checkpoint step 800
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 100544848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:57030a464a2bbde8bdd6e7dd1ed3be522e7c8b805d1b3c611fab90dbde7f3614
|
| 3 |
size 100544848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 51712981
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee2d5d03b58094add6d0d27d82d89ae7198e669b91b0ad47bc1f421f94f6e84f
|
| 3 |
size 51712981
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14917
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:09522a0dbba1eba6f84e9fb31fdb88b7a12bfffe1724f542d3c6fa739922ca17
|
| 3 |
size 14917
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14917
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ebd8d581ded1e82b8a0b4c3297f16419013df0dc4e303f8514618777eeb2464b
|
| 3 |
size 14917
|
last-checkpoint/scaler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1383
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6fef122931c86c2d2736773be787da21ac6460d41580735381e953556fb410be
|
| 3 |
size 1383
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da29d1c16c6c0fd4634593f0dc46470b2b3f11215c045642d97ddd81263900ea
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,10 +1,10 @@
|
|
| 1 |
{
|
| 2 |
-
"best_global_step":
|
| 3 |
-
"best_metric": 0.
|
| 4 |
-
"best_model_checkpoint": "/kaggle/working/run/checkpoint-
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 200,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -4232,6 +4232,1414 @@
|
|
| 4232 |
"eval_samples_per_second": 6.154,
|
| 4233 |
"eval_steps_per_second": 0.769,
|
| 4234 |
"step": 600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4235 |
}
|
| 4236 |
],
|
| 4237 |
"logging_steps": 1,
|
|
@@ -4260,7 +5668,7 @@
|
|
| 4260 |
"attributes": {}
|
| 4261 |
}
|
| 4262 |
},
|
| 4263 |
-
"total_flos":
|
| 4264 |
"train_batch_size": 1,
|
| 4265 |
"trial_name": null,
|
| 4266 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_global_step": 800,
|
| 3 |
+
"best_metric": 0.47066086530685425,
|
| 4 |
+
"best_model_checkpoint": "/kaggle/working/run/checkpoint-800",
|
| 5 |
+
"epoch": 0.21908804600848966,
|
| 6 |
"eval_steps": 200,
|
| 7 |
+
"global_step": 800,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 4232 |
"eval_samples_per_second": 6.154,
|
| 4233 |
"eval_steps_per_second": 0.769,
|
| 4234 |
"step": 600
|
| 4235 |
+
},
|
| 4236 |
+
{
|
| 4237 |
+
"epoch": 0.16458989456387785,
|
| 4238 |
+
"grad_norm": 0.4925611913204193,
|
| 4239 |
+
"learning_rate": 4.767588251267121e-05,
|
| 4240 |
+
"loss": 0.4466903507709503,
|
| 4241 |
+
"step": 601
|
| 4242 |
+
},
|
| 4243 |
+
{
|
| 4244 |
+
"epoch": 0.16486375462138847,
|
| 4245 |
+
"grad_norm": 0.479430228471756,
|
| 4246 |
+
"learning_rate": 4.766653718854742e-05,
|
| 4247 |
+
"loss": 0.5347510576248169,
|
| 4248 |
+
"step": 602
|
| 4249 |
+
},
|
| 4250 |
+
{
|
| 4251 |
+
"epoch": 0.1651376146788991,
|
| 4252 |
+
"grad_norm": 0.4862011969089508,
|
| 4253 |
+
"learning_rate": 4.765717403292763e-05,
|
| 4254 |
+
"loss": 0.46406984329223633,
|
| 4255 |
+
"step": 603
|
| 4256 |
+
},
|
| 4257 |
+
{
|
| 4258 |
+
"epoch": 0.16541147473640969,
|
| 4259 |
+
"grad_norm": 0.45232805609703064,
|
| 4260 |
+
"learning_rate": 4.7647793053177734e-05,
|
| 4261 |
+
"loss": 0.5002888441085815,
|
| 4262 |
+
"step": 604
|
| 4263 |
+
},
|
| 4264 |
+
{
|
| 4265 |
+
"epoch": 0.1656853347939203,
|
| 4266 |
+
"grad_norm": 0.4939963221549988,
|
| 4267 |
+
"learning_rate": 4.763839425667763e-05,
|
| 4268 |
+
"loss": 0.5311185121536255,
|
| 4269 |
+
"step": 605
|
| 4270 |
+
},
|
| 4271 |
+
{
|
| 4272 |
+
"epoch": 0.16595919485143093,
|
| 4273 |
+
"grad_norm": 0.5195974707603455,
|
| 4274 |
+
"learning_rate": 4.762897765082124e-05,
|
| 4275 |
+
"loss": 0.4355415403842926,
|
| 4276 |
+
"step": 606
|
| 4277 |
+
},
|
| 4278 |
+
{
|
| 4279 |
+
"epoch": 0.16623305490894152,
|
| 4280 |
+
"grad_norm": 0.5340760946273804,
|
| 4281 |
+
"learning_rate": 4.7619543243016495e-05,
|
| 4282 |
+
"loss": 0.5772802829742432,
|
| 4283 |
+
"step": 607
|
| 4284 |
+
},
|
| 4285 |
+
{
|
| 4286 |
+
"epoch": 0.16650691496645215,
|
| 4287 |
+
"grad_norm": 0.4266352653503418,
|
| 4288 |
+
"learning_rate": 4.761009104068533e-05,
|
| 4289 |
+
"loss": 0.5621449947357178,
|
| 4290 |
+
"step": 608
|
| 4291 |
+
},
|
| 4292 |
+
{
|
| 4293 |
+
"epoch": 0.16678077502396277,
|
| 4294 |
+
"grad_norm": 0.4664492905139923,
|
| 4295 |
+
"learning_rate": 4.760062105126369e-05,
|
| 4296 |
+
"loss": 0.510725736618042,
|
| 4297 |
+
"step": 609
|
| 4298 |
+
},
|
| 4299 |
+
{
|
| 4300 |
+
"epoch": 0.16705463508147336,
|
| 4301 |
+
"grad_norm": 0.4611157774925232,
|
| 4302 |
+
"learning_rate": 4.7591133282201504e-05,
|
| 4303 |
+
"loss": 0.45114099979400635,
|
| 4304 |
+
"step": 610
|
| 4305 |
+
},
|
| 4306 |
+
{
|
| 4307 |
+
"epoch": 0.16732849513898398,
|
| 4308 |
+
"grad_norm": 0.5299663543701172,
|
| 4309 |
+
"learning_rate": 4.7581627740962676e-05,
|
| 4310 |
+
"loss": 0.601570725440979,
|
| 4311 |
+
"step": 611
|
| 4312 |
+
},
|
| 4313 |
+
{
|
| 4314 |
+
"epoch": 0.16760235519649458,
|
| 4315 |
+
"grad_norm": 0.49501556158065796,
|
| 4316 |
+
"learning_rate": 4.757210443502511e-05,
|
| 4317 |
+
"loss": 0.3980218768119812,
|
| 4318 |
+
"step": 612
|
| 4319 |
+
},
|
| 4320 |
+
{
|
| 4321 |
+
"epoch": 0.1678762152540052,
|
| 4322 |
+
"grad_norm": 0.4982666075229645,
|
| 4323 |
+
"learning_rate": 4.756256337188067e-05,
|
| 4324 |
+
"loss": 0.4965696930885315,
|
| 4325 |
+
"step": 613
|
| 4326 |
+
},
|
| 4327 |
+
{
|
| 4328 |
+
"epoch": 0.16815007531151582,
|
| 4329 |
+
"grad_norm": 0.46974438428878784,
|
| 4330 |
+
"learning_rate": 4.755300455903521e-05,
|
| 4331 |
+
"loss": 0.517380952835083,
|
| 4332 |
+
"step": 614
|
| 4333 |
+
},
|
| 4334 |
+
{
|
| 4335 |
+
"epoch": 0.16842393536902642,
|
| 4336 |
+
"grad_norm": 0.47566041350364685,
|
| 4337 |
+
"learning_rate": 4.754342800400852e-05,
|
| 4338 |
+
"loss": 0.42522090673446655,
|
| 4339 |
+
"step": 615
|
| 4340 |
+
},
|
| 4341 |
+
{
|
| 4342 |
+
"epoch": 0.16869779542653704,
|
| 4343 |
+
"grad_norm": 0.4294373691082001,
|
| 4344 |
+
"learning_rate": 4.753383371433437e-05,
|
| 4345 |
+
"loss": 0.40655940771102905,
|
| 4346 |
+
"step": 616
|
| 4347 |
+
},
|
| 4348 |
+
{
|
| 4349 |
+
"epoch": 0.16897165548404766,
|
| 4350 |
+
"grad_norm": 0.46585121750831604,
|
| 4351 |
+
"learning_rate": 4.752422169756048e-05,
