Training in progress, step 110000, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +1554 -5
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 962707376
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc49b3d9b0679248cd15f65908e78a9fcfd11827d5389ba5cb0b2ebbdaf749cd
|
| 3 |
size 962707376
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 61873722
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f366356ff3cc9f946b2c19095ca1bd83dff58b6c74b4b7f4158a23546ef43c24
|
| 3 |
size 61873722
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f85fa1b5bcd6943c9f9c56d121f3e4dacc1a19e6d2b661b5087ef101c441a108
|
| 3 |
size 15024
|
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 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fcdc426f94a24625315b78c0cd3d5550941c2561f271869bd8aa30687554b3ef
|
| 3 |
size 15024
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0e95c808f21d4e49616ae42a547aa089261876cc46d0c0199694ad4579fe5b25
|
| 3 |
size 15024
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5d8d792fb2135a74aa62ac4384051c5b425a53ddaef874b5e263803483382478
|
| 3 |
size 15024
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:411cd335eac4c1185a1ba16ecc82a91f7993f0100f715329705d14d99f0dda72
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "/workspace/plateer_classifier_v0.1_result/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 55000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1592,6 +1592,1555 @@
|
|
| 1592 |
"eval_samples_per_second": 210.262,
|
| 1593 |
"eval_steps_per_second": 6.571,
|
| 1594 |
"step": 55000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1595 |
}
|
| 1596 |
],
|
| 1597 |
"logging_steps": 250,
|
|
@@ -1611,7 +3160,7 @@
|
|
| 1611 |
"attributes": {}
|
| 1612 |
}
|
| 1613 |
},
|
| 1614 |
-
"total_flos":
|
| 1615 |
"train_batch_size": 8,
|
| 1616 |
"trial_name": null,
|
| 1617 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.3242824375629425,
|
| 3 |
+
"best_model_checkpoint": "/workspace/plateer_classifier_v0.1_result/checkpoint-110000",
|
| 4 |
+
"epoch": 0.6441270979878347,
|
| 5 |
"eval_steps": 55000,
|
| 6 |
+
"global_step": 110000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1592 |
"eval_samples_per_second": 210.262,
|
| 1593 |
"eval_steps_per_second": 6.571,
|
| 1594 |
"step": 55000
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.32352747421661704,
|
| 1598 |
+
"grad_norm": 7.128218650817871,
|
| 1599 |
+
"learning_rate": 0.00014375796931076737,
|
| 1600 |
+
"loss": 0.357,
|
| 1601 |
+
"step": 55250
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.32499139943931665,
|
| 1605 |
+
"grad_norm": 4.943136692047119,
|
| 1606 |
+
"learning_rate": 0.00014344697181740715,
|
| 1607 |
+
"loss": 0.3576,
|
| 1608 |
+
"step": 55500
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.32645532466201627,
|
| 1612 |
+
"grad_norm": 7.633016109466553,
|
| 1613 |
+
"learning_rate": 0.00014313597432404694,
|
| 1614 |
+
"loss": 0.3655,
|
| 1615 |
+
"step": 55750
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.3279192498847159,
|
| 1619 |
+
"grad_norm": 9.49149227142334,
|
| 1620 |
+
"learning_rate": 0.00014282497683068673,
|
| 1621 |
+
"loss": 0.3687,
|
| 1622 |
+
"step": 56000
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.3293831751074155,
|
| 1626 |
+
"grad_norm": 7.4215521812438965,
|
| 1627 |
+
"learning_rate": 0.00014251397933732655,
|
| 1628 |
+
"loss": 0.3705,
|
| 1629 |
+
"step": 56250
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.33084710033011516,
|
| 1633 |
+
"grad_norm": 5.638499736785889,
|
| 1634 |
+
"learning_rate": 0.00014220298184396634,
|
| 1635 |
+
"loss": 0.3709,
|
| 1636 |
+
"step": 56500
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.3323110255528148,
|
| 1640 |
+
"grad_norm": 9.440450668334961,
|
| 1641 |
+
"learning_rate": 0.00014189198435060613,
|
| 1642 |
+
"loss": 0.35,
|
| 1643 |
+
"step": 56750
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.3337749507755144,
|
| 1647 |
+
"grad_norm": 7.706991195678711,
|
| 1648 |
+
"learning_rate": 0.00014158098685724594,
|
| 1649 |
+
"loss": 0.3601,
|
| 1650 |
+
"step": 57000
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.335238875998214,
|
| 1654 |
+
"grad_norm": 8.154605865478516,
|
| 1655 |
+
"learning_rate": 0.00014126998936388573,
|
| 1656 |
+
"loss": 0.3625,
|
| 1657 |
+
"step": 57250
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.3367028012209136,
|
| 1661 |
+
"grad_norm": 7.608438491821289,
|
| 1662 |
+
"learning_rate": 0.00014095899187052552,
|
| 1663 |
+
"loss": 0.3588,
|
| 1664 |
+
"step": 57500
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.3381667264436133,
|
| 1668 |
+
"grad_norm": 5.466573715209961,
|
| 1669 |
+
"learning_rate": 0.00014064799437716534,
|
| 1670 |
+
"loss": 0.3528,
|
| 1671 |
+
"step": 57750
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.3396306516663129,
|
| 1675 |
+
"grad_norm": 7.514803409576416,
|
| 1676 |
+
"learning_rate": 0.00014033699688380512,
|
| 1677 |
+
"loss": 0.3624,
|
| 1678 |
+
"step": 58000
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.3410945768890125,
|
| 1682 |
+
"grad_norm": 4.846391677856445,
|
| 1683 |
+
"learning_rate": 0.00014002599939044491,
|
| 1684 |
+
"loss": 0.3525,
|
| 1685 |
+
"step": 58250
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.3425585021117121,
|
| 1689 |
+
"grad_norm": 6.116271018981934,
|
| 1690 |
+
"learning_rate": 0.0001397150018970847,
|
| 1691 |
+
"loss": 0.3556,
|
| 1692 |
+
"step": 58500
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.34402242733441174,
|
| 1696 |
+
"grad_norm": 7.234938621520996,
|
| 1697 |
+
"learning_rate": 0.00013940400440372452,
|
| 1698 |
+
"loss": 0.3723,
|
| 1699 |
+
"step": 58750
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.3454863525571114,
|
| 1703 |
+
"grad_norm": 8.690266609191895,
|
| 1704 |
+
"learning_rate": 0.0001390930069103643,
|
| 1705 |
+
"loss": 0.3671,
|
| 1706 |
+
"step": 59000
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.346950277779811,
|
| 1710 |
+
"grad_norm": 5.558066368103027,
|
| 1711 |
+
"learning_rate": 0.0001387820094170041,
|
| 1712 |
+
"loss": 0.3563,
|
| 1713 |
+
"step": 59250
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.34841420300251064,
|
| 1717 |
+
"grad_norm": 5.277857303619385,
|
| 1718 |
+
"learning_rate": 0.0001384710119236439,
|
| 1719 |
+
"loss": 0.3633,
|
| 1720 |
+
"step": 59500
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.34987812822521025,
|
| 1724 |
+
"grad_norm": 4.810859680175781,
|
| 1725 |
+
"learning_rate": 0.00013816125842025712,
|
| 1726 |
+
"loss": 0.3615,
|
| 1727 |
+
"step": 59750
