Training in progress, step 37000
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 1001681601
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b33b3be63cbd0335040a8c1bb6b3cb05f8c2e44b26569734d235d880530edae5
|
3 |
size 1001681601
|
last-checkpoint/pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 551185105
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f4d19fdd863446395dc91e11322e4ba2bd5e19c6839cdde13579bfa74843369c
|
3 |
size 551185105
|
last-checkpoint/scaler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 559
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f1412dc44808e0012038d6ef3250d9de2a8b53cc8a108758aa3627a04013a4e3
|
3 |
size 559
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0065743802f2e7e946829c7ac32f69a986cbf893237df4b371adb40695fa0498
|
3 |
size 623
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 0.
|
5 |
-
"global_step":
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
@@ -438,11 +438,23 @@
|
|
438 |
"learning_rate": 9.102272727272727e-06,
|
439 |
"loss": 0.0287,
|
440 |
"step": 36000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
441 |
}
|
442 |
],
|
443 |
"max_steps": 44000,
|
444 |
"num_train_epochs": 1,
|
445 |
-
"total_flos": 3.
|
446 |
"trial_name": null,
|
447 |
"trial_params": null
|
448 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.5019126943216039,
|
5 |
+
"global_step": 37000,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
|
|
438 |
"learning_rate": 9.102272727272727e-06,
|
439 |
"loss": 0.0287,
|
440 |
"step": 36000
|
441 |
+
},
|
442 |
+
{
|
443 |
+
"epoch": 0.5,
|
444 |
+
"learning_rate": 8.53409090909091e-06,
|
445 |
+
"loss": 0.0309,
|
446 |
+
"step": 36500
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 0.5,
|
450 |
+
"learning_rate": 7.965909090909092e-06,
|
451 |
+
"loss": 0.0295,
|
452 |
+
"step": 37000
|
453 |
}
|
454 |
],
|
455 |
"max_steps": 44000,
|
456 |
"num_train_epochs": 1,
|
457 |
+
"total_flos": 3.866074006369075e+16,
|
458 |
"trial_name": null,
|
459 |
"trial_params": null
|
460 |
}
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 551185105
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f4d19fdd863446395dc91e11322e4ba2bd5e19c6839cdde13579bfa74843369c
|
3 |
size 551185105
|
runs/Oct19_12-48-30_40567d3bb726/events.out.tfevents.1666183950.40567d3bb726.23.0
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 sha256:77669a6f63e715aacb00d48fd8091bcb5fd0026a1961151fbc16f7b53ea8bc56
|
3 |
+
size 57390
|