|
| 4352 |
+
"loss": 0.5373576879501343,
|
| 4353 |
+
"step": 617
|
| 4354 |
+
},
|
| 4355 |
+
{
|
| 4356 |
+
"epoch": 0.16924551554155826,
|
| 4357 |
+
"grad_norm": 0.5104280710220337,
|
| 4358 |
+
"learning_rate": 4.75145919612485e-05,
|
| 4359 |
+
"loss": 0.37722110748291016,
|
| 4360 |
+
"step": 618
|
| 4361 |
+
},
|
| 4362 |
+
{
|
| 4363 |
+
"epoch": 0.16951937559906888,
|
| 4364 |
+
"grad_norm": 0.5375677347183228,
|
| 4365 |
+
"learning_rate": 4.7504944512974036e-05,
|
| 4366 |
+
"loss": 0.5337779521942139,
|
| 4367 |
+
"step": 619
|
| 4368 |
+
},
|
| 4369 |
+
{
|
| 4370 |
+
"epoch": 0.1697932356565795,
|
| 4371 |
+
"grad_norm": 0.5285770297050476,
|
| 4372 |
+
"learning_rate": 4.749527936032662e-05,
|
| 4373 |
+
"loss": 0.5558523535728455,
|
| 4374 |
+
"step": 620
|
| 4375 |
+
},
|
| 4376 |
+
{
|
| 4377 |
+
"epoch": 0.1700670957140901,
|
| 4378 |
+
"grad_norm": 0.493772953748703,
|
| 4379 |
+
"learning_rate": 4.74855965109097e-05,
|
| 4380 |
+
"loss": 0.42491260170936584,
|
| 4381 |
+
"step": 621
|
| 4382 |
+
},
|
| 4383 |
+
{
|
| 4384 |
+
"epoch": 0.17034095577160072,
|
| 4385 |
+
"grad_norm": 0.4404139220714569,
|
| 4386 |
+
"learning_rate": 4.747589597234068e-05,
|
| 4387 |
+
"loss": 0.37691378593444824,
|
| 4388 |
+
"step": 622
|
| 4389 |
+
},
|
| 4390 |
+
{
|
| 4391 |
+
"epoch": 0.1706148158291113,
|
| 4392 |
+
"grad_norm": 0.47126510739326477,
|
| 4393 |
+
"learning_rate": 4.746617775225084e-05,
|
| 4394 |
+
"loss": 0.525524377822876,
|
| 4395 |
+
"step": 623
|
| 4396 |
+
},
|
| 4397 |
+
{
|
| 4398 |
+
"epoch": 0.17088867588662193,
|
| 4399 |
+
"grad_norm": 0.4621577262878418,
|
| 4400 |
+
"learning_rate": 4.74564418582854e-05,
|
| 4401 |
+
"loss": 0.46691370010375977,
|
| 4402 |
+
"step": 624
|
| 4403 |
+
},
|
| 4404 |
+
{
|
| 4405 |
+
"epoch": 0.17116253594413255,
|
| 4406 |
+
"grad_norm": 0.46230247616767883,
|
| 4407 |
+
"learning_rate": 4.744668829810347e-05,
|
| 4408 |
+
"loss": 0.4473961591720581,
|
| 4409 |
+
"step": 625
|
| 4410 |
+
},
|
| 4411 |
+
{
|
| 4412 |
+
"epoch": 0.17143639600164315,
|
| 4413 |
+
"grad_norm": 0.5121751427650452,
|
| 4414 |
+
"learning_rate": 4.7436917079378066e-05,
|
| 4415 |
+
"loss": 0.44627809524536133,
|
| 4416 |
+
"step": 626
|
| 4417 |
+
},
|
| 4418 |
+
{
|
| 4419 |
+
"epoch": 0.17171025605915377,
|
| 4420 |
+
"grad_norm": 0.4475019872188568,
|
| 4421 |
+
"learning_rate": 4.742712820979607e-05,
|
| 4422 |
+
"loss": 0.42251938581466675,
|
| 4423 |
+
"step": 627
|
| 4424 |
+
},
|
| 4425 |
+
{
|
| 4426 |
+
"epoch": 0.1719841161166644,
|
| 4427 |
+
"grad_norm": 0.48283877968788147,
|
| 4428 |
+
"learning_rate": 4.741732169705829e-05,
|
| 4429 |
+
"loss": 0.4411700665950775,
|
| 4430 |
+
"step": 628
|
| 4431 |
+
},
|
| 4432 |
+
{
|
| 4433 |
+
"epoch": 0.172257976174175,
|
| 4434 |
+
"grad_norm": 0.5352658629417419,
|
| 4435 |
+
"learning_rate": 4.7407497548879384e-05,
|
| 4436 |
+
"loss": 0.4714464545249939,
|
| 4437 |
+
"step": 629
|
| 4438 |
+
},
|
| 4439 |
+
{
|
| 4440 |
+
"epoch": 0.1725318362316856,
|
| 4441 |
+
"grad_norm": 0.5437986254692078,
|
| 4442 |
+
"learning_rate": 4.7397655772987894e-05,
|
| 4443 |
+
"loss": 0.5160949230194092,
|
| 4444 |
+
"step": 630
|
| 4445 |
+
},
|
| 4446 |
+
{
|
| 4447 |
+
"epoch": 0.17280569628919623,
|
| 4448 |
+
"grad_norm": 0.5346359014511108,
|
| 4449 |
+
"learning_rate": 4.738779637712624e-05,
|
| 4450 |
+
"loss": 0.5248135328292847,
|
| 4451 |
+
"step": 631
|
| 4452 |
+
},
|
| 4453 |
+
{
|
| 4454 |
+
"epoch": 0.17307955634670683,
|
| 4455 |
+
"grad_norm": 0.5378316640853882,
|
| 4456 |
+
"learning_rate": 4.737791936905067e-05,
|
| 4457 |
+
"loss": 0.48193275928497314,
|
| 4458 |
+
"step": 632
|
| 4459 |
+
},
|
| 4460 |
+
{
|
| 4461 |
+
"epoch": 0.17335341640421745,
|
| 4462 |
+
"grad_norm": 0.5049542784690857,
|
| 4463 |
+
"learning_rate": 4.7368024756531316e-05,
|
| 4464 |
+
"loss": 0.5049288272857666,
|
| 4465 |
+
"step": 633
|
| 4466 |
+
},
|
| 4467 |
+
{
|
| 4468 |
+
"epoch": 0.17362727646172807,
|
| 4469 |
+
"grad_norm": 0.4460134506225586,
|
| 4470 |
+
"learning_rate": 4.7358112547352165e-05,
|
| 4471 |
+
"loss": 0.4830750524997711,
|
| 4472 |
+
"step": 634
|
| 4473 |
+
},
|
| 4474 |
+
{
|
| 4475 |
+
"epoch": 0.17390113651923866,
|
| 4476 |
+
"grad_norm": 0.5035218000411987,
|
| 4477 |
+
"learning_rate": 4.734818274931102e-05,
|
| 4478 |
+
"loss": 0.5668973922729492,
|
| 4479 |
+
"step": 635
|
| 4480 |
+
},
|
| 4481 |
+
{
|
| 4482 |
+
"epoch": 0.17417499657674929,
|
| 4483 |
+
"grad_norm": 0.5235883593559265,
|
| 4484 |
+
"learning_rate": 4.733823537021955e-05,
|
| 4485 |
+
"loss": 0.4753448963165283,
|
| 4486 |
+
"step": 636
|
| 4487 |
+
},
|
| 4488 |
+
{
|
| 4489 |
+
"epoch": 0.17444885663425988,
|
| 4490 |
+
"grad_norm": 0.5088533163070679,
|
| 4491 |
+
"learning_rate": 4.7328270417903224e-05,
|
| 4492 |
+
"loss": 0.566279411315918,
|
| 4493 |
+
"step": 637
|
| 4494 |
+
},
|
| 4495 |
+
{
|
| 4496 |
+
"epoch": 0.1747227166917705,
|
| 4497 |
+
"grad_norm": 0.48553016781806946,
|
| 4498 |
+
"learning_rate": 4.731828790020136e-05,
|
| 4499 |
+
"loss": 0.412465363740921,
|
| 4500 |
+
"step": 638
|
| 4501 |
+
},
|
| 4502 |
+
{
|
| 4503 |
+
"epoch": 0.17499657674928112,
|
| 4504 |
+
"grad_norm": 0.45916467905044556,
|
| 4505 |
+
"learning_rate": 4.730828782496709e-05,
|
| 4506 |
+
"loss": 0.44021153450012207,
|
| 4507 |
+
"step": 639
|
| 4508 |
+
},
|
| 4509 |
+
{
|
| 4510 |
+
"epoch": 0.17527043680679172,
|
| 4511 |
+
"grad_norm": 0.5229712128639221,
|
| 4512 |
+
"learning_rate": 4.729827020006735e-05,
|
| 4513 |
+
"loss": 0.5698124766349792,
|
| 4514 |
+
"step": 640
|
| 4515 |
+
},
|
| 4516 |
+
{
|
| 4517 |
+
"epoch": 0.17554429686430234,
|
| 4518 |
+
"grad_norm": 0.5298092365264893,
|
| 4519 |
+
"learning_rate": 4.7288235033382884e-05,
|
| 4520 |
+
"loss": 0.48957616090774536,
|
| 4521 |
+
"step": 641
|
| 4522 |
+
},
|
| 4523 |
+
{
|
| 4524 |
+
"epoch": 0.17581815692181296,
|
| 4525 |
+
"grad_norm": 0.48031046986579895,
|
| 4526 |
+
"learning_rate": 4.727818233280824e-05,
|
| 4527 |
+
"loss": 0.4863559305667877,
|
| 4528 |
+
"step": 642
|
| 4529 |
+
},
|
| 4530 |
+
{
|
| 4531 |
+
"epoch": 0.17609201697932356,
|
| 4532 |
+
"grad_norm": 0.464865505695343,
|
| 4533 |
+
"learning_rate": 4.726811210625176e-05,
|
| 4534 |
+
"loss": 0.3629401922225952,
|
| 4535 |
+
"step": 643
|
| 4536 |
+
},
|
| 4537 |
+
{
|
| 4538 |
+
