Commit ·
daee7a4
1
Parent(s): 67ccd85
Training in progress, step 200
Browse files- all_results.json +20 -0
- eval_results.json +9 -0
- pytorch_model.bin +2 -2
- runs/Feb01_01-41-55_node1/1643651046.7833238/events.out.tfevents.1643651046.node1 +3 -0
- runs/Feb01_01-41-55_node1/events.out.tfevents.1643651046.node1 +3 -0
- runs/Jan31_21-18-51_node1/events.out.tfevents.1643635211.node1 +2 -2
- runs/Jan31_21-18-51_node1/events.out.tfevents.1643640886.node1 +3 -0
- test_results.json +9 -0
- train_results.json +8 -0
- trainer_state.json +841 -0
- training_args.bin +1 -1
- training_config.json +1 -1
all_results.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_average_metrics": 90.63138686131387,
|
| 4 |
+
"eval_em": 90.36496350364963,
|
| 5 |
+
"eval_f1": 90.89781021897811,
|
| 6 |
+
"eval_loss": 0.2141435146331787,
|
| 7 |
+
"eval_runtime": 15.7609,
|
| 8 |
+
"eval_samples_per_second": 86.924,
|
| 9 |
+
"test_average_metrics": 74.47490712152019,
|
| 10 |
+
"test_em": 73.99182920400632,
|
| 11 |
+
"test_f1": 74.95798503903406,
|
| 12 |
+
"test_loss": 0.48021578788757324,
|
| 13 |
+
"test_runtime": 161.768,
|
| 14 |
+
"test_samples_per_second": 93.813,
|
| 15 |
+
"train_loss": 0.3297047741756515,
|
| 16 |
+
"train_runtime": 5659.5381,
|
| 17 |
+
"train_samples": 137484,
|
| 18 |
+
"train_samples_per_second": 72.877,
|
| 19 |
+
"train_steps_per_second": 4.555
|
| 20 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_average_metrics": 90.63138686131387,
|
| 4 |
+
"eval_em": 90.36496350364963,
|
| 5 |
+
"eval_f1": 90.89781021897811,
|
| 6 |
+
"eval_loss": 0.2141435146331787,
|
| 7 |
+
"eval_runtime": 15.7609,
|
| 8 |
+
"eval_samples_per_second": 86.924
|
| 9 |
+
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:697ff0d656a9f6f3d37dede6c1c838d5f15888d87469d9928e6ec6074dfe795e
|
| 3 |
+
size 7551621
|
runs/Feb01_01-41-55_node1/1643651046.7833238/events.out.tfevents.1643651046.node1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dbbcd9faac5e4993dcf01bf6ac455a92358d719291b762a30f5fab378511fae3
|
| 3 |
+
size 5047
|
runs/Feb01_01-41-55_node1/events.out.tfevents.1643651046.node1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:63b5dc9b6a0c88a10fc9623f05948f0aa16f106821d7816c1ca7fc9670c2e518
|
| 3 |
+
size 4697
|
runs/Jan31_21-18-51_node1/events.out.tfevents.1643635211.node1
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:727d08a4747fbb17bdf0b16cc330708f4fa29121ddda29452ff97cdb175ca11f
|
| 3 |
+
size 31354
|
runs/Jan31_21-18-51_node1/events.out.tfevents.1643640886.node1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7a782fda4e1bf1c1b52e696f156c6c097509af59397fe673f5bd30f45333b22d
|
| 3 |
+
size 778
|
test_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"test_average_metrics": 74.47490712152019,
|
| 4 |
+
"test_em": 73.99182920400632,
|
| 5 |
+
"test_f1": 74.95798503903406,
|
| 6 |
+
"test_loss": 0.48021578788757324,
|
| 7 |
+
"test_runtime": 161.768,
|
| 8 |
+
"test_samples_per_second": 93.813
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"train_loss": 0.3297047741756515,
|
| 4 |
+
"train_runtime": 5659.5381,
|
| 5 |
+
"train_samples": 137484,
|
| 6 |
+
"train_samples_per_second": 72.877,
|
| 7 |
+
"train_steps_per_second": 4.555
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,841 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 90.63138686131387,
|
| 3 |
+
"best_model_checkpoint": "outputs/bitfit/t5-base/superglue-record/checkpoint-21500",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 25779,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.06,
|
| 12 |
+
"learning_rate": 0.0002941813103689049,
|
| 13 |
+
"loss": 0.3506,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"eval_average_metrics": 90.1404240528328,
|
| 19 |
+
"eval_em": 89.85401459854015,
|
| 20 |
+
"eval_f1": 90.42683350712547,
|
| 21 |
+
"eval_loss": 0.23184123635292053,
|
| 22 |
+
"eval_runtime": 15.5789,
|
| 23 |
+
"eval_samples_per_second": 87.939,
|
| 24 |