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.35134205344790986,
|
| 1731 |
+
"grad_norm": 6.860721111297607,
|
| 1732 |
+
"learning_rate": 0.00013785026092689694,
|
| 1733 |
+
"loss": 0.3561,
|
| 1734 |
+
"step": 60000
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.35280597867060953,
|
| 1738 |
+
"grad_norm": 6.673612117767334,
|
| 1739 |
+
"learning_rate": 0.00013753926343353673,
|
| 1740 |
+
"loss": 0.3513,
|
| 1741 |
+
"step": 60250
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.35426990389330915,
|
| 1745 |
+
"grad_norm": 6.9296956062316895,
|
| 1746 |
+
"learning_rate": 0.00013722826594017652,
|
| 1747 |
+
"loss": 0.3563,
|
| 1748 |
+
"step": 60500
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.35573382911600876,
|
| 1752 |
+
"grad_norm": 6.235531806945801,
|
| 1753 |
+
"learning_rate": 0.0001369172684468163,
|
| 1754 |
+
"loss": 0.3586,
|
| 1755 |
+
"step": 60750
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.3571977543387084,
|
| 1759 |
+
"grad_norm": 6.549998760223389,
|
| 1760 |
+
"learning_rate": 0.00013660627095345612,
|
| 1761 |
+
"loss": 0.3572,
|
| 1762 |
+
"step": 61000
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.358661679561408,
|
| 1766 |
+
"grad_norm": 6.800797939300537,
|
| 1767 |
+
"learning_rate": 0.0001362952734600959,
|
| 1768 |
+
"loss": 0.3687,
|
| 1769 |
+
"step": 61250
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.3601256047841076,
|
| 1773 |
+
"grad_norm": 5.545276641845703,
|
| 1774 |
+
"learning_rate": 0.0001359842759667357,
|
| 1775 |
+
"loss": 0.3539,
|
| 1776 |
+
"step": 61500
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.36158953000680727,
|
| 1780 |
+
"grad_norm": 8.63070011138916,
|
| 1781 |
+
"learning_rate": 0.00013567327847337552,
|
| 1782 |
+
"loss": 0.3605,
|
| 1783 |
+
"step": 61750
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.3630534552295069,
|
| 1787 |
+
"grad_norm": 5.199543476104736,
|
| 1788 |
+
"learning_rate": 0.0001353622809800153,
|
| 1789 |
+
"loss": 0.3559,
|
| 1790 |
+
"step": 62000
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.3645173804522065,
|
| 1794 |
+
"grad_norm": 27.297420501708984,
|
| 1795 |
+
"learning_rate": 0.0001350512834866551,
|
| 1796 |
+
"loss": 0.3676,
|
| 1797 |
+
"step": 62250
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.3659813056749061,
|
| 1801 |
+
"grad_norm": 8.235854148864746,
|
| 1802 |
+
"learning_rate": 0.00013474152998326833,
|
| 1803 |
+
"loss": 0.3583,
|
| 1804 |
+
"step": 62500
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.3674452308976057,
|
| 1808 |
+
"grad_norm": 6.224372386932373,
|
| 1809 |
+
"learning_rate": 0.00013443053248990812,
|
| 1810 |
+
"loss": 0.3623,
|
| 1811 |
+
"step": 62750
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.3689091561203054,
|
| 1815 |
+
"grad_norm": 8.013957977294922,
|
| 1816 |
+
"learning_rate": 0.0001341195349965479,
|
| 1817 |
+
"loss": 0.3619,
|
| 1818 |
+
"step": 63000
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.370373081343005,
|
| 1822 |
+
"grad_norm": 6.442314147949219,
|
| 1823 |
+
"learning_rate": 0.00013380853750318773,
|
| 1824 |
+
"loss": 0.3586,
|
| 1825 |
+
"step": 63250
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.3718370065657046,
|
| 1829 |
+
"grad_norm": 6.883063793182373,
|
| 1830 |
+
"learning_rate": 0.00013349754000982752,
|
| 1831 |
+
"loss": 0.3635,
|
| 1832 |
+
"step": 63500
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.37330093178840423,
|
| 1836 |
+
"grad_norm": 5.502562999725342,
|
| 1837 |
+
"learning_rate": 0.0001331865425164673,
|
| 1838 |
+
"loss": 0.3525,
|
| 1839 |
+
"step": 63750
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.37476485701110385,
|
| 1843 |
+
"grad_norm": 6.841543197631836,
|
| 1844 |
+
"learning_rate": 0.00013287554502310712,
|
| 1845 |
+
"loss": 0.3564,
|
| 1846 |
+
"step": 64000
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.3762287822338035,
|
| 1850 |
+
"grad_norm": 6.850903034210205,
|
| 1851 |
+
"learning_rate": 0.0001325645475297469,
|
| 1852 |
+
"loss": 0.3549,
|
| 1853 |
+
"step": 64250
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.37769270745650313,
|
| 1857 |
+
"grad_norm": 5.823826313018799,
|
| 1858 |
+
"learning_rate": 0.00013225479402636015,
|
| 1859 |
+
"loss": 0.3488,
|
| 1860 |
+
"step": 64500
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.37915663267920274,
|
| 1864 |
+
"grad_norm": 9.849250793457031,
|
| 1865 |
+
"learning_rate": 0.00013194379653299997,
|
| 1866 |
+
"loss": 0.3526,
|
| 1867 |
+
"step": 64750
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.38062055790190236,
|
| 1871 |
+
"grad_norm": 7.8498992919921875,
|
| 1872 |
+
"learning_rate": 0.00013163279903963975,
|
| 1873 |
+
"loss": 0.3596,
|
| 1874 |
+
"step": 65000
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.38208448312460197,
|
| 1878 |
+
"grad_norm": 7.845436096191406,
|
| 1879 |
+
"learning_rate": 0.00013132180154627954,
|
| 1880 |
+
"loss": 0.3497,
|
| 1881 |
+
"step": 65250
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.38354840834730164,
|
| 1885 |
+
"grad_norm": 10.533845901489258,
|
| 1886 |
+
"learning_rate": 0.00013101080405291933,
|
| 1887 |
+
"loss": 0.3523,
|
| 1888 |
+
"step": 65500
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.38501233357000125,
|
| 1892 |
+
"grad_norm": 9.09399127960205,
|
| 1893 |
+
"learning_rate": 0.00013069980655955912,
|
| 1894 |
+
"loss": 0.347,
|
| 1895 |
+
"step": 65750
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.38647625879270087,
|
| 1899 |
+
"grad_norm": 7.205333232879639,
|
| 1900 |
+
"learning_rate": 0.00013038880906619894,
|
| 1901 |
+
"loss": 0.355,
|
| 1902 |
+
"step": 66000
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.3879401840154005,
|
| 1906 |
+
"grad_norm": 6.770249843597412,
|
| 1907 |
+
"learning_rate": 0.00013007781157283873,
|
| 1908 |
+
"loss": 0.3549,
|
| 1909 |
+
"step": 66250
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.3894041092381001,
|
| 1913 |
+
"grad_norm": 8.14482593536377,
|
| 1914 |
+
"learning_rate": 0.00012976681407947851,
|
| 1915 |
+
"loss": 0.3537,
|
| 1916 |
+
"step": 66500
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.39086803446079976,
|
| 1920 |
+
"grad_norm": 5.998184680938721,
|
| 1921 |
+
"learning_rate": 0.0001294558165861183,
|
| 1922 |
+
"loss": 0.3562,
|
| 1923 |
+
"step": 66750
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.3923319596834994,
|
| 1927 |
+
"grad_norm": 5.583696365356445,
|
| 1928 |
+
"learning_rate": 0.00012914481909275812,
|
| 1929 |
+
"loss": 0.3499,
|
| 1930 |
+
"step": 67000
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.393795884906199,
