Training in progress, step 7500, 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:b0c398bc843eea803ee6d700d1537de79fdc32bede5d65c5c9f86d67c40a71de
|
| 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 sha256:09e62b87c8cdb6fbdf1ca09dea3f1e41c6f59a2ccbb87b7b025db5eb51b5fa0d
|
| 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 sha256:04cb5208648fd09a2e0403d51973f74ffbfd93cbd5da59e1e99c8df03769a86c
|
| 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 sha256:7034685b36b93a4dd3a50697b0b1c314b249b2189ec2cb96b757312b1514a579
|
| 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 sha256:6e851fe1c1de0057f4eecefed6a131fa9021334eb43f6e7e65fdb270a25ac864
|
| 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 sha256:978379030048e432baa510ec4fc9514faa08fe564ab964b3a4d05e8f60306495
|
| 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:bdbc75d90af112615b53d15931e8157a80e37bcd110aac9a3089f5f6f5344171
|
| 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 sha256:8c8a310f6ca2ca89570eb2cc68544656b30224f00b2d6d96eeda6e0cb8be50ab
|
| 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 sha256:7c5b8110fcf6e044b6860c6305be969cfe03129549b92dc6fc2394448e9265d6
|
| 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 sha256:6f936acaf5a2d5fe8c38d945450417facbf1577584c216908a396d3cc20bec88
|
| 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 sha256:7c9ee2be288a50938aa76c672a598bafcd789d6a5d6e08c069ef8e7d474b5cd2
|
| 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,
|
|
@@ -3547,6 +3547,1772 @@
|
|
| 3547 |
"eval_samples_per_second": 1798.056,
|
| 3548 |
"eval_steps_per_second": 56.19,
|
| 3549 |
"step": 5000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3550 |
}
|
| 3551 |
],
|
| 3552 |
"logging_steps": 10,
|
|
@@ -3566,7 +5332,7 @@
|
|
| 3566 |
"attributes": {}
|
| 3567 |
}
|
| 3568 |
},
|
| 3569 |
-
"total_flos":
|
| 3570 |
"train_batch_size": 4,
|
| 3571 |
"trial_name": null,
|
| 3572 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.3328506998185964,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 7500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3547 |
"eval_samples_per_second": 1798.056,
|
| 3548 |
"eval_steps_per_second": 56.19,
|
| 3549 |
"step": 5000
|
| 3550 |
+
},
|
| 3551 |
+
{
|
| 3552 |
+
"epoch": 0.22234426747882238,
|
| 3553 |
+
"grad_norm": 100.71528625488281,
|
| 3554 |
+
"learning_rate": 9.991314683074347e-06,
|
| 3555 |
+
"loss": 11.8638,
|
| 3556 |
+
"step": 5010
|
| 3557 |
+
},
|
| 3558 |
+
{
|
| 3559 |
+
"epoch": 0.22278806841191384,
|
| 3560 |
+
"grad_norm": 114.36324310302734,
|
| 3561 |
+
"learning_rate": 9.991297347112418e-06,
|
| 3562 |
+
"loss": 12.7659,
|
| 3563 |
+
"step": 5020
|
| 3564 |
+
},
|
| 3565 |
+
{
|
| 3566 |
+
"epoch": 0.2232318693450053,
|
| 3567 |
+
"grad_norm": 123.63908386230469,
|
| 3568 |
+
"learning_rate": 9.991280011150491e-06,
|
| 3569 |
+
"loss": 12.2795,
|
| 3570 |
+
"step": 5030
|
| 3571 |
+
},
|
| 3572 |
+
{
|
| 3573 |
+
"epoch": 0.22367567027809676,