+
"step": 500
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.12,
|
| 28 |
+
"learning_rate": 0.0002883626207378098,
|
| 29 |
+
"loss": 0.3505,
|
| 30 |
+
"step": 1000
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.12,
|
| 34 |
+
"eval_average_metrics": 89.19586374695865,
|
| 35 |
+
"eval_em": 88.75912408759125,
|
| 36 |
+
"eval_f1": 89.63260340632604,
|
| 37 |
+
"eval_loss": 0.2271701991558075,
|
| 38 |
+
"eval_runtime": 15.9901,
|
| 39 |
+
"eval_samples_per_second": 85.678,
|
| 40 |
+
"step": 1000
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.17,
|
| 44 |
+
"learning_rate": 0.00028254393110671474,
|
| 45 |
+
"loss": 0.3559,
|
| 46 |
+
"step": 1500
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.17,
|
| 50 |
+
"eval_average_metrics": 87.68126520681264,
|
| 51 |
+
"eval_em": 87.37226277372262,
|
| 52 |
+
"eval_f1": 87.99026763990267,
|
| 53 |
+
"eval_loss": 0.23542903363704681,
|
| 54 |
+
"eval_runtime": 15.5654,
|
| 55 |
+
"eval_samples_per_second": 88.016,
|
| 56 |
+
"step": 1500
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.23,
|
| 60 |
+
"learning_rate": 0.0002767252414756197,
|
| 61 |
+
"loss": 0.3487,
|
| 62 |
+
"step": 2000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.23,
|
| 66 |
+
"eval_average_metrics": 88.3946819603754,
|
| 67 |
+
"eval_em": 87.95620437956204,
|
| 68 |
+
"eval_f1": 88.83315954118875,
|
| 69 |
+
"eval_loss": 0.23398332297801971,
|
| 70 |
+
"eval_runtime": 15.718,
|
| 71 |
+
"eval_samples_per_second": 87.161,
|
| 72 |
+
"step": 2000
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.29,
|
| 76 |
+
"learning_rate": 0.0002709065518445246,
|
| 77 |
+
"loss": 0.3461,
|
| 78 |
+
"step": 2500
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.29,
|
| 82 |
+
"eval_average_metrics": 88.34428223844282,
|
| 83 |
+
"eval_em": 88.02919708029196,
|
| 84 |
+
"eval_f1": 88.65936739659367,
|
| 85 |
+
"eval_loss": 0.23554831743240356,
|
| 86 |
+
"eval_runtime": 15.4728,
|
| 87 |
+
"eval_samples_per_second": 88.542,
|
| 88 |
+
"step": 2500
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 0.35,
|
| 92 |
+
"learning_rate": 0.0002650878622134295,
|
| 93 |
+
"loss": 0.3514,
|
| 94 |
+
"step": 3000
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.35,
|
| 98 |
+
"eval_average_metrics": 88.32290580465764,
|
| 99 |
+
"eval_em": 87.88321167883211,
|
| 100 |
+
"eval_f1": 88.76259993048315,
|
| 101 |
+
"eval_loss": 0.23650217056274414,
|
| 102 |
+
"eval_runtime": 15.254,
|
| 103 |
+
"eval_samples_per_second": 89.812,
|
| 104 |
+
"step": 3000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.41,
|
| 108 |
+
"learning_rate": 0.00025926917258233445,
|
| 109 |
+
"loss": 0.3468,
|
| 110 |
+
"step": 3500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.41,
|
| 114 |
+
"eval_average_metrics": 89.25669099756689,
|
| 115 |
+
"eval_em": 88.97810218978101,
|
| 116 |
+
"eval_f1": 89.53527980535279,
|
| 117 |
+
"eval_loss": 0.22920837998390198,
|
| 118 |
+
"eval_runtime": 15.3356,
|
| 119 |
+
"eval_samples_per_second": 89.335,
|
| 120 |
+
"step": 3500
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.47,
|
| 124 |
+
"learning_rate": 0.0002534504829512394,
|
| 125 |
+
"loss": 0.3428,
|
| 126 |
+
"step": 4000
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.47,
|
| 130 |
+
"eval_average_metrics": 88.33211678832117,
|
| 131 |
+
"eval_em": 88.02919708029196,
|
| 132 |
+
"eval_f1": 88.63503649635038,
|
| 133 |
+
"eval_loss": 0.2317986935377121,
|
| 134 |
+
"eval_runtime": 15.4226,
|
| 135 |
+
"eval_samples_per_second": 88.831,
|
| 136 |
+
"step": 4000
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.52,
|
| 140 |
+
"learning_rate": 0.0002476317933201443,
|
| 141 |
+
"loss": 0.3485,
|
| 142 |
+
"step": 4500
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.52,
|
| 146 |
+
"eval_average_metrics": 87.90024330900243,
|
| 147 |
+
"eval_em": 87.66423357664233,
|
| 148 |
+
"eval_f1": 88.13625304136254,
|
| 149 |
