Training in progress, step 12500, checkpoint
Browse files- last-checkpoint/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 +1769 -3
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 737632172
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be61347e986df1813f8514099ca7495d01a0aaff2cb1086c995e9f56e4864f44
|
| 3 |
size 737632172
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1475354682
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3f13242f121b2c14586fd1d56f596180d4a87cf3c3e6a8b06d2aa80aac67af52
|
| 3 |
size 1475354682
|
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:b0169a1fcfdb795965595a4f242f88323799f2590a349006ef637e474b948bd8
|
| 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:c87b1e14f1bde64851a26e79d9e7529d68eb1e143f87cb05c0bdf4c84c676fd8
|
| 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 sha256:0bea1a1e24ce9ba044268ec704e9b5435b962ac6a2de09e9847caeff5397ad96
|
| 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 sha256:3ec0b58b8e0950d68e3cbb11f67305b0912f5521574e32c474ae22410e7fadc8
|
| 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 sha256:284664914b90ff8b8a0dc92bb3d3f63cfa784487322d7dc115bd6038f6758aca
|
| 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 sha256:0468917ff18a692d9dedf8d79fa5e11dd93feec07877799ca86aca3ed690d129
|
| 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 sha256:842bb0cfac2bb06e2e811dabf1e415d78f36184efada16b4098faa08d32c3580
|
| 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:ca857437aedb445e7bf3dd47069eecac538c7d7fd16d601188beef14a54e520e
|
| 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 1000
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:123bdb81188a6d8339925f843c784a6596cb7ed0221abdafc8c5e0e110c82c27
|
| 3 |
size 1000
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 1000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -7087,6 +7087,1772 @@
|
|
| 7087 |
"eval_samples_per_second": 1802.303,
|
| 7088 |
"eval_steps_per_second": 56.322,
|
| 7089 |
"step": 10000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7090 |
}
|
| 7091 |
],
|
| 7092 |
"logging_steps": 10,
|
|
@@ -7106,7 +8872,7 @@
|
|
| 7106 |
"attributes": {}
|
| 7107 |
}
|
| 7108 |
},
|
| 7109 |
-
"total_flos":
|
| 7110 |
"train_batch_size": 4,
|
| 7111 |
"trial_name": null,
|
| 7112 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.5547511663643273,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 12500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 7087 |
"eval_samples_per_second": 1802.303,
|
| 7088 |
"eval_steps_per_second": 56.322,
|
| 7089 |
"step": 10000
|
| 7090 |
+
},
|
| 7091 |
+
{
|
| 7092 |
+
"epoch": 0.4442447340245533,
|
| 7093 |
+
"grad_norm": 87.5740966796875,
|
| 7094 |
+
"learning_rate": 9.98264670211062e-06,
|
| 7095 |
+
"loss": 11.53,
|
| 7096 |
+
"step": 10010
|
| 7097 |
+
},
|
| 7098 |
+
{
|
| 7099 |
+
"epoch": 0.44468853495764477,
|
| 7100 |
+
"grad_norm": 79.04003143310547,
|
| 7101 |
+
"learning_rate": 9.982629366148691e-06,
|
| 7102 |
+
"loss": 11.2839,
|
| 7103 |
+
"step": 10020
|
| 7104 |
+
},
|
| 7105 |
+
{
|
| 7106 |
+
"epoch": 0.4451323358907362,
|
| 7107 |
+
"grad_norm": 93.09468841552734,
|
| 7108 |
+
"learning_rate": 9.982612030186765e-06,
|
| 7109 |
+
"loss": 11.685,
|
| 7110 |
+
"step": 10030
|
| 7111 |
+
},
|
| 7112 |
+
{
|
| 7113 |
+
"epoch": 0.4455761368238277,
|
| 7114 |
+
"grad_norm": 75.00133514404297,
|
| 7115 |
+
"learning_rate": 9.982594694224838e-06,
|
| 7116 |
+
"loss": 10.8954,
|
| 7117 |
+
"step": 10040
|
| 7118 |
+
},
|
| 7119 |
+
{
|
| 7120 |
+
"epoch": 0.4460199377569191,
|
| 7121 |
+
"grad_norm": 95.3443374633789,
|
| 7122 |
+
"learning_rate": 9.982577358262909e-06,
|
| 7123 |
+
"loss": 10.8405,
|
| 7124 |
+
"step": 10050
|
| 7125 |
+
},
|
| 7126 |
+
{
|
| 7127 |
+
"epoch": 0.4464637386900106,
|
| 7128 |
+
"grad_norm": 81.28312683105469,
|
| 7129 |
+
"learning_rate": 9.982560022300982e-06,
|
| 7130 |
+
"loss": 10.8736,
|
| 7131 |
+
"step": 10060
|
| 7132 |
+
},
|
| 7133 |
+
{
|
| 7134 |
+
"epoch": 0.44690753962310203,
|
| 7135 |
+
"grad_norm": 84.82075500488281,
|
| 7136 |
+
"learning_rate": 9.982542686339055e-06,
|
| 7137 |
+
"loss": 10.9681,
|
| 7138 |
+
"step": 10070
|
| 7139 |
+
},
|
| 7140 |
+
{
|
| 7141 |
+
"epoch": 0.4473513405561935,
|
| 7142 |
+
"grad_norm": 83.95282745361328,
|
| 7143 |
+
"learning_rate": 9.982525350377127e-06,
|
| 7144 |
+
"loss": 11.012,
|
| 7145 |
+
"step": 10080
|
| 7146 |
+
},
|
| 7147 |
+
{
|
| 7148 |
+
"epoch": 0.44779514148928495,
|
| 7149 |
+
"grad_norm": 77.86396026611328,
|
| 7150 |
+
"learning_rate": 9.9825080144152e-06,
|
| 7151 |
+
"loss": 11.2847,
|
| 7152 |
+
"step": 10090
|
| 7153 |
+
},
|
| 7154 |
+
{
|
| 7155 |
+
"epoch": 0.44823894242237644,
|
| 7156 |
+
"grad_norm": 87.20881652832031,
|
| 7157 |
+
"learning_rate": 9.982490678453273e-06,
|
| 7158 |
+
"loss": 11.2584,
|
| 7159 |
+
"step": 10100
|
| 7160 |
+
},
|
| 7161 |
+
{
|
| 7162 |
+
"epoch": 0.44868274335546793,
|
| 7163 |
+
"grad_norm": 100.79487609863281,
|
| 7164 |
+
"learning_rate": 9.982473342491344e-06,
|
| 7165 |
+
"loss": 11.8089,
|
| 7166 |
+
"step": 10110
|
| 7167 |
+
},
|
| 7168 |
+
{
|
| 7169 |
+
"epoch": 0.44912654428855936,
|
| 7170 |
+
"grad_norm": 81.89049530029297,
|
| 7171 |
+
"learning_rate": 9.982456006529417e-06,
|
| 7172 |
+
"loss": 11.5939,
|
| 7173 |
+
"step": 10120
|
| 7174 |
+
},
|
| 7175 |
+
{
|
| 7176 |
+
"epoch": 0.44957034522165085,
|
| 7177 |
+
"grad_norm": 91.9642562866211,
|
| 7178 |
+
"learning_rate": 9.98243867056749e-06,
|
| 7179 |
+
"loss": 10.6203,
|
| 7180 |
+
"step": 10130
|
| 7181 |
+
},
|
| 7182 |
+
{
|
| 7183 |
+
"epoch": 0.4500141461547423,
|
| 7184 |
+
"grad_norm": 98.22434997558594,
|
| 7185 |
+
"learning_rate": 9.982421334605562e-06,
|
| 7186 |
+
"loss": 10.6234,
|
| 7187 |
+
"step": 10140
|
| 7188 |
+
},
|
| 7189 |
+
{
|
| 7190 |
+
"epoch": 0.45045794708783377,
|
| 7191 |
+
"grad_norm": 92.3149185180664,
|
| 7192 |
+
"learning_rate": 9.982403998643635e-06,
|
| 7193 |
+
"loss": 11.1096,
|
| 7194 |
+
"step": 10150
|
| 7195 |
+
},
|
| 7196 |
+
{
|
| 7197 |
+
"epoch": 0.4509017480209252,
|
| 7198 |
+
"grad_norm": 96.91551971435547,
|
| 7199 |
+
"learning_rate": 9.982386662681708e-06,
|
| 7200 |
+
"loss": 11.3593,
|
| 7201 |
+
"step": 10160
|
| 7202 |
+
},
|
| 7203 |
+
{
|
| 7204 |
+
"epoch": 0.4513455489540167,
|
| 7205 |
+
"grad_norm": 77.00688934326172,
|
| 7206 |
+
"learning_rate": 9.982369326719781e-06,
|
| 7207 |
+
"loss": 11.0639,
|
| 7208 |
+
"step": 10170
|
| 7209 |
+
},
|
| 7210 |
+
{
|
| 7211 |
+
"epoch": 0.4517893498871081,
|
| 7212 |
+
"grad_norm": 81.16304016113281,
|
| 7213 |
+
"learning_rate": 9.982351990757852e-06,
|
| 7214 |
+
"loss": 11.2663,
|
| 7215 |
+
"step": 10180
|
| 7216 |
+
},
|
| 7217 |
+
{
|
| 7218 |
+
"epoch": 0.4522331508201996,
|
| 7219 |
+
"grad_norm": 79.47703552246094,
|
| 7220 |
+
"learning_rate": 9.982334654795925e-06,
|
| 7221 |
+
"loss": 10.8026,
|
| 7222 |
+
"step": 10190
|
| 7223 |
+
},
|
| 7224 |
+
{
|
| 7225 |
+
"epoch": 0.45267695175329103,
|
| 7226 |
+
"grad_norm": 95.82781219482422,
|
| 7227 |
+
"learning_rate": 9.982317318833998e-06,
|
| 7228 |
+
"loss": 10.9343,
|
| 7229 |
+
"step": 10200
|
| 7230 |
+
},
|
| 7231 |
+
{
|
| 7232 |
+
"epoch": 0.4531207526863825,
|
| 7233 |
+
"grad_norm": 86.30982208251953,
|
| 7234 |
+
"learning_rate": 9.98229998287207e-06,
|
| 7235 |
+
"loss": 10.8559,
|
| 7236 |
+
"step": 10210
|
| 7237 |
+
},
|
| 7238 |
+
{
|
| 7239 |
+
"epoch": 0.453564553619474,
|
| 7240 |
+
"grad_norm": 84.82122802734375,
|
| 7241 |
+
"learning_rate": 9.982282646910143e-06,
