Training in progress, epoch 2, 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/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +956 -4
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 671149168
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c009ddc1d94f455e9195d7d4be1972903acc93c240854a54f2b7dca65b168152
|
| 3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 341314644
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:15011166716c02678454558a3c1c8063eab1a441164609250a4f7a413c1a9a0c
|
| 3 |
size 341314644
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c9e823615e7791da434ea2bf3e0b21d7f2555e1b7211e05c760d0e064f52187
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c43446093e1628094db12fedaa5d20a058b0e404459bac127359c7aff7c16508
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c903c87f6cd97e09254fee79c1ec265e6625798f1d33889dd1dfc725ce5e997
|
| 3 |
size 15984
|
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 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cdebfbcdc58146d93665b48de36547123b030e8c298a5acfdb0be3eb482d6737
|
| 3 |
size 15984
|
last-checkpoint/rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f54cdade05a89000248252a038952b1cdc8afd44da2194ff958ed676c185eec3
|
| 3 |
size 15984
|
last-checkpoint/rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90f5450644fcd4b776ed8591ec6cf0c4bcd4ecf66ea9ac014125200926a408e5
|
| 3 |
size 15984
|
last-checkpoint/rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:66ecc6eb41950527408a67f0b15598d30cacff86b3cf28d7afc6a699306482df
|
| 3 |
size 15984
|
last-checkpoint/rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5f8af120907e390b29e30bc78422d8123341c5043c7535f1b5bddd4248bf24bf
|
| 3 |
size 15984
|
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:b4ac31dcfba988823188c44099086577dd09e162577217892c3f4fd5ad8489c2
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3955,6 +3955,958 @@
|
|
| 3955 |
"learning_rate": 2.4543499379172615e-05,
|
| 3956 |
"loss": 0.0016,
|
| 3957 |
"step": 564
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3958 |
}
|
| 3959 |
],
|
| 3960 |
"logging_steps": 1,
|
|
@@ -3969,12 +4921,12 @@
|
|
| 3969 |
"should_evaluate": false,
|
| 3970 |
"should_log": false,
|
| 3971 |
"should_save": true,
|
| 3972 |
-
"should_training_stop":
|
| 3973 |
},
|
| 3974 |
"attributes": {}
|
| 3975 |
}
|
| 3976 |
},
|
| 3977 |
-
"total_flos": 2.
|
| 3978 |
"train_batch_size": 4,
|
| 3979 |
"trial_name": null,
|
| 3980 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.4800708591674048,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 700,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3955 |
"learning_rate": 2.4543499379172615e-05,
|
| 3956 |
"loss": 0.0016,
|
| 3957 |
"step": 564
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 2.0017714791851198,
|
| 3961 |
+
"grad_norm": 0.08274471014738083,
|
| 3962 |
+
"learning_rate": 2.4198997474396877e-05,
|
| 3963 |
+
"loss": 0.0003,
|
| 3964 |
+
"step": 565
|
| 3965 |
+
},
|
| 3966 |
+
{
|
| 3967 |
+
"epoch": 2.005314437555359,
|
| 3968 |
+
"grad_norm": 0.060062225908041,
|
| 3969 |
+
"learning_rate": 2.3856601107901166e-05,
|
| 3970 |
+
"loss": 0.0003,
|
| 3971 |
+
"step": 566
|
| 3972 |
+
},
|
| 3973 |
+
{
|
| 3974 |
+
"epoch": 2.008857395925598,
|
| 3975 |
+
"grad_norm": 0.1853446364402771,
|
| 3976 |
+
"learning_rate": 2.351631966665476e-05,
|
| 3977 |
+
"loss": 0.0015,
|
| 3978 |
+
"step": 567
|
| 3979 |
+
},
|
| 3980 |
+
{
|
| 3981 |
+
"epoch": 2.012400354295837,
|
| 3982 |
+
"grad_norm": 0.0592714287340641,
|
| 3983 |
+
"learning_rate": 2.31781624796453e-05,
|
| 3984 |
+
"loss": 0.0001,
|
| 3985 |
+
"step": 568
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 2.015943312666076,
|
| 3989 |
+
"grad_norm": 0.13232074677944183,
|
| 3990 |
+
"learning_rate": 2.2842138817622883e-05,
|
| 3991 |
+
"loss": 0.0005,
|
| 3992 |
+
"step": 569
|
| 3993 |
+
},
|
| 3994 |
+
{
|
| 3995 |
+
"epoch": 2.0194862710363153,
|
| 3996 |
+
"grad_norm": 0.24643929302692413,
|
| 3997 |
+
"learning_rate": 2.250825789284594e-05,
|
| 3998 |
+
"loss": 0.0012,
|
| 3999 |
+
"step": 570
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 2.0230292294065544,
|
| 4003 |
+
"grad_norm": 0.17991432547569275,
|
| 4004 |
+