|
| 1934 |
+
"grad_norm": 6.899207592010498,
|
| 1935 |
+
"learning_rate": 0.0001288338215993979,
|
| 1936 |
+
"loss": 0.3506,
|
| 1937 |
+
"step": 67250
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.3952598101288986,
|
| 1941 |
+
"grad_norm": 6.205395221710205,
|
| 1942 |
+
"learning_rate": 0.0001285228241060377,
|
| 1943 |
+
"loss": 0.3512,
|
| 1944 |
+
"step": 67500
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.3967237353515982,
|
| 1948 |
+
"grad_norm": 9.125551223754883,
|
| 1949 |
+
"learning_rate": 0.0001282118266126775,
|
| 1950 |
+
"loss": 0.3585,
|
| 1951 |
+
"step": 67750
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.3981876605742979,
|
| 1955 |
+
"grad_norm": 6.943772792816162,
|
| 1956 |
+
"learning_rate": 0.0001279008291193173,
|
| 1957 |
+
"loss": 0.362,
|
| 1958 |
+
"step": 68000
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.3996515857969975,
|
| 1962 |
+
"grad_norm": 6.106304168701172,
|
| 1963 |
+
"learning_rate": 0.0001275898316259571,
|
| 1964 |
+
"loss": 0.3545,
|
| 1965 |
+
"step": 68250
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.4011155110196971,
|
| 1969 |
+
"grad_norm": 6.197811126708984,
|
| 1970 |
+
"learning_rate": 0.00012728007812257036,
|
| 1971 |
+
"loss": 0.3524,
|
| 1972 |
+
"step": 68500
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 0.4025794362423967,
|
| 1976 |
+
"grad_norm": 8.07652759552002,
|
| 1977 |
+
"learning_rate": 0.00012696908062921015,
|
| 1978 |
+
"loss": 0.3467,
|
| 1979 |
+
"step": 68750
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.40404336146509634,
|
| 1983 |
+
"grad_norm": 7.444363117218018,
|
| 1984 |
+
"learning_rate": 0.00012665808313584994,
|
| 1985 |
+
"loss": 0.3541,
|
| 1986 |
+
"step": 69000
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 0.405507286687796,
|
| 1990 |
+
"grad_norm": 6.2395782470703125,
|
| 1991 |
+
"learning_rate": 0.00012634708564248972,
|
| 1992 |
+
"loss": 0.3488,
|
| 1993 |
+
"step": 69250
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.4069712119104956,
|
| 1997 |
+
"grad_norm": 7.489956378936768,
|
| 1998 |
+
"learning_rate": 0.00012603608814912954,
|
| 1999 |
+
"loss": 0.3595,
|
| 2000 |
+
"step": 69500
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.40843513713319524,
|
| 2004 |
+
"grad_norm": 6.762283802032471,
|
| 2005 |
+
"learning_rate": 0.00012572509065576933,
|
| 2006 |
+
"loss": 0.3555,
|
| 2007 |
+
"step": 69750
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.40989906235589485,
|
| 2011 |
+
"grad_norm": 10.423229217529297,
|
| 2012 |
+
"learning_rate": 0.00012541409316240912,
|
| 2013 |
+
"loss": 0.3474,
|
| 2014 |
+
"step": 70000
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 0.41136298757859446,
|
| 2018 |
+
"grad_norm": 7.812709331512451,
|
| 2019 |
+
"learning_rate": 0.0001251030956690489,
|
| 2020 |
+
"loss": 0.3588,
|
| 2021 |
+
"step": 70250
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.41282691280129413,
|
| 2025 |
+
"grad_norm": 8.506246566772461,
|
| 2026 |
+
"learning_rate": 0.00012479334216566215,
|
| 2027 |
+
"loss": 0.3473,
|
| 2028 |
+
"step": 70500
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 0.41429083802399375,
|
| 2032 |
+
"grad_norm": 6.0005784034729,
|
| 2033 |
+
"learning_rate": 0.00012448234467230196,
|
| 2034 |
+
"loss": 0.3423,
|
| 2035 |
+
"step": 70750
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.41575476324669336,
|
| 2039 |
+
"grad_norm": 7.6112494468688965,
|
| 2040 |
+
"learning_rate": 0.00012417134717894175,
|
| 2041 |
+
"loss": 0.3469,
|
| 2042 |
+
"step": 71000
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.417218688469393,
|
| 2046 |
+
"grad_norm": 6.460068225860596,
|
| 2047 |
+
"learning_rate": 0.00012386034968558154,
|
| 2048 |
+
"loss": 0.3514,
|
| 2049 |
+
"step": 71250
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.4186826136920926,
|
| 2053 |
+
"grad_norm": 25.509037017822266,
|
| 2054 |
+
"learning_rate": 0.00012354935219222136,
|
| 2055 |
+
"loss": 0.3538,
|
| 2056 |
+
"step": 71500
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.42014653891479226,
|
| 2060 |
+
"grad_norm": 5.778562068939209,
|
| 2061 |
+
"learning_rate": 0.00012323835469886114,
|
| 2062 |
+
"loss": 0.3409,
|
| 2063 |
+
"step": 71750
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 0.42161046413749187,
|
| 2067 |
+
"grad_norm": 10.19543170928955,
|
| 2068 |
+
"learning_rate": 0.00012292735720550093,
|
| 2069 |
+
"loss": 0.3487,
|
| 2070 |
+
"step": 72000
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 0.4230743893601915,
|
| 2074 |
+
"grad_norm": 7.6341633796691895,
|
| 2075 |
+
"learning_rate": 0.00012261635971214072,
|
| 2076 |
+
"loss": 0.3477,
|
| 2077 |
+
"step": 72250
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 0.4245383145828911,
|
| 2081 |
+
"grad_norm": 5.656210422515869,
|
| 2082 |
+
"learning_rate": 0.00012230536221878054,
|
| 2083 |
+
"loss": 0.353,
|
| 2084 |
+
"step": 72500
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.4260022398055907,
|
| 2088 |
+
"grad_norm": 7.81094217300415,
|
| 2089 |
+
"learning_rate": 0.00012199436472542031,
|
| 2090 |
+
"loss": 0.3589,
|
| 2091 |
+
"step": 72750
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 0.4274661650282904,
|
| 2095 |
+
"grad_norm": 5.924116611480713,
|
| 2096 |
+
"learning_rate": 0.0001216833672320601,
|
| 2097 |
+
"loss": 0.346,
|
| 2098 |
+
"step": 73000
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 0.42893009025099,
|
| 2102 |
+
"grad_norm": 6.293444633483887,
|
| 2103 |
+
"learning_rate": 0.00012137236973869992,
|
| 2104 |
+
"loss": 0.3496,
|
| 2105 |
+
"step": 73250
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 0.4303940154736896,
|
| 2109 |
+
"grad_norm": 9.766921997070312,
|
| 2110 |
+
"learning_rate": 0.00012106137224533971,
|
| 2111 |
+
"loss": 0.347,
|
| 2112 |
+
"step": 73500
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 0.4318579406963892,
|
| 2116 |
+
"grad_norm": 5.998900890350342,
|
| 2117 |
+
"learning_rate": 0.0001207503747519795,
|
| 2118 |
+
"loss": 0.3465,
|
| 2119 |
+
"step": 73750
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 0.43332186591908883,
|
| 2123 |
+
"grad_norm": 8.364704132080078,
|
| 2124 |
+
"learning_rate": 0.00012043937725861929,
|
| 2125 |
+
"loss": 0.3429,
|
| 2126 |
+
"step": 74000
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 0.4347857911417885,
|
| 2130 |
+
"grad_norm": 5.508989334106445,
|
| 2131 |
+
"learning_rate": 0.0001201283797652591,
|
| 2132 |
+
"loss": 0.355,
|
| 2133 |
+
"step": 74250
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.4362497163644881,
|
| 2137 |
+