|
| 7242 |
+
"loss": 11.2375,
|
| 7243 |
+
"step": 10220
|
| 7244 |
+
},
|
| 7245 |
+
{
|
| 7246 |
+
"epoch": 0.45400835455256544,
|
| 7247 |
+
"grad_norm": 85.73993682861328,
|
| 7248 |
+
"learning_rate": 9.982265310948216e-06,
|
| 7249 |
+
"loss": 11.1234,
|
| 7250 |
+
"step": 10230
|
| 7251 |
+
},
|
| 7252 |
+
{
|
| 7253 |
+
"epoch": 0.45445215548565693,
|
| 7254 |
+
"grad_norm": 76.42994689941406,
|
| 7255 |
+
"learning_rate": 9.982247974986287e-06,
|
| 7256 |
+
"loss": 11.2003,
|
| 7257 |
+
"step": 10240
|
| 7258 |
+
},
|
| 7259 |
+
{
|
| 7260 |
+
"epoch": 0.45489595641874836,
|
| 7261 |
+
"grad_norm": 84.67989349365234,
|
| 7262 |
+
"learning_rate": 9.98223063902436e-06,
|
| 7263 |
+
"loss": 10.8626,
|
| 7264 |
+
"step": 10250
|
| 7265 |
+
},
|
| 7266 |
+
{
|
| 7267 |
+
"epoch": 0.45533975735183985,
|
| 7268 |
+
"grad_norm": 97.71283721923828,
|
| 7269 |
+
"learning_rate": 9.982213303062434e-06,
|
| 7270 |
+
"loss": 10.8125,
|
| 7271 |
+
"step": 10260
|
| 7272 |
+
},
|
| 7273 |
+
{
|
| 7274 |
+
"epoch": 0.4557835582849313,
|
| 7275 |
+
"grad_norm": 77.48724365234375,
|
| 7276 |
+
"learning_rate": 9.982195967100505e-06,
|
| 7277 |
+
"loss": 10.9105,
|
| 7278 |
+
"step": 10270
|
| 7279 |
+
},
|
| 7280 |
+
{
|
| 7281 |
+
"epoch": 0.45622735921802277,
|
| 7282 |
+
"grad_norm": 73.06230926513672,
|
| 7283 |
+
"learning_rate": 9.982178631138578e-06,
|
| 7284 |
+
"loss": 10.8443,
|
| 7285 |
+
"step": 10280
|
| 7286 |
+
},
|
| 7287 |
+
{
|
| 7288 |
+
"epoch": 0.4566711601511142,
|
| 7289 |
+
"grad_norm": 78.53364562988281,
|
| 7290 |
+
"learning_rate": 9.982161295176651e-06,
|
| 7291 |
+
"loss": 11.1985,
|
| 7292 |
+
"step": 10290
|
| 7293 |
+
},
|
| 7294 |
+
{
|
| 7295 |
+
"epoch": 0.4571149610842057,
|
| 7296 |
+
"grad_norm": 71.37035369873047,
|
| 7297 |
+
"learning_rate": 9.982143959214722e-06,
|
| 7298 |
+
"loss": 11.2463,
|
| 7299 |
+
"step": 10300
|
| 7300 |
+
},
|
| 7301 |
+
{
|
| 7302 |
+
"epoch": 0.4575587620172971,
|
| 7303 |
+
"grad_norm": 76.0063247680664,
|
| 7304 |
+
"learning_rate": 9.982126623252796e-06,
|
| 7305 |
+
"loss": 10.9098,
|
| 7306 |
+
"step": 10310
|
| 7307 |
+
},
|
| 7308 |
+
{
|
| 7309 |
+
"epoch": 0.4580025629503886,
|
| 7310 |
+
"grad_norm": 82.43836212158203,
|
| 7311 |
+
"learning_rate": 9.982109287290869e-06,
|
| 7312 |
+
"loss": 10.8438,
|
| 7313 |
+
"step": 10320
|
| 7314 |
+
},
|
| 7315 |
+
{
|
| 7316 |
+
"epoch": 0.4584463638834801,
|
| 7317 |
+
"grad_norm": 88.15770721435547,
|
| 7318 |
+
"learning_rate": 9.98209195132894e-06,
|
| 7319 |
+
"loss": 11.2367,
|
| 7320 |
+
"step": 10330
|
| 7321 |
+
},
|
| 7322 |
+
{
|
| 7323 |
+
"epoch": 0.4588901648165715,
|
| 7324 |
+
"grad_norm": 78.84662628173828,
|
| 7325 |
+
"learning_rate": 9.982074615367013e-06,
|
| 7326 |
+
"loss": 10.8668,
|
| 7327 |
+
"step": 10340
|
| 7328 |
+
},
|
| 7329 |
+
{
|
| 7330 |
+
"epoch": 0.459333965749663,
|
| 7331 |
+
"grad_norm": 93.12088012695312,
|
| 7332 |
+
"learning_rate": 9.982057279405086e-06,
|
| 7333 |
+
"loss": 11.3541,
|
| 7334 |
+
"step": 10350
|
| 7335 |
+
},
|
| 7336 |
+
{
|
| 7337 |
+
"epoch": 0.45977776668275444,
|
| 7338 |
+
"grad_norm": 80.99417114257812,
|
| 7339 |
+
"learning_rate": 9.982039943443158e-06,
|
| 7340 |
+
"loss": 11.7431,
|
| 7341 |
+
"step": 10360
|
| 7342 |
+
},
|
| 7343 |
+
{
|
| 7344 |
+
"epoch": 0.46022156761584593,
|
| 7345 |
+
"grad_norm": 78.97187805175781,
|
| 7346 |
+
"learning_rate": 9.98202260748123e-06,
|
| 7347 |
+
"loss": 11.2636,
|
| 7348 |
+
"step": 10370
|
| 7349 |
+
},
|
| 7350 |
+
{
|
| 7351 |
+
"epoch": 0.46066536854893736,
|
| 7352 |
+
"grad_norm": 77.9780502319336,
|
| 7353 |
+
"learning_rate": 9.982005271519304e-06,
|
| 7354 |
+
"loss": 11.3237,
|
| 7355 |
+
"step": 10380
|
| 7356 |
+
},
|
| 7357 |
+
{
|
| 7358 |
+
"epoch": 0.46110916948202885,
|
| 7359 |
+
"grad_norm": 94.6609115600586,
|
| 7360 |
+
"learning_rate": 9.981987935557377e-06,
|
| 7361 |
+
"loss": 11.1363,
|
| 7362 |
+
"step": 10390
|
| 7363 |
+
},
|
| 7364 |
+
{
|
| 7365 |
+
"epoch": 0.4615529704151203,
|
| 7366 |
+
"grad_norm": 94.51988220214844,
|
| 7367 |
+
"learning_rate": 9.981970599595448e-06,
|
| 7368 |
+
"loss": 11.1418,
|
| 7369 |
+
"step": 10400
|
| 7370 |
+
},
|
| 7371 |
+
{
|
| 7372 |
+
"epoch": 0.46199677134821177,
|
| 7373 |
+
"grad_norm": 89.58201599121094,
|
| 7374 |
+
"learning_rate": 9.981953263633521e-06,
|
| 7375 |
+
"loss": 11.0321,
|
| 7376 |
+
"step": 10410
|
| 7377 |
+
},
|
| 7378 |
+
{
|
| 7379 |
+
"epoch": 0.4624405722813032,
|
| 7380 |
+
"grad_norm": 88.75037384033203,
|
| 7381 |
+
"learning_rate": 9.981935927671594e-06,
|
| 7382 |
+
"loss": 11.1615,
|
| 7383 |
+
"step": 10420
|
| 7384 |
+
},
|
| 7385 |
+
{
|
| 7386 |
+
"epoch": 0.4628843732143947,
|
| 7387 |
+
"grad_norm": 72.32737731933594,
|
| 7388 |
+
"learning_rate": 9.981918591709666e-06,
|
| 7389 |
+
"loss": 10.7545,
|
| 7390 |
+
"step": 10430
|
| 7391 |
+
},
|
| 7392 |
+
{
|
| 7393 |
+
"epoch": 0.46332817414748617,
|
| 7394 |
+
"grad_norm": 94.58203887939453,
|
| 7395 |
+
"learning_rate": 9.981901255747739e-06,
|
| 7396 |
+
"loss": 10.8112,
|
| 7397 |
+
"step": 10440
|
| 7398 |
+
},
|
| 7399 |
+
{
|
| 7400 |
+
"epoch": 0.4637719750805776,
|
| 7401 |
+
"grad_norm": 88.47208404541016,
|
| 7402 |
+
"learning_rate": 9.981883919785812e-06,
|
| 7403 |
+
"loss": 11.0896,
|
| 7404 |
+
"step": 10450
|
| 7405 |
+
},
|
| 7406 |
+
{
|
| 7407 |
+
"epoch": 0.4642157760136691,
|
| 7408 |
+
"grad_norm": 86.8807601928711,
|
| 7409 |
+
"learning_rate": 9.981866583823883e-06,
|
| 7410 |
+
"loss": 10.6616,
|
| 7411 |
+
"step": 10460
|
| 7412 |
+
},
|
| 7413 |
+
{
|
| 7414 |
+
"epoch": 0.4646595769467605,
|
| 7415 |
+
"grad_norm": 67.94696807861328,
|
| 7416 |
+
"learning_rate": 9.981849247861956e-06,
|
| 7417 |
+
"loss": 10.5705,
|
| 7418 |
+
"step": 10470
|
| 7419 |
+
},
|
| 7420 |
+
{
|
| 7421 |
+
"epoch": 0.465103377879852,
|
| 7422 |
+
"grad_norm": 83.57070922851562,
|
| 7423 |
+
"learning_rate": 9.98183191190003e-06,
|
| 7424 |
+
"loss": 11.4698,
|
| 7425 |
+
"step": 10480
|
| 7426 |
+
},
|
| 7427 |
+
{
|
| 7428 |
+
"epoch": 0.46554717881294344,
|
| 7429 |
+
"grad_norm": 83.02405548095703,
|
| 7430 |
+
"learning_rate": 9.9818145759381e-06,
|
| 7431 |
+
"loss": 10.5724,
|
| 7432 |
+
"step": 10490
|
| 7433 |
+
},
|
| 7434 |
+
{
|
| 7435 |
+
"epoch": 0.46599097974603493,
|
| 7436 |
+
"grad_norm": 83.82646942138672,
|
| 7437 |
+
"learning_rate": 9.981797239976174e-06,
|
| 7438 |
+
"loss": 11.0542,
|
| 7439 |
+
"step": 10500
|
| 7440 |
+
},
|
| 7441 |
+
{
|
| 7442 |
+
"epoch": 0.46643478067912636,
|
| 7443 |
+
"grad_norm": 77.04849243164062,
|
| 7444 |
+
"learning_rate": 9.981779904014247e-06,
|
| 7445 |
+
"loss": 11.3302,
|
| 7446 |
+
"step": 10510
|
| 7447 |
+
},
|
| 7448 |
+
{
|
| 7449 |
+
"epoch": 0.46687858161221785,
|
| 7450 |
+
"grad_norm": 74.8342514038086,
|
| 7451 |
+
"learning_rate": 9.981762568052318e-06,
|
| 7452 |
+
"loss": 11.5481,
|
| 7453 |
+
"step": 10520
|
| 7454 |
+
},
|
| 7455 |
+
{
|
| 7456 |
+
"epoch": 0.4673223825453093,
|
| 7457 |
+
"grad_norm": 76.42134094238281,
|
| 7458 |
+
"learning_rate": 9.981745232090391e-06,
|
| 7459 |
+
"loss": 10.9186,
|
| 7460 |
+
"step": 10530
|
| 7461 |
+
},
|
| 7462 |
+
{
|
| 7463 |
+
"epoch": 0.46776618347840077,
|
| 7464 |
+
"grad_norm": 82.68692779541016,
|
| 7465 |
+
"learning_rate": 9.981727896128464e-06,
|
| 7466 |
+
"loss": 11.2538,
|
| 7467 |
+
"step": 10540
|
| 7468 |
+
},
|
| 7469 |
+
{
|
| 7470 |
+
"epoch": 0.46820998441149225,
|
| 7471 |
+
"grad_norm": 69.0387191772461,
|
| 7472 |
+
"learning_rate": 9.981710560166536e-06,
