Training in progress, step 1700, checkpoint
Browse files
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 160285149
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b967e2784a474c2b26a772d6672c7c1550fbc2adbfc9acc49598b5c489a5d89f
|
3 |
size 160285149
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14575
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:caa26cd40f8bd339159d1064efac06feb5cc588c22df170f65fdc862f4bca0ad
|
3 |
size 14575
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 627
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:119acaabeb3e83e43ca8b3d26532ead56181d0051d00810e5cb205761dea60f2
|
3 |
size 627
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 2.
|
5 |
"eval_steps": 100,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -215,13 +215,26 @@
|
|
215 |
"eval_samples_per_second": 1.903,
|
216 |
"eval_steps_per_second": 0.238,
|
217 |
"step": 1600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
218 |
}
|
219 |
],
|
220 |
"logging_steps": 100,
|
221 |
"max_steps": 2025,
|
222 |
"num_train_epochs": 3,
|
223 |
"save_steps": 100,
|
224 |
-
"total_flos": 3.
|
225 |
"trial_name": null,
|
226 |
"trial_params": null
|
227 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 2.5185185185185186,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 1700,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
215 |
"eval_samples_per_second": 1.903,
|
216 |
"eval_steps_per_second": 0.238,
|
217 |
"step": 1600
|
218 |
+
},
|
219 |
+
{
|
220 |
+
"epoch": 2.52,
|
221 |
+
"learning_rate": 0.0,
|
222 |
+
"loss": 8.7091,
|
223 |
+
"step": 1700
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 2.52,
|
227 |
+
"eval_runtime": 788.7408,
|
228 |
+
"eval_samples_per_second": 1.902,
|
229 |
+
"eval_steps_per_second": 0.238,
|
230 |
+
"step": 1700
|
231 |
}
|
232 |
],
|
233 |
"logging_steps": 100,
|
234 |
"max_steps": 2025,
|
235 |
"num_train_epochs": 3,
|
236 |
"save_steps": 100,
|
237 |
+
"total_flos": 3.54008678006784e+17,
|
238 |
"trial_name": null,
|
239 |
"trial_params": null
|
240 |
}
|