+
"eval_loss": 0.23495055735111237,
|
| 150 |
+
"eval_runtime": 15.4567,
|
| 151 |
+
"eval_samples_per_second": 88.635,
|
| 152 |
+
"step": 4500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.58,
|
| 156 |
+
"learning_rate": 0.0002418131036890492,
|
| 157 |
+
"loss": 0.3509,
|
| 158 |
+
"step": 5000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.58,
|
| 162 |
+
"eval_average_metrics": 88.830900243309,
|
| 163 |
+
"eval_em": 88.61313868613139,
|
| 164 |
+
"eval_f1": 89.04866180048661,
|
| 165 |
+
"eval_loss": 0.2319142073392868,
|
| 166 |
+
"eval_runtime": 15.5322,
|
| 167 |
+
"eval_samples_per_second": 88.204,
|
| 168 |
+
"step": 5000
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 0.64,
|
| 172 |
+
"learning_rate": 0.0002359944140579541,
|
| 173 |
+
"loss": 0.3427,
|
| 174 |
+
"step": 5500
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.64,
|
| 178 |
+
"eval_average_metrics": 86.61678832116789,
|
| 179 |
+
"eval_em": 86.27737226277372,
|
| 180 |
+
"eval_f1": 86.95620437956205,
|
| 181 |
+
"eval_loss": 0.2410917580127716,
|
| 182 |
+
"eval_runtime": 15.3636,
|
| 183 |
+
"eval_samples_per_second": 89.172,
|
| 184 |
+
"step": 5500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.7,
|
| 188 |
+
"learning_rate": 0.00023017572442685905,
|
| 189 |
+
"loss": 0.349,
|
| 190 |
+
"step": 6000
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.7,
|
| 194 |
+
"eval_average_metrics": 88.17570385818561,
|
| 195 |
+
"eval_em": 87.8102189781022,
|
| 196 |
+
"eval_f1": 88.54118873826903,
|
| 197 |
+
"eval_loss": 0.23763984441757202,
|
| 198 |
+
"eval_runtime": 15.6242,
|
| 199 |
+
"eval_samples_per_second": 87.684,
|
| 200 |
+
"step": 6000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.76,
|
| 204 |
+
"learning_rate": 0.00022435703479576396,
|
| 205 |
+
"loss": 0.3404,
|
| 206 |
+
"step": 6500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.76,
|
| 210 |
+
"eval_average_metrics": 88.27128953771289,
|
| 211 |
+
"eval_em": 88.02919708029196,
|
| 212 |
+
"eval_f1": 88.51338199513381,
|
| 213 |
+
"eval_loss": 0.23219799995422363,
|
| 214 |
+
"eval_runtime": 15.7753,
|
| 215 |
+
"eval_samples_per_second": 86.845,
|
| 216 |
+
"step": 6500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 0.81,
|
| 220 |
+
"learning_rate": 0.0002185383451646689,
|
| 221 |
+
"loss": 0.3369,
|
| 222 |
+
"step": 7000
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.81,
|
| 226 |
+
"eval_average_metrics": 89.59054570733403,
|
| 227 |
+
"eval_em": 89.12408759124088,
|
| 228 |
+
"eval_f1": 90.05700382342718,
|
| 229 |
+
"eval_loss": 0.2249041497707367,
|
| 230 |
+
"eval_runtime": 15.3646,
|
| 231 |
+
"eval_samples_per_second": 89.166,
|
| 232 |
+
"step": 7000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 0.87,
|
| 236 |
+
"learning_rate": 0.00021271965553357382,
|
| 237 |
+
"loss": 0.3348,
|
| 238 |
+
"step": 7500
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 0.87,
|
| 242 |
+
"eval_average_metrics": 89.87104622871047,
|
| 243 |
+
"eval_em": 89.56204379562044,
|
| 244 |
+
"eval_f1": 90.1800486618005,
|
| 245 |
+
"eval_loss": 0.22438949346542358,
|
| 246 |
+
"eval_runtime": 15.6922,
|
| 247 |
+
"eval_samples_per_second": 87.305,
|
| 248 |
+
"step": 7500
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.93,
|
| 252 |
+
"learning_rate": 0.00020690096590247876,
|
| 253 |
+
"loss": 0.34,
|
| 254 |
+
"step": 8000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.93,
|
| 258 |
+
"eval_average_metrics": 87.99148418491484,
|
| 259 |
+
"eval_em": 87.66423357664233,
|
| 260 |
+
"eval_f1": 88.31873479318735,
|
| 261 |
+
"eval_loss": 0.23190169036388397,
|
| 262 |
+
"eval_runtime": 15.0876,
|
| 263 |
+
"eval_samples_per_second": 90.803,
|
| 264 |
+
"step": 8000
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.99,
|
| 268 |
+
"learning_rate": 0.00020108227627138367,
|
| 269 |
+
"loss": 0.3413,
|
| 270 |
+
"step": 8500
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 0.99,