|
| 7473 |
+
"loss": 10.8114,
|
| 7474 |
+
"step": 10550
|
| 7475 |
+
},
|
| 7476 |
+
{
|
| 7477 |
+
"epoch": 0.4686537853445837,
|
| 7478 |
+
"grad_norm": 103.88386535644531,
|
| 7479 |
+
"learning_rate": 9.981693224204609e-06,
|
| 7480 |
+
"loss": 10.9474,
|
| 7481 |
+
"step": 10560
|
| 7482 |
+
},
|
| 7483 |
+
{
|
| 7484 |
+
"epoch": 0.46909758627767517,
|
| 7485 |
+
"grad_norm": 88.8089370727539,
|
| 7486 |
+
"learning_rate": 9.981675888242682e-06,
|
| 7487 |
+
"loss": 10.8754,
|
| 7488 |
+
"step": 10570
|
| 7489 |
+
},
|
| 7490 |
+
{
|
| 7491 |
+
"epoch": 0.4695413872107666,
|
| 7492 |
+
"grad_norm": 79.1522216796875,
|
| 7493 |
+
"learning_rate": 9.981658552280753e-06,
|
| 7494 |
+
"loss": 11.2002,
|
| 7495 |
+
"step": 10580
|
| 7496 |
+
},
|
| 7497 |
+
{
|
| 7498 |
+
"epoch": 0.4699851881438581,
|
| 7499 |
+
"grad_norm": 75.01962280273438,
|
| 7500 |
+
"learning_rate": 9.981641216318826e-06,
|
| 7501 |
+
"loss": 10.5971,
|
| 7502 |
+
"step": 10590
|
| 7503 |
+
},
|
| 7504 |
+
{
|
| 7505 |
+
"epoch": 0.4704289890769495,
|
| 7506 |
+
"grad_norm": 88.03787994384766,
|
| 7507 |
+
"learning_rate": 9.9816238803569e-06,
|
| 7508 |
+
"loss": 11.2155,
|
| 7509 |
+
"step": 10600
|
| 7510 |
+
},
|
| 7511 |
+
{
|
| 7512 |
+
"epoch": 0.470872790010041,
|
| 7513 |
+
"grad_norm": 88.27069854736328,
|
| 7514 |
+
"learning_rate": 9.981606544394973e-06,
|
| 7515 |
+
"loss": 11.1796,
|
| 7516 |
+
"step": 10610
|
| 7517 |
+
},
|
| 7518 |
+
{
|
| 7519 |
+
"epoch": 0.47131659094313244,
|
| 7520 |
+
"grad_norm": 91.5178451538086,
|
| 7521 |
+
"learning_rate": 9.981589208433044e-06,
|
| 7522 |
+
"loss": 10.8964,
|
| 7523 |
+
"step": 10620
|
| 7524 |
+
},
|
| 7525 |
+
{
|
| 7526 |
+
"epoch": 0.47176039187622393,
|
| 7527 |
+
"grad_norm": 89.97875213623047,
|
| 7528 |
+
"learning_rate": 9.981571872471117e-06,
|
| 7529 |
+
"loss": 11.0617,
|
| 7530 |
+
"step": 10630
|
| 7531 |
+
},
|
| 7532 |
+
{
|
| 7533 |
+
"epoch": 0.47220419280931536,
|
| 7534 |
+
"grad_norm": 77.71656036376953,
|
| 7535 |
+
"learning_rate": 9.98155453650919e-06,
|
| 7536 |
+
"loss": 11.2671,
|
| 7537 |
+
"step": 10640
|
| 7538 |
+
},
|
| 7539 |
+
{
|
| 7540 |
+
"epoch": 0.47264799374240685,
|
| 7541 |
+
"grad_norm": 90.4183120727539,
|
| 7542 |
+
"learning_rate": 9.981537200547262e-06,
|
| 7543 |
+
"loss": 10.8842,
|
| 7544 |
+
"step": 10650
|
| 7545 |
+
},
|
| 7546 |
+
{
|
| 7547 |
+
"epoch": 0.4730917946754983,
|
| 7548 |
+
"grad_norm": 109.40415954589844,
|
| 7549 |
+
"learning_rate": 9.981519864585335e-06,
|
| 7550 |
+
"loss": 10.8791,
|
| 7551 |
+
"step": 10660
|
| 7552 |
+
},
|
| 7553 |
+
{
|
| 7554 |
+
"epoch": 0.47353559560858977,
|
| 7555 |
+
"grad_norm": 79.65886688232422,
|
| 7556 |
+
"learning_rate": 9.981502528623408e-06,
|
| 7557 |
+
"loss": 11.3155,
|
| 7558 |
+
"step": 10670
|
| 7559 |
+
},
|
| 7560 |
+
{
|
| 7561 |
+
"epoch": 0.47397939654168125,
|
| 7562 |
+
"grad_norm": 75.3977279663086,
|
| 7563 |
+
"learning_rate": 9.981485192661479e-06,
|
| 7564 |
+
"loss": 10.6279,
|
| 7565 |
+
"step": 10680
|
| 7566 |
+
},
|
| 7567 |
+
{
|
| 7568 |
+
"epoch": 0.4744231974747727,
|
| 7569 |
+
"grad_norm": 90.0768051147461,
|
| 7570 |
+
"learning_rate": 9.981467856699552e-06,
|
| 7571 |
+
"loss": 10.8386,
|
| 7572 |
+
"step": 10690
|
| 7573 |
+
},
|
| 7574 |
+
{
|
| 7575 |
+
"epoch": 0.47486699840786417,
|
| 7576 |
+
"grad_norm": 78.82632446289062,
|
| 7577 |
+
"learning_rate": 9.981450520737625e-06,
|
| 7578 |
+
"loss": 10.8152,
|
| 7579 |
+
"step": 10700
|
| 7580 |
+
},
|
| 7581 |
+
{
|
| 7582 |
+
"epoch": 0.4753107993409556,
|
| 7583 |
+
"grad_norm": 82.59823608398438,
|
| 7584 |
+
"learning_rate": 9.981433184775697e-06,
|
| 7585 |
+
"loss": 11.0986,
|
| 7586 |
+
"step": 10710
|
| 7587 |
+
},
|
| 7588 |
+
{
|
| 7589 |
+
"epoch": 0.4757546002740471,
|
| 7590 |
+
"grad_norm": 82.44734954833984,
|
| 7591 |
+
"learning_rate": 9.98141584881377e-06,
|
| 7592 |
+
"loss": 10.8024,
|
| 7593 |
+
"step": 10720
|
| 7594 |
+
},
|
| 7595 |
+
{
|
| 7596 |
+
"epoch": 0.4761984012071385,
|
| 7597 |
+
"grad_norm": 91.71231079101562,
|
| 7598 |
+
"learning_rate": 9.981398512851843e-06,
|
| 7599 |
+
"loss": 10.8372,
|
| 7600 |
+
"step": 10730
|
| 7601 |
+
},
|
| 7602 |
+
{
|
| 7603 |
+
"epoch": 0.47664220214023,
|
| 7604 |
+
"grad_norm": 87.06108093261719,
|
| 7605 |
+
"learning_rate": 9.981381176889914e-06,
|
| 7606 |
+
"loss": 11.0266,
|
| 7607 |
+
"step": 10740
|
| 7608 |
+
},
|
| 7609 |
+
{
|
| 7610 |
+
"epoch": 0.47708600307332144,
|
| 7611 |
+
"grad_norm": 90.88961029052734,
|
| 7612 |
+
"learning_rate": 9.981363840927987e-06,
|
| 7613 |
+
"loss": 10.607,
|
| 7614 |
+
"step": 10750
|
| 7615 |
+
},
|
| 7616 |
+
{
|
| 7617 |
+
"epoch": 0.47752980400641293,
|
| 7618 |
+
"grad_norm": 90.09719848632812,
|
| 7619 |
+
"learning_rate": 9.98134650496606e-06,
|
| 7620 |
+
"loss": 11.0468,
|
| 7621 |
+
"step": 10760
|
| 7622 |
+
},
|
| 7623 |
+
{
|
| 7624 |
+
"epoch": 0.47797360493950436,
|
| 7625 |
+
"grad_norm": 87.11450958251953,
|
| 7626 |
+
"learning_rate": 9.981329169004132e-06,
|
| 7627 |
+
"loss": 10.7198,
|
| 7628 |
+
"step": 10770
|
| 7629 |
+
},
|
| 7630 |
+
{
|
| 7631 |
+
"epoch": 0.47841740587259585,
|
| 7632 |
+
"grad_norm": 80.5800552368164,
|
| 7633 |
+
"learning_rate": 9.981311833042205e-06,
|
| 7634 |
+
"loss": 11.1395,
|
| 7635 |
+
"step": 10780
|
| 7636 |
+
},
|
| 7637 |
+
{
|
| 7638 |
+
"epoch": 0.47886120680568733,
|
| 7639 |
+
"grad_norm": 90.17820739746094,
|
| 7640 |
+
"learning_rate": 9.981294497080278e-06,
|
| 7641 |
+
"loss": 11.3689,
|
| 7642 |
+
"step": 10790
|
| 7643 |
+
},
|
| 7644 |
+
{
|
| 7645 |
+
"epoch": 0.47930500773877877,
|
| 7646 |
+
"grad_norm": 84.49495697021484,
|
| 7647 |
+
"learning_rate": 9.981277161118351e-06,
|
| 7648 |
+
"loss": 10.9592,
|
| 7649 |
+
"step": 10800
|
| 7650 |
+
},
|
| 7651 |
+
{
|
| 7652 |
+
"epoch": 0.47974880867187025,
|
| 7653 |
+
"grad_norm": 94.85502624511719,
|
| 7654 |
+
"learning_rate": 9.981259825156422e-06,
|
| 7655 |
+
"loss": 10.6913,
|
| 7656 |
+
"step": 10810
|
| 7657 |
+
},
|
| 7658 |
+
{
|
| 7659 |
+
"epoch": 0.4801926096049617,
|
| 7660 |
+
"grad_norm": 79.14407348632812,
|
| 7661 |
+
"learning_rate": 9.981242489194495e-06,
|
| 7662 |
+
"loss": 11.8454,
|
| 7663 |
+
"step": 10820
|
| 7664 |
+
},
|
| 7665 |
+
{
|
| 7666 |
+
"epoch": 0.48063641053805317,
|
| 7667 |
+
"grad_norm": 91.49024200439453,
|
| 7668 |
+
"learning_rate": 9.981225153232568e-06,
|
| 7669 |
+
"loss": 11.292,
|
| 7670 |
+
"step": 10830
|
| 7671 |
+
},
|
| 7672 |
+
{
|
| 7673 |
+
"epoch": 0.4810802114711446,
|
| 7674 |
+
"grad_norm": 97.75911712646484,
|
| 7675 |
+
"learning_rate": 9.98120781727064e-06,
|
| 7676 |
+
"loss": 11.3453,
|
| 7677 |
+
"step": 10840
|
| 7678 |
+
},
|
| 7679 |
+
{
|
| 7680 |
+
"epoch": 0.4815240124042361,
|
| 7681 |
+
"grad_norm": 76.58828735351562,
|
| 7682 |
+
"learning_rate": 9.981190481308713e-06,
|
| 7683 |
+
"loss": 11.1765,
|
| 7684 |
+
"step": 10850
|
| 7685 |
+
},
|
| 7686 |
+
{
|
| 7687 |
+
"epoch": 0.4819678133373275,
|
| 7688 |
+
"grad_norm": 79.92506408691406,
|
| 7689 |
+
"learning_rate": 9.981173145346786e-06,
|
| 7690 |
+
"loss": 10.868,
|
| 7691 |
+
"step": 10860
|
| 7692 |
+
},
|
| 7693 |
+
{
|
| 7694 |
+
"epoch": 0.482411614270419,
|
| 7695 |
+
"grad_norm": 101.02202606201172,
|
| 7696 |
+
"learning_rate": 9.981155809384857e-06,
|
| 7697 |
+
"loss": 11.0264,
|
| 7698 |
+
"step": 10870
|
| 7699 |
+
},
|
| 7700 |
+
{
|
| 7701 |
+
"epoch": 0.48285541520351044,
|
| 7702 |
+
"grad_norm": 81.69430541992188,
|
| 7703 |
+
"learning_rate": 9.98113847342293e-06,
|
| 7704 |
+