|
| 3574 |
+
"grad_norm": 120.35507202148438,
|
| 3575 |
+
"learning_rate": 9.991262675188565e-06,
|
| 3576 |
+
"loss": 12.277,
|
| 3577 |
+
"step": 5040
|
| 3578 |
+
},
|
| 3579 |
+
{
|
| 3580 |
+
"epoch": 0.22411947121118822,
|
| 3581 |
+
"grad_norm": 109.37151336669922,
|
| 3582 |
+
"learning_rate": 9.991245339226636e-06,
|
| 3583 |
+
"loss": 11.6699,
|
| 3584 |
+
"step": 5050
|
| 3585 |
+
},
|
| 3586 |
+
{
|
| 3587 |
+
"epoch": 0.22456327214427968,
|
| 3588 |
+
"grad_norm": 105.27191925048828,
|
| 3589 |
+
"learning_rate": 9.991228003264709e-06,
|
| 3590 |
+
"loss": 12.0232,
|
| 3591 |
+
"step": 5060
|
| 3592 |
+
},
|
| 3593 |
+
{
|
| 3594 |
+
"epoch": 0.22500707307737114,
|
| 3595 |
+
"grad_norm": 129.59083557128906,
|
| 3596 |
+
"learning_rate": 9.991210667302782e-06,
|
| 3597 |
+
"loss": 11.7139,
|
| 3598 |
+
"step": 5070
|
| 3599 |
+
},
|
| 3600 |
+
{
|
| 3601 |
+
"epoch": 0.2254508740104626,
|
| 3602 |
+
"grad_norm": 102.64170837402344,
|
| 3603 |
+
"learning_rate": 9.991193331340853e-06,
|
| 3604 |
+
"loss": 12.3,
|
| 3605 |
+
"step": 5080
|
| 3606 |
+
},
|
| 3607 |
+
{
|
| 3608 |
+
"epoch": 0.22589467494355406,
|
| 3609 |
+
"grad_norm": 97.96053314208984,
|
| 3610 |
+
"learning_rate": 9.991175995378927e-06,
|
| 3611 |
+
"loss": 12.2807,
|
| 3612 |
+
"step": 5090
|
| 3613 |
+
},
|
| 3614 |
+
{
|
| 3615 |
+
"epoch": 0.22633847587664552,
|
| 3616 |
+
"grad_norm": 111.13671112060547,
|
| 3617 |
+
"learning_rate": 9.991158659417e-06,
|
| 3618 |
+
"loss": 12.2035,
|
| 3619 |
+
"step": 5100
|
| 3620 |
+
},
|
| 3621 |
+
{
|
| 3622 |
+
"epoch": 0.226782276809737,
|
| 3623 |
+
"grad_norm": 127.20590209960938,
|
| 3624 |
+
"learning_rate": 9.991141323455071e-06,
|
| 3625 |
+
"loss": 11.7232,
|
| 3626 |
+
"step": 5110
|
| 3627 |
+
},
|
| 3628 |
+
{
|
| 3629 |
+
"epoch": 0.22722607774282846,
|
| 3630 |
+
"grad_norm": 128.55152893066406,
|
| 3631 |
+
"learning_rate": 9.991123987493144e-06,
|
| 3632 |
+
"loss": 12.2339,
|
| 3633 |
+
"step": 5120
|
| 3634 |
+
},
|
| 3635 |
+
{
|
| 3636 |
+
"epoch": 0.22766987867591992,
|
| 3637 |
+
"grad_norm": 152.3295135498047,
|
| 3638 |
+
"learning_rate": 9.991106651531217e-06,
|
| 3639 |
+
"loss": 12.0261,
|
| 3640 |
+
"step": 5130
|
| 3641 |
+
},
|
| 3642 |
+
{
|
| 3643 |
+
"epoch": 0.22811367960901138,
|
| 3644 |
+
"grad_norm": 125.02354431152344,
|
| 3645 |
+
"learning_rate": 9.99108931556929e-06,
|
| 3646 |
+
"loss": 11.9501,
|
| 3647 |
+
"step": 5140
|
| 3648 |
+
},
|
| 3649 |
+
{
|
| 3650 |
+
"epoch": 0.22855748054210284,
|
| 3651 |
+
"grad_norm": 126.0276107788086,
|
| 3652 |
+
"learning_rate": 9.991071979607362e-06,
|
| 3653 |
+
"loss": 11.9311,
|
| 3654 |
+
"step": 5150
|
| 3655 |
+
},
|
| 3656 |
+
{
|
| 3657 |
+
"epoch": 0.2290012814751943,
|
| 3658 |
+
"grad_norm": 101.98572540283203,
|
| 3659 |
+
"learning_rate": 9.991054643645435e-06,
|
| 3660 |
+
"loss": 11.7741,
|
| 3661 |
+
"step": 5160
|
| 3662 |
+
},
|
| 3663 |
+
{
|
| 3664 |
+
"epoch": 0.22944508240828576,
|
| 3665 |