|
| 274 |
+
"eval_average_metrics": 89.0498783454988,
|
| 275 |
+
"eval_em": 88.75912408759125,
|
| 276 |
+
"eval_f1": 89.34063260340633,
|
| 277 |
+
"eval_loss": 0.2290135771036148,
|
| 278 |
+
"eval_runtime": 15.6237,
|
| 279 |
+
"eval_samples_per_second": 87.688,
|
| 280 |
+
"step": 8500
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.05,
|
| 284 |
+
"learning_rate": 0.00019526358664028858,
|
| 285 |
+
"loss": 0.3207,
|
| 286 |
+
"step": 9000
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 1.05,
|
| 290 |
+
"eval_average_metrics": 89.54257907542579,
|
| 291 |
+
"eval_em": 89.34306569343066,
|
| 292 |
+
"eval_f1": 89.74209245742092,
|
| 293 |
+
"eval_loss": 0.22638055682182312,
|
| 294 |
+
"eval_runtime": 15.2056,
|
| 295 |
+
"eval_samples_per_second": 90.098,
|
| 296 |
+
"step": 9000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.11,
|
| 300 |
+
"learning_rate": 0.00018944489700919352,
|
| 301 |
+
"loss": 0.333,
|
| 302 |
+
"step": 9500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.11,
|
| 306 |
+
"eval_average_metrics": 89.5912408759124,
|
| 307 |
+
"eval_em": 89.34306569343066,
|
| 308 |
+
"eval_f1": 89.83941605839416,
|
| 309 |
+
"eval_loss": 0.22846655547618866,
|
| 310 |
+
"eval_runtime": 15.7388,
|
| 311 |
+
"eval_samples_per_second": 87.046,
|
| 312 |
+
"step": 9500
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 1.16,
|
| 316 |
+
"learning_rate": 0.00018362620737809844,
|
| 317 |
+
"loss": 0.3313,
|
| 318 |
+
"step": 10000
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.16,
|
| 322 |
+
"eval_average_metrics": 88.52485227667711,
|
| 323 |
+
"eval_em": 88.17518248175182,
|
| 324 |
+
"eval_f1": 88.87452207160239,
|
| 325 |
+
"eval_loss": 0.23135380446910858,
|
| 326 |
+
"eval_runtime": 15.5843,
|
| 327 |
+
"eval_samples_per_second": 87.909,
|
| 328 |
+
"step": 10000
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 1.22,
|
| 332 |
+
"learning_rate": 0.00017780751774700338,
|
| 333 |
+
"loss": 0.3326,
|
| 334 |
+
"step": 10500
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 1.22,
|
| 338 |
+
"eval_average_metrics": 89.61730969760167,
|
| 339 |
+
"eval_em": 89.12408759124088,
|
| 340 |
+
"eval_f1": 90.11053180396246,
|
| 341 |
+
"eval_loss": 0.22540691494941711,
|
| 342 |
+
"eval_runtime": 15.4318,
|
| 343 |
+
"eval_samples_per_second": 88.777,
|
| 344 |
+
"step": 10500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.28,
|
| 348 |
+
"learning_rate": 0.0001719888281159083,
|
| 349 |
+
"loss": 0.3349,
|
| 350 |
+
"step": 11000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.28,
|
| 354 |
+
"eval_average_metrics": 88.49808828640946,
|
| 355 |
+
"eval_em": 88.17518248175182,
|
| 356 |
+
"eval_f1": 88.82099409106708,
|
| 357 |
+
"eval_loss": 0.2274966984987259,
|
| 358 |
+
"eval_runtime": 15.7884,
|
| 359 |
+
"eval_samples_per_second": 86.773,
|
| 360 |
+
"step": 11000
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 1.34,
|
| 364 |
+
"learning_rate": 0.00016617013848481323,
|
| 365 |
+
"loss": 0.3268,
|
| 366 |
+
"step": 11500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.34,
|
| 370 |
+
"eval_average_metrics": 90.12043795620437,
|
| 371 |
+
"eval_em": 89.7080291970803,
|
| 372 |
+
"eval_f1": 90.53284671532846,
|
| 373 |
+
"eval_loss": 0.22236795723438263,
|
| 374 |
+
"eval_runtime": 15.5257,
|
| 375 |
+
"eval_samples_per_second": 88.241,
|
| 376 |
+
"step": 11500
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 1.4,
|
| 380 |
+
"learning_rate": 0.00016035144885371815,
|
| 381 |
+
"loss": 0.3315,
|
| 382 |
+
"step": 12000
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 1.4,
|
| 386 |
+
"eval_average_metrics": 88.9403892944039,
|
| 387 |
+
"eval_em": 88.54014598540147,
|
| 388 |
+
"eval_f1": 89.34063260340633,
|
| 389 |
+
"eval_loss": 0.22285930812358856,
|
| 390 |
+
"eval_runtime": 15.103,
|
| 391 |
+
"eval_samples_per_second": 90.71,
|
| 392 |
+