"epoch": 0.17636587703683418,
|
| 4539 |
+
"grad_norm": 0.5245168805122375,
|
| 4540 |
+
"learning_rate": 4.725802436163558e-05,
|
| 4541 |
+
"loss": 0.45334750413894653,
|
| 4542 |
+
"step": 644
|
| 4543 |
+
},
|
| 4544 |
+
{
|
| 4545 |
+
"epoch": 0.1766397370943448,
|
| 4546 |
+
"grad_norm": 0.5137346982955933,
|
| 4547 |
+
"learning_rate": 4.724791910689559e-05,
|
| 4548 |
+
"loss": 0.4320746064186096,
|
| 4549 |
+
"step": 645
|
| 4550 |
+
},
|
| 4551 |
+
{
|
| 4552 |
+
"epoch": 0.1769135971518554,
|
| 4553 |
+
"grad_norm": 0.45216622948646545,
|
| 4554 |
+
"learning_rate": 4.723779634998149e-05,
|
| 4555 |
+
"loss": 0.4198233485221863,
|
| 4556 |
+
"step": 646
|
| 4557 |
+
},
|
| 4558 |
+
{
|
| 4559 |
+
"epoch": 0.17718745720936602,
|
| 4560 |
+
"grad_norm": 0.4594189524650574,
|
| 4561 |
+
"learning_rate": 4.722765609885672e-05,
|
| 4562 |
+
"loss": 0.42744123935699463,
|
| 4563 |
+
"step": 647
|
| 4564 |
+
},
|
| 4565 |
+
{
|
| 4566 |
+
"epoch": 0.17746131726687664,
|
| 4567 |
+
"grad_norm": 0.5114706754684448,
|
| 4568 |
+
"learning_rate": 4.721749836149851e-05,
|
| 4569 |
+
"loss": 0.5363755822181702,
|
| 4570 |
+
"step": 648
|
| 4571 |
+
},
|
| 4572 |
+
{
|
| 4573 |
+
"epoch": 0.17773517732438723,
|
| 4574 |
+
"grad_norm": 0.5669101476669312,
|
| 4575 |
+
"learning_rate": 4.720732314589783e-05,
|
| 4576 |
+
"loss": 0.5142309665679932,
|
| 4577 |
+
"step": 649
|
| 4578 |
+
},
|
| 4579 |
+
{
|
| 4580 |
+
"epoch": 0.17800903738189786,
|
| 4581 |
+
"grad_norm": 0.5608327984809875,
|
| 4582 |
+
"learning_rate": 4.719713046005938e-05,
|
| 4583 |
+
"loss": 0.49245917797088623,
|
| 4584 |
+
"step": 650
|
| 4585 |
+
},
|
| 4586 |
+
{
|
| 4587 |
+
"epoch": 0.17828289743940845,
|
| 4588 |
+
"grad_norm": 0.46120744943618774,
|
| 4589 |
+
"learning_rate": 4.7186920312001654e-05,
|
| 4590 |
+
"loss": 0.4803634285926819,
|
| 4591 |
+
"step": 651
|
| 4592 |
+
},
|
| 4593 |
+
{
|
| 4594 |
+
"epoch": 0.17855675749691907,
|
| 4595 |
+
"grad_norm": 0.4628528654575348,
|
| 4596 |
+
"learning_rate": 4.7176692709756835e-05,
|
| 4597 |
+
"loss": 0.4415169358253479,
|
| 4598 |
+
"step": 652
|
| 4599 |
+
},
|
| 4600 |
+
{
|
| 4601 |
+
"epoch": 0.1788306175544297,
|
| 4602 |
+
"grad_norm": 0.4840557873249054,
|
| 4603 |
+
"learning_rate": 4.7166447661370865e-05,
|
| 4604 |
+
"loss": 0.41893041133880615,
|
| 4605 |
+
"step": 653
|
| 4606 |
+
},
|
| 4607 |
+
{
|
| 4608 |
+
"epoch": 0.1791044776119403,
|
| 4609 |
+
"grad_norm": 0.5100110173225403,
|
| 4610 |
+
"learning_rate": 4.71561851749034e-05,
|
| 4611 |
+
"loss": 0.49889039993286133,
|
| 4612 |
+
"step": 654
|
| 4613 |
+
},
|
| 4614 |
+
{
|
| 4615 |
+
"epoch": 0.1793783376694509,
|
| 4616 |
+
"grad_norm": 0.46650663018226624,
|
| 4617 |
+
"learning_rate": 4.714590525842781e-05,
|
| 4618 |
+
"loss": 0.5377525687217712,
|
| 4619 |
+
"step": 655
|
| 4620 |
+
},
|
| 4621 |
+
{
|
| 4622 |
+
"epoch": 0.17965219772696153,
|
| 4623 |
+
"grad_norm": 0.46631065011024475,
|
| 4624 |
+
"learning_rate": 4.71356079200312e-05,
|
| 4625 |
+
"loss": 0.4596570134162903,
|
| 4626 |
+
"step": 656
|
| 4627 |
+
},
|
| 4628 |
+
{
|
| 4629 |
+
"epoch": 0.17992605778447213,
|
| 4630 |
+
"grad_norm": 0.46090149879455566,
|
| 4631 |
+
"learning_rate": 4.7125293167814345e-05,
|
| 4632 |
+
"loss": 0.485940158367157,
|
| 4633 |
+
"step": 657
|
| 4634 |
+
},
|
| 4635 |
+
{
|
| 4636 |
+
"epoch": 0.18019991784198275,
|
| 4637 |
+
"grad_norm": 0.4863240122795105,
|
| 4638 |
+
"learning_rate": 4.711496100989176e-05,
|
| 4639 |
+
"loss": 0.4165606200695038,
|
| 4640 |
+
"step": 658
|
| 4641 |
+
},
|
| 4642 |
+
{
|
| 4643 |
+
"epoch": 0.18047377789949337,
|
| 4644 |
+
"grad_norm": 0.6029651165008545,
|
| 4645 |
+
"learning_rate": 4.7104611454391604e-05,
|
| 4646 |
+
"loss": 0.5597526431083679,
|
| 4647 |
+
"step": 659
|
| 4648 |
+
},
|
| 4649 |
+
{
|
| 4650 |
+
"epoch": 0.18074763795700396,
|
| 4651 |
+
"grad_norm": 0.4787464439868927,
|
| 4652 |
+
"learning_rate": 4.7094244509455766e-05,
|
| 4653 |
+
"loss": 0.4288611114025116,
|
| 4654 |
+
"step": 660
|
| 4655 |
+
},
|
| 4656 |
+
{
|
| 4657 |
+
"epoch": 0.1810214980145146,
|
| 4658 |
+
"grad_norm": 0.48518380522727966,
|
| 4659 |
+
"learning_rate": 4.708386018323979e-05,
|
| 4660 |
+
"loss": 0.5409377813339233,
|
| 4661 |
+
"step": 661
|
| 4662 |
+
},
|
| 4663 |
+
{
|
| 4664 |
+
"epoch": 0.1812953580720252,
|
| 4665 |
+
"grad_norm": 0.47271570563316345,
|
| 4666 |
+
"learning_rate": 4.7073458483912926e-05,
|
| 4667 |
+
"loss": 0.37001827359199524,
|
| 4668 |
+
"step": 662
|
| 4669 |
+
},
|
| 4670 |
+
{
|
| 4671 |
+
"epoch": 0.1815692181295358,
|
| 4672 |
+
"grad_norm": 0.4767431616783142,
|
| 4673 |
+
"learning_rate": 4.7063039419658035e-05,
|
| 4674 |
+
"loss": 0.4279036819934845,
|
| 4675 |
+
"step": 663
|
| 4676 |
+
},
|
| 4677 |
+
{
|
| 4678 |
+
"epoch": 0.18184307818704643,
|
| 4679 |
+
"grad_norm": 0.4960452914237976,
|
| 4680 |
+
"learning_rate": 4.705260299867169e-05,
|
| 4681 |
+
"loss": 0.4352184534072876,
|
| 4682 |
+
"step": 664
|
| 4683 |
+
},
|
| 4684 |
+
{
|
| 4685 |
+
"epoch": 0.18211693824455702,
|
| 4686 |
+
"grad_norm": 0.4702984094619751,
|
| 4687 |
+
"learning_rate": 4.7042149229164095e-05,
|
| 4688 |
+
"loss": 0.47267651557922363,
|
| 4689 |
+
"step": 665
|
| 4690 |
+
},
|
| 4691 |
+
{
|
| 4692 |
+
"epoch": 0.18239079830206764,
|
| 4693 |
+
"grad_norm": 0.5084679126739502,
|
| 4694 |
+
"learning_rate": 4.7031678119359104e-05,
|
| 4695 |
+
"loss": 0.49344146251678467,
|
| 4696 |
+
"step": 666
|
| 4697 |
+
},
|
| 4698 |
+
{
|
| 4699 |
+
"epoch": 0.18266465835957826,
|
| 4700 |
+
"grad_norm": 0.5078561305999756,
|
| 4701 |
+
"learning_rate": 4.702118967749423e-05,
|
| 4702 |
+
"loss": 0.5684317350387573,
|
| 4703 |
+
"step": 667
|
| 4704 |
+
},
|
| 4705 |
+
{
|
| 4706 |
+
"epoch": 0.18293851841708886,
|
| 4707 |
+
"grad_norm": 0.5051935911178589,
|
| 4708 |
+
"learning_rate": 4.701068391182059e-05,
|
| 4709 |
+
"loss": 0.5679233074188232,
|
| 4710 |
+
"step": 668
|
| 4711 |
+
},
|
| 4712 |
+
{
|
| 4713 |
+
"epoch": 0.18321237847459948,
|
| 4714 |
+
"grad_norm": 0.5072962045669556,
|
| 4715 |
+
"learning_rate": 4.7000160830602936e-05,
|
| 4716 |
+
"loss": 0.4675022065639496,
|
| 4717 |
+
"step": 669
|
| 4718 |
+
},
|
| 4719 |
+
{
|
| 4720 |
+
"epoch": 0.1834862385321101,
|
| 4721 |
+
"grad_norm": 0.45282915234565735,
|
| 4722 |
+
"learning_rate": 4.698962044211968e-05,
|