+
"grad_norm": 93.39753723144531,
|
| 3666 |
+
"learning_rate": 9.991037307683508e-06,
|
| 3667 |
+
"loss": 12.3775,
|
| 3668 |
+
"step": 5170
|
| 3669 |
+
},
|
| 3670 |
+
{
|
| 3671 |
+
"epoch": 0.22988888334137722,
|
| 3672 |
+
"grad_norm": 119.41414642333984,
|
| 3673 |
+
"learning_rate": 9.99101997172158e-06,
|
| 3674 |
+
"loss": 12.3944,
|
| 3675 |
+
"step": 5180
|
| 3676 |
+
},
|
| 3677 |
+
{
|
| 3678 |
+
"epoch": 0.23033268427446868,
|
| 3679 |
+
"grad_norm": 102.53070068359375,
|
| 3680 |
+
"learning_rate": 9.991002635759652e-06,
|
| 3681 |
+
"loss": 11.7407,
|
| 3682 |
+
"step": 5190
|
| 3683 |
+
},
|
| 3684 |
+
{
|
| 3685 |
+
"epoch": 0.23077648520756014,
|
| 3686 |
+
"grad_norm": 114.39293670654297,
|
| 3687 |
+
"learning_rate": 9.990985299797725e-06,
|
| 3688 |
+
"loss": 12.0433,
|
| 3689 |
+
"step": 5200
|
| 3690 |
+
},
|
| 3691 |
+
{
|
| 3692 |
+
"epoch": 0.2312202861406516,
|
| 3693 |
+
"grad_norm": 95.0862808227539,
|
| 3694 |
+
"learning_rate": 9.990967963835797e-06,
|
| 3695 |
+
"loss": 11.8278,
|
| 3696 |
+
"step": 5210
|
| 3697 |
+
},
|
| 3698 |
+
{
|
| 3699 |
+
"epoch": 0.23166408707374309,
|
| 3700 |
+
"grad_norm": 95.9500732421875,
|
| 3701 |
+
"learning_rate": 9.99095062787387e-06,
|
| 3702 |
+
"loss": 12.0699,
|
| 3703 |
+
"step": 5220
|
| 3704 |
+
},
|
| 3705 |
+
{
|
| 3706 |
+
"epoch": 0.23210788800683455,
|
| 3707 |
+
"grad_norm": 108.53816223144531,
|
| 3708 |
+
"learning_rate": 9.990933291911943e-06,
|
| 3709 |
+
"loss": 12.0234,
|
| 3710 |
+
"step": 5230
|
| 3711 |
+
},
|
| 3712 |
+
{
|
| 3713 |
+
"epoch": 0.232551688939926,
|
| 3714 |
+
"grad_norm": 108.30989837646484,
|
| 3715 |
+
"learning_rate": 9.990915955950014e-06,
|
| 3716 |
+
"loss": 11.899,
|
| 3717 |
+
"step": 5240
|
| 3718 |
+
},
|
| 3719 |
+
{
|
| 3720 |
+
"epoch": 0.23299548987301746,
|
| 3721 |
+
"grad_norm": 103.17509460449219,
|
| 3722 |
+
"learning_rate": 9.990898619988087e-06,
|
| 3723 |
+
"loss": 12.0308,
|
| 3724 |
+
"step": 5250
|
| 3725 |
+
},
|
| 3726 |
+
{
|
| 3727 |
+
"epoch": 0.23343929080610892,
|
| 3728 |
+
"grad_norm": 115.72237396240234,
|
| 3729 |
+
"learning_rate": 9.99088128402616e-06,
|
| 3730 |
+
"loss": 11.7059,
|
| 3731 |
+
"step": 5260
|
| 3732 |
+
},
|
| 3733 |
+
{
|
| 3734 |
+
"epoch": 0.23388309173920038,
|
| 3735 |
+
"grad_norm": 120.50923919677734,
|
| 3736 |
+
"learning_rate": 9.990863948064232e-06,
|
| 3737 |
+
"loss": 12.0906,
|
| 3738 |
+
"step": 5270
|
| 3739 |
+
},
|
| 3740 |
+
{
|
| 3741 |
+
"epoch": 0.23432689267229184,
|
| 3742 |
+
"grad_norm": 103.17623138427734,
|
| 3743 |
+
"learning_rate": 9.990846612102305e-06,
|
| 3744 |
+
"loss": 11.5532,
|
| 3745 |
+
"step": 5280
|
| 3746 |
+
},
|
| 3747 |
+
{
|
| 3748 |
+
"epoch": 0.2347706936053833,
|
| 3749 |
+
"grad_norm": 116.489013671875,
|
| 3750 |
+
"learning_rate": 9.990829276140378e-06,
|
| 3751 |
+
"loss": 11.8999,
|
| 3752 |
+
"step": 5290
|
| 3753 |
+
},
|
| 3754 |
+
{
|
| 3755 |
+
"epoch": 0.23521449453847476,
|
| 3756 |
+
"grad_norm": 102.14614868164062,
