
| 11867 |
+
"step": 1694
|
| 11868 |
+
},
|
| 11869 |
+
{
|
| 11870 |
+
"epoch": 0.9877622377622378,
|
| 11871 |
+
"grad_norm": 0.0010577579960227013,
|
| 11872 |
+
"learning_rate": 1.2237762237762237e-05,
|
| 11873 |
+
"loss": 0.0052,
|
| 11874 |
+
"step": 1695
|
| 11875 |
+
},
|
| 11876 |
+
{
|
| 11877 |
+
"epoch": 0.9883449883449883,
|
| 11878 |
+
"grad_norm": 0.0009795171208679676,
|
| 11879 |
+
"learning_rate": 1.1655011655011657e-05,
|
| 11880 |
+
"loss": 0.0046,
|
| 11881 |
+
"step": 1696
|
| 11882 |
+
},
|
| 11883 |
+
{
|
| 11884 |
+
"epoch": 0.9889277389277389,
|
| 11885 |
+
"grad_norm": 0.0007197922095656395,
|
| 11886 |
+
"learning_rate": 1.1072261072261073e-05,
|
| 11887 |
+
"loss": 0.0031,
|
| 11888 |
+
"step": 1697
|
| 11889 |
+
},
|
| 11890 |
+
{
|
| 11891 |
+
"epoch": 0.9895104895104895,
|
| 11892 |
+
"grad_norm": 0.0010326344054192305,
|
| 11893 |
+
"learning_rate": 1.048951048951049e-05,
|
| 11894 |
+
"loss": 0.0028,
|
| 11895 |
+
"step": 1698
|
| 11896 |
+
},
|
| 11897 |
+
{
|
| 11898 |
+
"epoch": 0.9900932400932401,
|
| 11899 |
+
"grad_norm": 0.0008666233043186367,
|
| 11900 |
+
"learning_rate": 9.906759906759906e-06,
|
| 11901 |
+
"loss": 0.003,
|
| 11902 |
+
"step": 1699
|
| 11903 |
+
},
|
| 11904 |
+
{
|
| 11905 |
+
"epoch": 0.9906759906759907,
|
| 11906 |
+
"grad_norm": 0.0008774434681981802,
|
| 11907 |
+
"learning_rate": 9.324009324009324e-06,
|
| 11908 |
+
"loss": 0.003,
|
| 11909 |
+
"step": 1700
|
| 11910 |
}
|
| 11911 |
],
|
| 11912 |
"logging_steps": 1,
|
|
|
|
| 11926 |
"attributes": {}
|
| 11927 |
}
|
| 11928 |
},
|
| 11929 |
+
"total_flos": 7.109695330572042e+19,
|
| 11930 |
"train_batch_size": 4,
|
| 11931 |
"trial_name": null,
|
| 11932 |
"trial_params": null
|