Training in progress, step 8500
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 sha256:672917659646e43ec3c5bfdd29c6b49b10a73d6059f3cf909f92d4a916c96caf
|
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:f3985b34932049b4cf0916636d6f8dba717d09ea60b0e266b45e3ed15fa5a59c
|
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:b435baf134c80a8875f89ac2fcef60e437795aed8b32d8349a8bdc2ad663d117
|
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 sha256:848d14ef43c39f39761793dc949eeb32a79a1dc83a29edf9645f9f996cdb567b
|
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,
|
@@ -102,11 +102,17 @@
|
|
102 |
"learning_rate": 4.0909090909090915e-05,
|
103 |
"loss": 0.0355,
|
104 |
"step": 8000
|
|
|
|
|
|
|
|
|
|
|
|
|
105 |
}
|
106 |
],
|
107 |
"max_steps": 44000,
|
108 |
"num_train_epochs": 1,
|
109 |
-
"total_flos":
|
110 |
"trial_name": null,
|
111 |
"trial_params": null
|
112 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.11530426761442253,
|
5 |
+
"global_step": 8500,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
|
|
102 |
"learning_rate": 4.0909090909090915e-05,
|
103 |
"loss": 0.0355,
|
104 |
"step": 8000
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.12,
|
108 |
+
"learning_rate": 4.0342045454545456e-05,
|
109 |
+
"loss": 0.0382,
|
110 |
+
"step": 8500
|
111 |
}
|
112 |
],
|
113 |
"max_steps": 44000,
|
114 |
"num_train_epochs": 1,
|
115 |
+
"total_flos": 8883130966474752.0,
|
116 |
"trial_name": null,
|
117 |
"trial_params": null
|
118 |
}
|
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:f3985b34932049b4cf0916636d6f8dba717d09ea60b0e266b45e3ed15fa5a59c
|
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 sha256:55de3cb96334eb56f366afb7d44c1df942c6890a2c630c8bc180fab3b7802a09
|
3 |
+
size 48315
|