"grad_norm": 6.357595443725586,
|
| 2138 |
+
"learning_rate": 0.00011981738227189889,
|
| 2139 |
+
"loss": 0.3504,
|
| 2140 |
+
"step": 74500
|
| 2141 |
+
},
|
| 2142 |
+
{
|
| 2143 |
+
"epoch": 0.43771364158718773,
|
| 2144 |
+
"grad_norm": 8.691376686096191,
|
| 2145 |
+
"learning_rate": 0.00011950762876851213,
|
| 2146 |
+
"loss": 0.3471,
|
| 2147 |
+
"step": 74750
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 0.43917756680988734,
|
| 2151 |
+
"grad_norm": 11.246256828308105,
|
| 2152 |
+
"learning_rate": 0.00011919663127515193,
|
| 2153 |
+
"loss": 0.3487,
|
| 2154 |
+
"step": 75000
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 0.44064149203258696,
|
| 2158 |
+
"grad_norm": 6.3526811599731445,
|
| 2159 |
+
"learning_rate": 0.00011888563378179172,
|
| 2160 |
+
"loss": 0.3414,
|
| 2161 |
+
"step": 75250
|
| 2162 |
+
},
|
| 2163 |
+
{
|
| 2164 |
+
"epoch": 0.4421054172552866,
|
| 2165 |
+
"grad_norm": 9.6268310546875,
|
| 2166 |
+
"learning_rate": 0.00011857463628843152,
|
| 2167 |
+
"loss": 0.3457,
|
| 2168 |
+
"step": 75500
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 0.44356934247798624,
|
| 2172 |
+
"grad_norm": 8.093045234680176,
|
| 2173 |
+
"learning_rate": 0.00011826363879507131,
|
| 2174 |
+
"loss": 0.3515,
|
| 2175 |
+
"step": 75750
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 0.44503326770068585,
|
| 2179 |
+
"grad_norm": 7.497385025024414,
|
| 2180 |
+
"learning_rate": 0.00011795264130171111,
|
| 2181 |
+
"loss": 0.3361,
|
| 2182 |
+
"step": 76000
|
| 2183 |
+
},
|
| 2184 |
+
{
|
| 2185 |
+
"epoch": 0.44649719292338547,
|
| 2186 |
+
"grad_norm": 8.374622344970703,
|
| 2187 |
+
"learning_rate": 0.00011764164380835092,
|
| 2188 |
+
"loss": 0.3552,
|
| 2189 |
+
"step": 76250
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 0.4479611181460851,
|
| 2193 |
+
"grad_norm": 8.583603858947754,
|
| 2194 |
+
"learning_rate": 0.0001173306463149907,
|
| 2195 |
+
"loss": 0.3395,
|
| 2196 |
+
"step": 76500
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 0.44942504336878475,
|
| 2200 |
+
"grad_norm": 5.933279991149902,
|
| 2201 |
+
"learning_rate": 0.0001170196488216305,
|
| 2202 |
+
"loss": 0.3539,
|
| 2203 |
+
"step": 76750
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 0.45088896859148436,
|
| 2207 |
+
"grad_norm": 7.1400556564331055,
|
| 2208 |
+
"learning_rate": 0.00011670989531824375,
|
| 2209 |
+
"loss": 0.3556,
|
| 2210 |
+
"step": 77000
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 0.452352893814184,
|
| 2214 |
+
"grad_norm": 6.4177374839782715,
|
| 2215 |
+
"learning_rate": 0.00011639889782488354,
|
| 2216 |
+
"loss": 0.34,
|
| 2217 |
+
"step": 77250
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 0.4538168190368836,
|
| 2221 |
+
"grad_norm": 8.248872756958008,
|
| 2222 |
+
"learning_rate": 0.00011608790033152333,
|
| 2223 |
+
"loss": 0.3454,
|
| 2224 |
+
"step": 77500
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 0.4552807442595832,
|
| 2228 |
+
"grad_norm": 6.789691925048828,
|
| 2229 |
+
"learning_rate": 0.00011577690283816314,
|
| 2230 |
+
"loss": 0.3506,
|
| 2231 |
+
"step": 77750
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 0.4567446694822829,
|
| 2235 |
+
"grad_norm": 7.519604206085205,
|
| 2236 |
+
"learning_rate": 0.00011546590534480293,
|
| 2237 |
+
"loss": 0.3438,
|
| 2238 |
+
"step": 78000
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 0.4582085947049825,
|
| 2242 |
+
"grad_norm": 11.287620544433594,
|
| 2243 |
+
"learning_rate": 0.00011515490785144272,
|
| 2244 |
+
"loss": 0.3536,
|
| 2245 |
+
"step": 78250
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 0.4596725199276821,
|
| 2249 |
+
"grad_norm": 5.6864914894104,
|
| 2250 |
+
"learning_rate": 0.00011484391035808254,
|
| 2251 |
+
"loss": 0.348,
|
| 2252 |
+
"step": 78500
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 0.4611364451503817,
|
| 2256 |
+
"grad_norm": 7.405890941619873,
|
| 2257 |
+
"learning_rate": 0.00011453291286472232,
|
| 2258 |
+
"loss": 0.3395,
|
| 2259 |
+
"step": 78750
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 0.4626003703730813,
|
| 2263 |
+
"grad_norm": 5.379487991333008,
|
| 2264 |
+
"learning_rate": 0.00011422315936133556,
|
| 2265 |
+
"loss": 0.3463,
|
| 2266 |
+
"step": 79000
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 0.46406429559578094,
|
| 2270 |
+
"grad_norm": 7.769617080688477,
|
| 2271 |
+
"learning_rate": 0.00011391216186797535,
|
| 2272 |
+
"loss": 0.3458,
|
| 2273 |
+
"step": 79250
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 0.4655282208184806,
|
| 2277 |
+
"grad_norm": 9.26171875,
|
| 2278 |
+
"learning_rate": 0.00011360116437461514,
|
| 2279 |
+
"loss": 0.3394,
|
| 2280 |
+
"step": 79500
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 0.4669921460411802,
|
| 2284 |
+
"grad_norm": 9.037941932678223,
|
| 2285 |
+
"learning_rate": 0.00011329016688125493,
|
| 2286 |
+
"loss": 0.349,
|
| 2287 |
+
"step": 79750
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 0.46845607126387984,
|
| 2291 |
+
"grad_norm": 8.776792526245117,
|
| 2292 |
+
"learning_rate": 0.00011297916938789475,
|
| 2293 |
+
"loss": 0.3384,
|
| 2294 |
+
"step": 80000
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 0.46991999648657945,
|
| 2298 |
+
"grad_norm": 6.737313270568848,
|
| 2299 |
+
"learning_rate": 0.00011266817189453454,
|
| 2300 |
+
"loss": 0.3472,
|
| 2301 |
+
"step": 80250
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 0.47138392170927906,
|
| 2305 |
+
"grad_norm": 7.2374114990234375,
|
| 2306 |
+
"learning_rate": 0.00011235717440117432,
|
| 2307 |
+
"loss": 0.3434,
|
| 2308 |
+
"step": 80500
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 0.47284784693197873,
|
| 2312 |
+
"grad_norm": 6.939677715301514,
|
| 2313 |
+
"learning_rate": 0.00011204617690781414,
|
| 2314 |
+
"loss": 0.3451,
|
| 2315 |
+
"step": 80750
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 0.47431177215467835,
|
| 2319 |
+
"grad_norm": 4.702803611755371,
|
| 2320 |
+
"learning_rate": 0.00011173517941445393,
|
| 2321 |
+
"loss": 0.3508,
|
| 2322 |
+
"step": 81000
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 0.47577569737737796,
|
| 2326 |
+
"grad_norm": 7.359582901000977,
|
| 2327 |
+
"learning_rate": 0.00011142418192109372,
|
| 2328 |
+
"loss": 0.3415,
|
| 2329 |
+
"step": 81250
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 0.4772396226000776,
|
| 2333 |
+
"grad_norm": 8.404651641845703,
|
| 2334 |
+
"learning_rate": 0.00011111442841770696,
|
| 2335 |
+
"loss": 0.3438,
|
| 2336 |
+
"step": 81500
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 0.4787035478227772,
|