|
| 3757 |
+
"learning_rate": 9.99081194017845e-06,
|
| 3758 |
+
"loss": 12.2294,
|
| 3759 |
+
"step": 5300
|
| 3760 |
+
},
|
| 3761 |
+
{
|
| 3762 |
+
"epoch": 0.23565829547156622,
|
| 3763 |
+
"grad_norm": 127.21760559082031,
|
| 3764 |
+
"learning_rate": 9.990794604216522e-06,
|
| 3765 |
+
"loss": 12.0739,
|
| 3766 |
+
"step": 5310
|
| 3767 |
+
},
|
| 3768 |
+
{
|
| 3769 |
+
"epoch": 0.23610209640465768,
|
| 3770 |
+
"grad_norm": 129.3851318359375,
|
| 3771 |
+
"learning_rate": 9.990777268254595e-06,
|
| 3772 |
+
"loss": 12.2383,
|
| 3773 |
+
"step": 5320
|
| 3774 |
+
},
|
| 3775 |
+
{
|
| 3776 |
+
"epoch": 0.23654589733774914,
|
| 3777 |
+
"grad_norm": 107.56005096435547,
|
| 3778 |
+
"learning_rate": 9.990759932292667e-06,
|
| 3779 |
+
"loss": 11.7989,
|
| 3780 |
+
"step": 5330
|
| 3781 |
+
},
|
| 3782 |
+
{
|
| 3783 |
+
"epoch": 0.23698969827084063,
|
| 3784 |
+
"grad_norm": 96.8632583618164,
|
| 3785 |
+
"learning_rate": 9.99074259633074e-06,
|
| 3786 |
+
"loss": 12.1831,
|
| 3787 |
+
"step": 5340
|
| 3788 |
+
},
|
| 3789 |
+
{
|
| 3790 |
+
"epoch": 0.23743349920393209,
|
| 3791 |
+
"grad_norm": 106.97894287109375,
|
| 3792 |
+
"learning_rate": 9.990725260368813e-06,
|
| 3793 |
+
"loss": 12.0481,
|
| 3794 |
+
"step": 5350
|
| 3795 |
+
},
|
| 3796 |
+
{
|
| 3797 |
+
"epoch": 0.23787730013702355,
|
| 3798 |
+
"grad_norm": 91.43546295166016,
|
| 3799 |
+
"learning_rate": 9.990707924406886e-06,
|
| 3800 |
+
"loss": 11.3025,
|
| 3801 |
+
"step": 5360
|
| 3802 |
+
},
|
| 3803 |
+
{
|
| 3804 |
+
"epoch": 0.238321101070115,
|
| 3805 |
+
"grad_norm": 104.5177001953125,
|
| 3806 |
+
"learning_rate": 9.990690588444957e-06,
|
| 3807 |
+
"loss": 11.5902,
|
| 3808 |
+
"step": 5370
|
| 3809 |
+
},
|
| 3810 |
+
{
|
| 3811 |
+
"epoch": 0.23876490200320646,
|
| 3812 |
+
"grad_norm": 117.70657348632812,
|
| 3813 |
+
"learning_rate": 9.99067325248303e-06,
|
| 3814 |
+
"loss": 11.7063,
|
| 3815 |
+
"step": 5380
|
| 3816 |
+
},
|
| 3817 |
+
{
|
| 3818 |
+
"epoch": 0.23920870293629792,
|
| 3819 |
+
"grad_norm": 113.05216979980469,
|
| 3820 |
+
"learning_rate": 9.990655916521104e-06,
|
| 3821 |
+
"loss": 12.0053,
|
| 3822 |
+
"step": 5390
|
| 3823 |
+
},
|
| 3824 |
+
{
|
| 3825 |
+
"epoch": 0.23965250386938938,
|
| 3826 |
+
"grad_norm": 115.97754669189453,
|
| 3827 |
+
"learning_rate": 9.990638580559175e-06,
|
| 3828 |
+
"loss": 12.3194,
|
| 3829 |
+
"step": 5400
|
| 3830 |
+
},
|
| 3831 |
+
{
|
| 3832 |
+
"epoch": 0.24009630480248084,
|
| 3833 |
+
"grad_norm": 110.10850524902344,
|
| 3834 |
+
"learning_rate": 9.990621244597248e-06,
|
| 3835 |
+
"loss": 11.2671,
|
| 3836 |
+
"step": 5410
|
| 3837 |
+
},
|
| 3838 |
+
{
|
| 3839 |
+
"epoch": 0.2405401057355723,
|
| 3840 |
+
"grad_norm": 110.06661987304688,
|
| 3841 |
+
"learning_rate": 9.990603908635321e-06,
|
| 3842 |
+
"loss": 12.2819,
|
| 3843 |
+
"step": 5420
|
| 3844 |
+
},
|
| 3845 |
+
{
|
| 3846 |
+
"epoch": 0.24098390666866376,
|
| 3847 |
+
"grad_norm": 126.98220825195312,