"learning_rate": 2.217652885882869e-05,
|
| 4005 |
+
"loss": 0.0004,
|
| 4006 |
+
"step": 571
|
| 4007 |
+
},
|
| 4008 |
+
{
|
| 4009 |
+
"epoch": 2.0265721877767935,
|
| 4010 |
+
"grad_norm": 0.144062802195549,
|
| 4011 |
+
"learning_rate": 2.1846960810090188e-05,
|
| 4012 |
+
"loss": 0.0005,
|
| 4013 |
+
"step": 572
|
| 4014 |
+
},
|
| 4015 |
+
{
|
| 4016 |
+
"epoch": 2.0301151461470326,
|
| 4017 |
+
"grad_norm": 0.5604081153869629,
|
| 4018 |
+
"learning_rate": 2.151956278190494e-05,
|
| 4019 |
+
"loss": 0.0017,
|
| 4020 |
+
"step": 573
|
| 4021 |
+
},
|
| 4022 |
+
{
|
| 4023 |
+
"epoch": 2.033658104517272,
|
| 4024 |
+
"grad_norm": 0.4855101406574249,
|
| 4025 |
+
"learning_rate": 2.119434375005527e-05,
|
| 4026 |
+
"loss": 0.0008,
|
| 4027 |
+
"step": 574
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 2.0372010628875112,
|
| 4031 |
+
"grad_norm": 0.9981550574302673,
|
| 4032 |
+
"learning_rate": 2.087131263058526e-05,
|
| 4033 |
+
"loss": 0.004,
|
| 4034 |
+
"step": 575
|
| 4035 |
+
},
|
| 4036 |
+
{
|
| 4037 |
+
"epoch": 2.0407440212577503,
|
| 4038 |
+
"grad_norm": 0.12432882189750671,
|
| 4039 |
+
"learning_rate": 2.055047827955618e-05,
|
| 4040 |
+
"loss": 0.0003,
|
| 4041 |
+
"step": 576
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 2.0442869796279894,
|
| 4045 |
+
"grad_norm": 0.009903721511363983,
|
| 4046 |
+
"learning_rate": 2.0231849492803852e-05,
|
| 4047 |
+
"loss": 0.0001,
|
| 4048 |
+
"step": 577
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 2.0478299379982285,
|
| 4052 |
+
"grad_norm": 0.044433970004320145,
|
| 4053 |
+
"learning_rate": 1.991543500569745e-05,
|
| 4054 |
+
"loss": 0.0002,
|
| 4055 |
+
"step": 578
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 2.0513728963684676,
|
| 4059 |
+
"grad_norm": 0.2867472171783447,
|
| 4060 |
+
"learning_rate": 1.960124349289992e-05,
|
| 4061 |
+
"loss": 0.0019,
|
| 4062 |
+
"step": 579
|
| 4063 |
+
},
|
| 4064 |
+
{
|
| 4065 |
+
"epoch": 2.0549158547387067,
|
| 4066 |
+
"grad_norm": 0.371579110622406,
|
| 4067 |
+
"learning_rate": 1.928928356813032e-05,
|
| 4068 |
+
"loss": 0.0012,
|
| 4069 |
+
"step": 580
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 2.058458813108946,
|
| 4073 |
+
"grad_norm": 0.07519207894802094,
|
| 4074 |
+
"learning_rate": 1.8979563783927565e-05,
|
| 4075 |
+
"loss": 0.0002,
|
| 4076 |
+
"step": 581
|
| 4077 |
+
},
|
| 4078 |
+
{
|
| 4079 |
+
"epoch": 2.062001771479185,
|
| 4080 |
+
"grad_norm": 0.20066216588020325,
|
| 4081 |
+
"learning_rate": 1.8672092631416013e-05,
|
| 4082 |
+
"loss": 0.0006,
|
| 4083 |
+
"step": 582
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 2.065544729849424,
|
| 4087 |
+
"grad_norm": 0.19873085618019104,
|
| 4088 |
+
"learning_rate": 1.8366878540072614e-05,
|
| 4089 |
+
"loss": 0.0008,
|
| 4090 |
+
"step": 583
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 2.0690876882196636,
|
| 4094 |
+
"grad_norm": 0.014256162568926811,
|
| 4095 |
+
"learning_rate": 1.8063929877495892e-05,
|
| 4096 |
+
"loss": 0.0001,
|
| 4097 |
+
"step": 584
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 2.0726306465899027,
|
| 4101 |
+
"grad_norm": 0.03786277398467064,
|
| 4102 |
+
"learning_rate": 1.7763254949176414e-05,
|
| 4103 |
+
"loss": 0.0001,
|
| 4104 |
+
"step": 585
|
| 4105 |
+
},
|
| 4106 |
+
{
|
| 4107 |
+
"epoch": 2.076173604960142,
|
| 4108 |
+
"grad_norm": 0.15015755593776703,
|
| 4109 |
+
"learning_rate": 1.7464861998269243e-05,
|
| 4110 |
+
"loss": 0.0003,
|
| 4111 |
+
"step": 586
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 2.079716563330381,
|
| 4115 |
+
"grad_norm": 0.3068810999393463,
|
| 4116 |
+
"learning_rate": 1.7168759205367893e-05,
|
| 4117 |
+
"loss": 0.0003,
|
| 4118 |
+
"step": 587
|
| 4119 |
+
},
|
| 4120 |
+
{
|
| 4121 |
+
"epoch": 2.08325952170062,
|
| 4122 |
+
"grad_norm": 0.32651999592781067,
|
| 4123 |
+
"learning_rate": 1.6874954688279956e-05,
|
| 4124 |
+
"loss": 0.0007,
|
| 4125 |
+
"step": 588
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 2.086802480070859,
|
| 4129 |
+
"grad_norm": 0.5359931588172913,
|
| 4130 |
+
"learning_rate": 1.6583456501804725e-05,
|
| 4131 |
+
"loss": 0.0018,
|
| 4132 |
+
"step": 589
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 2.090345438441098,
|
| 4136 |