"loss": 11.2259,
|
| 7705 |
+
"step": 10880
|
| 7706 |
+
},
|
| 7707 |
+
{
|
| 7708 |
+
"epoch": 0.48329921613660193,
|
| 7709 |
+
"grad_norm": 86.8892822265625,
|
| 7710 |
+
"learning_rate": 9.981121137461004e-06,
|
| 7711 |
+
"loss": 11.0128,
|
| 7712 |
+
"step": 10890
|
| 7713 |
+
},
|
| 7714 |
+
{
|
| 7715 |
+
"epoch": 0.4837430170696934,
|
| 7716 |
+
"grad_norm": 72.10417938232422,
|
| 7717 |
+
"learning_rate": 9.981103801499075e-06,
|
| 7718 |
+
"loss": 11.3905,
|
| 7719 |
+
"step": 10900
|
| 7720 |
+
},
|
| 7721 |
+
{
|
| 7722 |
+
"epoch": 0.48418681800278485,
|
| 7723 |
+
"grad_norm": 85.03720092773438,
|
| 7724 |
+
"learning_rate": 9.981086465537148e-06,
|
| 7725 |
+
"loss": 11.1796,
|
| 7726 |
+
"step": 10910
|
| 7727 |
+
},
|
| 7728 |
+
{
|
| 7729 |
+
"epoch": 0.48463061893587633,
|
| 7730 |
+
"grad_norm": 88.04219818115234,
|
| 7731 |
+
"learning_rate": 9.981069129575221e-06,
|
| 7732 |
+
"loss": 11.2465,
|
| 7733 |
+
"step": 10920
|
| 7734 |
+
},
|
| 7735 |
+
{
|
| 7736 |
+
"epoch": 0.48507441986896777,
|
| 7737 |
+
"grad_norm": 73.39752960205078,
|
| 7738 |
+
"learning_rate": 9.981051793613294e-06,
|
| 7739 |
+
"loss": 10.9952,
|
| 7740 |
+
"step": 10930
|
| 7741 |
+
},
|
| 7742 |
+
{
|
| 7743 |
+
"epoch": 0.48551822080205925,
|
| 7744 |
+
"grad_norm": 91.22920227050781,
|
| 7745 |
+
"learning_rate": 9.981034457651366e-06,
|
| 7746 |
+
"loss": 10.9643,
|
| 7747 |
+
"step": 10940
|
| 7748 |
+
},
|
| 7749 |
+
{
|
| 7750 |
+
"epoch": 0.4859620217351507,
|
| 7751 |
+
"grad_norm": 67.8987045288086,
|
| 7752 |
+
"learning_rate": 9.981017121689439e-06,
|
| 7753 |
+
"loss": 11.1133,
|
| 7754 |
+
"step": 10950
|
| 7755 |
+
},
|
| 7756 |
+
{
|
| 7757 |
+
"epoch": 0.48640582266824217,
|
| 7758 |
+
"grad_norm": 78.0125732421875,
|
| 7759 |
+
"learning_rate": 9.980999785727512e-06,
|
| 7760 |
+
"loss": 11.5816,
|
| 7761 |
+
"step": 10960
|
| 7762 |
+
},
|
| 7763 |
+
{
|
| 7764 |
+
"epoch": 0.4868496236013336,
|
| 7765 |
+
"grad_norm": 90.03052520751953,
|
| 7766 |
+
"learning_rate": 9.980982449765583e-06,
|
| 7767 |
+
"loss": 11.2881,
|
| 7768 |
+
"step": 10970
|
| 7769 |
+
},
|
| 7770 |
+
{
|
| 7771 |
+
"epoch": 0.4872934245344251,
|
| 7772 |
+
"grad_norm": 90.72000122070312,
|
| 7773 |
+
"learning_rate": 9.980965113803656e-06,
|
| 7774 |
+
"loss": 11.1198,
|
| 7775 |
+
"step": 10980
|
| 7776 |
+
},
|
| 7777 |
+
{
|
| 7778 |
+
"epoch": 0.4877372254675165,
|
| 7779 |
+
"grad_norm": 72.88054656982422,
|
| 7780 |
+
"learning_rate": 9.98094777784173e-06,
|
| 7781 |
+
"loss": 11.3579,
|
| 7782 |
+
"step": 10990
|
| 7783 |
+
},
|
| 7784 |
+
{
|
| 7785 |
+
"epoch": 0.488181026400608,
|
| 7786 |
+
"grad_norm": 84.08674621582031,
|
| 7787 |
+
"learning_rate": 9.9809304418798e-06,
|
| 7788 |
+
"loss": 10.871,
|
| 7789 |
+
"step": 11000
|
| 7790 |
+
},
|
| 7791 |
+
{
|
| 7792 |
+
"epoch": 0.488181026400608,
|
| 7793 |
+
"eval_loss": 0.34528353810310364,
|
| 7794 |
+
"eval_runtime": 674.0977,
|
| 7795 |
+
"eval_samples_per_second": 1801.506,
|
| 7796 |
+
"eval_steps_per_second": 56.297,
|
| 7797 |
+
"step": 11000
|
| 7798 |
+
},
|
| 7799 |
+
{
|
| 7800 |
+
"epoch": 0.4886248273336995,
|
| 7801 |
+
"grad_norm": 74.27079010009766,
|
| 7802 |
+
"learning_rate": 9.980913105917874e-06,
|
| 7803 |
+
"loss": 11.5738,
|
| 7804 |
+
"step": 11010
|
| 7805 |
+
},
|
| 7806 |
+
{
|
| 7807 |
+
"epoch": 0.48906862826679093,
|
| 7808 |
+
"grad_norm": 79.84703063964844,
|
| 7809 |
+
"learning_rate": 9.980895769955947e-06,
|
| 7810 |
+
"loss": 10.8113,
|
| 7811 |
+
"step": 11020
|
| 7812 |
+
},
|
| 7813 |
+
{
|
| 7814 |
+
"epoch": 0.4895124291998824,
|
| 7815 |
+
"grad_norm": 78.84832000732422,
|
| 7816 |
+
"learning_rate": 9.980878433994018e-06,
|
| 7817 |
+
"loss": 11.1961,
|
| 7818 |
+
"step": 11030
|
| 7819 |
+
},
|
| 7820 |
+
{
|
| 7821 |
+
"epoch": 0.48995623013297385,
|
| 7822 |
+
"grad_norm": 77.97138977050781,
|
| 7823 |
+
"learning_rate": 9.980861098032091e-06,
|
| 7824 |
+
"loss": 10.747,
|
| 7825 |
+
"step": 11040
|
| 7826 |
+
},
|
| 7827 |
+
{
|
| 7828 |
+
"epoch": 0.49040003106606533,
|
| 7829 |
+
"grad_norm": 68.38355255126953,
|
| 7830 |
+
"learning_rate": 9.980843762070164e-06,
|
| 7831 |
+
"loss": 10.9581,
|
| 7832 |
+
"step": 11050
|
| 7833 |
+
},
|
| 7834 |
+
{
|
| 7835 |
+
"epoch": 0.49084383199915677,
|
| 7836 |
+
"grad_norm": 90.01298522949219,
|
| 7837 |
+
"learning_rate": 9.980826426108237e-06,
|
| 7838 |
+
"loss": 11.2473,
|
| 7839 |
+
"step": 11060
|
| 7840 |
+
},
|
| 7841 |
+
{
|
| 7842 |
+
"epoch": 0.49128763293224825,
|
| 7843 |
+
"grad_norm": 69.46513366699219,
|
| 7844 |
+
"learning_rate": 9.980809090146309e-06,
|
| 7845 |
+
"loss": 10.7865,
|
| 7846 |
+
"step": 11070
|
| 7847 |
+
},
|
| 7848 |
+
{
|
| 7849 |
+
"epoch": 0.4917314338653397,
|
| 7850 |
+
"grad_norm": 77.08251190185547,
|
| 7851 |
+
"learning_rate": 9.980791754184382e-06,
|
| 7852 |
+
"loss": 10.7613,
|
| 7853 |
+
"step": 11080
|
| 7854 |
+
},
|
| 7855 |
+
{
|
| 7856 |
+
"epoch": 0.49217523479843117,
|
| 7857 |
+
"grad_norm": 86.79558563232422,
|
| 7858 |
+
"learning_rate": 9.980774418222455e-06,
|
| 7859 |
+
"loss": 10.4478,
|
| 7860 |
+
"step": 11090
|
| 7861 |
+
},
|
| 7862 |
+
{
|
| 7863 |
+
"epoch": 0.4926190357315226,
|
| 7864 |
+
"grad_norm": 78.37332153320312,
|
| 7865 |
+
"learning_rate": 9.980757082260526e-06,
|
| 7866 |
+
"loss": 11.3574,
|
| 7867 |
+
"step": 11100
|
| 7868 |
+
},
|
| 7869 |
+
{
|
| 7870 |
+
"epoch": 0.4930628366646141,
|
| 7871 |
+
"grad_norm": 77.12301635742188,
|
| 7872 |
+
"learning_rate": 9.9807397462986e-06,
|
| 7873 |
+
"loss": 10.9563,
|
| 7874 |
+
"step": 11110
|
| 7875 |
+
},
|
| 7876 |
+
{
|
| 7877 |
+
"epoch": 0.4935066375977055,
|
| 7878 |
+
"grad_norm": 84.5125961303711,
|
| 7879 |
+
"learning_rate": 9.980722410336672e-06,
|
| 7880 |
+
"loss": 10.7285,
|
| 7881 |
+
"step": 11120
|
| 7882 |
+
},
|
| 7883 |
+
{
|
| 7884 |
+
"epoch": 0.493950438530797,
|
| 7885 |
+
"grad_norm": 72.76651763916016,
|
| 7886 |
+
"learning_rate": 9.980705074374744e-06,
|
| 7887 |
+
"loss": 10.8235,
|
| 7888 |
+
"step": 11130
|
| 7889 |
+
},
|
| 7890 |
+
{
|
| 7891 |
+
"epoch": 0.4943942394638885,
|
| 7892 |
+
"grad_norm": 65.90115356445312,
|
| 7893 |
+
"learning_rate": 9.980687738412817e-06,
|
| 7894 |
+
"loss": 11.2566,
|
| 7895 |
+
"step": 11140
|
| 7896 |
+
},
|
| 7897 |
+
{
|
| 7898 |
+
"epoch": 0.49483804039697993,
|
| 7899 |
+
"grad_norm": 93.98876190185547,
|
| 7900 |
+
"learning_rate": 9.98067040245089e-06,
|
| 7901 |
+
"loss": 11.1416,
|
| 7902 |
+
"step": 11150
|
| 7903 |
+
},
|
| 7904 |
+
{
|
| 7905 |
+
"epoch": 0.4952818413300714,
|
| 7906 |
+
"grad_norm": 84.59708404541016,
|
| 7907 |
+
"learning_rate": 9.980653066488961e-06,
|
| 7908 |
+
"loss": 10.9902,
|
| 7909 |
+
"step": 11160
|
| 7910 |
+
},
|
| 7911 |
+
{
|
| 7912 |
+
"epoch": 0.49572564226316285,
|
| 7913 |
+
"grad_norm": 82.25418090820312,
|
| 7914 |
+
"learning_rate": 9.980635730527034e-06,
|
| 7915 |
+
"loss": 10.9595,
|
| 7916 |
+
"step": 11170
|
| 7917 |
+
},
|
| 7918 |
+
{
|
| 7919 |
+
"epoch": 0.49616944319625433,
|
| 7920 |
+
"grad_norm": 77.34684753417969,
|
| 7921 |
+
"learning_rate": 9.980618394565108e-06,
|
| 7922 |
+
"loss": 11.4648,
|
| 7923 |
+
"step": 11180
|
| 7924 |
+
},
|
| 7925 |
+
{
|
| 7926 |
+
"epoch": 0.49661324412934577,
|
| 7927 |
+
"grad_norm": 93.93881225585938,
|
| 7928 |
+
"learning_rate": 9.98060105860318e-06,
|
| 7929 |
+
"loss": 10.8638,
|
| 7930 |
+
"step": 11190
|
| 7931 |
+
},
|
| 7932 |
+
{
|
| 7933 |
+
"epoch": 0.49705704506243725,
|
| 7934 |
+
"grad_norm": 70.26705932617188,
|
| 7935 |
+
"learning_rate": 9.980583722641252e-06,