|
| 3848 |
+
"learning_rate": 9.990586572673393e-06,
|
| 3849 |
+
"loss": 12.0827,
|
| 3850 |
+
"step": 5430
|
| 3851 |
+
},
|
| 3852 |
+
{
|
| 3853 |
+
"epoch": 0.24142770760175522,
|
| 3854 |
+
"grad_norm": 101.59259033203125,
|
| 3855 |
+
"learning_rate": 9.990569236711466e-06,
|
| 3856 |
+
"loss": 11.5167,
|
| 3857 |
+
"step": 5440
|
| 3858 |
+
},
|
| 3859 |
+
{
|
| 3860 |
+
"epoch": 0.2418715085348467,
|
| 3861 |
+
"grad_norm": 91.08474731445312,
|
| 3862 |
+
"learning_rate": 9.990551900749539e-06,
|
| 3863 |
+
"loss": 11.4829,
|
| 3864 |
+
"step": 5450
|
| 3865 |
+
},
|
| 3866 |
+
{
|
| 3867 |
+
"epoch": 0.24231530946793817,
|
| 3868 |
+
"grad_norm": 100.78240966796875,
|
| 3869 |
+
"learning_rate": 9.99053456478761e-06,
|
| 3870 |
+
"loss": 12.0629,
|
| 3871 |
+
"step": 5460
|
| 3872 |
+
},
|
| 3873 |
+
{
|
| 3874 |
+
"epoch": 0.24275911040102963,
|
| 3875 |
+
"grad_norm": 121.29551696777344,
|
| 3876 |
+
"learning_rate": 9.990517228825683e-06,
|
| 3877 |
+
"loss": 11.9303,
|
| 3878 |
+
"step": 5470
|
| 3879 |
+
},
|
| 3880 |
+
{
|
| 3881 |
+
"epoch": 0.24320291133412109,
|
| 3882 |
+
"grad_norm": 100.81685638427734,
|
| 3883 |
+
"learning_rate": 9.990499892863756e-06,
|
| 3884 |
+
"loss": 11.4778,
|
| 3885 |
+
"step": 5480
|
| 3886 |
+
},
|
| 3887 |
+
{
|
| 3888 |
+
"epoch": 0.24364671226721255,
|
| 3889 |
+
"grad_norm": 134.68348693847656,
|
| 3890 |
+
"learning_rate": 9.990482556901828e-06,
|
| 3891 |
+
"loss": 12.5336,
|
| 3892 |
+
"step": 5490
|
| 3893 |
+
},
|
| 3894 |
+
{
|
| 3895 |
+
"epoch": 0.244090513200304,
|
| 3896 |
+
"grad_norm": 103.34284973144531,
|
| 3897 |
+
"learning_rate": 9.9904652209399e-06,
|
| 3898 |
+
"loss": 11.9692,
|
| 3899 |
+
"step": 5500
|
| 3900 |
+
},
|
| 3901 |
+
{
|
| 3902 |
+
"epoch": 0.24453431413339546,
|
| 3903 |
+
"grad_norm": 99.77964782714844,
|
| 3904 |
+
"learning_rate": 9.990447884977974e-06,
|
| 3905 |
+
"loss": 11.5971,
|
| 3906 |
+
"step": 5510
|
| 3907 |
+
},
|
| 3908 |
+
{
|
| 3909 |
+
"epoch": 0.24497811506648692,
|
| 3910 |
+
"grad_norm": 93.10983276367188,
|
| 3911 |
+
"learning_rate": 9.990430549016045e-06,
|
| 3912 |
+
"loss": 12.3927,
|
| 3913 |
+
"step": 5520
|
| 3914 |
+
},
|
| 3915 |
+
{
|
| 3916 |
+
"epoch": 0.24542191599957838,
|
| 3917 |
+
"grad_norm": 109.86616516113281,
|
| 3918 |
+
"learning_rate": 9.990413213054118e-06,
|
| 3919 |
+
"loss": 11.945,
|
| 3920 |
+
"step": 5530
|
| 3921 |
+
},
|
| 3922 |
+
{
|
| 3923 |
+
"epoch": 0.24586571693266984,
|
| 3924 |
+
"grad_norm": 112.61761474609375,
|
| 3925 |
+
"learning_rate": 9.990395877092191e-06,
|
| 3926 |
+
"loss": 11.7543,
|
| 3927 |
+
"step": 5540
|
| 3928 |
+
},
|
| 3929 |
+
{
|
| 3930 |
+
"epoch": 0.2463095178657613,
|
| 3931 |
+
"grad_norm": 115.5595703125,
|
| 3932 |
+
"learning_rate": 9.990378541130263e-06,
|
| 3933 |
+
"loss": 12.0871,
|
| 3934 |
+
"step": 5550
|
| 3935 |
+
},
|
| 3936 |
+
{
|
| 3937 |
+
"epoch": 0.24675331879885276,
|
| 3938 |
+
"grad_norm": 108.1959228515625,
|
| 3939 |
+