+
"grad_norm": 0.08258194476366043,
|
| 4137 |
+
"learning_rate": 1.6294272637512183e-05,
|
| 4138 |
+
"loss": 0.0003,
|
| 4139 |
+
"step": 590
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 2.0938883968113373,
|
| 4143 |
+
"grad_norm": 0.09279931336641312,
|
| 4144 |
+
"learning_rate": 1.600741102352409e-05,
|
| 4145 |
+
"loss": 0.0009,
|
| 4146 |
+
"step": 591
|
| 4147 |
+
},
|
| 4148 |
+
{
|
| 4149 |
+
"epoch": 2.0974313551815764,
|
| 4150 |
+
"grad_norm": 0.7681455016136169,
|
| 4151 |
+
"learning_rate": 1.57228795242965e-05,
|
| 4152 |
+
"loss": 0.0008,
|
| 4153 |
+
"step": 592
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 2.100974313551816,
|
| 4157 |
+
"grad_norm": 0.5113418698310852,
|
| 4158 |
+
"learning_rate": 1.544068594040417e-05,
|
| 4159 |
+
"loss": 0.0006,
|
| 4160 |
+
"step": 593
|
| 4161 |
+
},
|
| 4162 |
+
{
|
| 4163 |
+
"epoch": 2.104517271922055,
|
| 4164 |
+
"grad_norm": 0.11657733470201492,
|
| 4165 |
+
"learning_rate": 1.516083800832676e-05,
|
| 4166 |
+
"loss": 0.0002,
|
| 4167 |
+
"step": 594
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 2.108060230292294,
|
| 4171 |
+
"grad_norm": 0.06766359508037567,
|
| 4172 |
+
"learning_rate": 1.488334340023669e-05,
|
| 4173 |
+
"loss": 0.0002,
|
| 4174 |
+
"step": 595
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 2.1116031886625333,
|
| 4178 |
+
"grad_norm": 0.01443282887339592,
|
| 4179 |
+
"learning_rate": 1.4608209723788835e-05,
|
| 4180 |
+
"loss": 0.0001,
|
| 4181 |
+
"step": 596
|
| 4182 |
+
},
|
| 4183 |
+
{
|
| 4184 |
+
"epoch": 2.1151461470327724,
|
| 4185 |
+
"grad_norm": 0.005601163953542709,
|
| 4186 |
+
"learning_rate": 1.4335444521911899e-05,
|
| 4187 |
+
"loss": 0.0,
|
| 4188 |
+
"step": 597
|
| 4189 |
+
},
|
| 4190 |
+
{
|
| 4191 |
+
"epoch": 2.1186891054030115,
|
| 4192 |
+
"grad_norm": 0.032016199082136154,
|
| 4193 |
+
"learning_rate": 1.4065055272601703e-05,
|
| 4194 |
+
"loss": 0.0001,
|
| 4195 |
+
"step": 598
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 2.1222320637732506,
|
| 4199 |
+
"grad_norm": 0.04691418632864952,
|
| 4200 |
+
"learning_rate": 1.3797049388716065e-05,
|
| 4201 |
+
"loss": 0.0001,
|
| 4202 |
+
"step": 599
|
| 4203 |
+
},
|
| 4204 |
+
{
|
| 4205 |
+
"epoch": 2.1257750221434897,
|
| 4206 |
+
"grad_norm": 0.037374719977378845,
|
| 4207 |
+
"learning_rate": 1.3531434217771692e-05,
|
| 4208 |
+
"loss": 0.0002,
|
| 4209 |
+
"step": 600
|
| 4210 |
+
},
|
| 4211 |
+
{
|
| 4212 |
+
"epoch": 2.129317980513729,
|
| 4213 |
+
"grad_norm": 0.0651448667049408,
|
| 4214 |
+
"learning_rate": 1.3268217041742701e-05,
|
| 4215 |
+
"loss": 0.0001,
|
| 4216 |
+
"step": 601
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 2.132860938883968,
|
| 4220 |
+
"grad_norm": 0.053037695586681366,
|
| 4221 |
+
"learning_rate": 1.3007405076860875e-05,
|
| 4222 |
+
"loss": 0.0001,
|
| 4223 |
+
"step": 602
|
| 4224 |
+
},
|
| 4225 |
+
{
|
| 4226 |
+
"epoch": 2.1364038972542074,
|
| 4227 |
+
"grad_norm": 0.056240785866975784,
|
| 4228 |
+
"learning_rate": 1.2749005473418015e-05,
|
| 4229 |
+
"loss": 0.0002,
|
| 4230 |
+
"step": 603
|
| 4231 |
+
},
|
| 4232 |
+
{
|
| 4233 |
+
"epoch": 2.1399468556244465,
|
| 4234 |
+
"grad_norm": 0.21008822321891785,
|
| 4235 |
+
"learning_rate": 1.2493025315569801e-05,
|
| 4236 |
+
"loss": 0.001,
|
| 4237 |
+
"step": 604
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 2.1434898139946856,
|
| 4241 |
+
"grad_norm": 0.1465017944574356,
|
| 4242 |
+
"learning_rate": 1.2239471621141508e-05,
|
| 4243 |
+
"loss": 0.0005,
|
| 4244 |
+
"step": 605
|
| 4245 |
+
},
|
| 4246 |
+
{
|
| 4247 |
+
"epoch": 2.1470327723649247,
|
| 4248 |
+
"grad_norm": 0.017339682206511497,
|
| 4249 |
+
"learning_rate": 1.1988351341435792e-05,
|
| 4250 |
+
"loss": 0.0001,
|
| 4251 |
+
"step": 606
|
| 4252 |
+
},
|
| 4253 |
+
{
|
| 4254 |
+
"epoch": 2.150575730735164,
|
| 4255 |
+
"grad_norm": 0.5210103392601013,
|
| 4256 |
+
"learning_rate": 1.173967136104196e-05,
|
| 4257 |
+
"loss": 0.0013,
|
| 4258 |
+
"step": 607
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 2.154118689105403,
|
| 4262 |
+
"grad_norm": 0.004422559402883053,
|
| 4263 |
+
"learning_rate": 1.1493438497647313e-05,
|
| 4264 |
+
"loss": 0.0,
|
| 4265 |
+
"step": 608
|
| 4266 |
+
},