"step": 12000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 1.45,
|
| 396 |
+
"learning_rate": 0.00015453275922262303,
|
| 397 |
+
"loss": 0.3262,
|
| 398 |
+
"step": 12500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.45,
|
| 402 |
+
"eval_average_metrics": 86.66718804310045,
|
| 403 |
+
"eval_em": 86.27737226277372,
|
| 404 |
+
"eval_f1": 87.0570038234272,
|
| 405 |
+
"eval_loss": 0.23210552334785461,
|
| 406 |
+
"eval_runtime": 15.5946,
|
| 407 |
+
"eval_samples_per_second": 87.851,
|
| 408 |
+
"step": 12500
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.51,
|
| 412 |
+
"learning_rate": 0.00014871406959152797,
|
| 413 |
+
"loss": 0.3389,
|
| 414 |
+
"step": 13000
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.51,
|
| 418 |
+
"eval_average_metrics": 89.97445255474452,
|
| 419 |
+
"eval_em": 89.7080291970803,
|
| 420 |
+
"eval_f1": 90.24087591240875,
|
| 421 |
+
"eval_loss": 0.22476908564567566,
|
| 422 |
+
"eval_runtime": 15.3804,
|
| 423 |
+
"eval_samples_per_second": 89.075,
|
| 424 |
+
"step": 13000
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 1.57,
|
| 428 |
+
"learning_rate": 0.00014289537996043292,
|
| 429 |
+
"loss": 0.324,
|
| 430 |
+
"step": 13500
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.57,
|
| 434 |
+
"eval_average_metrics": 88.98470629127564,
|
| 435 |
+
"eval_em": 88.54014598540147,
|
| 436 |
+
"eval_f1": 89.42926659714982,
|
| 437 |
+
"eval_loss": 0.22203609347343445,
|
| 438 |
+
"eval_runtime": 15.4489,
|
| 439 |
+
"eval_samples_per_second": 88.679,
|
| 440 |
+
"step": 13500
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 1.63,
|
| 444 |
+
"learning_rate": 0.00013707669032933783,
|
| 445 |
+
"loss": 0.327,
|
| 446 |
+
"step": 14000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 1.63,
|
| 450 |
+
"eval_average_metrics": 89.79370872436566,
|
| 451 |
+
"eval_em": 89.41605839416059,
|
| 452 |
+
"eval_f1": 90.17135905457073,
|
| 453 |
+
"eval_loss": 0.22265847027301788,
|
| 454 |
+
"eval_runtime": 15.2397,
|
| 455 |
+
"eval_samples_per_second": 89.897,
|
| 456 |
+
"step": 14000
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 1.69,
|
| 460 |
+
"learning_rate": 0.00013125800069824274,
|
| 461 |
+
"loss": 0.3315,
|
| 462 |
+
"step": 14500
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 1.69,
|
| 466 |
+
"eval_average_metrics": 89.11678832116787,
|
| 467 |
+
"eval_em": 88.83211678832117,
|
| 468 |
+
"eval_f1": 89.40145985401458,
|
| 469 |
+
"eval_loss": 0.22402139008045197,
|
| 470 |
+
"eval_runtime": 15.9941,
|
| 471 |
+
"eval_samples_per_second": 85.657,
|
| 472 |
+
"step": 14500
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.75,
|
| 476 |
+
"learning_rate": 0.00012543931106714766,
|
| 477 |
+
"loss": 0.3236,
|
| 478 |
+
"step": 15000
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.75,
|
| 482 |
+
"eval_average_metrics": 89.00729927007299,
|
| 483 |
+
"eval_em": 88.6861313868613,
|
| 484 |
+
"eval_f1": 89.32846715328466,
|
| 485 |
+
"eval_loss": 0.22223517298698425,
|
| 486 |
+
"eval_runtime": 15.5255,
|
| 487 |
+
"eval_samples_per_second": 88.242,
|
| 488 |
+
"step": 15000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.8,
|
| 492 |
+
"learning_rate": 0.00011962062143605258,
|
| 493 |
+
"loss": 0.3214,
|
| 494 |
+
"step": 15500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.8,
|
| 498 |
+
"eval_average_metrics": 88.32603406326032,
|
| 499 |
+
"eval_em": 88.02919708029196,
|
| 500 |
+
"eval_f1": 88.6228710462287,
|
| 501 |
+
"eval_loss": 0.22402779757976532,
|
| 502 |
+
"eval_runtime": 15.9032,
|
| 503 |
+
"eval_samples_per_second": 86.146,
|
| 504 |
+
"step": 15500
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 1.86,
|
| 508 |
+
"learning_rate": 0.00011380193180495751,
|
| 509 |
+
"loss": 0.3254,
|
| 510 |
+
"step": 16000
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.86,
|
| 514 |
+
"eval_average_metrics": 89.72506082725062,
|
| 515 |
+
"eval_em": 89.48905109489051,
|