|
| 7936 |
+
"loss": 10.9112,
|
| 7937 |
+
"step": 11200
|
| 7938 |
+
},
|
| 7939 |
+
{
|
| 7940 |
+
"epoch": 0.4975008459955287,
|
| 7941 |
+
"grad_norm": 82.66825103759766,
|
| 7942 |
+
"learning_rate": 9.980566386679325e-06,
|
| 7943 |
+
"loss": 11.5958,
|
| 7944 |
+
"step": 11210
|
| 7945 |
+
},
|
| 7946 |
+
{
|
| 7947 |
+
"epoch": 0.49794464692862017,
|
| 7948 |
+
"grad_norm": 86.82162475585938,
|
| 7949 |
+
"learning_rate": 9.980549050717398e-06,
|
| 7950 |
+
"loss": 10.8187,
|
| 7951 |
+
"step": 11220
|
| 7952 |
+
},
|
| 7953 |
+
{
|
| 7954 |
+
"epoch": 0.4983884478617116,
|
| 7955 |
+
"grad_norm": 79.9798355102539,
|
| 7956 |
+
"learning_rate": 9.98053171475547e-06,
|
| 7957 |
+
"loss": 11.2143,
|
| 7958 |
+
"step": 11230
|
| 7959 |
+
},
|
| 7960 |
+
{
|
| 7961 |
+
"epoch": 0.4988322487948031,
|
| 7962 |
+
"grad_norm": 99.23787689208984,
|
| 7963 |
+
"learning_rate": 9.980514378793543e-06,
|
| 7964 |
+
"loss": 10.9515,
|
| 7965 |
+
"step": 11240
|
| 7966 |
+
},
|
| 7967 |
+
{
|
| 7968 |
+
"epoch": 0.4992760497278946,
|
| 7969 |
+
"grad_norm": 70.37545013427734,
|
| 7970 |
+
"learning_rate": 9.980497042831616e-06,
|
| 7971 |
+
"loss": 11.0358,
|
| 7972 |
+
"step": 11250
|
| 7973 |
+
},
|
| 7974 |
+
{
|
| 7975 |
+
"epoch": 0.499719850660986,
|
| 7976 |
+
"grad_norm": 91.39663696289062,
|
| 7977 |
+
"learning_rate": 9.980479706869687e-06,
|
| 7978 |
+
"loss": 10.7406,
|
| 7979 |
+
"step": 11260
|
| 7980 |
+
},
|
| 7981 |
+
{
|
| 7982 |
+
"epoch": 0.5001636515940775,
|
| 7983 |
+
"grad_norm": 81.30348205566406,
|
| 7984 |
+
"learning_rate": 9.98046237090776e-06,
|
| 7985 |
+
"loss": 11.2646,
|
| 7986 |
+
"step": 11270
|
| 7987 |
+
},
|
| 7988 |
+
{
|
| 7989 |
+
"epoch": 0.500607452527169,
|
| 7990 |
+
"grad_norm": 83.41373443603516,
|
| 7991 |
+
"learning_rate": 9.980445034945833e-06,
|
| 7992 |
+
"loss": 10.5715,
|
| 7993 |
+
"step": 11280
|
| 7994 |
+
},
|
| 7995 |
+
{
|
| 7996 |
+
"epoch": 0.5010512534602604,
|
| 7997 |
+
"grad_norm": 84.33602142333984,
|
| 7998 |
+
"learning_rate": 9.980427698983906e-06,
|
| 7999 |
+
"loss": 10.9119,
|
| 8000 |
+
"step": 11290
|
| 8001 |
+
},
|
| 8002 |
+
{
|
| 8003 |
+
"epoch": 0.5014950543933518,
|
| 8004 |
+
"grad_norm": 76.0499038696289,
|
| 8005 |
+
"learning_rate": 9.980410363021978e-06,
|
| 8006 |
+
"loss": 11.1319,
|
| 8007 |
+
"step": 11300
|
| 8008 |
+
},
|
| 8009 |
+
{
|
| 8010 |
+
"epoch": 0.5019388553264433,
|
| 8011 |
+
"grad_norm": 87.45086669921875,
|
| 8012 |
+
"learning_rate": 9.98039302706005e-06,
|
| 8013 |
+
"loss": 10.3676,
|
| 8014 |
+
"step": 11310
|
| 8015 |
+
},
|
| 8016 |
+
{
|
| 8017 |
+
"epoch": 0.5023826562595348,
|
| 8018 |
+
"grad_norm": 88.60616302490234,
|
| 8019 |
+
"learning_rate": 9.980375691098124e-06,
|
| 8020 |
+
"loss": 11.0152,
|
| 8021 |
+
"step": 11320
|
| 8022 |
+
},
|
| 8023 |
+
{
|
| 8024 |
+
"epoch": 0.5028264571926262,
|
| 8025 |
+
"grad_norm": 91.6775894165039,
|
| 8026 |
+
"learning_rate": 9.980358355136195e-06,
|
| 8027 |
+
"loss": 10.8913,
|
| 8028 |
+
"step": 11330
|
| 8029 |
+
},
|
| 8030 |
+
{
|
| 8031 |
+
"epoch": 0.5032702581257177,
|
| 8032 |
+
"grad_norm": 83.63784790039062,
|
| 8033 |
+
"learning_rate": 9.980341019174268e-06,
|
| 8034 |
+
"loss": 10.4357,
|
| 8035 |
+
"step": 11340
|
| 8036 |
+
},
|
| 8037 |
+
{
|
| 8038 |
+
"epoch": 0.5037140590588092,
|
| 8039 |
+
"grad_norm": 93.07415008544922,
|
| 8040 |
+
"learning_rate": 9.980323683212341e-06,
|
| 8041 |
+
"loss": 10.3792,
|
| 8042 |
+
"step": 11350
|
| 8043 |
+
},
|
| 8044 |
+
{
|
| 8045 |
+
"epoch": 0.5041578599919007,
|
| 8046 |
+
"grad_norm": 88.19844818115234,
|
| 8047 |
+
"learning_rate": 9.980306347250413e-06,
|
| 8048 |
+
"loss": 10.886,
|
| 8049 |
+
"step": 11360
|
| 8050 |
+
},
|
| 8051 |
+
{
|
| 8052 |
+
"epoch": 0.5046016609249921,
|
| 8053 |
+
"grad_norm": 79.55142211914062,
|
| 8054 |
+
"learning_rate": 9.980289011288486e-06,
|
| 8055 |
+
"loss": 10.8876,
|
| 8056 |
+
"step": 11370
|
| 8057 |
+
},
|
| 8058 |
+
{
|
| 8059 |
+
"epoch": 0.5050454618580835,
|
| 8060 |
+
"grad_norm": 85.38150787353516,
|
| 8061 |
+
"learning_rate": 9.980271675326559e-06,
|
| 8062 |
+
"loss": 11.1223,
|
| 8063 |
+
"step": 11380
|
| 8064 |
+
},
|
| 8065 |
+
{
|
| 8066 |
+
"epoch": 0.505489262791175,
|
| 8067 |
+
"grad_norm": 69.45784759521484,
|
| 8068 |
+
"learning_rate": 9.98025433936463e-06,
|
| 8069 |
+
"loss": 10.688,
|
| 8070 |
+
"step": 11390
|
| 8071 |
+
},
|
| 8072 |
+
{
|
| 8073 |
+
"epoch": 0.5059330637242665,
|
| 8074 |
+
"grad_norm": 83.27287292480469,
|
| 8075 |
+
"learning_rate": 9.980237003402703e-06,
|
| 8076 |
+
"loss": 11.482,
|
| 8077 |
+
"step": 11400
|
| 8078 |
+
},
|
| 8079 |
+
{
|
| 8080 |
+
"epoch": 0.506376864657358,
|
| 8081 |
+
"grad_norm": 87.00122833251953,
|
| 8082 |
+
"learning_rate": 9.980219667440777e-06,
|
| 8083 |
+
"loss": 11.2689,
|
| 8084 |
+
"step": 11410
|
| 8085 |
+
},
|
| 8086 |
+
{
|
| 8087 |
+
"epoch": 0.5068206655904494,
|
| 8088 |
+
"grad_norm": 78.08293914794922,
|
| 8089 |
+
"learning_rate": 9.98020233147885e-06,
|
| 8090 |
+
"loss": 10.4533,
|
| 8091 |
+
"step": 11420
|
| 8092 |
+
},
|
| 8093 |
+
{
|
| 8094 |
+
"epoch": 0.5072644665235408,
|
| 8095 |
+
"grad_norm": 71.37157440185547,
|
| 8096 |
+
"learning_rate": 9.980184995516921e-06,
|
| 8097 |
+
"loss": 10.856,
|
| 8098 |
+
"step": 11430
|
| 8099 |
+
},
|
| 8100 |
+
{
|
| 8101 |
+
"epoch": 0.5077082674566323,
|
| 8102 |
+
"grad_norm": 78.55634307861328,
|
| 8103 |
+
"learning_rate": 9.980167659554994e-06,
|
| 8104 |
+
"loss": 11.0349,
|
| 8105 |
+
"step": 11440
|
| 8106 |
+
},
|
| 8107 |
+
{
|
| 8108 |
+
"epoch": 0.5081520683897238,
|
| 8109 |
+
"grad_norm": 83.2050552368164,
|
| 8110 |
+
"learning_rate": 9.980150323593067e-06,
|
| 8111 |
+
"loss": 11.0373,
|
| 8112 |
+
"step": 11450
|
| 8113 |
+
},
|
| 8114 |
+
{
|
| 8115 |
+
"epoch": 0.5085958693228152,
|
| 8116 |
+
"grad_norm": 94.2475357055664,
|
| 8117 |
+
"learning_rate": 9.980132987631139e-06,
|
| 8118 |
+
"loss": 10.5705,
|
| 8119 |
+
"step": 11460
|
| 8120 |
+
},
|
| 8121 |
+
{
|
| 8122 |
+
"epoch": 0.5090396702559067,
|
| 8123 |
+
"grad_norm": 75.6529769897461,
|
| 8124 |
+
"learning_rate": 9.980115651669212e-06,
|
| 8125 |
+
"loss": 10.8084,
|
| 8126 |
+
"step": 11470
|
| 8127 |
+
},
|
| 8128 |
+
{
|
| 8129 |
+
"epoch": 0.5094834711889982,
|
| 8130 |
+
"grad_norm": 68.75302124023438,
|
| 8131 |
+
"learning_rate": 9.980098315707285e-06,
|
| 8132 |
+
"loss": 10.4888,
|
| 8133 |
+
"step": 11480
|
| 8134 |
+
},
|
| 8135 |
+
{
|
| 8136 |
+
"epoch": 0.5099272721220897,
|
| 8137 |
+
"grad_norm": 88.40451049804688,
|
| 8138 |
+
"learning_rate": 9.980080979745356e-06,
|
| 8139 |
+
"loss": 11.4719,
|
| 8140 |
+
"step": 11490
|
| 8141 |
+
},
|
| 8142 |
+
{
|
| 8143 |
+
"epoch": 0.5103710730551811,
|
| 8144 |
+
"grad_norm": 85.7581558227539,
|
| 8145 |
+
"learning_rate": 9.980063643783429e-06,
|
| 8146 |
+
"loss": 11.1014,
|
| 8147 |
+
"step": 11500
|
| 8148 |
+
},
|
| 8149 |
+
{
|
| 8150 |
+
"epoch": 0.5108148739882725,
|
| 8151 |
+
"grad_norm": 79.09970092773438,
|
| 8152 |
+
"learning_rate": 9.980046307821502e-06,
|
| 8153 |
+
"loss": 11.0976,
|
| 8154 |
+
"step": 11510
|
| 8155 |
+
},
|
| 8156 |
+
{
|
| 8157 |
+
"epoch": 0.511258674921364,
|
| 8158 |
+
"grad_norm": 83.1121597290039,
|
| 8159 |
+
"learning_rate": 9.980028971859574e-06,
|
| 8160 |
+
"loss": 10.8814,
|
| 8161 |
+
"step": 11520
|
| 8162 |
+
},
|
| 8163 |
+
{
|
| 8164 |
+
"epoch": 0.5117024758544555,
|
| 8165 |
+
"grad_norm": 77.43012237548828,
|
| 8166 |
+
"learning_rate": 9.980011635897647e-06,
|
| 8167 |
+
"loss": 10.775,