|
| 4267 |
+
{
|
| 4268 |
+
"epoch": 2.157661647475642,
|
| 4269 |
+
"grad_norm": 0.03284211456775665,
|
| 4270 |
+
"learning_rate": 1.1249659501850155e-05,
|
| 4271 |
+
"loss": 0.0002,
|
| 4272 |
+
"step": 609
|
| 4273 |
+
},
|
| 4274 |
+
{
|
| 4275 |
+
"epoch": 2.161204605845881,
|
| 4276 |
+
"grad_norm": 0.03831435367465019,
|
| 4277 |
+
"learning_rate": 1.1008341056974854e-05,
|
| 4278 |
+
"loss": 0.0001,
|
| 4279 |
+
"step": 610
|
| 4280 |
+
},
|
| 4281 |
+
{
|
| 4282 |
+
"epoch": 2.1647475642161202,
|
| 4283 |
+
"grad_norm": 0.15469998121261597,
|
| 4284 |
+
"learning_rate": 1.0769489778888405e-05,
|
| 4285 |
+
"loss": 0.0003,
|
| 4286 |
+
"step": 611
|
| 4287 |
+
},
|
| 4288 |
+
{
|
| 4289 |
+
"epoch": 2.16829052258636,
|
| 4290 |
+
"grad_norm": 0.01614706963300705,
|
| 4291 |
+
"learning_rate": 1.0533112215819298e-05,
|
| 4292 |
+
"loss": 0.0,
|
| 4293 |
+
"step": 612
|
| 4294 |
+
},
|
| 4295 |
+
{
|
| 4296 |
+
"epoch": 2.171833480956599,
|
| 4297 |
+
"grad_norm": 0.10824459046125412,
|
| 4298 |
+
"learning_rate": 1.029921484817783e-05,
|
| 4299 |
+
"loss": 0.0004,
|
| 4300 |
+
"step": 613
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 2.175376439326838,
|
| 4304 |
+
"grad_norm": 0.2749079465866089,
|
| 4305 |
+
"learning_rate": 1.0067804088378455e-05,
|
| 4306 |
+
"loss": 0.0013,
|
| 4307 |
+
"step": 614
|
| 4308 |
+
},
|
| 4309 |
+
{
|
| 4310 |
+
"epoch": 2.178919397697077,
|
| 4311 |
+
"grad_norm": 0.00494280643761158,
|
| 4312 |
+
"learning_rate": 9.8388862806641e-06,
|
| 4313 |
+
"loss": 0.0,
|
| 4314 |
+
"step": 615
|
| 4315 |
+
},
|
| 4316 |
+
{
|
| 4317 |
+
"epoch": 2.182462356067316,
|
| 4318 |
+
"grad_norm": 0.008706189692020416,
|
| 4319 |
+
"learning_rate": 9.612467700932045e-06,
|
| 4320 |
+
"loss": 0.0001,
|
| 4321 |
+
"step": 616
|
| 4322 |
+
},
|
| 4323 |
+
{
|
| 4324 |
+
"epoch": 2.1860053144375553,
|
| 4325 |
+
"grad_norm": 0.021209556609392166,
|
| 4326 |
+
"learning_rate": 9.388554556562049e-06,
|
| 4327 |
+
"loss": 0.0001,
|
| 4328 |
+
"step": 617
|
| 4329 |
+
},
|
| 4330 |
+
{
|
| 4331 |
+
"epoch": 2.1895482728077944,
|
| 4332 |
+
"grad_norm": 0.6736172437667847,
|
| 4333 |
+
"learning_rate": 9.167152986246078e-06,
|
| 4334 |
+
"loss": 0.001,
|
| 4335 |
+
"step": 618
|
| 4336 |
+
},
|
| 4337 |
+
{
|
| 4338 |
+
"epoch": 2.1930912311780335,
|
| 4339 |
+
"grad_norm": 0.09763076156377792,
|
| 4340 |
+
"learning_rate": 8.948269059820025e-06,
|
| 4341 |
+
"loss": 0.0007,
|
| 4342 |
+
"step": 619
|
| 4343 |
+
},
|
| 4344 |
+
{
|
| 4345 |
+
"epoch": 2.1966341895482726,
|
| 4346 |
+
"grad_norm": 0.38925012946128845,
|
| 4347 |
+
"learning_rate": 8.731908778097302e-06,
|
| 4348 |
+
"loss": 0.0025,
|
| 4349 |
+
"step": 620
|
| 4350 |
+
},
|
| 4351 |
+
{
|
| 4352 |
+
"epoch": 2.200177147918512,
|
| 4353 |
+
"grad_norm": 0.11619476974010468,
|
| 4354 |
+
"learning_rate": 8.518078072704338e-06,
|
| 4355 |
+
"loss": 0.0005,
|
| 4356 |
+
"step": 621
|
| 4357 |
+
},
|
| 4358 |
+
{
|
| 4359 |
+
"epoch": 2.2037201062887513,
|
| 4360 |
+
"grad_norm": 0.24150855839252472,
|
| 4361 |
+
"learning_rate": 8.306782805917904e-06,
|
| 4362 |
+
"loss": 0.0025,
|
| 4363 |
+
"step": 622
|
| 4364 |
+
},
|
| 4365 |
+
{
|
| 4366 |
+
"epoch": 2.2072630646589904,
|
| 4367 |
+
"grad_norm": 0.04961521923542023,
|
| 4368 |
+
"learning_rate": 8.098028770504494e-06,
|
| 4369 |
+
"loss": 0.0001,
|
| 4370 |
+
"step": 623
|
| 4371 |
+
},
|
| 4372 |
+
{
|
| 4373 |
+
"epoch": 2.2108060230292295,
|
| 4374 |
+
"grad_norm": 0.0838765874505043,
|
| 4375 |
+
"learning_rate": 7.891821689561459e-06,
|
| 4376 |
+
"loss": 0.0006,
|
| 4377 |
+
"step": 624
|
| 4378 |
+
},
|
| 4379 |
+
{
|
| 4380 |
+
"epoch": 2.2143489813994686,
|
| 4381 |
+
"grad_norm": 0.9362369179725647,
|
| 4382 |
+
"learning_rate": 7.68816721636004e-06,
|
| 4383 |
+
"loss": 0.0019,
|
| 4384 |
+
"step": 625
|
| 4385 |
+
},
|
| 4386 |
+
{
|
| 4387 |
+
"epoch": 2.2178919397697077,
|
| 4388 |
+
"grad_norm": 0.03214077651500702,
|
| 4389 |
+
"learning_rate": 7.487070934190532e-06,
|
| 4390 |
+
"loss": 0.0002,
|
| 4391 |
+
"step": 626
|
| 4392 |
+
},
|
| 4393 |
+
{
|
| 4394 |
+
"epoch": 2.2214348981399468,
|
| 4395 |
+
"grad_norm": 0.1560969352722168,
|
| 4396 |
+
"learning_rate": 7.288538356209092e-06,
|
| 4397 |
+
"loss": 0.0003,