| 4723 |
+
"loss": 0.5042353272438049,
|
| 4724 |
+
"step": 670
|
| 4725 |
+
},
|
| 4726 |
+
{
|
| 4727 |
+
"epoch": 0.1837600985896207,
|
| 4728 |
+
"grad_norm": 0.48417022824287415,
|
| 4729 |
+
"learning_rate": 4.697906275466279e-05,
|
| 4730 |
+
"loss": 0.41886505484580994,
|
| 4731 |
+
"step": 671
|
| 4732 |
+
},
|
| 4733 |
+
{
|
| 4734 |
+
"epoch": 0.18403395864713132,
|
| 4735 |
+
"grad_norm": 0.4952453374862671,
|
| 4736 |
+
"learning_rate": 4.6968487776537906e-05,
|
| 4737 |
+
"loss": 0.4531964659690857,
|
| 4738 |
+
"step": 672
|
| 4739 |
+
},
|
| 4740 |
+
{
|
| 4741 |
+
"epoch": 0.18430781870464194,
|
| 4742 |
+
"grad_norm": 0.5625466108322144,
|
| 4743 |
+
"learning_rate": 4.695789551606421e-05,
|
| 4744 |
+
"loss": 0.49242186546325684,
|
| 4745 |
+
"step": 673
|
| 4746 |
+
},
|
| 4747 |
+
{
|
| 4748 |
+
"epoch": 0.18458167876215253,
|
| 4749 |
+
"grad_norm": 0.5095588564872742,
|
| 4750 |
+
"learning_rate": 4.694728598157452e-05,
|
| 4751 |
+
"loss": 0.48174357414245605,
|
| 4752 |
+
"step": 674
|
| 4753 |
+
},
|
| 4754 |
+
{
|
| 4755 |
+
"epoch": 0.18485553881966316,
|
| 4756 |
+
"grad_norm": 0.48543933033943176,
|
| 4757 |
+
"learning_rate": 4.693665918141523e-05,
|
| 4758 |
+
"loss": 0.4756564199924469,
|
| 4759 |
+
"step": 675
|
| 4760 |
+
},
|
| 4761 |
+
{
|
| 4762 |
+
"epoch": 0.18512939887717375,
|
| 4763 |
+
"grad_norm": 0.5054110884666443,
|
| 4764 |
+
"learning_rate": 4.692601512394631e-05,
|
| 4765 |
+
"loss": 0.5012005567550659,
|
| 4766 |
+
"step": 676
|
| 4767 |
+
},
|
| 4768 |
+
{
|
| 4769 |
+
"epoch": 0.18540325893468437,
|
| 4770 |
+
"grad_norm": 0.4390055537223816,
|
| 4771 |
+
"learning_rate": 4.691535381754133e-05,
|
| 4772 |
+
"loss": 0.4927503764629364,
|
| 4773 |
+
"step": 677
|
| 4774 |
+
},
|
| 4775 |
+
{
|
| 4776 |
+
"epoch": 0.185677118992195,
|
| 4777 |
+
"grad_norm": 0.4409768283367157,
|
| 4778 |
+
"learning_rate": 4.69046752705874e-05,
|
| 4779 |
+
"loss": 0.4275025725364685,
|
| 4780 |
+
"step": 678
|
| 4781 |
+
},
|
| 4782 |
+
{
|
| 4783 |
+
"epoch": 0.1859509790497056,
|
| 4784 |
+
"grad_norm": 0.48191842436790466,
|
| 4785 |
+
"learning_rate": 4.689397949148521e-05,
|
| 4786 |
+
"loss": 0.45680519938468933,
|
| 4787 |
+
"step": 679
|
| 4788 |
+
},
|
| 4789 |
+
{
|
| 4790 |
+
"epoch": 0.1862248391072162,
|
| 4791 |
+
"grad_norm": 0.5262720584869385,
|
| 4792 |
+
"learning_rate": 4.6883266488649004e-05,
|
| 4793 |
+
"loss": 0.4728527069091797,
|
| 4794 |
+
"step": 680
|
| 4795 |
+
},
|
| 4796 |
+
{
|
| 4797 |
+
"epoch": 0.18649869916472683,
|
| 4798 |
+
"grad_norm": 0.4898194670677185,
|
| 4799 |
+
"learning_rate": 4.687253627050657e-05,
|
| 4800 |
+
"loss": 0.49217599630355835,
|
| 4801 |
+
"step": 681
|
| 4802 |
+
},
|
| 4803 |
+
{
|
| 4804 |
+
"epoch": 0.18677255922223743,
|
| 4805 |
+
"grad_norm": 0.4904230833053589,
|
| 4806 |
+
"learning_rate": 4.686178884549925e-05,
|
| 4807 |
+
"loss": 0.49807068705558777,
|
| 4808 |
+
"step": 682
|
| 4809 |
+
},
|
| 4810 |
+
{
|
| 4811 |
+
"epoch": 0.18704641927974805,
|
| 4812 |
+
"grad_norm": 0.5072822570800781,
|
| 4813 |
+
"learning_rate": 4.685102422208191e-05,
|
| 4814 |
+
"loss": 0.4758497178554535,
|
| 4815 |
+
"step": 683
|
| 4816 |
+
},
|
| 4817 |
+
{
|
| 4818 |
+
"epoch": 0.18732027933725867,
|
| 4819 |
+
"grad_norm": 0.49742990732192993,
|
| 4820 |
+
"learning_rate": 4.6840242408722956e-05,
|
| 4821 |
+
"loss": 0.5167051553726196,
|
| 4822 |
+
"step": 684
|
| 4823 |
+
},
|
| 4824 |
+
{
|
| 4825 |
+
"epoch": 0.18759413939476927,
|
| 4826 |
+
"grad_norm": 0.5105602145195007,
|
| 4827 |
+
"learning_rate": 4.68294434139043e-05,
|
| 4828 |
+
"loss": 0.41443032026290894,
|
| 4829 |
+
"step": 685
|
| 4830 |
+
},
|
| 4831 |
+
{
|
| 4832 |
+
"epoch": 0.1878679994522799,
|
| 4833 |
+
"grad_norm": 0.48594075441360474,
|
| 4834 |
+
"learning_rate": 4.681862724612141e-05,
|
| 4835 |
+
"loss": 0.466337114572525,
|
| 4836 |
+
"step": 686
|
| 4837 |
+
},
|
| 4838 |
+
{
|
| 4839 |
+
"epoch": 0.1881418595097905,
|
| 4840 |
+
"grad_norm": 0.487560510635376,
|
| 4841 |
+
"learning_rate": 4.680779391388321e-05,
|
| 4842 |
+
"loss": 0.46644315123558044,
|
| 4843 |
+
"step": 687
|
| 4844 |
+
},
|
| 4845 |
+
{
|
| 4846 |
+
"epoch": 0.1884157195673011,
|
| 4847 |
+
"grad_norm": 0.5286445021629333,
|
| 4848 |
+
"learning_rate": 4.679694342571216e-05,
|
| 4849 |
+
"loss": 0.506851851940155,
|
| 4850 |
+
"step": 688
|
| 4851 |
+
},
|
| 4852 |
+
{
|
| 4853 |
+
"epoch": 0.18868957962481173,
|
| 4854 |
+
"grad_norm": 0.4647861123085022,
|
| 4855 |
+
"learning_rate": 4.678607579014422e-05,
|
| 4856 |
+
"loss": 0.461232453584671,
|
| 4857 |
+
"step": 689
|
| 4858 |
+
},
|
| 4859 |
+
{
|
| 4860 |
+
"epoch": 0.18896343968232232,
|
| 4861 |
+
"grad_norm": 0.4583735764026642,
|
| 4862 |
+
"learning_rate": 4.6775191015728834e-05,
|
| 4863 |
+
"loss": 0.46333104372024536,
|
| 4864 |
+
"step": 690
|
| 4865 |
+
},
|
| 4866 |
+
{
|
| 4867 |
+
"epoch": 0.18923729973983294,
|
| 4868 |
+
"grad_norm": 0.4145049452781677,
|
| 4869 |
+
"learning_rate": 4.67642891110289e-05,
|
| 4870 |
+
"loss": 0.3542670011520386,
|
| 4871 |
+
"step": 691
|
| 4872 |
+
},
|
| 4873 |
+
{
|
| 4874 |
+
"epoch": 0.18951115979734356,
|
| 4875 |
+
"grad_norm": 0.4818451702594757,
|
| 4876 |
+
"learning_rate": 4.6753370084620844e-05,
|
| 4877 |
+
"loss": 0.4092712998390198,
|
| 4878 |
+
"step": 692
|
| 4879 |
+
},
|
| 4880 |
+
{
|
| 4881 |
+
"epoch": 0.18978501985485416,
|
| 4882 |
+
"grad_norm": 0.45267534255981445,
|
| 4883 |
+
"learning_rate": 4.674243394509452e-05,
|
| 4884 |
+
"loss": 0.4984908103942871,
|
| 4885 |
+
"step": 693
|
| 4886 |
+
},
|
| 4887 |
+
{
|
| 4888 |
+
"epoch": 0.19005887991236478,
|
| 4889 |
+
"grad_norm": 0.5062838196754456,
|
| 4890 |
+
"learning_rate": 4.6731480701053266e-05,
|
| 4891 |
+
"loss": 0.5047342777252197,
|
| 4892 |
+
"step": 694
|
| 4893 |
+
},
|
| 4894 |
+
{
|
| 4895 |
+
"epoch": 0.1903327399698754,
|
| 4896 |
+
"grad_norm": 0.49547243118286133,
|
| 4897 |
+
"learning_rate": 4.6720510361113864e-05,
|
| 4898 |
+
"loss": 0.48536020517349243,
|
| 4899 |
+
"step": 695
|
| 4900 |
+
},
|
| 4901 |
+
{
|
| 4902 |
+
"epoch": 0.190606600027386,
|
| 4903 |
+
"grad_norm": 0.5105172991752625,
|
| 4904 |
+
"learning_rate": 4.670952293390655e-05,
|
| 4905 |
+
"loss": 0.5215955376625061,
|
| 4906 |
+
"step": 696
|
| 4907 |
+
},
|
| 4908 |
+
{
|
| 4909 |
+
"epoch": 0.19088046008489662,