| 516 |
+
"eval_f1": 89.96107055961072,
|
| 517 |
+
"eval_loss": 0.21747739613056183,
|
| 518 |
+
"eval_runtime": 15.403,
|
| 519 |
+
"eval_samples_per_second": 88.944,
|
| 520 |
+
"step": 16000
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.92,
|
| 524 |
+
"learning_rate": 0.00010798324217386244,
|
| 525 |
+
"loss": 0.3216,
|
| 526 |
+
"step": 16500
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 1.92,
|
| 530 |
+
"eval_average_metrics": 90.24209245742091,
|
| 531 |
+
"eval_em": 90.0,
|
| 532 |
+
"eval_f1": 90.48418491484183,
|
| 533 |
+
"eval_loss": 0.21551567316055298,
|
| 534 |
+
"eval_runtime": 15.0664,
|
| 535 |
+
"eval_samples_per_second": 90.931,
|
| 536 |
+
"step": 16500
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.98,
|
| 540 |
+
"learning_rate": 0.00010216455254276737,
|
| 541 |
+
"loss": 0.327,
|
| 542 |
+
"step": 17000
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.98,
|
| 546 |
+
"eval_average_metrics": 89.62773722627736,
|
| 547 |
+
"eval_em": 89.34306569343066,
|
| 548 |
+
"eval_f1": 89.91240875912406,
|
| 549 |
+
"eval_loss": 0.22303573787212372,
|
| 550 |
+
"eval_runtime": 15.6955,
|
| 551 |
+
"eval_samples_per_second": 87.286,
|
| 552 |
+
"step": 17000
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 2.04,
|
| 556 |
+
"learning_rate": 9.634586291167229e-05,
|
| 557 |
+
"loss": 0.3216,
|
| 558 |
+
"step": 17500
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 2.04,
|
| 562 |
+
"eval_average_metrics": 89.6581508515815,
|
| 563 |
+
"eval_em": 89.34306569343066,
|
| 564 |
+
"eval_f1": 89.97323600973235,
|
| 565 |
+
"eval_loss": 0.21931786835193634,
|
| 566 |
+
"eval_runtime": 16.0777,
|
| 567 |
+
"eval_samples_per_second": 85.211,
|
| 568 |
+
"step": 17500
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 2.09,
|
| 572 |
+
"learning_rate": 9.052717328057722e-05,
|
| 573 |
+
"loss": 0.3166,
|
| 574 |
+
"step": 18000
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 2.09,
|
| 578 |
+
"eval_average_metrics": 88.78397636426834,
|
| 579 |
+
"eval_em": 88.39416058394161,
|
| 580 |
+
"eval_f1": 89.17379214459508,
|
| 581 |
+
"eval_loss": 0.2221771478652954,
|
| 582 |
+
"eval_runtime": 15.5735,
|
| 583 |
+
"eval_samples_per_second": 87.97,
|
| 584 |
+
"step": 18000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.15,
|
| 588 |
+
"learning_rate": 8.470848364948212e-05,
|
| 589 |
+
"loss": 0.321,
|
| 590 |
+
"step": 18500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.15,
|
| 594 |
+
"eval_average_metrics": 90.1508515815085,
|
| 595 |
+
"eval_em": 89.92700729927007,
|
| 596 |
+
"eval_f1": 90.37469586374695,
|
| 597 |
+
"eval_loss": 0.2213568538427353,
|
| 598 |
+
"eval_runtime": 15.5146,
|
| 599 |
+
"eval_samples_per_second": 88.304,
|
| 600 |
+
"step": 18500
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 2.21,
|
| 604 |
+
"learning_rate": 7.888979401838705e-05,
|
| 605 |
+
"loss": 0.3099,
|
| 606 |
+
"step": 19000
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 2.21,
|
| 610 |
+
"eval_average_metrics": 90.41240875912409,
|
| 611 |
+
"eval_em": 90.14598540145985,
|
| 612 |
+
"eval_f1": 90.67883211678833,
|
| 613 |
+
"eval_loss": 0.21670033037662506,
|
| 614 |
+
"eval_runtime": 15.2815,
|
| 615 |
+
"eval_samples_per_second": 89.651,
|
| 616 |
+
"step": 19000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 2.27,
|
| 620 |
+
"learning_rate": 7.307110438729197e-05,
|
| 621 |
+
"loss": 0.3178,
|
| 622 |
+
"step": 19500
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 2.27,
|
| 626 |
+
"eval_average_metrics": 89.41484184914842,
|
| 627 |
+
"eval_em": 89.19708029197079,
|
| 628 |
+
"eval_f1": 89.63260340632604,
|
| 629 |
+
"eval_loss": 0.21756400167942047,
|
| 630 |
+
"eval_runtime": 15.5668,
|
| 631 |
+
"eval_samples_per_second": 88.008,
|
| 632 |
+
"step": 19500
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.33,
|
| 636 |
+
"learning_rate": 6.72524147561969e-05,
|
| 637 |
+
"loss": 0.3082,