|
| 4398 |
+
"step": 627
|
| 4399 |
+
},
|
| 4400 |
+
{
|
| 4401 |
+
"epoch": 2.224977856510186,
|
| 4402 |
+
"grad_norm": 0.05944027379155159,
|
| 4403 |
+
"learning_rate": 7.092574925286614e-06,
|
| 4404 |
+
"loss": 0.0002,
|
| 4405 |
+
"step": 628
|
| 4406 |
+
},
|
| 4407 |
+
{
|
| 4408 |
+
"epoch": 2.228520814880425,
|
| 4409 |
+
"grad_norm": 0.269961416721344,
|
| 4410 |
+
"learning_rate": 6.899186013859561e-06,
|
| 4411 |
+
"loss": 0.001,
|
| 4412 |
+
"step": 629
|
| 4413 |
+
},
|
| 4414 |
+
{
|
| 4415 |
+
"epoch": 2.2320637732506645,
|
| 4416 |
+
"grad_norm": 0.005772658158093691,
|
| 4417 |
+
"learning_rate": 6.708376923782635e-06,
|
| 4418 |
+
"loss": 0.0,
|
| 4419 |
+
"step": 630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 2.2356067316209036,
|
| 4423 |
+
"grad_norm": 0.0038862484507262707,
|
| 4424 |
+
"learning_rate": 6.520152886183406e-06,
|
| 4425 |
+
"loss": 0.0,
|
| 4426 |
+
"step": 631
|
| 4427 |
+
},
|
| 4428 |
+
{
|
| 4429 |
+
"epoch": 2.2391496899911427,
|
| 4430 |
+
"grad_norm": 0.055060192942619324,
|
| 4431 |
+
"learning_rate": 6.3345190613189635e-06,
|
| 4432 |
+
"loss": 0.0001,
|
| 4433 |
+
"step": 632
|
| 4434 |
+
},
|
| 4435 |
+
{
|
| 4436 |
+
"epoch": 2.242692648361382,
|
| 4437 |
+
"grad_norm": 0.020991992205381393,
|
| 4438 |
+
"learning_rate": 6.151480538434382e-06,
|
| 4439 |
+
"loss": 0.0001,
|
| 4440 |
+
"step": 633
|
| 4441 |
+
},
|
| 4442 |
+
{
|
| 4443 |
+
"epoch": 2.246235606731621,
|
| 4444 |
+
"grad_norm": 0.008830989710986614,
|
| 4445 |
+
"learning_rate": 5.971042335623229e-06,
|
| 4446 |
+
"loss": 0.0,
|
| 4447 |
+
"step": 634
|
| 4448 |
+
},
|
| 4449 |
+
{
|
| 4450 |
+
"epoch": 2.24977856510186,
|
| 4451 |
+
"grad_norm": 0.2199835330247879,
|
| 4452 |
+
"learning_rate": 5.793209399689978e-06,
|
| 4453 |
+
"loss": 0.0004,
|
| 4454 |
+
"step": 635
|
| 4455 |
+
},
|
| 4456 |
+
{
|
| 4457 |
+
"epoch": 2.253321523472099,
|
| 4458 |
+
"grad_norm": 0.08529902994632721,
|
| 4459 |
+
"learning_rate": 5.617986606014419e-06,
|
| 4460 |
+
"loss": 0.0004,
|
| 4461 |
+
"step": 636
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 2.2568644818423382,
|
| 4465 |
+
"grad_norm": 0.19139614701271057,
|
| 4466 |
+
"learning_rate": 5.445378758417925e-06,
|
| 4467 |
+
"loss": 0.0016,
|
| 4468 |
+
"step": 637
|
| 4469 |
+
},
|
| 4470 |
+
{
|
| 4471 |
+
"epoch": 2.2604074402125773,
|
| 4472 |
+
"grad_norm": 0.009827593341469765,
|
| 4473 |
+
"learning_rate": 5.275390589031859e-06,
|
| 4474 |
+
"loss": 0.0001,
|
| 4475 |
+
"step": 638
|
| 4476 |
+
},
|
| 4477 |
+
{
|
| 4478 |
+
"epoch": 2.263950398582817,
|
| 4479 |
+
"grad_norm": 0.1290542483329773,
|
| 4480 |
+
"learning_rate": 5.108026758167719e-06,
|
| 4481 |
+
"loss": 0.0002,
|
| 4482 |
+
"step": 639
|
| 4483 |
+
},
|
| 4484 |
+
{
|
| 4485 |
+
"epoch": 2.267493356953056,
|
| 4486 |
+
"grad_norm": 0.17350342869758606,
|
| 4487 |
+
"learning_rate": 4.943291854189493e-06,
|
| 4488 |
+
"loss": 0.0005,
|
| 4489 |
+
"step": 640
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 2.271036315323295,
|
| 4493 |
+
"grad_norm": 0.3507162928581238,
|
| 4494 |
+
"learning_rate": 4.781190393387796e-06,
|
| 4495 |
+
"loss": 0.0003,
|
| 4496 |
+
"step": 641
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 2.274579273693534,
|
| 4500 |
+
"grad_norm": 0.02020534686744213,
|
| 4501 |
+
"learning_rate": 4.6217268198560404e-06,
|
| 4502 |
+
"loss": 0.0001,
|
| 4503 |
+
"step": 642
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 2.2781222320637733,
|
| 4507 |
+
"grad_norm": 0.025796467438340187,
|
| 4508 |
+
"learning_rate": 4.464905505368658e-06,
|
| 4509 |
+
"loss": 0.0001,
|
| 4510 |
+
"step": 643
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 2.2816651904340124,
|
| 4514 |
+
"grad_norm": 0.07381541281938553,
|
| 4515 |
+
"learning_rate": 4.3107307492612086e-06,
|
| 4516 |
+
"loss": 0.0003,
|
| 4517 |
+
"step": 644
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 2.2852081488042515,
|
| 4521 |
+
"grad_norm": 0.03729734942317009,
|
| 4522 |
+
"learning_rate": 4.1592067783125015e-06,
|
| 4523 |
+
"loss": 0.0001,
|
| 4524 |
+
"step": 645
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 2.2887511071744906,
|
| 4528 |
+
"grad_norm": 0.009395123459398746,
|
| 4529 |
+