| 2340 |
+
"grad_norm": 6.176925182342529,
|
| 2341 |
+
"learning_rate": 0.00011080343092434675,
|
| 2342 |
+
"loss": 0.3484,
|
| 2343 |
+
"step": 81750
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 0.48016747304547686,
|
| 2347 |
+
"grad_norm": 8.614276885986328,
|
| 2348 |
+
"learning_rate": 0.00011049243343098655,
|
| 2349 |
+
"loss": 0.3525,
|
| 2350 |
+
"step": 82000
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 0.48163139826817647,
|
| 2354 |
+
"grad_norm": 5.756929874420166,
|
| 2355 |
+
"learning_rate": 0.00011018143593762635,
|
| 2356 |
+
"loss": 0.3432,
|
| 2357 |
+
"step": 82250
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 0.4830953234908761,
|
| 2361 |
+
"grad_norm": 7.686267852783203,
|
| 2362 |
+
"learning_rate": 0.00010987043844426614,
|
| 2363 |
+
"loss": 0.3508,
|
| 2364 |
+
"step": 82500
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 0.4845592487135757,
|
| 2368 |
+
"grad_norm": 6.590146541595459,
|
| 2369 |
+
"learning_rate": 0.00010955944095090593,
|
| 2370 |
+
"loss": 0.3357,
|
| 2371 |
+
"step": 82750
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 0.4860231739362753,
|
| 2375 |
+
"grad_norm": 7.363981246948242,
|
| 2376 |
+
"learning_rate": 0.00010924968744751918,
|
| 2377 |
+
"loss": 0.3469,
|
| 2378 |
+
"step": 83000
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 0.487487099158975,
|
| 2382 |
+
"grad_norm": 5.942411422729492,
|
| 2383 |
+
"learning_rate": 0.00010893868995415897,
|
| 2384 |
+
"loss": 0.3464,
|
| 2385 |
+
"step": 83250
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 0.4889510243816746,
|
| 2389 |
+
"grad_norm": 8.531744003295898,
|
| 2390 |
+
"learning_rate": 0.00010862769246079879,
|
| 2391 |
+
"loss": 0.3349,
|
| 2392 |
+
"step": 83500
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 0.4904149496043742,
|
| 2396 |
+
"grad_norm": 20.821125030517578,
|
| 2397 |
+
"learning_rate": 0.00010831669496743858,
|
| 2398 |
+
"loss": 0.3434,
|
| 2399 |
+
"step": 83750
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 0.4918788748270738,
|
| 2403 |
+
"grad_norm": 9.569067001342773,
|
| 2404 |
+
"learning_rate": 0.00010800569747407836,
|
| 2405 |
+
"loss": 0.3421,
|
| 2406 |
+
"step": 84000
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 0.49334280004977343,
|
| 2410 |
+
"grad_norm": 7.6851725578308105,
|
| 2411 |
+
"learning_rate": 0.00010769469998071815,
|
| 2412 |
+
"loss": 0.3407,
|
| 2413 |
+
"step": 84250
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 0.4948067252724731,
|
| 2417 |
+
"grad_norm": 9.591890335083008,
|
| 2418 |
+
"learning_rate": 0.00010738370248735797,
|
| 2419 |
+
"loss": 0.347,
|
| 2420 |
+
"step": 84500
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 0.4962706504951727,
|
| 2424 |
+
"grad_norm": 5.16259765625,
|
| 2425 |
+
"learning_rate": 0.00010707270499399776,
|
| 2426 |
+
"loss": 0.3383,
|
| 2427 |
+
"step": 84750
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 0.49773457571787233,
|
| 2431 |
+
"grad_norm": 4.6993794441223145,
|
| 2432 |
+
"learning_rate": 0.00010676170750063755,
|
| 2433 |
+
"loss": 0.3392,
|
| 2434 |
+
"step": 85000
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 0.49919850094057194,
|
| 2438 |
+
"grad_norm": 6.331507682800293,
|
| 2439 |
+
"learning_rate": 0.00010645071000727735,
|
| 2440 |
+
"loss": 0.351,
|
| 2441 |
+
"step": 85250
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 0.5006624261632716,
|
| 2445 |
+
"grad_norm": 7.329137325286865,
|
| 2446 |
+
"learning_rate": 0.00010613971251391714,
|
| 2447 |
+
"loss": 0.3486,
|
| 2448 |
+
"step": 85500
|
| 2449 |
+
},
|
| 2450 |
+
{
|
| 2451 |
+
"epoch": 0.5021263513859712,
|
| 2452 |
+
"grad_norm": 6.907947540283203,
|
| 2453 |
+
"learning_rate": 0.00010582871502055694,
|
| 2454 |
+
"loss": 0.3443,
|
| 2455 |
+
"step": 85750
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 0.5035902766086708,
|
| 2459 |
+
"grad_norm": 4.780885696411133,
|
| 2460 |
+
"learning_rate": 0.00010551771752719674,
|
| 2461 |
+
"loss": 0.3401,
|
| 2462 |
+
"step": 86000
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 0.5050542018313705,
|
| 2466 |
+
"grad_norm": 9.042526245117188,
|
| 2467 |
+
"learning_rate": 0.00010520672003383653,
|
| 2468 |
+
"loss": 0.3402,
|
| 2469 |
+
"step": 86250
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 0.5065181270540701,
|
| 2473 |
+
"grad_norm": 5.397533416748047,
|
| 2474 |
+
"learning_rate": 0.00010489572254047632,
|
| 2475 |
+
"loss": 0.3392,
|
| 2476 |
+
"step": 86500
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 0.5079820522767697,
|
| 2480 |
+
"grad_norm": 7.72251033782959,
|
| 2481 |
+
"learning_rate": 0.00010458472504711612,
|
| 2482 |
+
"loss": 0.3337,
|
| 2483 |
+
"step": 86750
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 0.5094459774994693,
|
| 2487 |
+
"grad_norm": 7.379674434661865,
|
| 2488 |
+
"learning_rate": 0.00010427497154372936,
|
| 2489 |
+
"loss": 0.3457,
|
| 2490 |
+
"step": 87000
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 0.510909902722169,
|
| 2494 |
+
"grad_norm": 7.123027801513672,
|
| 2495 |
+
"learning_rate": 0.00010396397405036915,
|
| 2496 |
+
"loss": 0.3311,
|
| 2497 |
+
"step": 87250
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 0.5123738279448685,
|
| 2501 |
+
"grad_norm": 6.388451099395752,
|
| 2502 |
+
"learning_rate": 0.00010365297655700897,
|
| 2503 |
+
"loss": 0.3386,
|
| 2504 |
+
"step": 87500
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 0.5138377531675682,
|
| 2508 |
+
"grad_norm": 8.933717727661133,
|
| 2509 |
+
"learning_rate": 0.00010334197906364876,
|
| 2510 |
+
"loss": 0.3377,
|
| 2511 |
+
"step": 87750
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 0.5153016783902679,
|
| 2515 |
+
"grad_norm": 5.813757419586182,
|
| 2516 |
+
"learning_rate": 0.000103032225560262,
|
| 2517 |
+
"loss": 0.3368,
|
| 2518 |
+
"step": 88000
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"epoch": 0.5167656036129674,
|
| 2522 |
+
"grad_norm": 10.707741737365723,
|
| 2523 |
+
"learning_rate": 0.00010272122806690178,
|
| 2524 |
+
"loss": 0.3429,
|
| 2525 |
+
"step": 88250
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 0.5182295288356671,
|
| 2529 |
+
"grad_norm": 7.433245658874512,
|
| 2530 |
+
"learning_rate": 0.00010241023057354157,
|
| 2531 |
+
"loss": 0.3457,
|
| 2532 |
+
"step": 88500
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 0.5196934540583666,
|
| 2536 |
+
"grad_norm": 6.408331394195557,
|
| 2537 |
+
"learning_rate": 0.00010209923308018139,
|
| 2538 |
+
"loss": 0.3409,
|
| 2539 |
+
"step": 88750
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 0.5211573792810663,