|
| 4910 |
+
"grad_norm": 0.513580322265625,
|
| 4911 |
+
"learning_rate": 4.6698518428075004e-05,
|
| 4912 |
+
"loss": 0.45753979682922363,
|
| 4913 |
+
"step": 697
|
| 4914 |
+
},
|
| 4915 |
+
{
|
| 4916 |
+
"epoch": 0.19115432014240724,
|
| 4917 |
+
"grad_norm": 0.49127158522605896,
|
| 4918 |
+
"learning_rate": 4.6687496852276344e-05,
|
| 4919 |
+
"loss": 0.4460606575012207,
|
| 4920 |
+
"step": 698
|
| 4921 |
+
},
|
| 4922 |
+
{
|
| 4923 |
+
"epoch": 0.19142818019991784,
|
| 4924 |
+
"grad_norm": 0.4289352297782898,
|
| 4925 |
+
"learning_rate": 4.667645821518111e-05,
|
| 4926 |
+
"loss": 0.44201815128326416,
|
| 4927 |
+
"step": 699
|
| 4928 |
+
},
|
| 4929 |
+
{
|
| 4930 |
+
"epoch": 0.19170204025742846,
|
| 4931 |
+
"grad_norm": 0.4497436583042145,
|
| 4932 |
+
"learning_rate": 4.666540252547327e-05,
|
| 4933 |
+
"loss": 0.4204838275909424,
|
| 4934 |
+
"step": 700
|
| 4935 |
+
},
|
| 4936 |
+
{
|
| 4937 |
+
"epoch": 0.19197590031493908,
|
| 4938 |
+
"grad_norm": 0.4641953706741333,
|
| 4939 |
+
"learning_rate": 4.665432979185019e-05,
|
| 4940 |
+
"loss": 0.45025739073753357,
|
| 4941 |
+
"step": 701
|
| 4942 |
+
},
|
| 4943 |
+
{
|
| 4944 |
+
"epoch": 0.19224976037244967,
|
| 4945 |
+
"grad_norm": 0.4756088852882385,
|
| 4946 |
+
"learning_rate": 4.6643240023022684e-05,
|
| 4947 |
+
"loss": 0.40468281507492065,
|
| 4948 |
+
"step": 702
|
| 4949 |
+
},
|
| 4950 |
+
{
|
| 4951 |
+
"epoch": 0.1925236204299603,
|
| 4952 |
+
"grad_norm": 0.5222814679145813,
|
| 4953 |
+
"learning_rate": 4.6632133227714915e-05,
|
| 4954 |
+
"loss": 0.5164584517478943,
|
| 4955 |
+
"step": 703
|
| 4956 |
+
},
|
| 4957 |
+
{
|
| 4958 |
+
"epoch": 0.1927974804874709,
|
| 4959 |
+
"grad_norm": 0.504379391670227,
|
| 4960 |
+
"learning_rate": 4.662100941466448e-05,
|
| 4961 |
+
"loss": 0.44075268507003784,
|
| 4962 |
+
"step": 704
|
| 4963 |
+
},
|
| 4964 |
+
{
|
| 4965 |
+
"epoch": 0.1930713405449815,
|
| 4966 |
+
"grad_norm": 0.4763759672641754,
|
| 4967 |
+
"learning_rate": 4.660986859262236e-05,
|
| 4968 |
+
"loss": 0.5072387456893921,
|
| 4969 |
+
"step": 705
|
| 4970 |
+
},
|
| 4971 |
+
{
|
| 4972 |
+
"epoch": 0.19334520060249213,
|
| 4973 |
+
"grad_norm": 0.4972545802593231,
|
| 4974 |
+
"learning_rate": 4.6598710770352897e-05,
|
| 4975 |
+
"loss": 0.4284913241863251,
|
| 4976 |
+
"step": 706
|
| 4977 |
+
},
|
| 4978 |
+
{
|
| 4979 |
+
"epoch": 0.19361906066000273,
|
| 4980 |
+
"grad_norm": 0.4820251762866974,
|
| 4981 |
+
"learning_rate": 4.65875359566338e-05,
|
| 4982 |
+
"loss": 0.44016969203948975,
|
| 4983 |
+
"step": 707
|
| 4984 |
+
},
|
| 4985 |
+
{
|
| 4986 |
+
"epoch": 0.19389292071751335,
|
| 4987 |
+
"grad_norm": 0.496611624956131,
|
| 4988 |
+
"learning_rate": 4.65763441602562e-05,
|
| 4989 |
+
"loss": 0.5227499008178711,
|
| 4990 |
+
"step": 708
|
| 4991 |
+
},
|
| 4992 |
+
{
|
| 4993 |
+
"epoch": 0.19416678077502397,
|
| 4994 |
+
"grad_norm": 0.5342353582382202,
|
| 4995 |
+
"learning_rate": 4.6565135390024515e-05,
|
| 4996 |
+
"loss": 0.4097047448158264,
|
| 4997 |
+
"step": 709
|
| 4998 |
+
},
|
| 4999 |
+
{
|
| 5000 |
+
"epoch": 0.19444064083253457,
|
| 5001 |
+
"grad_norm": 0.4659307301044464,
|
| 5002 |
+
"learning_rate": 4.6553909654756576e-05,
|
| 5003 |
+
"loss": 0.4587275981903076,
|
| 5004 |
+
"step": 710
|
| 5005 |
+
},
|
| 5006 |
+
{
|
| 5007 |
+
"epoch": 0.1947145008900452,
|
| 5008 |
+
"grad_norm": 0.4939675033092499,
|
| 5009 |
+
"learning_rate": 4.6542666963283535e-05,
|
| 5010 |
+
"loss": 0.4996609091758728,
|
| 5011 |
+
"step": 711
|
| 5012 |
+
},
|
| 5013 |
+
{
|
| 5014 |
+
"epoch": 0.1949883609475558,
|
| 5015 |
+
"grad_norm": 0.47300779819488525,
|
| 5016 |
+
"learning_rate": 4.6531407324449864e-05,
|
| 5017 |
+
"loss": 0.43784981966018677,
|
| 5018 |
+
"step": 712
|
| 5019 |
+
},
|
| 5020 |
+
{
|
| 5021 |
+
"epoch": 0.1952622210050664,
|
| 5022 |
+
"grad_norm": 0.46679016947746277,
|
| 5023 |
+
"learning_rate": 4.65201307471134e-05,
|
| 5024 |
+
"loss": 0.4550427794456482,
|
| 5025 |
+
"step": 713
|
| 5026 |
+
},
|
| 5027 |
+
{
|
| 5028 |
+
"epoch": 0.19553608106257703,
|
| 5029 |
+
"grad_norm": 0.5156670212745667,
|
| 5030 |
+
"learning_rate": 4.65088372401453e-05,
|
| 5031 |
+
"loss": 0.4418240785598755,
|
| 5032 |
+
"step": 714
|
| 5033 |
+
},
|
| 5034 |
+
{
|
| 5035 |
+
"epoch": 0.19580994112008762,
|
| 5036 |
+
"grad_norm": 0.5111796259880066,
|
| 5037 |
+
"learning_rate": 4.6497526812430016e-05,
|
| 5038 |
+
"loss": 0.5044018030166626,
|
| 5039 |
+
"step": 715
|
| 5040 |
+
},
|
| 5041 |
+
{
|
| 5042 |
+
"epoch": 0.19608380117759824,
|
| 5043 |
+
"grad_norm": 0.46974053978919983,
|
| 5044 |
+
"learning_rate": 4.6486199472865344e-05,
|
| 5045 |
+
"loss": 0.49730271100997925,
|
| 5046 |
+
"step": 716
|
| 5047 |
+
},
|
| 5048 |
+
{
|
| 5049 |
+
"epoch": 0.19635766123510887,
|
| 5050 |
+
"grad_norm": 0.47753018140792847,
|
| 5051 |
+
"learning_rate": 4.6474855230362355e-05,
|
| 5052 |
+
"loss": 0.5762851238250732,
|
| 5053 |
+
"step": 717
|
| 5054 |
+
},
|
| 5055 |
+
{
|
| 5056 |
+
"epoch": 0.19663152129261946,
|
| 5057 |
+
"grad_norm": 0.47892236709594727,
|
| 5058 |
+
"learning_rate": 4.6463494093845427e-05,
|
| 5059 |
+
"loss": 0.541034460067749,
|
| 5060 |
+
"step": 718
|
| 5061 |
+
},
|
| 5062 |
+
{
|
| 5063 |
+
"epoch": 0.19690538135013008,
|
| 5064 |
+
"grad_norm": 0.4789484739303589,
|
| 5065 |
+
"learning_rate": 4.645211607225225e-05,
|
| 5066 |
+
"loss": 0.5369265675544739,
|
| 5067 |
+
"step": 719
|
| 5068 |
+
},
|
| 5069 |
+
{
|
| 5070 |
+
"epoch": 0.1971792414076407,
|
| 5071 |
+
"grad_norm": 0.4913516938686371,
|
| 5072 |
+
"learning_rate": 4.644072117453376e-05,
|
| 5073 |
+
"loss": 0.5272049307823181,
|
| 5074 |
+
"step": 720
|
| 5075 |
+
},
|
| 5076 |
+
{
|
| 5077 |
+
"epoch": 0.1974531014651513,
|
| 5078 |
+
"grad_norm": 0.4601913392543793,
|
| 5079 |
+
"learning_rate": 4.64293094096542e-05,
|
| 5080 |
+
"loss": 0.39497241377830505,
|
| 5081 |
+
"step": 721
|
| 5082 |
+
},
|
| 5083 |
+
{
|
| 5084 |
+
"epoch": 0.19772696152266192,
|
| 5085 |
+
"grad_norm": 0.4708000421524048,
|
| 5086 |
+
"learning_rate": 4.641788078659106e-05,
|
| 5087 |
+
"loss": 0.4594809412956238,
|
| 5088 |
+
"step": 722
|
| 5089 |
+
},
|
| 5090 |
+
{
|
| 5091 |
+
"epoch": 0.19800082158017254,
|
| 5092 |
+
"grad_norm": 0.5330820679664612,
|
| 5093 |
+
"learning_rate": 4.640643531433512e-05,
|
| 5094 |