|
| 638 |
+
"step": 20000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.33,
|
| 642 |
+
"eval_average_metrics": 90.38807785888078,
|
| 643 |
+
"eval_em": 90.14598540145985,
|
| 644 |
+
"eval_f1": 90.6301703163017,
|
| 645 |
+
"eval_loss": 0.213778555393219,
|
| 646 |
+
"eval_runtime": 15.3113,
|
| 647 |
+
"eval_samples_per_second": 89.476,
|
| 648 |
+
"step": 20000
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 2.39,
|
| 652 |
+
"learning_rate": 6.143372512510183e-05,
|
| 653 |
+
"loss": 0.3185,
|
| 654 |
+
"step": 20500
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 2.39,
|
| 658 |
+
"eval_average_metrics": 89.44091067083768,
|
| 659 |
+
"eval_em": 89.05109489051095,
|
| 660 |
+
"eval_f1": 89.83072645116441,
|
| 661 |
+
"eval_loss": 0.21920597553253174,
|
| 662 |
+
"eval_runtime": 15.0292,
|
| 663 |
+
"eval_samples_per_second": 91.156,
|
| 664 |
+
"step": 20500
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 2.44,
|
| 668 |
+
"learning_rate": 5.561503549400675e-05,
|
| 669 |
+
"loss": 0.3214,
|
| 670 |
+
"step": 21000
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 2.44,
|
| 674 |
+
"eval_average_metrics": 89.58515815085158,
|
| 675 |
+
"eval_em": 89.27007299270073,
|
| 676 |
+
"eval_f1": 89.90024330900242,
|
| 677 |
+
"eval_loss": 0.21777711808681488,
|
| 678 |
+
"eval_runtime": 15.5974,
|
| 679 |
+
"eval_samples_per_second": 87.835,
|
| 680 |
+
"step": 21000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.5,
|
| 684 |
+
"learning_rate": 4.9796345862911663e-05,
|
| 685 |
+
"loss": 0.3204,
|
| 686 |
+
"step": 21500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.5,
|
| 690 |
+
"eval_average_metrics": 90.63138686131387,
|
| 691 |
+
"eval_em": 90.36496350364963,
|
| 692 |
+
"eval_f1": 90.89781021897811,
|
| 693 |
+
"eval_loss": 0.2141435146331787,
|
| 694 |
+
"eval_runtime": 15.7201,
|
| 695 |
+
"eval_samples_per_second": 87.149,
|
| 696 |
+
"step": 21500
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 2.56,
|
| 700 |
+
"learning_rate": 4.397765623181659e-05,
|
| 701 |
+
"loss": 0.3219,
|
| 702 |
+
"step": 22000
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 2.56,
|
| 706 |
+
"eval_average_metrics": 89.92579075425792,
|
| 707 |
+
"eval_em": 89.56204379562044,
|
| 708 |
+
"eval_f1": 90.28953771289538,
|
| 709 |
+
"eval_loss": 0.21593248844146729,
|
| 710 |
+
"eval_runtime": 15.5402,
|
| 711 |
+
"eval_samples_per_second": 88.158,
|
| 712 |
+
"step": 22000
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 2.62,
|
| 716 |
+
"learning_rate": 3.815896660072152e-05,
|
| 717 |
+
"loss": 0.3082,
|
| 718 |
+
"step": 22500
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 2.62,
|
| 722 |
+
"eval_average_metrics": 90.19343065693431,
|
| 723 |
+
"eval_em": 89.92700729927007,
|
| 724 |
+
"eval_f1": 90.45985401459853,
|
| 725 |
+
"eval_loss": 0.2151171863079071,
|
| 726 |
+
"eval_runtime": 15.4013,
|
| 727 |
+
"eval_samples_per_second": 88.953,
|
| 728 |
+
"step": 22500
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.68,
|
| 732 |
+
"learning_rate": 3.234027696962644e-05,
|
| 733 |
+
"loss": 0.3139,
|
| 734 |
+
"step": 23000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.68,
|
| 738 |
+
"eval_average_metrics": 90.58272506082724,
|
| 739 |
+
"eval_em": 90.2919708029197,
|
| 740 |
+
"eval_f1": 90.8734793187348,
|
| 741 |
+
"eval_loss": 0.21431832015514374,
|
| 742 |
+
"eval_runtime": 15.4496,
|
| 743 |
+
"eval_samples_per_second": 88.675,
|
| 744 |
+
"step": 23000
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.73,
|
| 748 |
+
"learning_rate": 2.6521587338531362e-05,
|
| 749 |
+
"loss": 0.3111,
|
| 750 |
+
"step": 23500
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 2.73,
|
| 754 |
+
"eval_average_metrics": 89.7433090024331,
|
| 755 |
+
"eval_em": 89.41605839416059,
|
| 756 |
+
"eval_f1": 90.0705596107056,
|
| 757 |
+
"eval_loss": 0.21385075151920319,
|
| 758 |
+
"eval_runtime": 15.7553,
|
| 759 |