|
| 2543 |
+
"grad_norm": 7.5843987464904785,
|
| 2544 |
+
"learning_rate": 0.00010178823558682118,
|
| 2545 |
+
"loss": 0.3347,
|
| 2546 |
+
"step": 89000
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 0.522621304503766,
|
| 2550 |
+
"grad_norm": 9.049858093261719,
|
| 2551 |
+
"learning_rate": 0.00010147723809346097,
|
| 2552 |
+
"loss": 0.3392,
|
| 2553 |
+
"step": 89250
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 0.5240852297264655,
|
| 2557 |
+
"grad_norm": 8.207107543945312,
|
| 2558 |
+
"learning_rate": 0.00010116624060010076,
|
| 2559 |
+
"loss": 0.334,
|
| 2560 |
+
"step": 89500
|
| 2561 |
+
},
|
| 2562 |
+
{
|
| 2563 |
+
"epoch": 0.5255491549491652,
|
| 2564 |
+
"grad_norm": 6.511790752410889,
|
| 2565 |
+
"learning_rate": 0.00010085648709671401,
|
| 2566 |
+
"loss": 0.3462,
|
| 2567 |
+
"step": 89750
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 0.5270130801718648,
|
| 2571 |
+
"grad_norm": 5.541443824768066,
|
| 2572 |
+
"learning_rate": 0.0001005454896033538,
|
| 2573 |
+
"loss": 0.3318,
|
| 2574 |
+
"step": 90000
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 0.5284770053945644,
|
| 2578 |
+
"grad_norm": 6.216821670532227,
|
| 2579 |
+
"learning_rate": 0.0001002344921099936,
|
| 2580 |
+
"loss": 0.338,
|
| 2581 |
+
"step": 90250
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 0.5299409306172641,
|
| 2585 |
+
"grad_norm": 5.138360977172852,
|
| 2586 |
+
"learning_rate": 9.992349461663339e-05,
|
| 2587 |
+
"loss": 0.3457,
|
| 2588 |
+
"step": 90500
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 0.5314048558399637,
|
| 2592 |
+
"grad_norm": 8.401073455810547,
|
| 2593 |
+
"learning_rate": 9.961249712327319e-05,
|
| 2594 |
+
"loss": 0.3523,
|
| 2595 |
+
"step": 90750
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 0.5328687810626633,
|
| 2599 |
+
"grad_norm": 8.749157905578613,
|
| 2600 |
+
"learning_rate": 9.930149962991298e-05,
|
| 2601 |
+
"loss": 0.3391,
|
| 2602 |
+
"step": 91000
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 0.5343327062853629,
|
| 2606 |
+
"grad_norm": 7.809004783630371,
|
| 2607 |
+
"learning_rate": 9.899050213655278e-05,
|
| 2608 |
+
"loss": 0.3422,
|
| 2609 |
+
"step": 91250
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 0.5357966315080626,
|
| 2613 |
+
"grad_norm": 7.649618148803711,
|
| 2614 |
+
"learning_rate": 9.867950464319257e-05,
|
| 2615 |
+
"loss": 0.3512,
|
| 2616 |
+
"step": 91500
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 0.5372605567307622,
|
| 2620 |
+
"grad_norm": 8.770468711853027,
|
| 2621 |
+
"learning_rate": 9.836850714983237e-05,
|
| 2622 |
+
"loss": 0.3367,
|
| 2623 |
+
"step": 91750
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 0.5387244819534618,
|
| 2627 |
+
"grad_norm": 8.32112979888916,
|
| 2628 |
+
"learning_rate": 9.805750965647216e-05,
|
| 2629 |
+
"loss": 0.3384,
|
| 2630 |
+
"step": 92000
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 0.5401884071761615,
|
| 2634 |
+
"grad_norm": 9.602888107299805,
|
| 2635 |
+
"learning_rate": 9.774651216311197e-05,
|
| 2636 |
+
"loss": 0.3344,
|
| 2637 |
+
"step": 92250
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 0.541652332398861,
|
| 2641 |
+
"grad_norm": 3.2295093536376953,
|
| 2642 |
+
"learning_rate": 9.743551466975177e-05,
|
| 2643 |
+
"loss": 0.3314,
|
| 2644 |
+
"step": 92500
|
| 2645 |
+
},
|
| 2646 |
+
{
|
| 2647 |
+
"epoch": 0.5431162576215607,
|
| 2648 |
+
"grad_norm": 5.456012725830078,
|
| 2649 |
+
"learning_rate": 9.712451717639156e-05,
|
| 2650 |
+
"loss": 0.3313,
|
| 2651 |
+
"step": 92750
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 0.5445801828442604,
|
| 2655 |
+
"grad_norm": 7.777164936065674,
|
| 2656 |
+
"learning_rate": 9.681351968303136e-05,
|
| 2657 |
+
"loss": 0.3417,
|
| 2658 |
+
"step": 93000
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 0.5460441080669599,
|
| 2662 |
+
"grad_norm": 10.10175895690918,
|
| 2663 |
+
"learning_rate": 9.650252218967115e-05,
|
| 2664 |
+
"loss": 0.3357,
|
| 2665 |
+
"step": 93250
|
| 2666 |
+
},
|
| 2667 |
+
{
|
| 2668 |
+
"epoch": 0.5475080332896596,
|
| 2669 |
+
"grad_norm": 8.296233177185059,
|
| 2670 |
+
"learning_rate": 9.619152469631095e-05,
|
| 2671 |
+
"loss": 0.3368,
|
| 2672 |
+
"step": 93500
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 0.5489719585123591,
|
| 2676 |
+
"grad_norm": 5.55683708190918,
|
| 2677 |
+
"learning_rate": 9.588052720295075e-05,
|
| 2678 |
+
"loss": 0.3338,
|
| 2679 |
+
"step": 93750
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 0.5504358837350588,
|
| 2683 |
+
"grad_norm": 5.92700719833374,
|
| 2684 |
+
"learning_rate": 9.556952970959054e-05,
|
| 2685 |
+
"loss": 0.3431,
|
| 2686 |
+
"step": 94000
|
| 2687 |
+
},
|
| 2688 |
+
{
|
| 2689 |
+
"epoch": 0.5518998089577585,
|
| 2690 |
+
"grad_norm": 5.411899089813232,
|
| 2691 |
+
"learning_rate": 9.525853221623034e-05,
|
| 2692 |
+
"loss": 0.3393,
|
| 2693 |
+
"step": 94250
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 0.553363734180458,
|
| 2697 |
+
"grad_norm": 6.517271995544434,
|
| 2698 |
+
"learning_rate": 9.494753472287013e-05,
|
| 2699 |
+
"loss": 0.3332,
|
| 2700 |
+
"step": 94500
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 0.5548276594031577,
|
| 2704 |
+
"grad_norm": 9.099715232849121,
|
| 2705 |
+
"learning_rate": 9.463653722950994e-05,
|
| 2706 |
+
"loss": 0.3343,
|
| 2707 |
+
"step": 94750
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 0.5562915846258573,
|
| 2711 |
+
"grad_norm": 4.845067501068115,
|
| 2712 |
+
"learning_rate": 9.432553973614972e-05,
|
| 2713 |
+
"loss": 0.3344,
|
| 2714 |
+
"step": 95000
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 0.5577555098485569,
|
| 2718 |
+
"grad_norm": 8.56153392791748,
|
| 2719 |
+
"learning_rate": 9.401454224278953e-05,
|
| 2720 |
+
"loss": 0.33,
|
| 2721 |
+
"step": 95250
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 0.5592194350712566,
|
| 2725 |
+
"grad_norm": 7.1542439460754395,
|
| 2726 |
+
"learning_rate": 9.370354474942933e-05,
|
| 2727 |
+
"loss": 0.3186,
|
| 2728 |
+
"step": 95500
|
| 2729 |
+
},
|
| 2730 |
+
{
|
| 2731 |
+
"epoch": 0.5606833602939562,
|
| 2732 |
+
"grad_norm": 7.00217342376709,
|
| 2733 |
+
"learning_rate": 9.339254725606912e-05,
|
| 2734 |
+
"loss": 0.335,
|
| 2735 |
+
"step": 95750
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 0.5621472855166558,
|
| 2739 |
+
"grad_norm": 7.365664482116699,
|
| 2740 |
+
"learning_rate": 9.308279375268236e-05,
|
| 2741 |
+
"loss": 0.3303,
|
| 2742 |
+
"step": 96000
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 0.5636112107393554,
|
| 2746 |