+
"loss": 0.5015921592712402,
|
| 5095 |
+
"step": 723
|
| 5096 |
+
},
|
| 5097 |
+
{
|
| 5098 |
+
"epoch": 0.19827468163768314,
|
| 5099 |
+
"grad_norm": 0.4795275628566742,
|
| 5100 |
+
"learning_rate": 4.639497300189037e-05,
|
| 5101 |
+
"loss": 0.4719768762588501,
|
| 5102 |
+
"step": 724
|
| 5103 |
+
},
|
| 5104 |
+
{
|
| 5105 |
+
"epoch": 0.19854854169519376,
|
| 5106 |
+
"grad_norm": 0.5057756900787354,
|
| 5107 |
+
"learning_rate": 4.63834938582741e-05,
|
| 5108 |
+
"loss": 0.4283849000930786,
|
| 5109 |
+
"step": 725
|
| 5110 |
+
},
|
| 5111 |
+
{
|
| 5112 |
+
"epoch": 0.19882240175270438,
|
| 5113 |
+
"grad_norm": 0.49866625666618347,
|
| 5114 |
+
"learning_rate": 4.6371997892516795e-05,
|
| 5115 |
+
"loss": 0.4195783734321594,
|
| 5116 |
+
"step": 726
|
| 5117 |
+
},
|
| 5118 |
+
{
|
| 5119 |
+
"epoch": 0.19909626181021498,
|
| 5120 |
+
"grad_norm": 0.5645210146903992,
|
| 5121 |
+
"learning_rate": 4.6360485113662216e-05,
|
| 5122 |
+
"loss": 0.6437886953353882,
|
| 5123 |
+
"step": 727
|
| 5124 |
+
},
|
| 5125 |
+
{
|
| 5126 |
+
"epoch": 0.1993701218677256,
|
| 5127 |
+
"grad_norm": 0.5750858187675476,
|
| 5128 |
+
"learning_rate": 4.6348955530767314e-05,
|
| 5129 |
+
"loss": 0.5785911679267883,
|
| 5130 |
+
"step": 728
|
| 5131 |
+
},
|
| 5132 |
+
{
|
| 5133 |
+
"epoch": 0.1996439819252362,
|
| 5134 |
+
"grad_norm": 0.495633989572525,
|
| 5135 |
+
"learning_rate": 4.633740915290229e-05,
|
| 5136 |
+
"loss": 0.4707483947277069,
|
| 5137 |
+
"step": 729
|
| 5138 |
+
},
|
| 5139 |
+
{
|
| 5140 |
+
"epoch": 0.1999178419827468,
|
| 5141 |
+
"grad_norm": 0.521964430809021,
|
| 5142 |
+
"learning_rate": 4.6325845989150526e-05,
|
| 5143 |
+
"loss": 0.46005406975746155,
|
| 5144 |
+
"step": 730
|
| 5145 |
+
},
|
| 5146 |
+
{
|
| 5147 |
+
"epoch": 0.20019170204025744,
|
| 5148 |
+
"grad_norm": 0.47659438848495483,
|
| 5149 |
+
"learning_rate": 4.631426604860864e-05,
|
| 5150 |
+
"loss": 0.46020495891571045,
|
| 5151 |
+
"step": 731
|
| 5152 |
+
},
|
| 5153 |
+
{
|
| 5154 |
+
"epoch": 0.20046556209776803,
|
| 5155 |
+
"grad_norm": 0.4953257739543915,
|
| 5156 |
+
"learning_rate": 4.630266934038642e-05,
|
| 5157 |
+
"loss": 0.45418301224708557,
|
| 5158 |
+
"step": 732
|
| 5159 |
+
},
|
| 5160 |
+
{
|
| 5161 |
+
"epoch": 0.20073942215527865,
|
| 5162 |
+
"grad_norm": 0.5367743968963623,
|
| 5163 |
+
"learning_rate": 4.6291055873606856e-05,
|
| 5164 |
+
"loss": 0.5839319825172424,
|
| 5165 |
+
"step": 733
|
| 5166 |
+
},
|
| 5167 |
+
{
|
| 5168 |
+
"epoch": 0.20101328221278927,
|
| 5169 |
+
"grad_norm": 0.499185711145401,
|
| 5170 |
+
"learning_rate": 4.627942565740615e-05,
|
| 5171 |
+
"loss": 0.47755926847457886,
|
| 5172 |
+
"step": 734
|
| 5173 |
+
},
|
| 5174 |
+
{
|
| 5175 |
+
"epoch": 0.20128714227029987,
|
| 5176 |
+
"grad_norm": 0.511204719543457,
|
| 5177 |
+
"learning_rate": 4.626777870093364e-05,
|
| 5178 |
+
"loss": 0.4478704333305359,
|
| 5179 |
+
"step": 735
|
| 5180 |
+
},
|
| 5181 |
+
{
|
| 5182 |
+
"epoch": 0.2015610023278105,
|
| 5183 |
+
"grad_norm": 0.5286024808883667,
|
| 5184 |
+
"learning_rate": 4.625611501335184e-05,
|
| 5185 |
+
"loss": 0.5261520743370056,
|
| 5186 |
+
"step": 736
|
| 5187 |
+
},
|
| 5188 |
+
{
|
| 5189 |
+
"epoch": 0.2018348623853211,
|
| 5190 |
+
"grad_norm": 0.45660191774368286,
|
| 5191 |
+
"learning_rate": 4.624443460383645e-05,
|
| 5192 |
+
"loss": 0.528244137763977,
|
| 5193 |
+
"step": 737
|
| 5194 |
+
},
|
| 5195 |
+
{
|
| 5196 |
+
"epoch": 0.2021087224428317,
|
| 5197 |
+
"grad_norm": 0.5103270411491394,
|
| 5198 |
+
"learning_rate": 4.6232737481576296e-05,
|
| 5199 |
+
"loss": 0.4677639901638031,
|
| 5200 |
+
"step": 738
|
| 5201 |
+
},
|
| 5202 |
+
{
|
| 5203 |
+
"epoch": 0.20238258250034233,
|
| 5204 |
+
"grad_norm": 0.48876386880874634,
|
| 5205 |
+
"learning_rate": 4.6221023655773375e-05,
|
| 5206 |
+
"loss": 0.3837924897670746,
|
| 5207 |
+
"step": 739
|
| 5208 |
+
},
|
| 5209 |
+
{
|
| 5210 |
+
"epoch": 0.20265644255785295,
|
| 5211 |
+
"grad_norm": 0.5333821773529053,
|
| 5212 |
+
"learning_rate": 4.620929313564282e-05,
|
| 5213 |
+
"loss": 0.5168666839599609,
|
| 5214 |
+
"step": 740
|
| 5215 |
+
},
|
| 5216 |
+
{
|
| 5217 |
+
"epoch": 0.20293030261536354,
|
| 5218 |
+
"grad_norm": 0.47938454151153564,
|
| 5219 |
+
"learning_rate": 4.6197545930412874e-05,
|
| 5220 |
+
"loss": 0.5067731142044067,
|
| 5221 |
+
"step": 741
|
| 5222 |
+
},
|
| 5223 |
+
{
|
| 5224 |
+
"epoch": 0.20320416267287417,
|
| 5225 |
+
"grad_norm": 0.44225993752479553,
|
| 5226 |
+
"learning_rate": 4.6185782049324936e-05,
|
| 5227 |
+
"loss": 0.5088138580322266,
|
| 5228 |
+
"step": 742
|
| 5229 |
+
},
|
| 5230 |
+
{
|
| 5231 |
+
"epoch": 0.20347802273038476,
|
| 5232 |
+
"grad_norm": 0.4763149917125702,
|
| 5233 |
+
"learning_rate": 4.617400150163351e-05,
|
| 5234 |
+
"loss": 0.4639149308204651,
|
| 5235 |
+
"step": 743
|
| 5236 |
+
},
|
| 5237 |
+
{
|
| 5238 |
+
"epoch": 0.20375188278789538,
|
| 5239 |
+
"grad_norm": 0.5303117036819458,
|
| 5240 |
+
"learning_rate": 4.616220429660621e-05,
|
| 5241 |
+
"loss": 0.5655274391174316,
|
| 5242 |
+
"step": 744
|
| 5243 |
+
},
|
| 5244 |
+
{
|
| 5245 |
+
"epoch": 0.204025742845406,
|
| 5246 |
+
"grad_norm": 0.49112215638160706,
|
| 5247 |
+
"learning_rate": 4.615039044352376e-05,
|
| 5248 |
+
"loss": 0.4511549472808838,
|
| 5249 |
+
"step": 745
|
| 5250 |
+
},
|
| 5251 |
+
{
|
| 5252 |
+
"epoch": 0.2042996029029166,
|
| 5253 |
+
"grad_norm": 0.4602307081222534,
|
| 5254 |
+
"learning_rate": 4.6138559951679985e-05,
|
| 5255 |
+
"loss": 0.45093655586242676,
|
| 5256 |
+
"step": 746
|
| 5257 |
+
},
|
| 5258 |
+
{
|
| 5259 |
+
"epoch": 0.20457346296042722,
|
| 5260 |
+
"grad_norm": 0.4816972017288208,
|
| 5261 |
+
"learning_rate": 4.6126712830381777e-05,
|
| 5262 |
+
"loss": 0.42840924859046936,
|
| 5263 |
+
"step": 747
|
| 5264 |
+
},
|
| 5265 |
+
{
|
| 5266 |
+
"epoch": 0.20484732301793784,
|
| 5267 |
+
"grad_norm": 0.4697628915309906,
|
| 5268 |
+
"learning_rate": 4.611484908894914e-05,
|
| 5269 |
+
"loss": 0.5758565068244934,
|
| 5270 |
+
"step": 748
|
| 5271 |
+
},
|
| 5272 |
+
{
|
| 5273 |
+
"epoch": 0.20512118307544844,
|
| 5274 |
+
"grad_norm": 0.4721646010875702,
|
| 5275 |
+
"learning_rate": 4.6102968736715134e-05,
|
| 5276 |
+
"loss": 0.37101873755455017,
|
| 5277 |
+
"step": 749
|
| 5278 |
+
},
|
| 5279 |
+
{
|
| 5280 |
+
"epoch": 0.20539504313295906,