+
"eval_samples_per_second": 86.955,
|
| 760 |
+
"step": 23500
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 2.79,
|
| 764 |
+
"learning_rate": 2.0702897707436283e-05,
|
| 765 |
+
"loss": 0.3168,
|
| 766 |
+
"step": 24000
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 2.79,
|
| 770 |
+
"eval_average_metrics": 89.81630170316302,
|
| 771 |
+
"eval_em": 89.56204379562044,
|
| 772 |
+
"eval_f1": 90.0705596107056,
|
| 773 |
+
"eval_loss": 0.2146376520395279,
|
| 774 |
+
"eval_runtime": 15.4091,
|
| 775 |
+
"eval_samples_per_second": 88.908,
|
| 776 |
+
"step": 24000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 2.85,
|
| 780 |
+
"learning_rate": 1.4884208076341207e-05,
|
| 781 |
+
"loss": 0.3115,
|
| 782 |
+
"step": 24500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 2.85,
|
| 786 |
+
"eval_average_metrics": 89.26885644768856,
|
| 787 |
+
"eval_em": 88.97810218978101,
|
| 788 |
+
"eval_f1": 89.5596107055961,
|
| 789 |
+
"eval_loss": 0.21739177405834198,
|
| 790 |
+
"eval_runtime": 15.7449,
|
| 791 |
+
"eval_samples_per_second": 87.012,
|
| 792 |
+
"step": 24500
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 2.91,
|
| 796 |
+
"learning_rate": 9.065518445246129e-06,
|
| 797 |
+
"loss": 0.3144,
|
| 798 |
+
"step": 25000
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 2.91,
|
| 802 |
+
"eval_average_metrics": 90.08394160583941,
|
| 803 |
+
"eval_em": 89.78102189781022,
|
| 804 |
+
"eval_f1": 90.38686131386861,
|
| 805 |
+
"eval_loss": 0.21506215631961823,
|
| 806 |
+
"eval_runtime": 15.5375,
|
| 807 |
+
"eval_samples_per_second": 88.174,
|
| 808 |
+
"step": 25000
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 2.97,
|
| 812 |
+
"learning_rate": 3.2468288141510528e-06,
|
| 813 |
+
"loss": 0.3119,
|
| 814 |
+
"step": 25500
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 2.97,
|
| 818 |
+
"eval_average_metrics": 90.10827250608273,
|
| 819 |
+
"eval_em": 89.85401459854015,
|
| 820 |
+
"eval_f1": 90.36253041362531,
|
| 821 |
+
"eval_loss": 0.21424712240695953,
|
| 822 |
+
"eval_runtime": 15.4741,
|
| 823 |
+
"eval_samples_per_second": 88.535,
|
| 824 |
+
"step": 25500
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 3.0,
|
| 828 |
+
"step": 25779,
|
| 829 |
+
"total_flos": 2.1011901579256954e+17,
|
| 830 |
+
"train_loss": 0.3297047741756515,
|
| 831 |
+
"train_runtime": 5659.5381,
|
| 832 |
+
"train_samples_per_second": 72.877,
|
| 833 |
+
"train_steps_per_second": 4.555
|
| 834 |
+
}
|
| 835 |
+
],
|
| 836 |
+
"max_steps": 25779,
|
| 837 |
+
"num_train_epochs": 3,
|
| 838 |
+
"total_flos": 2.1011901579256954e+17,
|
| 839 |
+
"trial_name": null,
|
| 840 |
+
"trial_params": null
|
| 841 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3183
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9fc0b2c802946c2d87bb004854bb82ca322e9997328e616e2973d3f37a9dacc
|
| 3 |
size 3183
|
training_config.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"dataset_config_name": ["en"], "delta_type": "
|
|
|
|
| 1 |
+
{"bottleneck_dim": 24, "dataset_config_name": ["en"], "delta_type": "adapter", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "superglue-record", "eval_steps": 200, "evaluation_strategy": "steps", "greater_is_better": true, "learning_rate": 0.0003, "load_best_model_at_end": true, "max_source_length": 512, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 3, "output_dir": "outputs/bitfit/t5-base/superglue-record", "overwrite_output_dir": true, "per_device_eval_batch_size": 16, "per_device_train_batch_size": 16, "predict_with_generate": true, "push_to_hub": true, "save_steps": 200, "save_strategy": "steps", "save_total_limit": 1, "seed": 42, "split_validation_test": true, "task_name": "superglue-record", "test_dataset_config_name": ["en"], "test_dataset_name": "superglue-record", "tokenizer_name": "../../../../plm_cache/t5-base", "unfrozen_modules": ["deltas", "layer_norm", "final_layer_norm"], "warmup_steps": 0}
|