+
"grad_norm": 8.063042640686035,
|
| 2747 |
+
"learning_rate": 9.277179625932215e-05,
|
| 2748 |
+
"loss": 0.3441,
|
| 2749 |
+
"step": 96250
|
| 2750 |
+
},
|
| 2751 |
+
{
|
| 2752 |
+
"epoch": 0.565075135962055,
|
| 2753 |
+
"grad_norm": 5.403791904449463,
|
| 2754 |
+
"learning_rate": 9.246079876596195e-05,
|
| 2755 |
+
"loss": 0.3318,
|
| 2756 |
+
"step": 96500
|
| 2757 |
+
},
|
| 2758 |
+
{
|
| 2759 |
+
"epoch": 0.5665390611847547,
|
| 2760 |
+
"grad_norm": 5.911950588226318,
|
| 2761 |
+
"learning_rate": 9.215104526257519e-05,
|
| 2762 |
+
"loss": 0.3327,
|
| 2763 |
+
"step": 96750
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 0.5680029864074543,
|
| 2767 |
+
"grad_norm": 5.484018802642822,
|
| 2768 |
+
"learning_rate": 9.184004776921499e-05,
|
| 2769 |
+
"loss": 0.3384,
|
| 2770 |
+
"step": 97000
|
| 2771 |
+
},
|
| 2772 |
+
{
|
| 2773 |
+
"epoch": 0.569466911630154,
|
| 2774 |
+
"grad_norm": 4.785627365112305,
|
| 2775 |
+
"learning_rate": 9.152905027585478e-05,
|
| 2776 |
+
"loss": 0.3437,
|
| 2777 |
+
"step": 97250
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 0.5709308368528535,
|
| 2781 |
+
"grad_norm": 7.17230749130249,
|
| 2782 |
+
"learning_rate": 9.121805278249458e-05,
|
| 2783 |
+
"loss": 0.3331,
|
| 2784 |
+
"step": 97500
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 0.5723947620755532,
|
| 2788 |
+
"grad_norm": 7.777104377746582,
|
| 2789 |
+
"learning_rate": 9.090705528913437e-05,
|
| 2790 |
+
"loss": 0.3371,
|
| 2791 |
+
"step": 97750
|
| 2792 |
+
},
|
| 2793 |
+
{
|
| 2794 |
+
"epoch": 0.5738586872982528,
|
| 2795 |
+
"grad_norm": 6.8572001457214355,
|
| 2796 |
+
"learning_rate": 9.059605779577417e-05,
|
| 2797 |
+
"loss": 0.3397,
|
| 2798 |
+
"step": 98000
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 0.5753226125209524,
|
| 2802 |
+
"grad_norm": 9.132293701171875,
|
| 2803 |
+
"learning_rate": 9.028506030241398e-05,
|
| 2804 |
+
"loss": 0.3421,
|
| 2805 |
+
"step": 98250
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 0.5767865377436521,
|
| 2809 |
+
"grad_norm": 7.351444244384766,
|
| 2810 |
+
"learning_rate": 8.997406280905376e-05,
|
| 2811 |
+
"loss": 0.3315,
|
| 2812 |
+
"step": 98500
|
| 2813 |
+
},
|
| 2814 |
+
{
|
| 2815 |
+
"epoch": 0.5782504629663516,
|
| 2816 |
+
"grad_norm": 5.444695949554443,
|
| 2817 |
+
"learning_rate": 8.966306531569357e-05,
|
| 2818 |
+
"loss": 0.3313,
|
| 2819 |
+
"step": 98750
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 0.5797143881890513,
|
| 2823 |
+
"grad_norm": 6.229501724243164,
|
| 2824 |
+
"learning_rate": 8.935206782233336e-05,
|
| 2825 |
+
"loss": 0.3321,
|
| 2826 |
+
"step": 99000
|
| 2827 |
+
},
|
| 2828 |
+
{
|
| 2829 |
+
"epoch": 0.581178313411751,
|
| 2830 |
+
"grad_norm": 4.431236743927002,
|
| 2831 |
+
"learning_rate": 8.904107032897316e-05,
|
| 2832 |
+
"loss": 0.3326,
|
| 2833 |
+
"step": 99250
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 0.5826422386344505,
|
| 2837 |
+
"grad_norm": 4.78348445892334,
|
| 2838 |
+
"learning_rate": 8.873007283561296e-05,
|
| 2839 |
+
"loss": 0.3362,
|
| 2840 |
+
"step": 99500
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 0.5841061638571502,
|
| 2844 |
+
"grad_norm": 5.964051723480225,
|
| 2845 |
+
"learning_rate": 8.841907534225275e-05,
|
| 2846 |
+
"loss": 0.3408,
|
| 2847 |
+
"step": 99750
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 0.5855700890798498,
|
| 2851 |
+
"grad_norm": 5.310559272766113,
|
| 2852 |
+
"learning_rate": 8.810807784889255e-05,
|
| 2853 |
+
"loss": 0.3328,
|
| 2854 |
+
"step": 100000
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 0.5870340143025494,
|
| 2858 |
+
"grad_norm": 4.985818862915039,
|
| 2859 |
+
"learning_rate": 8.779708035553234e-05,
|
| 2860 |
+
"loss": 0.337,
|
| 2861 |
+
"step": 100250
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.5884979395252491,
|
| 2865 |
+
"grad_norm": 4.851356506347656,
|
| 2866 |
+
"learning_rate": 8.748608286217213e-05,
|
| 2867 |
+
"loss": 0.3314,
|
| 2868 |
+
"step": 100500
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 0.5899618647479486,
|
| 2872 |
+
"grad_norm": 6.863201141357422,
|
| 2873 |
+
"learning_rate": 8.717508536881193e-05,
|
| 2874 |
+
"loss": 0.3231,
|
| 2875 |
+
"step": 100750
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 0.5914257899706483,
|
| 2879 |
+
"grad_norm": 6.387337684631348,
|
| 2880 |
+
"learning_rate": 8.686533186542517e-05,
|
| 2881 |
+
"loss": 0.322,
|
| 2882 |
+
"step": 101000
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 0.5928897151933479,
|
| 2886 |
+
"grad_norm": 7.897363662719727,
|
| 2887 |
+
"learning_rate": 8.655433437206496e-05,
|
| 2888 |
+
"loss": 0.3361,
|
| 2889 |
+
"step": 101250
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.5943536404160475,
|
| 2893 |
+
"grad_norm": 5.876019477844238,
|
| 2894 |
+
"learning_rate": 8.624333687870476e-05,
|
| 2895 |
+
"loss": 0.3211,
|
| 2896 |
+
"step": 101500
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.5958175656387472,
|
| 2900 |
+
"grad_norm": 4.175768852233887,
|
| 2901 |
+
"learning_rate": 8.593233938534457e-05,
|
| 2902 |
+
"loss": 0.3317,
|
| 2903 |
+
"step": 101750
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.5972814908614468,
|
| 2907 |
+
"grad_norm": 6.496226787567139,
|
| 2908 |
+
"learning_rate": 8.562134189198435e-05,
|
| 2909 |
+
"loss": 0.3289,
|
| 2910 |
+
"step": 102000
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.5987454160841464,
|
| 2914 |
+
"grad_norm": 7.092103004455566,
|
| 2915 |
+
"learning_rate": 8.531034439862416e-05,
|
| 2916 |
+
"loss": 0.3329,
|
| 2917 |
+
"step": 102250
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.600209341306846,
|
| 2921 |
+
"grad_norm": 7.335963726043701,
|
| 2922 |
+
"learning_rate": 8.499934690526395e-05,
|
| 2923 |
+
"loss": 0.3305,
|
| 2924 |
+
"step": 102500
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.6016732665295457,
|
| 2928 |
+
"grad_norm": 6.620415687561035,
|
| 2929 |
+
"learning_rate": 8.468834941190375e-05,
|
| 2930 |
+
"loss": 0.3324,
|
| 2931 |
+
"step": 102750
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.6031371917522453,
|
| 2935 |
+
"grad_norm": 6.866759777069092,
|
| 2936 |
+
"learning_rate": 8.437735191854355e-05,
|
| 2937 |
+
"loss": 0.3395,
|
| 2938 |
+
"step": 103000
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.6046011169749449,
|
| 2942 |
+
"grad_norm": 7.7242045402526855,
|
| 2943 |
+
"learning_rate": 8.406759841515678e-05,
|
| 2944 |
+
"loss": 0.3368,
|
| 2945 |
+
"step": 103250
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.6060650421976446,
|
| 2949 |
+
"grad_norm": 6.402958869934082,