|
| 8168 |
+
"step": 11530
|
| 8169 |
+
},
|
| 8170 |
+
{
|
| 8171 |
+
"epoch": 0.512146276787547,
|
| 8172 |
+
"grad_norm": 68.88916778564453,
|
| 8173 |
+
"learning_rate": 9.97999429993572e-06,
|
| 8174 |
+
"loss": 11.1833,
|
| 8175 |
+
"step": 11540
|
| 8176 |
+
},
|
| 8177 |
+
{
|
| 8178 |
+
"epoch": 0.5125900777206384,
|
| 8179 |
+
"grad_norm": 70.91609954833984,
|
| 8180 |
+
"learning_rate": 9.979976963973793e-06,
|
| 8181 |
+
"loss": 10.6464,
|
| 8182 |
+
"step": 11550
|
| 8183 |
+
},
|
| 8184 |
+
{
|
| 8185 |
+
"epoch": 0.5130338786537298,
|
| 8186 |
+
"grad_norm": 83.58845520019531,
|
| 8187 |
+
"learning_rate": 9.979959628011864e-06,
|
| 8188 |
+
"loss": 10.9326,
|
| 8189 |
+
"step": 11560
|
| 8190 |
+
},
|
| 8191 |
+
{
|
| 8192 |
+
"epoch": 0.5134776795868213,
|
| 8193 |
+
"grad_norm": 81.35511016845703,
|
| 8194 |
+
"learning_rate": 9.979942292049937e-06,
|
| 8195 |
+
"loss": 11.163,
|
| 8196 |
+
"step": 11570
|
| 8197 |
+
},
|
| 8198 |
+
{
|
| 8199 |
+
"epoch": 0.5139214805199128,
|
| 8200 |
+
"grad_norm": 77.6694564819336,
|
| 8201 |
+
"learning_rate": 9.97992495608801e-06,
|
| 8202 |
+
"loss": 10.8365,
|
| 8203 |
+
"step": 11580
|
| 8204 |
+
},
|
| 8205 |
+
{
|
| 8206 |
+
"epoch": 0.5143652814530043,
|
| 8207 |
+
"grad_norm": 87.32294464111328,
|
| 8208 |
+
"learning_rate": 9.979907620126082e-06,
|
| 8209 |
+
"loss": 10.8808,
|
| 8210 |
+
"step": 11590
|
| 8211 |
+
},
|
| 8212 |
+
{
|
| 8213 |
+
"epoch": 0.5148090823860957,
|
| 8214 |
+
"grad_norm": 79.58880615234375,
|
| 8215 |
+
"learning_rate": 9.979890284164155e-06,
|
| 8216 |
+
"loss": 10.7793,
|
| 8217 |
+
"step": 11600
|
| 8218 |
+
},
|
| 8219 |
+
{
|
| 8220 |
+
"epoch": 0.5152528833191872,
|
| 8221 |
+
"grad_norm": 70.23893737792969,
|
| 8222 |
+
"learning_rate": 9.979872948202228e-06,
|
| 8223 |
+
"loss": 10.9845,
|
| 8224 |
+
"step": 11610
|
| 8225 |
+
},
|
| 8226 |
+
{
|
| 8227 |
+
"epoch": 0.5156966842522787,
|
| 8228 |
+
"grad_norm": 72.99483489990234,
|
| 8229 |
+
"learning_rate": 9.9798556122403e-06,
|
| 8230 |
+
"loss": 10.8218,
|
| 8231 |
+
"step": 11620
|
| 8232 |
+
},
|
| 8233 |
+
{
|
| 8234 |
+
"epoch": 0.5161404851853701,
|
| 8235 |
+
"grad_norm": 96.56175994873047,
|
| 8236 |
+
"learning_rate": 9.979838276278372e-06,
|
| 8237 |
+
"loss": 11.0247,
|
| 8238 |
+
"step": 11630
|
| 8239 |
+
},
|
| 8240 |
+
{
|
| 8241 |
+
"epoch": 0.5165842861184615,
|
| 8242 |
+
"grad_norm": 76.31138610839844,
|
| 8243 |
+
"learning_rate": 9.979820940316445e-06,
|
| 8244 |
+
"loss": 11.0729,
|
| 8245 |
+
"step": 11640
|
| 8246 |
+
},
|
| 8247 |
+
{
|
| 8248 |
+
"epoch": 0.517028087051553,
|
| 8249 |
+
"grad_norm": 77.18230438232422,
|
| 8250 |
+
"learning_rate": 9.979803604354517e-06,
|
| 8251 |
+
"loss": 10.3969,
|
| 8252 |
+
"step": 11650
|
| 8253 |
+
},
|
| 8254 |
+
{
|
| 8255 |
+
"epoch": 0.5174718879846445,
|
| 8256 |
+
"grad_norm": 90.09929656982422,
|
| 8257 |
+
"learning_rate": 9.97978626839259e-06,
|
| 8258 |
+
"loss": 11.3239,
|
| 8259 |
+
"step": 11660
|
| 8260 |
+
},
|
| 8261 |
+
{
|
| 8262 |
+
"epoch": 0.517915688917736,
|
| 8263 |
+
"grad_norm": 69.90628051757812,
|
| 8264 |
+
"learning_rate": 9.979768932430663e-06,
|
| 8265 |
+
"loss": 10.7297,
|
| 8266 |
+
"step": 11670
|
| 8267 |
+
},
|
| 8268 |
+
{
|
| 8269 |
+
"epoch": 0.5183594898508274,
|
| 8270 |
+
"grad_norm": 90.91302490234375,
|
| 8271 |
+
"learning_rate": 9.979751596468736e-06,
|
| 8272 |
+
"loss": 11.3096,
|
| 8273 |
+
"step": 11680
|
| 8274 |
+
},
|
| 8275 |
+
{
|
| 8276 |
+
"epoch": 0.5188032907839188,
|
| 8277 |
+
"grad_norm": 81.82889556884766,
|
| 8278 |
+
"learning_rate": 9.979734260506807e-06,
|
| 8279 |
+
"loss": 10.9807,
|
| 8280 |
+
"step": 11690
|
| 8281 |
+
},
|
| 8282 |
+
{
|
| 8283 |
+
"epoch": 0.5192470917170103,
|
| 8284 |
+
"grad_norm": 86.98174285888672,
|
| 8285 |
+
"learning_rate": 9.97971692454488e-06,
|
| 8286 |
+
"loss": 11.2847,
|
| 8287 |
+
"step": 11700
|
| 8288 |
+
},
|
| 8289 |
+
{
|
| 8290 |
+
"epoch": 0.5196908926501018,
|
| 8291 |
+
"grad_norm": 72.4999771118164,
|
| 8292 |
+
"learning_rate": 9.979699588582954e-06,
|
| 8293 |
+
"loss": 10.7933,
|
| 8294 |
+
"step": 11710
|
| 8295 |
+
},
|
| 8296 |
+
{
|
| 8297 |
+
"epoch": 0.5201346935831933,
|
| 8298 |
+
"grad_norm": 82.73175048828125,
|
| 8299 |
+
"learning_rate": 9.979682252621025e-06,
|
| 8300 |
+
"loss": 10.8406,
|
| 8301 |
+
"step": 11720
|
| 8302 |
+
},
|
| 8303 |
+
{
|
| 8304 |
+
"epoch": 0.5205784945162847,
|
| 8305 |
+
"grad_norm": 68.61962127685547,
|
| 8306 |
+
"learning_rate": 9.979664916659098e-06,
|
| 8307 |
+
"loss": 10.7406,
|
| 8308 |
+
"step": 11730
|
| 8309 |
+
},
|
| 8310 |
+
{
|
| 8311 |
+
"epoch": 0.5210222954493762,
|
| 8312 |
+
"grad_norm": 80.85415649414062,
|
| 8313 |
+
"learning_rate": 9.979647580697171e-06,
|
| 8314 |
+
"loss": 11.0003,
|
| 8315 |
+
"step": 11740
|
| 8316 |
+
},
|
| 8317 |
+
{
|
| 8318 |
+
"epoch": 0.5214660963824677,
|
| 8319 |
+
"grad_norm": 81.90907287597656,
|
| 8320 |
+
"learning_rate": 9.979630244735243e-06,
|
| 8321 |
+
"loss": 11.4761,
|
| 8322 |
+
"step": 11750
|
| 8323 |
+
},
|
| 8324 |
+
{
|
| 8325 |
+
"epoch": 0.5219098973155591,
|
| 8326 |
+
"grad_norm": 70.7921142578125,
|
| 8327 |
+
"learning_rate": 9.979612908773316e-06,
|
| 8328 |
+
"loss": 10.7498,
|
| 8329 |
+
"step": 11760
|
| 8330 |
+
},
|
| 8331 |
+
{
|
| 8332 |
+
"epoch": 0.5223536982486505,
|
| 8333 |
+
"grad_norm": 74.50489044189453,
|
| 8334 |
+
"learning_rate": 9.979595572811389e-06,
|
| 8335 |
+
"loss": 11.7802,
|
| 8336 |
+
"step": 11770
|
| 8337 |
+
},
|
| 8338 |
+
{
|
| 8339 |
+
"epoch": 0.522797499181742,
|
| 8340 |
+
"grad_norm": 75.52880859375,
|
| 8341 |
+
"learning_rate": 9.97957823684946e-06,
|
| 8342 |
+
"loss": 10.7195,
|
| 8343 |
+
"step": 11780
|
| 8344 |
+
},
|
| 8345 |
+
{
|
| 8346 |
+
"epoch": 0.5232413001148335,
|
| 8347 |
+
"grad_norm": 80.11971282958984,
|
| 8348 |
+
"learning_rate": 9.979560900887533e-06,
|
| 8349 |
+
"loss": 10.747,
|
| 8350 |
+
"step": 11790
|
| 8351 |
+
},
|
| 8352 |
+
{
|
| 8353 |
+
"epoch": 0.523685101047925,
|
| 8354 |
+
"grad_norm": 84.3338851928711,
|
| 8355 |
+
"learning_rate": 9.979543564925606e-06,
|
| 8356 |
+
"loss": 11.132,
|
| 8357 |
+
"step": 11800
|
| 8358 |
+
},
|
| 8359 |
+
{
|
| 8360 |
+
"epoch": 0.5241289019810164,
|
| 8361 |
+
"grad_norm": 80.16667938232422,
|
| 8362 |
+
"learning_rate": 9.979526228963678e-06,
|
| 8363 |
+
"loss": 11.2093,
|
| 8364 |
+
"step": 11810
|
| 8365 |
+
},
|
| 8366 |
+
{
|
| 8367 |
+
"epoch": 0.5245727029141078,
|
| 8368 |
+
"grad_norm": 70.78595733642578,
|
| 8369 |
+
"learning_rate": 9.97950889300175e-06,
|
| 8370 |
+
"loss": 10.7076,
|
| 8371 |
+
"step": 11820
|
| 8372 |
+
},
|
| 8373 |
+
{
|
| 8374 |
+
"epoch": 0.5250165038471993,
|
| 8375 |
+
"grad_norm": 75.8795394897461,
|
| 8376 |
+
"learning_rate": 9.979491557039824e-06,
|
| 8377 |
+
"loss": 10.7308,
|
| 8378 |
+
"step": 11830
|
| 8379 |
+
},
|
| 8380 |
+
{
|
| 8381 |
+
"epoch": 0.5254603047802908,
|
| 8382 |
+
"grad_norm": 87.91172790527344,
|
| 8383 |
+
"learning_rate": 9.979474221077895e-06,
|
| 8384 |
+
"loss": 10.3702,
|
| 8385 |
+
"step": 11840
|
| 8386 |
+
},
|
| 8387 |
+
{
|
| 8388 |
+
"epoch": 0.5259041057133823,
|
| 8389 |
+
"grad_norm": 86.1435317993164,
|
| 8390 |
+
"learning_rate": 9.979456885115968e-06,
|
| 8391 |
+
"loss": 10.6837,
|
| 8392 |
+
"step": 11850
|
| 8393 |
+
},
|
| 8394 |
+
{
|
| 8395 |
+
"epoch": 0.5263479066464737,
|
| 8396 |
+
"grad_norm": 74.28446960449219,
|
| 8397 |
+
"learning_rate": 9.979439549154041e-06,
|
| 8398 |
+
"loss": 11.3425,
|
| 8399 |
+
"step": 11860
|
| 8400 |
+