"learning_rate": 4.010337746628751e-06,
|
| 4530 |
+
"loss": 0.0,
|
| 4531 |
+
"step": 646
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 2.2922940655447297,
|
| 4535 |
+
"grad_norm": 0.1595831662416458,
|
| 4536 |
+
"learning_rate": 3.864127735529656e-06,
|
| 4537 |
+
"loss": 0.0004,
|
| 4538 |
+
"step": 647
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 2.2958370239149692,
|
| 4542 |
+
"grad_norm": 0.008252732455730438,
|
| 4543 |
+
"learning_rate": 3.7205807534365315e-06,
|
| 4544 |
+
"loss": 0.0,
|
| 4545 |
+
"step": 648
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 2.299379982285208,
|
| 4549 |
+
"grad_norm": 0.517535924911499,
|
| 4550 |
+
"learning_rate": 3.5797007357623945e-06,
|
| 4551 |
+
"loss": 0.0005,
|
| 4552 |
+
"step": 649
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 2.3029229406554474,
|
| 4556 |
+
"grad_norm": 0.012122491374611855,
|
| 4557 |
+
"learning_rate": 3.441491544804112e-06,
|
| 4558 |
+
"loss": 0.0001,
|
| 4559 |
+
"step": 650
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 2.3064658990256866,
|
| 4563 |
+
"grad_norm": 0.09782838821411133,
|
| 4564 |
+
"learning_rate": 3.3059569696364502e-06,
|
| 4565 |
+
"loss": 0.0004,
|
| 4566 |
+
"step": 651
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 2.3100088573959257,
|
| 4570 |
+
"grad_norm": 0.19513198733329773,
|
| 4571 |
+
"learning_rate": 3.1731007260082616e-06,
|
| 4572 |
+
"loss": 0.0012,
|
| 4573 |
+
"step": 652
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 2.3135518157661648,
|
| 4577 |
+
"grad_norm": 0.017000757157802582,
|
| 4578 |
+
"learning_rate": 3.0429264562405776e-06,
|
| 4579 |
+
"loss": 0.0001,
|
| 4580 |
+
"step": 653
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 2.317094774136404,
|
| 4584 |
+
"grad_norm": 0.0052710892632603645,
|
| 4585 |
+
"learning_rate": 2.9154377291267674e-06,
|
| 4586 |
+
"loss": 0.0,
|
| 4587 |
+
"step": 654
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 2.320637732506643,
|
| 4591 |
+
"grad_norm": 0.06341676414012909,
|
| 4592 |
+
"learning_rate": 2.790638039834668e-06,
|
| 4593 |
+
"loss": 0.0002,
|
| 4594 |
+
"step": 655
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 2.324180690876882,
|
| 4598 |
+
"grad_norm": 0.018487611785531044,
|
| 4599 |
+
"learning_rate": 2.6685308098108106e-06,
|
| 4600 |
+
"loss": 0.0001,
|
| 4601 |
+
"step": 656
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 2.327723649247121,
|
| 4605 |
+
"grad_norm": 0.19819696247577667,
|
| 4606 |
+
"learning_rate": 2.5491193866866025e-06,
|
| 4607 |
+
"loss": 0.0008,
|
| 4608 |
+
"step": 657
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 2.3312666076173603,
|
| 4612 |
+
"grad_norm": 0.9627290964126587,
|
| 4613 |
+
"learning_rate": 2.432407044186509e-06,
|
| 4614 |
+
"loss": 0.0032,
|
| 4615 |
+
"step": 658
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 2.3348095659876,
|
| 4619 |
+
"grad_norm": 0.26399293541908264,
|
| 4620 |
+
"learning_rate": 2.3183969820383735e-06,
|
| 4621 |
+
"loss": 0.0002,
|
| 4622 |
+
"step": 659
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 2.338352524357839,
|
| 4626 |
+
"grad_norm": 0.0048317075707018375,
|
| 4627 |
+
"learning_rate": 2.2070923258856255e-06,
|
| 4628 |
+
"loss": 0.0,
|
| 4629 |
+
"step": 660
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 2.341895482728078,
|
| 4633 |
+
"grad_norm": 0.019670935347676277,
|
| 4634 |
+
"learning_rate": 2.098496127201648e-06,
|
| 4635 |
+
"loss": 0.0001,
|
| 4636 |
+
"step": 661
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 2.345438441098317,
|
| 4640 |
+
"grad_norm": 0.0395568385720253,
|
| 4641 |
+
"learning_rate": 1.992611363206103e-06,
|
| 4642 |
+
"loss": 0.0001,
|
| 4643 |
+
"step": 662
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 2.348981399468556,
|
| 4647 |
+
"grad_norm": 0.2325470745563507,
|
| 4648 |
+
"learning_rate": 1.889440936783242e-06,
|
| 4649 |
+
"loss": 0.0026,
|
| 4650 |
+
"step": 663
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 2.3525243578387953,
|
| 4654 |
+
"grad_norm": 0.019612450152635574,
|
| 4655 |
+
"learning_rate": 1.7889876764024505e-06,
|
| 4656 |
+
"loss": 0.0001,
|
| 4657 |
+
"step": 664
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 2.3560673162090344,
|
| 4661 |