|
| 2950 |
+
"learning_rate": 8.375660092179658e-05,
|
| 2951 |
+
"loss": 0.3366,
|
| 2952 |
+
"step": 103500
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.6075289674203441,
|
| 2956 |
+
"grad_norm": 6.456150531768799,
|
| 2957 |
+
"learning_rate": 8.344560342843637e-05,
|
| 2958 |
+
"loss": 0.3372,
|
| 2959 |
+
"step": 103750
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.6089928926430438,
|
| 2963 |
+
"grad_norm": 7.6825971603393555,
|
| 2964 |
+
"learning_rate": 8.313460593507617e-05,
|
| 2965 |
+
"loss": 0.3331,
|
| 2966 |
+
"step": 104000
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.6104568178657435,
|
| 2970 |
+
"grad_norm": 11.974824905395508,
|
| 2971 |
+
"learning_rate": 8.282360844171596e-05,
|
| 2972 |
+
"loss": 0.3317,
|
| 2973 |
+
"step": 104250
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.611920743088443,
|
| 2977 |
+
"grad_norm": 5.445409774780273,
|
| 2978 |
+
"learning_rate": 8.251261094835576e-05,
|
| 2979 |
+
"loss": 0.3303,
|
| 2980 |
+
"step": 104500
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.6133846683111427,
|
| 2984 |
+
"grad_norm": 8.099034309387207,
|
| 2985 |
+
"learning_rate": 8.220161345499555e-05,
|
| 2986 |
+
"loss": 0.3317,
|
| 2987 |
+
"step": 104750
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.6148485935338422,
|
| 2991 |
+
"grad_norm": 21.789043426513672,
|
| 2992 |
+
"learning_rate": 8.189061596163535e-05,
|
| 2993 |
+
"loss": 0.3146,
|
| 2994 |
+
"step": 105000
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.6163125187565419,
|
| 2998 |
+
"grad_norm": 6.879361152648926,
|
| 2999 |
+
"learning_rate": 8.158086245824859e-05,
|
| 3000 |
+
"loss": 0.3346,
|
| 3001 |
+
"step": 105250
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.6177764439792416,
|
| 3005 |
+
"grad_norm": 5.477085113525391,
|
| 3006 |
+
"learning_rate": 8.126986496488838e-05,
|
| 3007 |
+
"loss": 0.3274,
|
| 3008 |
+
"step": 105500
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.6192403692019411,
|
| 3012 |
+
"grad_norm": 6.2816667556762695,
|
| 3013 |
+
"learning_rate": 8.095886747152818e-05,
|
| 3014 |
+
"loss": 0.3271,
|
| 3015 |
+
"step": 105750
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.6207042944246408,
|
| 3019 |
+
"grad_norm": 9.089285850524902,
|
| 3020 |
+
"learning_rate": 8.064786997816797e-05,
|
| 3021 |
+
"loss": 0.3351,
|
| 3022 |
+
"step": 106000
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.6221682196473404,
|
| 3026 |
+
"grad_norm": 6.114886283874512,
|
| 3027 |
+
"learning_rate": 8.033687248480777e-05,
|
| 3028 |
+
"loss": 0.3296,
|
| 3029 |
+
"step": 106250
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.62363214487004,
|
| 3033 |
+
"grad_norm": 7.2542548179626465,
|
| 3034 |
+
"learning_rate": 8.002587499144756e-05,
|
| 3035 |
+
"loss": 0.3246,
|
| 3036 |
+
"step": 106500
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.6250960700927397,
|
| 3040 |
+
"grad_norm": 5.58528995513916,
|
| 3041 |
+
"learning_rate": 7.971487749808737e-05,
|
| 3042 |
+
"loss": 0.3327,
|
| 3043 |
+
"step": 106750
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.6265599953154393,
|
| 3047 |
+
"grad_norm": 3.898178815841675,
|
| 3048 |
+
"learning_rate": 7.940388000472715e-05,
|
| 3049 |
+
"loss": 0.3291,
|
| 3050 |
+
"step": 107000
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 0.6280239205381389,
|
| 3054 |
+
"grad_norm": 5.644820690155029,
|
| 3055 |
+
"learning_rate": 7.909288251136696e-05,
|
| 3056 |
+
"loss": 0.3281,
|
| 3057 |
+
"step": 107250
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 0.6294878457608385,
|
| 3061 |
+
"grad_norm": 6.363776206970215,
|
| 3062 |
+
"learning_rate": 7.878188501800676e-05,
|
| 3063 |
+
"loss": 0.3304,
|
| 3064 |
+
"step": 107500
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.6309517709835382,
|
| 3068 |
+
"grad_norm": 5.209687232971191,
|
| 3069 |
+
"learning_rate": 7.847213151462e-05,
|
| 3070 |
+
"loss": 0.3224,
|
| 3071 |
+
"step": 107750
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 0.6324156962062378,
|
| 3075 |
+
"grad_norm": 6.911553382873535,
|
| 3076 |
+
"learning_rate": 7.81611340212598e-05,
|
| 3077 |
+
"loss": 0.3246,
|
| 3078 |
+
"step": 108000
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 0.6338796214289374,
|
| 3082 |
+
"grad_norm": 7.6557111740112305,
|
| 3083 |
+
"learning_rate": 7.785013652789959e-05,
|
| 3084 |
+
"loss": 0.322,
|
| 3085 |
+
"step": 108250
|
| 3086 |
+
},
|
| 3087 |
+
{
|
| 3088 |
+
"epoch": 0.6353435466516371,
|
| 3089 |
+
"grad_norm": 7.857481002807617,
|
| 3090 |
+
"learning_rate": 7.753913903453939e-05,
|
| 3091 |
+
"loss": 0.3318,
|
| 3092 |
+
"step": 108500
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 0.6368074718743366,
|
| 3096 |
+
"grad_norm": 5.911120891571045,
|
| 3097 |
+
"learning_rate": 7.722814154117918e-05,
|
| 3098 |
+
"loss": 0.325,
|
| 3099 |
+
"step": 108750
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 0.6382713970970363,
|
| 3103 |
+
"grad_norm": 8.592209815979004,
|
| 3104 |
+
"learning_rate": 7.691714404781898e-05,
|
| 3105 |
+
"loss": 0.3209,
|
| 3106 |
+
"step": 109000
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.639735322319736,
|
| 3110 |
+
"grad_norm": 6.824602127075195,
|
| 3111 |
+
"learning_rate": 7.660614655445879e-05,
|
| 3112 |
+
"loss": 0.3331,
|
| 3113 |
+
"step": 109250
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 0.6411992475424355,
|
| 3117 |
+
"grad_norm": 6.813981056213379,
|
| 3118 |
+
"learning_rate": 7.629514906109858e-05,
|
| 3119 |
+
"loss": 0.3313,
|
| 3120 |
+
"step": 109500
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 0.6426631727651352,
|
| 3124 |
+
"grad_norm": 5.7169671058654785,
|
| 3125 |
+
"learning_rate": 7.598539555771181e-05,
|
| 3126 |
+
"loss": 0.3206,
|
| 3127 |
+
"step": 109750
|
| 3128 |
+
},
|
| 3129 |
+
{
|
| 3130 |
+
"epoch": 0.6441270979878347,
|
| 3131 |
+
"grad_norm": 5.429720401763916,
|
| 3132 |
+
"learning_rate": 7.56743980643516e-05,
|
| 3133 |
+
"loss": 0.3192,
|
| 3134 |
+
"step": 110000
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 0.6441270979878347,
|
| 3138 |
+
"eval_accuracy": 0.8997983351325891,
|
| 3139 |
+
"eval_loss": 0.3242824375629425,
|
| 3140 |
+
"eval_runtime": 11546.6804,
|
| 3141 |
+
"eval_samples_per_second": 210.345,
|
| 3142 |
+
"eval_steps_per_second": 6.573,
|
| 3143 |
+
"step": 110000
|
| 3144 |
}
|
| 3145 |
],
|
| 3146 |
"logging_steps": 250,
|
|
|
|
| 3160 |
"attributes": {}
|
| 3161 |
}
|
| 3162 |
},
|
| 3163 |
+
"total_flos": 2.8505890873482936e+19,
|
| 3164 |
"train_batch_size": 8,
|
| 3165 |
"trial_name": null,
|
| 3166 |
"trial_params": null
|