},
|
| 8401 |
+
{
|
| 8402 |
+
"epoch": 0.5267917075795652,
|
| 8403 |
+
"grad_norm": 77.42523956298828,
|
| 8404 |
+
"learning_rate": 9.979422213192113e-06,
|
| 8405 |
+
"loss": 10.7451,
|
| 8406 |
+
"step": 11870
|
| 8407 |
+
},
|
| 8408 |
+
{
|
| 8409 |
+
"epoch": 0.5272355085126567,
|
| 8410 |
+
"grad_norm": 71.55403900146484,
|
| 8411 |
+
"learning_rate": 9.979404877230186e-06,
|
| 8412 |
+
"loss": 11.2173,
|
| 8413 |
+
"step": 11880
|
| 8414 |
+
},
|
| 8415 |
+
{
|
| 8416 |
+
"epoch": 0.5276793094457481,
|
| 8417 |
+
"grad_norm": 79.82381439208984,
|
| 8418 |
+
"learning_rate": 9.979387541268259e-06,
|
| 8419 |
+
"loss": 10.9905,
|
| 8420 |
+
"step": 11890
|
| 8421 |
+
},
|
| 8422 |
+
{
|
| 8423 |
+
"epoch": 0.5281231103788395,
|
| 8424 |
+
"grad_norm": 90.72064208984375,
|
| 8425 |
+
"learning_rate": 9.979370205306332e-06,
|
| 8426 |
+
"loss": 10.8288,
|
| 8427 |
+
"step": 11900
|
| 8428 |
+
},
|
| 8429 |
+
{
|
| 8430 |
+
"epoch": 0.528566911311931,
|
| 8431 |
+
"grad_norm": 68.74069213867188,
|
| 8432 |
+
"learning_rate": 9.979352869344403e-06,
|
| 8433 |
+
"loss": 11.3552,
|
| 8434 |
+
"step": 11910
|
| 8435 |
+
},
|
| 8436 |
+
{
|
| 8437 |
+
"epoch": 0.5290107122450225,
|
| 8438 |
+
"grad_norm": 79.70250701904297,
|
| 8439 |
+
"learning_rate": 9.979335533382476e-06,
|
| 8440 |
+
"loss": 11.0605,
|
| 8441 |
+
"step": 11920
|
| 8442 |
+
},
|
| 8443 |
+
{
|
| 8444 |
+
"epoch": 0.529454513178114,
|
| 8445 |
+
"grad_norm": 77.91078186035156,
|
| 8446 |
+
"learning_rate": 9.97931819742055e-06,
|
| 8447 |
+
"loss": 10.8511,
|
| 8448 |
+
"step": 11930
|
| 8449 |
+
},
|
| 8450 |
+
{
|
| 8451 |
+
"epoch": 0.5298983141112055,
|
| 8452 |
+
"grad_norm": 75.19290924072266,
|
| 8453 |
+
"learning_rate": 9.979300861458621e-06,
|
| 8454 |
+
"loss": 11.1562,
|
| 8455 |
+
"step": 11940
|
| 8456 |
+
},
|
| 8457 |
+
{
|
| 8458 |
+
"epoch": 0.5303421150442968,
|
| 8459 |
+
"grad_norm": 86.28694152832031,
|
| 8460 |
+
"learning_rate": 9.979283525496694e-06,
|
| 8461 |
+
"loss": 10.9518,
|
| 8462 |
+
"step": 11950
|
| 8463 |
+
},
|
| 8464 |
+
{
|
| 8465 |
+
"epoch": 0.5307859159773883,
|
| 8466 |
+
"grad_norm": 83.85566711425781,
|
| 8467 |
+
"learning_rate": 9.979266189534767e-06,
|
| 8468 |
+
"loss": 10.5831,
|
| 8469 |
+
"step": 11960
|
| 8470 |
+
},
|
| 8471 |
+
{
|
| 8472 |
+
"epoch": 0.5312297169104798,
|
| 8473 |
+
"grad_norm": 89.48451232910156,
|
| 8474 |
+
"learning_rate": 9.979248853572838e-06,
|
| 8475 |
+
"loss": 11.1542,
|
| 8476 |
+
"step": 11970
|
| 8477 |
+
},
|
| 8478 |
+
{
|
| 8479 |
+
"epoch": 0.5316735178435713,
|
| 8480 |
+
"grad_norm": 74.22183227539062,
|
| 8481 |
+
"learning_rate": 9.979231517610911e-06,
|
| 8482 |
+
"loss": 11.742,
|
| 8483 |
+
"step": 11980
|
| 8484 |
+
},
|
| 8485 |
+
{
|
| 8486 |
+
"epoch": 0.5321173187766627,
|
| 8487 |
+
"grad_norm": 74.32493591308594,
|
| 8488 |
+
"learning_rate": 9.979214181648985e-06,
|
| 8489 |
+
"loss": 11.5062,
|
| 8490 |
+
"step": 11990
|
| 8491 |
+
},
|
| 8492 |
+
{
|
| 8493 |
+
"epoch": 0.5325611197097542,
|
| 8494 |
+
"grad_norm": 84.4752197265625,
|
| 8495 |
+
"learning_rate": 9.979196845687056e-06,
|
| 8496 |
+
"loss": 11.24,
|
| 8497 |
+
"step": 12000
|
| 8498 |
+
},
|
| 8499 |
+
{
|
| 8500 |
+
"epoch": 0.5325611197097542,
|
| 8501 |
+
"eval_loss": 0.3403577506542206,
|
| 8502 |
+
"eval_runtime": 674.8568,
|
| 8503 |
+
"eval_samples_per_second": 1799.479,
|
| 8504 |
+
"eval_steps_per_second": 56.234,
|
| 8505 |
+
"step": 12000
|
| 8506 |
+
},
|
| 8507 |
+
{
|
| 8508 |
+
"epoch": 0.5330049206428457,
|
| 8509 |
+
"grad_norm": 70.7219467163086,
|
| 8510 |
+
"learning_rate": 9.979179509725129e-06,
|
| 8511 |
+
"loss": 10.7596,
|
| 8512 |
+
"step": 12010
|
| 8513 |
+
},
|
| 8514 |
+
{
|
| 8515 |
+
"epoch": 0.5334487215759371,
|
| 8516 |
+
"grad_norm": 81.48099517822266,
|
| 8517 |
+
"learning_rate": 9.979162173763202e-06,
|
| 8518 |
+
"loss": 10.9401,
|
| 8519 |
+
"step": 12020
|
| 8520 |
+
},
|
| 8521 |
+
{
|
| 8522 |
+
"epoch": 0.5338925225090285,
|
| 8523 |
+
"grad_norm": 90.189208984375,
|
| 8524 |
+
"learning_rate": 9.979144837801273e-06,
|
| 8525 |
+
"loss": 11.0275,
|
| 8526 |
+
"step": 12030
|
| 8527 |
+
},
|
| 8528 |
+
{
|
| 8529 |
+
"epoch": 0.53433632344212,
|
| 8530 |
+
"grad_norm": 77.25494384765625,
|
| 8531 |
+
"learning_rate": 9.979127501839347e-06,
|
| 8532 |
+
"loss": 10.5871,
|
| 8533 |
+
"step": 12040
|
| 8534 |
+
},
|
| 8535 |
+
{
|
| 8536 |
+
"epoch": 0.5347801243752115,
|
| 8537 |
+
"grad_norm": 83.02628326416016,
|
| 8538 |
+
"learning_rate": 9.97911016587742e-06,
|
| 8539 |
+
"loss": 11.6188,
|
| 8540 |
+
"step": 12050
|
| 8541 |
+
},
|
| 8542 |
+
{
|
| 8543 |
+
"epoch": 0.535223925308303,
|
| 8544 |
+
"grad_norm": 73.51893615722656,
|
| 8545 |
+
"learning_rate": 9.979092829915491e-06,
|
| 8546 |
+
"loss": 11.0726,
|
| 8547 |
+
"step": 12060
|
| 8548 |
+
},
|
| 8549 |
+
{
|
| 8550 |
+
"epoch": 0.5356677262413945,
|
| 8551 |
+
"grad_norm": 84.49578857421875,
|
| 8552 |
+
"learning_rate": 9.979075493953564e-06,
|
| 8553 |
+
"loss": 11.4908,
|
| 8554 |
+
"step": 12070
|
| 8555 |
+
},
|
| 8556 |
+
{
|
| 8557 |
+
"epoch": 0.5361115271744858,
|
| 8558 |
+
"grad_norm": 71.82061767578125,
|
| 8559 |
+
"learning_rate": 9.979058157991637e-06,
|
| 8560 |
+
"loss": 10.8268,
|
| 8561 |
+
"step": 12080
|
| 8562 |
+
},
|
| 8563 |
+
{
|
| 8564 |
+
"epoch": 0.5365553281075773,
|
| 8565 |
+
"grad_norm": 79.56192016601562,
|
| 8566 |
+
"learning_rate": 9.979040822029709e-06,
|
| 8567 |
+
"loss": 10.6541,
|
| 8568 |
+
"step": 12090
|
| 8569 |
+
},
|
| 8570 |
+
{
|
| 8571 |
+
"epoch": 0.5369991290406688,
|
| 8572 |
+
"grad_norm": 92.98332977294922,
|
| 8573 |
+
"learning_rate": 9.979023486067782e-06,
|
| 8574 |
+
"loss": 10.8328,
|
| 8575 |
+
"step": 12100
|
| 8576 |
+
},
|
| 8577 |
+
{
|
| 8578 |
+
"epoch": 0.5374429299737603,
|
| 8579 |
+
"grad_norm": 78.83235931396484,
|
| 8580 |
+
"learning_rate": 9.979006150105855e-06,
|
| 8581 |
+
"loss": 10.7446,
|
| 8582 |
+
"step": 12110
|
| 8583 |
+
},
|
| 8584 |
+
{
|
| 8585 |
+
"epoch": 0.5378867309068517,
|
| 8586 |
+
"grad_norm": 85.79434204101562,
|
| 8587 |
+
"learning_rate": 9.978988814143928e-06,
|
| 8588 |
+
"loss": 11.2323,
|
| 8589 |
+
"step": 12120
|
| 8590 |
+
},
|
| 8591 |
+
{
|
| 8592 |
+
"epoch": 0.5383305318399432,
|
| 8593 |
+
"grad_norm": 74.76625061035156,
|
| 8594 |
+
"learning_rate": 9.978971478182e-06,
|
| 8595 |
+
"loss": 11.2963,
|
| 8596 |
+
"step": 12130
|
| 8597 |
+
},
|
| 8598 |
+
{
|
| 8599 |
+
"epoch": 0.5387743327730347,
|
| 8600 |
+
"grad_norm": 75.79556274414062,
|
| 8601 |
+
"learning_rate": 9.978954142220072e-06,
|
| 8602 |
+
"loss": 11.4193,
|
| 8603 |
+
"step": 12140
|
| 8604 |
+
},
|
| 8605 |
+
{
|
| 8606 |
+
"epoch": 0.5392181337061261,
|
| 8607 |
+
"grad_norm": 71.63494110107422,
|
| 8608 |
+
"learning_rate": 9.978936806258145e-06,
|
| 8609 |
+
"loss": 10.7985,
|
| 8610 |
+
"step": 12150
|
| 8611 |
+
},
|
| 8612 |
+
{
|
| 8613 |
+
"epoch": 0.5396619346392176,
|
| 8614 |
+
"grad_norm": 85.59138488769531,
|
| 8615 |
+
"learning_rate": 9.978919470296217e-06,
|
| 8616 |
+
"loss": 11.4485,
|
| 8617 |
+
"step": 12160
|
| 8618 |
+
},
|
| 8619 |
+
{
|
| 8620 |
+
"epoch": 0.540105735572309,
|
| 8621 |
+
"grad_norm": 76.133544921875,
|
| 8622 |
+
"learning_rate": 9.97890213433429e-06,
|
| 8623 |
+
"loss": 11.4676,
|
| 8624 |
+
"step": 12170
|
| 8625 |
+
},
|
| 8626 |
+
{
|
| 8627 |
+
"epoch": 0.5405495365054005,
|
| 8628 |
+
"grad_norm": 74.8232650756836,
|
| 8629 |
+
"learning_rate": 9.978884798372363e-06,
|
| 8630 |
+
"loss": 10.934,
|
| 8631 |
+
"step": 12180
|