+
"grad_norm": 0.11476591974496841,
|
| 4662 |
+
"learning_rate": 1.691254336040595e-06,
|
| 4663 |
+
"loss": 0.0004,
|
| 4664 |
+
"step": 665
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 2.3596102745792735,
|
| 4668 |
+
"grad_norm": 0.022656535729765892,
|
| 4669 |
+
"learning_rate": 1.59624359510657e-06,
|
| 4670 |
+
"loss": 0.0001,
|
| 4671 |
+
"step": 666
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 2.3631532329495126,
|
| 4675 |
+
"grad_norm": 0.04269755259156227,
|
| 4676 |
+
"learning_rate": 1.5039580583678393e-06,
|
| 4677 |
+
"loss": 0.0003,
|
| 4678 |
+
"step": 667
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 2.366696191319752,
|
| 4682 |
+
"grad_norm": 0.6594712734222412,
|
| 4683 |
+
"learning_rate": 1.414400255879008e-06,
|
| 4684 |
+
"loss": 0.0011,
|
| 4685 |
+
"step": 668
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 2.3702391496899913,
|
| 4689 |
+
"grad_norm": 0.10037083178758621,
|
| 4690 |
+
"learning_rate": 1.327572642912468e-06,
|
| 4691 |
+
"loss": 0.0002,
|
| 4692 |
+
"step": 669
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 2.3737821080602304,
|
| 4696 |
+
"grad_norm": 0.21993687748908997,
|
| 4697 |
+
"learning_rate": 1.2434775998910964e-06,
|
| 4698 |
+
"loss": 0.0002,
|
| 4699 |
+
"step": 670
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 2.3773250664304695,
|
| 4703 |
+
"grad_norm": 0.018617277964949608,
|
| 4704 |
+
"learning_rate": 1.1621174323229612e-06,
|
| 4705 |
+
"loss": 0.0001,
|
| 4706 |
+
"step": 671
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 2.3808680248007086,
|
| 4710 |
+
"grad_norm": 0.019156964495778084,
|
| 4711 |
+
"learning_rate": 1.0834943707381784e-06,
|
| 4712 |
+
"loss": 0.0001,
|
| 4713 |
+
"step": 672
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 2.3844109831709477,
|
| 4717 |
+
"grad_norm": 0.017957083880901337,
|
| 4718 |
+
"learning_rate": 1.0076105706276888e-06,
|
| 4719 |
+
"loss": 0.0001,
|
| 4720 |
+
"step": 673
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 2.387953941541187,
|
| 4724 |
+
"grad_norm": 0.0515175499022007,
|
| 4725 |
+
"learning_rate": 9.344681123841967e-07,
|
| 4726 |
+
"loss": 0.0002,
|
| 4727 |
+
"step": 674
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 2.391496899911426,
|
| 4731 |
+
"grad_norm": 0.05943327769637108,
|
| 4732 |
+
"learning_rate": 8.640690012451515e-07,
|
| 4733 |
+
"loss": 0.0004,
|
| 4734 |
+
"step": 675
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 2.395039858281665,
|
| 4738 |
+
"grad_norm": 0.24372032284736633,
|
| 4739 |
+
"learning_rate": 7.964151672377458e-07,
|
| 4740 |
+
"loss": 0.0015,
|
| 4741 |
+
"step": 676
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 2.3985828166519045,
|
| 4745 |
+
"grad_norm": 0.07449764758348465,
|
| 4746 |
+
"learning_rate": 7.315084651260009e-07,
|
| 4747 |
+
"loss": 0.0002,
|
| 4748 |
+
"step": 677
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 2.4021257750221436,
|
| 4752 |
+
"grad_norm": 0.021373214200139046,
|
| 4753 |
+
"learning_rate": 6.69350674359959e-07,
|
| 4754 |
+
"loss": 0.0001,
|
| 4755 |
+
"step": 678
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 2.4056687333923827,
|
| 4759 |
+
"grad_norm": 0.008149687200784683,
|
| 4760 |
+
"learning_rate": 6.099434990268609e-07,
|
| 4761 |
+
"loss": 0.0001,
|
| 4762 |
+
"step": 679
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 2.409211691762622,
|
| 4766 |
+
"grad_norm": 0.10549762845039368,
|
| 4767 |
+
"learning_rate": 5.532885678043977e-07,
|
| 4768 |
+
"loss": 0.0003,
|
| 4769 |
+
"step": 680
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 2.412754650132861,
|
| 4773 |
+
"grad_norm": 0.08827023208141327,
|
| 4774 |
+
"learning_rate": 4.9938743391615e-07,
|
| 4775 |
+
"loss": 0.0003,
|
| 4776 |
+
"step": 681
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 2.4162976085031,
|
| 4780 |
+
"grad_norm": 0.011174162849783897,
|
| 4781 |
+
"learning_rate": 4.482415750889204e-07,
|
| 4782 |
+
"loss": 0.0,
|
| 4783 |
+
"step": 682
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 2.419840566873339,
|
| 4787 |
+
"grad_norm": 0.0026113265193998814,
|
| 4788 |
+
"learning_rate": 3.998523935122772e-07,
|
| 4789 |
+
"loss": 0.0,
|
| 4790 |
+
"step": 683