"learning_rate": 9.990361205168336e-06,
|
| 3940 |
+
"loss": 12.3206,
|
| 3941 |
+
"step": 5560
|
| 3942 |
+
},
|
| 3943 |
+
{
|
| 3944 |
+
"epoch": 0.24719711973194425,
|
| 3945 |
+
"grad_norm": 102.17387390136719,
|
| 3946 |
+
"learning_rate": 9.990343869206409e-06,
|
| 3947 |
+
"loss": 11.5629,
|
| 3948 |
+
"step": 5570
|
| 3949 |
+
},
|
| 3950 |
+
{
|
| 3951 |
+
"epoch": 0.2476409206650357,
|
| 3952 |
+
"grad_norm": 107.28736114501953,
|
| 3953 |
+
"learning_rate": 9.990326533244482e-06,
|
| 3954 |
+
"loss": 11.5035,
|
| 3955 |
+
"step": 5580
|
| 3956 |
+
},
|
| 3957 |
+
{
|
| 3958 |
+
"epoch": 0.24808472159812717,
|
| 3959 |
+
"grad_norm": 101.89647674560547,
|
| 3960 |
+
"learning_rate": 9.990309197282553e-06,
|
| 3961 |
+
"loss": 12.1624,
|
| 3962 |
+
"step": 5590
|
| 3963 |
+
},
|
| 3964 |
+
{
|
| 3965 |
+
"epoch": 0.24852852253121863,
|
| 3966 |
+
"grad_norm": 86.52133178710938,
|
| 3967 |
+
"learning_rate": 9.990291861320626e-06,
|
| 3968 |
+
"loss": 12.3599,
|
| 3969 |
+
"step": 5600
|
| 3970 |
+
},
|
| 3971 |
+
{
|
| 3972 |
+
"epoch": 0.24897232346431009,
|
| 3973 |
+
"grad_norm": 118.91706848144531,
|
| 3974 |
+
"learning_rate": 9.9902745253587e-06,
|
| 3975 |
+
"loss": 12.4713,
|
| 3976 |
+
"step": 5610
|
| 3977 |
+
},
|
| 3978 |
+
{
|
| 3979 |
+
"epoch": 0.24941612439740155,
|
| 3980 |
+
"grad_norm": 92.26959991455078,
|
| 3981 |
+
"learning_rate": 9.990257189396771e-06,
|
| 3982 |
+
"loss": 11.8632,
|
| 3983 |
+
"step": 5620
|
| 3984 |
+
},
|
| 3985 |
+
{
|
| 3986 |
+
"epoch": 0.249859925330493,
|
| 3987 |
+
"grad_norm": 96.55217742919922,
|
| 3988 |
+
"learning_rate": 9.990239853434844e-06,
|
| 3989 |
+
"loss": 11.7402,
|
| 3990 |
+
"step": 5630
|
| 3991 |
+
},
|
| 3992 |
+
{
|
| 3993 |
+
"epoch": 0.2503037262635845,
|
| 3994 |
+
"grad_norm": 108.4555892944336,
|
| 3995 |
+
"learning_rate": 9.990222517472917e-06,
|
| 3996 |
+
"loss": 11.7055,
|
| 3997 |
+
"step": 5640
|
| 3998 |
+
},
|
| 3999 |
+
{
|
| 4000 |
+
"epoch": 0.2507475271966759,
|
| 4001 |
+
"grad_norm": 103.9815902709961,
|
| 4002 |
+
"learning_rate": 9.990205181510988e-06,
|
| 4003 |
+
"loss": 12.0786,
|
| 4004 |
+
"step": 5650
|
| 4005 |
+
},
|
| 4006 |
+
{
|
| 4007 |
+
"epoch": 0.2511913281297674,
|
| 4008 |
+
"grad_norm": 112.2682876586914,
|
| 4009 |
+
"learning_rate": 9.990187845549062e-06,
|
| 4010 |
+
"loss": 12.11,
|
| 4011 |
+
"step": 5660
|
| 4012 |
+
},
|
| 4013 |
+
{
|
| 4014 |
+
"epoch": 0.25163512906285884,
|
| 4015 |
+
"grad_norm": 90.49382019042969,
|
| 4016 |
+
"learning_rate": 9.990170509587135e-06,
|
| 4017 |
+
"loss": 12.3861,
|
| 4018 |
+
"step": 5670
|
| 4019 |
+
},
|
| 4020 |
+
{
|
| 4021 |
+
"epoch": 0.25207892999595033,
|
| 4022 |
+
"grad_norm": 115.12799072265625,
|
| 4023 |
+
"learning_rate": 9.990153173625206e-06,
|
| 4024 |
+
"loss": 12.0726,
|
| 4025 |
+
"step": 5680
|
| 4026 |
+
},
|
| 4027 |
+
{
|
| 4028 |
+
"epoch": 0.25252273092904176,
|
| 4029 |
+
"grad_norm": 102.43321990966797,
|
| 4030 |
+
"learning_rate": 9.990135837663279e-06,