| 8632 |
+
},
|
| 8633 |
+
{
|
| 8634 |
+
"epoch": 0.540993337438492,
|
| 8635 |
+
"grad_norm": 82.42066955566406,
|
| 8636 |
+
"learning_rate": 9.978867462410434e-06,
|
| 8637 |
+
"loss": 11.0111,
|
| 8638 |
+
"step": 12190
|
| 8639 |
+
},
|
| 8640 |
+
{
|
| 8641 |
+
"epoch": 0.5414371383715835,
|
| 8642 |
+
"grad_norm": 80.84729766845703,
|
| 8643 |
+
"learning_rate": 9.978850126448507e-06,
|
| 8644 |
+
"loss": 10.4483,
|
| 8645 |
+
"step": 12200
|
| 8646 |
+
},
|
| 8647 |
+
{
|
| 8648 |
+
"epoch": 0.5418809393046748,
|
| 8649 |
+
"grad_norm": 83.81796264648438,
|
| 8650 |
+
"learning_rate": 9.97883279048658e-06,
|
| 8651 |
+
"loss": 10.8341,
|
| 8652 |
+
"step": 12210
|
| 8653 |
+
},
|
| 8654 |
+
{
|
| 8655 |
+
"epoch": 0.5423247402377663,
|
| 8656 |
+
"grad_norm": 73.58375549316406,
|
| 8657 |
+
"learning_rate": 9.978815454524652e-06,
|
| 8658 |
+
"loss": 10.5816,
|
| 8659 |
+
"step": 12220
|
| 8660 |
+
},
|
| 8661 |
+
{
|
| 8662 |
+
"epoch": 0.5427685411708578,
|
| 8663 |
+
"grad_norm": 94.62816619873047,
|
| 8664 |
+
"learning_rate": 9.978798118562725e-06,
|
| 8665 |
+
"loss": 11.0264,
|
| 8666 |
+
"step": 12230
|
| 8667 |
+
},
|
| 8668 |
+
{
|
| 8669 |
+
"epoch": 0.5432123421039493,
|
| 8670 |
+
"grad_norm": 80.29084014892578,
|
| 8671 |
+
"learning_rate": 9.978780782600798e-06,
|
| 8672 |
+
"loss": 11.1186,
|
| 8673 |
+
"step": 12240
|
| 8674 |
+
},
|
| 8675 |
+
{
|
| 8676 |
+
"epoch": 0.5436561430370407,
|
| 8677 |
+
"grad_norm": 86.72004699707031,
|
| 8678 |
+
"learning_rate": 9.97876344663887e-06,
|
| 8679 |
+
"loss": 10.6926,
|
| 8680 |
+
"step": 12250
|
| 8681 |
+
},
|
| 8682 |
+
{
|
| 8683 |
+
"epoch": 0.5440999439701322,
|
| 8684 |
+
"grad_norm": 78.45811462402344,
|
| 8685 |
+
"learning_rate": 9.978746110676942e-06,
|
| 8686 |
+
"loss": 10.6757,
|
| 8687 |
+
"step": 12260
|
| 8688 |
+
},
|
| 8689 |
+
{
|
| 8690 |
+
"epoch": 0.5445437449032237,
|
| 8691 |
+
"grad_norm": 81.1561050415039,
|
| 8692 |
+
"learning_rate": 9.978728774715015e-06,
|
| 8693 |
+
"loss": 10.9724,
|
| 8694 |
+
"step": 12270
|
| 8695 |
+
},
|
| 8696 |
+
{
|
| 8697 |
+
"epoch": 0.5449875458363151,
|
| 8698 |
+
"grad_norm": 81.57537841796875,
|
| 8699 |
+
"learning_rate": 9.978711438753087e-06,
|
| 8700 |
+
"loss": 10.9973,
|
| 8701 |
+
"step": 12280
|
| 8702 |
+
},
|
| 8703 |
+
{
|
| 8704 |
+
"epoch": 0.5454313467694066,
|
| 8705 |
+
"grad_norm": 71.61268615722656,
|
| 8706 |
+
"learning_rate": 9.97869410279116e-06,
|
| 8707 |
+
"loss": 10.961,
|
| 8708 |
+
"step": 12290
|
| 8709 |
+
},
|
| 8710 |
+
{
|
| 8711 |
+
"epoch": 0.545875147702498,
|
| 8712 |
+
"grad_norm": 69.37332153320312,
|
| 8713 |
+
"learning_rate": 9.978676766829233e-06,
|
| 8714 |
+
"loss": 10.6944,
|
| 8715 |
+
"step": 12300
|
| 8716 |
+
},
|
| 8717 |
+
{
|
| 8718 |
+
"epoch": 0.5463189486355895,
|
| 8719 |
+
"grad_norm": 75.34327697753906,
|
| 8720 |
+
"learning_rate": 9.978659430867304e-06,
|
| 8721 |
+
"loss": 10.9383,
|
| 8722 |
+
"step": 12310
|
| 8723 |
+
},
|
| 8724 |
+
{
|
| 8725 |
+
"epoch": 0.546762749568681,
|
| 8726 |
+
"grad_norm": 81.64777374267578,
|
| 8727 |
+
"learning_rate": 9.978642094905377e-06,
|
| 8728 |
+
"loss": 10.8668,
|
| 8729 |
+
"step": 12320
|
| 8730 |
+
},
|
| 8731 |
+
{
|
| 8732 |
+
"epoch": 0.5472065505017725,
|
| 8733 |
+
"grad_norm": 77.13945770263672,
|
| 8734 |
+
"learning_rate": 9.97862475894345e-06,
|
| 8735 |
+
"loss": 10.5414,
|
| 8736 |
+
"step": 12330
|
| 8737 |
+
},
|
| 8738 |
+
{
|
| 8739 |
+
"epoch": 0.5476503514348638,
|
| 8740 |
+
"grad_norm": 68.74790954589844,
|
| 8741 |
+
"learning_rate": 9.978607422981524e-06,
|
| 8742 |
+
"loss": 10.5574,
|
| 8743 |
+
"step": 12340
|
| 8744 |
+
},
|
| 8745 |
+
{
|
| 8746 |
+
"epoch": 0.5480941523679553,
|
| 8747 |
+
"grad_norm": 73.5103988647461,
|
| 8748 |
+
"learning_rate": 9.978590087019595e-06,
|
| 8749 |
+
"loss": 10.6581,
|
| 8750 |
+
"step": 12350
|
| 8751 |
+
},
|
| 8752 |
+
{
|
| 8753 |
+
"epoch": 0.5485379533010468,
|
| 8754 |
+
"grad_norm": 74.06155395507812,
|
| 8755 |
+
"learning_rate": 9.978572751057668e-06,
|
| 8756 |
+
"loss": 10.9724,
|
| 8757 |
+
"step": 12360
|
| 8758 |
+
},
|
| 8759 |
+
{
|
| 8760 |
+
"epoch": 0.5489817542341383,
|
| 8761 |
+
"grad_norm": 83.41978454589844,
|
| 8762 |
+
"learning_rate": 9.978555415095741e-06,
|
| 8763 |
+
"loss": 10.8189,
|
| 8764 |
+
"step": 12370
|
| 8765 |
+
},
|
| 8766 |
+
{
|
| 8767 |
+
"epoch": 0.5494255551672297,
|
| 8768 |
+
"grad_norm": 75.58580780029297,
|
| 8769 |
+
"learning_rate": 9.978538079133813e-06,
|
| 8770 |
+
"loss": 10.9859,
|
| 8771 |
+
"step": 12380
|
| 8772 |
+
},
|
| 8773 |
+
{
|
| 8774 |
+
"epoch": 0.5498693561003212,
|
| 8775 |
+
"grad_norm": 88.49089050292969,
|
| 8776 |
+
"learning_rate": 9.978520743171886e-06,
|
| 8777 |
+
"loss": 10.7513,
|
| 8778 |
+
"step": 12390
|
| 8779 |
+
},
|
| 8780 |
+
{
|
| 8781 |
+
"epoch": 0.5503131570334127,
|
| 8782 |
+
"grad_norm": 58.82784652709961,
|
| 8783 |
+
"learning_rate": 9.978503407209959e-06,
|
| 8784 |
+
"loss": 10.5162,
|
| 8785 |
+
"step": 12400
|
| 8786 |
+
},
|
| 8787 |
+
{
|
| 8788 |
+
"epoch": 0.5507569579665041,
|
| 8789 |
+
"grad_norm": 83.45179748535156,
|
| 8790 |
+
"learning_rate": 9.97848607124803e-06,
|
| 8791 |
+
"loss": 10.8963,
|
| 8792 |
+
"step": 12410
|
| 8793 |
+
},
|
| 8794 |
+
{
|
| 8795 |
+
"epoch": 0.5512007588995956,
|
| 8796 |
+
"grad_norm": 78.52918243408203,
|
| 8797 |
+
"learning_rate": 9.978468735286103e-06,
|
| 8798 |
+
"loss": 11.3171,
|
| 8799 |
+
"step": 12420
|
| 8800 |
+
},
|
| 8801 |
+
{
|
| 8802 |
+
"epoch": 0.551644559832687,
|
| 8803 |
+
"grad_norm": 92.01652526855469,
|
| 8804 |
+
"learning_rate": 9.978451399324176e-06,
|
| 8805 |
+
"loss": 10.9977,
|
| 8806 |
+
"step": 12430
|
| 8807 |
+
},
|
| 8808 |
+
{
|
| 8809 |
+
"epoch": 0.5520883607657785,
|
| 8810 |
+
"grad_norm": 81.95563507080078,
|
| 8811 |
+
"learning_rate": 9.978434063362248e-06,
|
| 8812 |
+
"loss": 10.8333,
|
| 8813 |
+
"step": 12440
|
| 8814 |
+
},
|
| 8815 |
+
{
|
| 8816 |
+
"epoch": 0.55253216169887,
|
| 8817 |
+
"grad_norm": 77.15538024902344,
|
| 8818 |
+
"learning_rate": 9.97841672740032e-06,
|
| 8819 |
+
"loss": 11.1091,
|
| 8820 |
+
"step": 12450
|
| 8821 |
+
},
|
| 8822 |
+
{
|
| 8823 |
+
"epoch": 0.5529759626319615,
|
| 8824 |
+
"grad_norm": 77.65177917480469,
|
| 8825 |
+
"learning_rate": 9.978399391438394e-06,
|
| 8826 |
+
"loss": 10.9557,
|
| 8827 |
+
"step": 12460
|
| 8828 |
+
},
|
| 8829 |
+
{
|
| 8830 |
+
"epoch": 0.5534197635650528,
|
| 8831 |
+
"grad_norm": 78.06676483154297,
|
| 8832 |
+
"learning_rate": 9.978382055476465e-06,
|
| 8833 |
+
"loss": 10.6794,
|
| 8834 |
+
"step": 12470
|
| 8835 |
+
},
|
| 8836 |
+
{
|
| 8837 |
+
"epoch": 0.5538635644981443,
|
| 8838 |
+
"grad_norm": 74.32503509521484,
|
| 8839 |
+
"learning_rate": 9.978364719514538e-06,
|
| 8840 |
+
"loss": 11.1013,
|
| 8841 |
+
"step": 12480
|
| 8842 |
+
},
|
| 8843 |
+
{
|
| 8844 |
+
"epoch": 0.5543073654312358,
|
| 8845 |
+
"grad_norm": 81.63748931884766,
|
| 8846 |
+
"learning_rate": 9.978347383552611e-06,
|
| 8847 |
+
"loss": 11.074,
|
| 8848 |
+
"step": 12490
|
| 8849 |
+
},
|
| 8850 |
+
{
|
| 8851 |
+
"epoch": 0.5547511663643273,
|
| 8852 |
+
"grad_norm": 81.70726776123047,
|
| 8853 |
+
"learning_rate": 9.978330047590683e-06,
|
| 8854 |
+
"loss": 10.3868,
|
| 8855 |
+
"step": 12500
|
| 8856 |
}
|
| 8857 |
],
|
| 8858 |
"logging_steps": 10,
|
|
|
|
| 8872 |
"attributes": {}
|
| 8873 |
}
|
| 8874 |
},
|
| 8875 |
+
"total_flos": 4.36215400628224e+18,
|
| 8876 |
"train_batch_size": 4,
|
| 8877 |
"trial_name": null,
|
| 8878 |
"trial_params": null
|