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 2.4233835252435783,
|
| 4794 |
+
"grad_norm": 0.03312141075730324,
|
| 4795 |
+
"learning_rate": 3.5422121580005864e-07,
|
| 4796 |
+
"loss": 0.0001,
|
| 4797 |
+
"step": 684
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 2.4269264836138174,
|
| 4801 |
+
"grad_norm": 0.08379507064819336,
|
| 4802 |
+
"learning_rate": 3.1134929295407564e-07,
|
| 4803 |
+
"loss": 0.0003,
|
| 4804 |
+
"step": 685
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 2.430469441984057,
|
| 4808 |
+
"grad_norm": 0.025159165263175964,
|
| 4809 |
+
"learning_rate": 2.7123780032973235e-07,
|
| 4810 |
+
"loss": 0.0001,
|
| 4811 |
+
"step": 686
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 2.434012400354296,
|
| 4815 |
+
"grad_norm": 0.025299694389104843,
|
| 4816 |
+
"learning_rate": 2.3388783760386601e-07,
|
| 4817 |
+
"loss": 0.0001,
|
| 4818 |
+
"step": 687
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 2.437555358724535,
|
| 4822 |
+
"grad_norm": 0.6148785352706909,
|
| 4823 |
+
"learning_rate": 1.9930042874457254e-07,
|
| 4824 |
+
"loss": 0.0163,
|
| 4825 |
+
"step": 688
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 2.441098317094774,
|
| 4829 |
+
"grad_norm": 0.0018735540797933936,
|
| 4830 |
+
"learning_rate": 1.6747652198313957e-07,
|
| 4831 |
+
"loss": 0.0,
|
| 4832 |
+
"step": 689
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 2.4446412754650133,
|
| 4836 |
+
"grad_norm": 0.11185324192047119,
|
| 4837 |
+
"learning_rate": 1.3841698978804285e-07,
|
| 4838 |
+
"loss": 0.0003,
|
| 4839 |
+
"step": 690
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 2.4481842338352524,
|
| 4843 |
+
"grad_norm": 0.47679388523101807,
|
| 4844 |
+
"learning_rate": 1.1212262884103974e-07,
|
| 4845 |
+
"loss": 0.001,
|
| 4846 |
+
"step": 691
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 2.4517271922054915,
|
| 4850 |
+
"grad_norm": 0.017726508900523186,
|
| 4851 |
+
"learning_rate": 8.85941600153033e-08,
|
| 4852 |
+
"loss": 0.0001,
|
| 4853 |
+
"step": 692
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 2.4552701505757306,
|
| 4857 |
+
"grad_norm": 0.40522223711013794,
|
| 4858 |
+
"learning_rate": 6.783222835572055e-08,
|
| 4859 |
+
"loss": 0.0018,
|
| 4860 |
+
"step": 693
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 2.4588131089459697,
|
| 4864 |
+
"grad_norm": 0.01428899448364973,
|
| 4865 |
+
"learning_rate": 4.98374030611084e-08,
|
| 4866 |
+
"loss": 0.0,
|
| 4867 |
+
"step": 694
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 2.4623560673162093,
|
| 4871 |
+
"grad_norm": 1.708270788192749,
|
| 4872 |
+
"learning_rate": 3.461017746871675e-08,
|
| 4873 |
+
"loss": 0.0083,
|
| 4874 |
+
"step": 695
|
| 4875 |
+
},
|
| 4876 |
+
{
|
| 4877 |
+
"epoch": 2.4658990256864484,
|
| 4878 |
+
"grad_norm": 0.04217607527971268,
|
| 4879 |
+
"learning_rate": 2.215096904060454e-08,
|
| 4880 |
+
"loss": 0.0001,
|
| 4881 |
+
"step": 696
|
| 4882 |
+
},
|
| 4883 |
+
{
|
| 4884 |
+
"epoch": 2.4694419840566875,
|
| 4885 |
+
"grad_norm": 0.007940036244690418,
|
| 4886 |
+
"learning_rate": 1.246011935228064e-08,
|
| 4887 |
+
"loss": 0.0,
|
| 4888 |
+
"step": 697
|
| 4889 |
+
},
|
| 4890 |
+
{
|
| 4891 |
+
"epoch": 2.4729849424269266,
|
| 4892 |
+
"grad_norm": 0.14349356293678284,
|
| 4893 |
+
"learning_rate": 5.537894083273543e-09,
|
| 4894 |
+
"loss": 0.0007,
|
| 4895 |
+
"step": 698
|
| 4896 |
+
},
|
| 4897 |
+
{
|
| 4898 |
+
"epoch": 2.4765279007971657,
|
| 4899 |
+
"grad_norm": 0.017691675573587418,
|
| 4900 |
+
"learning_rate": 1.384483009898796e-09,
|
| 4901 |
+
"loss": 0.0001,
|
| 4902 |
+
"step": 699
|
| 4903 |
+
},
|
| 4904 |
+
{
|
| 4905 |
+
"epoch": 2.4800708591674048,
|
| 4906 |
+
"grad_norm": 0.371564656496048,
|
| 4907 |
+
"learning_rate": 0.0,
|
| 4908 |
+
"loss": 0.0009,
|
| 4909 |
+
"step": 700
|
| 4910 |
}
|
| 4911 |
],
|
| 4912 |
"logging_steps": 1,
|
|
|
|
| 4921 |
"should_evaluate": false,
|
| 4922 |
"should_log": false,
|
| 4923 |
"should_save": true,
|
| 4924 |
+
"should_training_stop": true
|
| 4925 |
},
|
| 4926 |
"attributes": {}
|
| 4927 |
}
|
| 4928 |
},
|
| 4929 |
+
"total_flos": 2.77336688291532e+18,
|
| 4930 |
"train_batch_size": 4,
|
| 4931 |
"trial_name": null,
|
| 4932 |
"trial_params": null
|