|
| 4031 |
+
"loss": 11.9243,
|
| 4032 |
+
"step": 5690
|
| 4033 |
+
},
|
| 4034 |
+
{
|
| 4035 |
+
"epoch": 0.25296653186213325,
|
| 4036 |
+
"grad_norm": 133.68328857421875,
|
| 4037 |
+
"learning_rate": 9.990118501701352e-06,
|
| 4038 |
+
"loss": 12.2475,
|
| 4039 |
+
"step": 5700
|
| 4040 |
+
},
|
| 4041 |
+
{
|
| 4042 |
+
"epoch": 0.2534103327952247,
|
| 4043 |
+
"grad_norm": 103.25040435791016,
|
| 4044 |
+
"learning_rate": 9.990101165739424e-06,
|
| 4045 |
+
"loss": 11.596,
|
| 4046 |
+
"step": 5710
|
| 4047 |
+
},
|
| 4048 |
+
{
|
| 4049 |
+
"epoch": 0.25385413372831617,
|
| 4050 |
+
"grad_norm": 106.11602783203125,
|
| 4051 |
+
"learning_rate": 9.990083829777497e-06,
|
| 4052 |
+
"loss": 12.3109,
|
| 4053 |
+
"step": 5720
|
| 4054 |
+
},
|
| 4055 |
+
{
|
| 4056 |
+
"epoch": 0.2542979346614076,
|
| 4057 |
+
"grad_norm": 99.80020904541016,
|
| 4058 |
+
"learning_rate": 9.99006649381557e-06,
|
| 4059 |
+
"loss": 11.2771,
|
| 4060 |
+
"step": 5730
|
| 4061 |
+
},
|
| 4062 |
+
{
|
| 4063 |
+
"epoch": 0.2547417355944991,
|
| 4064 |
+
"grad_norm": 115.24018859863281,
|
| 4065 |
+
"learning_rate": 9.990049157853641e-06,
|
| 4066 |
+
"loss": 11.7818,
|
| 4067 |
+
"step": 5740
|
| 4068 |
+
},
|
| 4069 |
+
{
|
| 4070 |
+
"epoch": 0.2551855365275906,
|
| 4071 |
+
"grad_norm": 93.93407440185547,
|
| 4072 |
+
"learning_rate": 9.990031821891714e-06,
|
| 4073 |
+
"loss": 12.2078,
|
| 4074 |
+
"step": 5750
|
| 4075 |
+
},
|
| 4076 |
+
{
|
| 4077 |
+
"epoch": 0.255629337460682,
|
| 4078 |
+
"grad_norm": 97.67652130126953,
|
| 4079 |
+
"learning_rate": 9.990014485929787e-06,
|
| 4080 |
+
"loss": 11.739,
|
| 4081 |
+
"step": 5760
|
| 4082 |
+
},
|
| 4083 |
+
{
|
| 4084 |
+
"epoch": 0.2560731383937735,
|
| 4085 |
+
"grad_norm": 117.71260070800781,
|
| 4086 |
+
"learning_rate": 9.989997149967859e-06,
|
| 4087 |
+
"loss": 11.7933,
|
| 4088 |
+
"step": 5770
|
| 4089 |
+
},
|
| 4090 |
+
{
|
| 4091 |
+
"epoch": 0.2565169393268649,
|
| 4092 |
+
"grad_norm": 85.63137817382812,
|
| 4093 |
+
"learning_rate": 9.989979814005932e-06,
|
| 4094 |
+
"loss": 11.5123,
|
| 4095 |
+
"step": 5780
|
| 4096 |
+
},
|
| 4097 |
+
{
|
| 4098 |
+
"epoch": 0.2569607402599564,
|
| 4099 |
+
"grad_norm": 91.07071685791016,
|
| 4100 |
+
"learning_rate": 9.989962478044005e-06,
|
| 4101 |
+
"loss": 11.7954,
|
| 4102 |
+
"step": 5790
|
| 4103 |
+
},
|
| 4104 |
+
{
|
| 4105 |
+
"epoch": 0.25740454119304784,
|
| 4106 |
+
"grad_norm": 94.04301452636719,
|
| 4107 |
+
"learning_rate": 9.989945142082078e-06,
|
| 4108 |
+
"loss": 11.4023,
|
| 4109 |
+
"step": 5800
|
| 4110 |
+
},
|
| 4111 |
+
{
|
| 4112 |
+
"epoch": 0.25784834212613933,
|
| 4113 |
+
"grad_norm": 98.91206359863281,
|
| 4114 |
+
"learning_rate": 9.98992780612015e-06,
|
| 4115 |
+
"loss": 12.2875,
|
| 4116 |
+
"step": 5810
|
| 4117 |
+
},
|
| 4118 |
+
{
|
| 4119 |
+
"epoch": 0.25829214305923076,
|
| 4120 |
+
"grad_norm": 94.439208984375,
|
| 4121 |
+
"learning_rate": 9.989910470158222e-06,
|
| 4122 |
+