|
| 5281 |
+
"grad_norm": 0.47777965664863586,
|
| 5282 |
+
"learning_rate": 4.609107178302589e-05,
|
| 5283 |
+
"loss": 0.48357635736465454,
|
| 5284 |
+
"step": 750
|
| 5285 |
+
},
|
| 5286 |
+
{
|
| 5287 |
+
"epoch": 0.20566890319046968,
|
| 5288 |
+
"grad_norm": 0.5223643183708191,
|
| 5289 |
+
"learning_rate": 4.60791582372406e-05,
|
| 5290 |
+
"loss": 0.46817511320114136,
|
| 5291 |
+
"step": 751
|
| 5292 |
+
},
|
| 5293 |
+
{
|
| 5294 |
+
"epoch": 0.20594276324798028,
|
| 5295 |
+
"grad_norm": 0.5534898042678833,
|
| 5296 |
+
"learning_rate": 4.606722810873151e-05,
|
| 5297 |
+
"loss": 0.5060561299324036,
|
| 5298 |
+
"step": 752
|
| 5299 |
+
},
|
| 5300 |
+
{
|
| 5301 |
+
"epoch": 0.2062166233054909,
|
| 5302 |
+
"grad_norm": 0.4662359654903412,
|
| 5303 |
+
"learning_rate": 4.605528140688391e-05,
|
| 5304 |
+
"loss": 0.4500129222869873,
|
| 5305 |
+
"step": 753
|
| 5306 |
+
},
|
| 5307 |
+
{
|
| 5308 |
+
"epoch": 0.2064904833630015,
|
| 5309 |
+
"grad_norm": 0.46944305300712585,
|
| 5310 |
+
"learning_rate": 4.6043318141096135e-05,
|
| 5311 |
+
"loss": 0.39894068241119385,
|
| 5312 |
+
"step": 754
|
| 5313 |
+
},
|
| 5314 |
+
{
|
| 5315 |
+
"epoch": 0.20676434342051211,
|
| 5316 |
+
"grad_norm": 0.49894019961357117,
|
| 5317 |
+
"learning_rate": 4.6031338320779534e-05,
|
| 5318 |
+
"loss": 0.39345842599868774,
|
| 5319 |
+
"step": 755
|
| 5320 |
+
},
|
| 5321 |
+
{
|
| 5322 |
+
"epoch": 0.20703820347802274,
|
| 5323 |
+
"grad_norm": 0.5299913883209229,
|
| 5324 |
+
"learning_rate": 4.6019341955358494e-05,
|
| 5325 |
+
"loss": 0.4625664949417114,
|
| 5326 |
+
"step": 756
|
| 5327 |
+
},
|
| 5328 |
+
{
|
| 5329 |
+
"epoch": 0.20731206353553333,
|
| 5330 |
+
"grad_norm": 0.4864354133605957,
|
| 5331 |
+
"learning_rate": 4.600732905427041e-05,
|
| 5332 |
+
"loss": 0.5182598233222961,
|
| 5333 |
+
"step": 757
|
| 5334 |
+
},
|
| 5335 |
+
{
|
| 5336 |
+
"epoch": 0.20758592359304395,
|
| 5337 |
+
"grad_norm": 0.4955335259437561,
|
| 5338 |
+
"learning_rate": 4.59952996269657e-05,
|
| 5339 |
+
"loss": 0.5308426022529602,
|
| 5340 |
+
"step": 758
|
| 5341 |
+
},
|
| 5342 |
+
{
|
| 5343 |
+
"epoch": 0.20785978365055457,
|
| 5344 |
+
"grad_norm": 0.46057236194610596,
|
| 5345 |
+
"learning_rate": 4.598325368290776e-05,
|
| 5346 |
+
"loss": 0.476673424243927,
|
| 5347 |
+
"step": 759
|
| 5348 |
+
},
|
| 5349 |
+
{
|
| 5350 |
+
"epoch": 0.20813364370806517,
|
| 5351 |
+
"grad_norm": 0.4926740229129791,
|
| 5352 |
+
"learning_rate": 4.5971191231572994e-05,
|
| 5353 |
+
"loss": 0.42311763763427734,
|
| 5354 |
+
"step": 760
|
| 5355 |
+
},
|
| 5356 |
+
{
|
| 5357 |
+
"epoch": 0.2084075037655758,
|
| 5358 |
+
"grad_norm": 0.5065831542015076,
|
| 5359 |
+
"learning_rate": 4.595911228245079e-05,
|
| 5360 |
+
"loss": 0.4618191123008728,
|
| 5361 |
+
"step": 761
|
| 5362 |
+
},
|
| 5363 |
+
{
|
| 5364 |
+
"epoch": 0.2086813638230864,
|
| 5365 |
+
"grad_norm": 0.5033929347991943,
|
| 5366 |
+
"learning_rate": 4.594701684504352e-05,
|
| 5367 |
+
"loss": 0.48177337646484375,
|
| 5368 |
+
"step": 762
|
| 5369 |
+
},
|
| 5370 |
+
{
|
| 5371 |
+
"epoch": 0.208955223880597,
|
| 5372 |
+
"grad_norm": 0.4465022087097168,
|
| 5373 |
+
"learning_rate": 4.593490492886652e-05,
|
| 5374 |
+
"loss": 0.4690370261669159,
|
| 5375 |
+
"step": 763
|
| 5376 |
+
},
|
| 5377 |
+
{
|
| 5378 |
+
"epoch": 0.20922908393810763,
|
| 5379 |
+
"grad_norm": 0.4558869004249573,
|
| 5380 |
+
"learning_rate": 4.592277654344807e-05,
|
| 5381 |
+
"loss": 0.46224331855773926,
|
| 5382 |
+
"step": 764
|
| 5383 |
+
},
|
| 5384 |
+
{
|
| 5385 |
+
"epoch": 0.20950294399561825,
|
| 5386 |
+
"grad_norm": 0.5098778605461121,
|
| 5387 |
+
"learning_rate": 4.591063169832947e-05,
|
| 5388 |
+
"loss": 0.5044959187507629,
|
| 5389 |
+
"step": 765
|
| 5390 |
+
},
|
| 5391 |
+
{
|
| 5392 |
+
"epoch": 0.20977680405312885,
|
| 5393 |
+
"grad_norm": 0.47884607315063477,
|
| 5394 |
+
"learning_rate": 4.5898470403064876e-05,
|
| 5395 |
+
"loss": 0.48726868629455566,
|
| 5396 |
+
"step": 766
|
| 5397 |
+
},
|
| 5398 |
+
{
|
| 5399 |
+
"epoch": 0.21005066411063947,
|
| 5400 |
+
"grad_norm": 0.4915436804294586,
|
| 5401 |
+
"learning_rate": 4.588629266722147e-05,
|
| 5402 |
+
"loss": 0.42750534415245056,
|
| 5403 |
+
"step": 767
|
| 5404 |
+
},
|
| 5405 |
+
{
|
| 5406 |
+
"epoch": 0.21032452416815006,
|
| 5407 |
+
"grad_norm": 0.47338443994522095,
|
| 5408 |
+
"learning_rate": 4.587409850037932e-05,
|
| 5409 |
+
"loss": 0.45149001479148865,
|
| 5410 |
+
"step": 768
|
| 5411 |
+
},
|
| 5412 |
+
{
|
| 5413 |
+
"epoch": 0.21059838422566068,
|
| 5414 |
+
"grad_norm": 0.4632510244846344,
|
| 5415 |
+
"learning_rate": 4.586188791213143e-05,
|
| 5416 |
+
"loss": 0.44037431478500366,
|
| 5417 |
+
"step": 769
|
| 5418 |
+
},
|
| 5419 |
+
{
|
| 5420 |
+
"epoch": 0.2108722442831713,
|
| 5421 |
+
"grad_norm": 0.4755762219429016,
|
| 5422 |
+
"learning_rate": 4.584966091208373e-05,
|
| 5423 |
+
"loss": 0.40315476059913635,
|
| 5424 |
+
"step": 770
|
| 5425 |
+
},
|
| 5426 |
+
{
|
| 5427 |
+
"epoch": 0.2111461043406819,
|
| 5428 |
+
"grad_norm": 0.5018426179885864,
|
| 5429 |
+
"learning_rate": 4.583741750985505e-05,
|
| 5430 |
+
"loss": 0.4870222806930542,
|
| 5431 |
+
"step": 771
|
| 5432 |
+
},
|
| 5433 |
+
{
|
| 5434 |
+
"epoch": 0.21141996439819252,
|
| 5435 |
+
"grad_norm": 0.48830169439315796,
|
| 5436 |
+
"learning_rate": 4.582515771507714e-05,
|
| 5437 |
+
"loss": 0.4371567666530609,
|
| 5438 |
+
"step": 772
|
| 5439 |
+
},
|
| 5440 |
+
{
|
| 5441 |
+
"epoch": 0.21169382445570314,
|
| 5442 |
+
"grad_norm": 0.442939817905426,
|
| 5443 |
+
"learning_rate": 4.581288153739462e-05,
|
| 5444 |
+
"loss": 0.4059355854988098,
|
| 5445 |
+
"step": 773
|
| 5446 |
+
},
|
| 5447 |
+
{
|
| 5448 |
+
"epoch": 0.21196768451321374,
|
| 5449 |
+
"grad_norm": 0.4912157654762268,
|
| 5450 |
+
"learning_rate": 4.580058898646503e-05,
|
| 5451 |
+
"loss": 0.5478442907333374,
|
| 5452 |
+
"step": 774
|
| 5453 |
+
},
|
| 5454 |
+
{
|
| 5455 |
+
"epoch": 0.21224154457072436,
|
| 5456 |
+
"grad_norm": 0.49508124589920044,
|
| 5457 |
+
"learning_rate": 4.5788280071958764e-05,
|
| 5458 |
+
"loss": 0.43974995613098145,
|
| 5459 |
+
"step": 775
|
| 5460 |
+
},
|
| 5461 |
+
{
|
| 5462 |
+
"epoch": 0.21251540462823498,
|
| 5463 |
+
"grad_norm": 0.49629858136177063,
|
| 5464 |
+
"learning_rate": 4.577595480355911e-05,
|
| 5465 |
