Training in progress, step 1400, 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 sha256:1b3fc1519a80ff09d34e614c9dcdfe5447dd304b116bc3936ecf5d7f466c41c5
|
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 sha256:00c70fec5855430ce0afb62afb7a7522f878bca813147b4cb565f553e75c1fd5
|
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 sha256:426b58c51fde03deb12bf7d7b53933d1a534d1c5cd0baca806ba16d364acbc1f
|
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":
|
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,
|
@@ -176,13 +176,26 @@
|
|
176 |
"eval_samples_per_second": 1.904,
|
177 |
"eval_steps_per_second": 0.239,
|
178 |
"step": 1300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
179 |
}
|
180 |
],
|
181 |
"logging_steps": 100,
|
182 |
"max_steps": 2025,
|
183 |
"num_train_epochs": 3,
|
184 |
"save_steps": 100,
|
185 |
-
"total_flos": 2.
|
186 |
"trial_name": null,
|
187 |
"trial_params": null
|
188 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 2.074074074074074,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 1400,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
176 |
"eval_samples_per_second": 1.904,
|
177 |
"eval_steps_per_second": 0.239,
|
178 |
"step": 1300
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"epoch": 2.07,
|
182 |
+
"learning_rate": 0.0,
|
183 |
+
"loss": 8.4373,
|
184 |
+
"step": 1400
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 2.07,
|
188 |
+
"eval_runtime": 788.547,
|
189 |
+
"eval_samples_per_second": 1.902,
|
190 |
+
"eval_steps_per_second": 0.238,
|
191 |
+
"step": 1400
|
192 |
}
|
193 |
],
|
194 |
"logging_steps": 100,
|
195 |
"max_steps": 2025,
|
196 |
"num_train_epochs": 3,
|
197 |
"save_steps": 100,
|
198 |
+
"total_flos": 2.91536558358528e+17,
|
199 |
"trial_name": null,
|
200 |
"trial_params": null
|
201 |
}
|