+
"loss": 0.41481247544288635,
|
| 5466 |
+
"step": 776
|
| 5467 |
+
},
|
| 5468 |
+
{
|
| 5469 |
+
"epoch": 0.21278926468574558,
|
| 5470 |
+
"grad_norm": 0.5357978940010071,
|
| 5471 |
+
"learning_rate": 4.576361319096219e-05,
|
| 5472 |
+
"loss": 0.48343920707702637,
|
| 5473 |
+
"step": 777
|
| 5474 |
+
},
|
| 5475 |
+
{
|
| 5476 |
+
"epoch": 0.2130631247432562,
|
| 5477 |
+
"grad_norm": 0.5325136184692383,
|
| 5478 |
+
"learning_rate": 4.5751255243877015e-05,
|
| 5479 |
+
"loss": 0.5431987047195435,
|
| 5480 |
+
"step": 778
|
| 5481 |
+
},
|
| 5482 |
+
{
|
| 5483 |
+
"epoch": 0.21333698480076682,
|
| 5484 |
+
"grad_norm": 0.5254838466644287,
|
| 5485 |
+
"learning_rate": 4.573888097202544e-05,
|
| 5486 |
+
"loss": 0.4898068904876709,
|
| 5487 |
+
"step": 779
|
| 5488 |
+
},
|
| 5489 |
+
{
|
| 5490 |
+
"epoch": 0.21361084485827742,
|
| 5491 |
+
"grad_norm": 0.6639240384101868,
|
| 5492 |
+
"learning_rate": 4.572649038514216e-05,
|
| 5493 |
+
"loss": 0.5662462711334229,
|
| 5494 |
+
"step": 780
|
| 5495 |
+
},
|
| 5496 |
+
{
|
| 5497 |
+
"epoch": 0.21388470491578804,
|
| 5498 |
+
"grad_norm": 0.450163334608078,
|
| 5499 |
+
"learning_rate": 4.571408349297468e-05,
|
| 5500 |
+
"loss": 0.5176421999931335,
|
| 5501 |
+
"step": 781
|
| 5502 |
+
},
|
| 5503 |
+
{
|
| 5504 |
+
"epoch": 0.21415856497329863,
|
| 5505 |
+
"grad_norm": 0.5275413393974304,
|
| 5506 |
+
"learning_rate": 4.570166030528339e-05,
|
| 5507 |
+
"loss": 0.6069135665893555,
|
| 5508 |
+
"step": 782
|
| 5509 |
+
},
|
| 5510 |
+
{
|
| 5511 |
+
"epoch": 0.21443242503080925,
|
| 5512 |
+
"grad_norm": 0.47154977917671204,
|
| 5513 |
+
"learning_rate": 4.568922083184144e-05,
|
| 5514 |
+
"loss": 0.4319678544998169,
|
| 5515 |
+
"step": 783
|
| 5516 |
+
},
|
| 5517 |
+
{
|
| 5518 |
+
"epoch": 0.21470628508831988,
|
| 5519 |
+
"grad_norm": 0.4944733679294586,
|
| 5520 |
+
"learning_rate": 4.567676508243482e-05,
|
| 5521 |
+
"loss": 0.3476209044456482,
|
| 5522 |
+
"step": 784
|
| 5523 |
+
},
|
| 5524 |
+
{
|
| 5525 |
+
"epoch": 0.21498014514583047,
|
| 5526 |
+
"grad_norm": 0.6367828249931335,
|
| 5527 |
+
"learning_rate": 4.5664293066862324e-05,
|
| 5528 |
+
"loss": 0.5427706837654114,
|
| 5529 |
+
"step": 785
|
| 5530 |
+
},
|
| 5531 |
+
{
|
| 5532 |
+
"epoch": 0.2152540052033411,
|
| 5533 |
+
"grad_norm": 0.48862558603286743,
|
| 5534 |
+
"learning_rate": 4.5651804794935535e-05,
|
| 5535 |
+
"loss": 0.4620243310928345,
|
| 5536 |
+
"step": 786
|
| 5537 |
+
},
|
| 5538 |
+
{
|
| 5539 |
+
"epoch": 0.21552786526085171,
|
| 5540 |
+
"grad_norm": 0.4793543219566345,
|
| 5541 |
+
"learning_rate": 4.5639300276478834e-05,
|
| 5542 |
+
"loss": 0.3833208680152893,
|
| 5543 |
+
"step": 787
|
| 5544 |
+
},
|
| 5545 |
+
{
|
| 5546 |
+
"epoch": 0.2158017253183623,
|
| 5547 |
+
"grad_norm": 0.4692493677139282,
|
| 5548 |
+
"learning_rate": 4.5626779521329376e-05,
|
| 5549 |
+
"loss": 0.39024579524993896,
|
| 5550 |
+
"step": 788
|
| 5551 |
+
},
|
| 5552 |
+
{
|
| 5553 |
+
"epoch": 0.21607558537587293,
|
| 5554 |
+
"grad_norm": 0.5044723153114319,
|
| 5555 |
+
"learning_rate": 4.5614242539337085e-05,
|
| 5556 |
+
"loss": 0.46899715065956116,
|
| 5557 |
+
"step": 789
|
| 5558 |
+
},
|
| 5559 |
+
{
|
| 5560 |
+
"epoch": 0.21634944543338355,
|
| 5561 |
+
"grad_norm": 0.482030987739563,
|
| 5562 |
+
"learning_rate": 4.560168934036466e-05,
|
| 5563 |
+
"loss": 0.4870491325855255,
|
| 5564 |
+
"step": 790
|
| 5565 |
+
},
|
| 5566 |
+
{
|
| 5567 |
+
"epoch": 0.21662330549089415,
|
| 5568 |
+
"grad_norm": 0.5300486087799072,
|
| 5569 |
+
"learning_rate": 4.558911993428756e-05,
|
| 5570 |
+
"loss": 0.4511997103691101,
|
| 5571 |
+
"step": 791
|
| 5572 |
+
},
|
| 5573 |
+
{
|
| 5574 |
+
"epoch": 0.21689716554840477,
|
| 5575 |
+
"grad_norm": 0.4578883945941925,
|
| 5576 |
+
"learning_rate": 4.557653433099398e-05,
|
| 5577 |
+
"loss": 0.4333620071411133,
|
| 5578 |
+
"step": 792
|
| 5579 |
+
},
|
| 5580 |
+
{
|
| 5581 |
+
"epoch": 0.2171710256059154,
|
| 5582 |
+
"grad_norm": 0.4444334805011749,
|
| 5583 |
+
"learning_rate": 4.556393254038486e-05,
|
| 5584 |
+
"loss": 0.39333802461624146,
|
| 5585 |
+
"step": 793
|
| 5586 |
+
},
|
| 5587 |
+
{
|
| 5588 |
+
"epoch": 0.21744488566342599,
|
| 5589 |
+
"grad_norm": 0.4724874198436737,
|
| 5590 |
+
"learning_rate": 4.555131457237389e-05,
|
| 5591 |
+
"loss": 0.5250375270843506,
|
| 5592 |
+
"step": 794
|
| 5593 |
+
},
|
| 5594 |
+
{
|
| 5595 |
+
"epoch": 0.2177187457209366,
|
| 5596 |
+
"grad_norm": 0.4619012176990509,
|
| 5597 |
+
"learning_rate": 4.553868043688747e-05,
|
| 5598 |
+
"loss": 0.41184496879577637,
|
| 5599 |
+
"step": 795
|
| 5600 |
+
},
|
| 5601 |
+
{
|
| 5602 |
+
"epoch": 0.2179926057784472,
|
| 5603 |
+
"grad_norm": 0.4536142647266388,
|
| 5604 |
+
"learning_rate": 4.552603014386473e-05,
|
| 5605 |
+
"loss": 0.3886784613132477,
|
| 5606 |
+
"step": 796
|
| 5607 |
+
},
|
| 5608 |
+
{
|
| 5609 |
+
"epoch": 0.21826646583595782,
|
| 5610 |
+
"grad_norm": 0.5302965044975281,
|
| 5611 |
+
"learning_rate": 4.5513363703257496e-05,
|
| 5612 |
+
"loss": 0.4028399586677551,
|
| 5613 |
+
"step": 797
|
| 5614 |
+
},
|
| 5615 |
+
{
|
| 5616 |
+
"epoch": 0.21854032589346845,
|
| 5617 |
+
"grad_norm": 0.510245144367218,
|
| 5618 |
+
"learning_rate": 4.550068112503032e-05,
|
| 5619 |
+
"loss": 0.5373148322105408,
|
| 5620 |
+
"step": 798
|
| 5621 |
+
},
|
| 5622 |
+
{
|
| 5623 |
+
"epoch": 0.21881418595097904,
|
| 5624 |
+
"grad_norm": 0.4829505980014801,
|
| 5625 |
+
"learning_rate": 4.548798241916042e-05,
|
| 5626 |
+
"loss": 0.5038117170333862,
|
| 5627 |
+
"step": 799
|
| 5628 |
+
},
|
| 5629 |
+
{
|
| 5630 |
+
"epoch": 0.21908804600848966,
|
| 5631 |
+
"grad_norm": 0.49203822016716003,
|
| 5632 |
+
"learning_rate": 4.547526759563773e-05,
|
| 5633 |
+
"loss": 0.44997766613960266,
|
| 5634 |
+
"step": 800
|
| 5635 |
+
},
|
| 5636 |
+
{
|
| 5637 |
+
"epoch": 0.21908804600848966,
|
| 5638 |
+
"eval_loss": 0.47066086530685425,
|
| 5639 |
+
"eval_runtime": 32.4252,
|
| 5640 |
+
"eval_samples_per_second": 6.168,
|
| 5641 |
+
"eval_steps_per_second": 0.771,
|
| 5642 |
+
"step": 800
|
| 5643 |
}
|
| 5644 |
],
|
| 5645 |
"logging_steps": 1,
|
|
|
|
| 5668 |
"attributes": {}
|
| 5669 |
}
|
| 5670 |
},
|
| 5671 |
+
"total_flos": 1.0840359422237082e+17,
|
| 5672 |
"train_batch_size": 1,
|
| 5673 |
"trial_name": null,
|
| 5674 |
"trial_params": null
|