"loss": 11.4681,
|
| 4123 |
+
"step": 5820
|
| 4124 |
+
},
|
| 4125 |
+
{
|
| 4126 |
+
"epoch": 0.25873594399232225,
|
| 4127 |
+
"grad_norm": 96.2354507446289,
|
| 4128 |
+
"learning_rate": 9.989893134196295e-06,
|
| 4129 |
+
"loss": 11.2651,
|
| 4130 |
+
"step": 5830
|
| 4131 |
+
},
|
| 4132 |
+
{
|
| 4133 |
+
"epoch": 0.2591797449254137,
|
| 4134 |
+
"grad_norm": 99.5677490234375,
|
| 4135 |
+
"learning_rate": 9.989875798234367e-06,
|
| 4136 |
+
"loss": 11.7862,
|
| 4137 |
+
"step": 5840
|
| 4138 |
+
},
|
| 4139 |
+
{
|
| 4140 |
+
"epoch": 0.25962354585850517,
|
| 4141 |
+
"grad_norm": 117.41152954101562,
|
| 4142 |
+
"learning_rate": 9.98985846227244e-06,
|
| 4143 |
+
"loss": 12.0639,
|
| 4144 |
+
"step": 5850
|
| 4145 |
+
},
|
| 4146 |
+
{
|
| 4147 |
+
"epoch": 0.26006734679159665,
|
| 4148 |
+
"grad_norm": 92.46183013916016,
|
| 4149 |
+
"learning_rate": 9.989841126310513e-06,
|
| 4150 |
+
"loss": 11.5852,
|
| 4151 |
+
"step": 5860
|
| 4152 |
+
},
|
| 4153 |
+
{
|
| 4154 |
+
"epoch": 0.2605111477246881,
|
| 4155 |
+
"grad_norm": 116.44026947021484,
|
| 4156 |
+
"learning_rate": 9.989823790348584e-06,
|
| 4157 |
+
"loss": 11.847,
|
| 4158 |
+
"step": 5870
|
| 4159 |
+
},
|
| 4160 |
+
{
|
| 4161 |
+
"epoch": 0.2609549486577796,
|
| 4162 |
+
"grad_norm": 109.089111328125,
|
| 4163 |
+
"learning_rate": 9.989806454386657e-06,
|
| 4164 |
+
"loss": 11.9711,
|
| 4165 |
+
"step": 5880
|
| 4166 |
+
},
|
| 4167 |
+
{
|
| 4168 |
+
"epoch": 0.261398749590871,
|
| 4169 |
+
"grad_norm": 126.8874740600586,
|
| 4170 |
+
"learning_rate": 9.98978911842473e-06,
|
| 4171 |
+
"loss": 11.7392,
|
| 4172 |
+
"step": 5890
|
| 4173 |
+
},
|
| 4174 |
+
{
|
| 4175 |
+
"epoch": 0.2618425505239625,
|
| 4176 |
+
"grad_norm": 121.22462463378906,
|
| 4177 |
+
"learning_rate": 9.989771782462802e-06,
|
| 4178 |
+
"loss": 11.8794,
|
| 4179 |
+
"step": 5900
|
| 4180 |
+
},
|
| 4181 |
+
{
|
| 4182 |
+
"epoch": 0.2622863514570539,
|
| 4183 |
+
"grad_norm": 87.884521484375,
|
| 4184 |
+
"learning_rate": 9.989754446500875e-06,
|
| 4185 |
+
"loss": 11.6205,
|
| 4186 |
+
"step": 5910
|
| 4187 |
+
},
|
| 4188 |
+
{
|
| 4189 |
+
"epoch": 0.2627301523901454,
|
| 4190 |
+
"grad_norm": 102.1272201538086,
|
| 4191 |
+
"learning_rate": 9.989737110538948e-06,
|
| 4192 |
+
"loss": 11.8747,
|
| 4193 |
+
"step": 5920
|
| 4194 |
+
},
|
| 4195 |
+
{
|
| 4196 |
+
"epoch": 0.26317395332323684,
|
| 4197 |
+
"grad_norm": 100.64566802978516,
|
| 4198 |
+
"learning_rate": 9.989719774577021e-06,
|
| 4199 |
+
"loss": 12.3826,
|
| 4200 |
+
"step": 5930
|
| 4201 |
+
},
|
| 4202 |
+
{
|
| 4203 |
+
"epoch": 0.26361775425632833,
|
| 4204 |
+
"grad_norm": 94.28556823730469,
|
| 4205 |
+
"learning_rate": 9.989702438615092e-06,
|
| 4206 |
+
"loss": 11.7302,
|
| 4207 |
+
"step": 5940
|
| 4208 |
+
},
|
| 4209 |
+
{
|
| 4210 |
+
"epoch": 0.26406155518941976,
|
| 4211 |
+
"grad_norm": 109.41671752929688,
|
| 4212 |
+
"learning_rate": 9.989685102653166e-06,
|
| 4213 |
+
"loss": 11.0984,
|
| 4214 |
+
"step": 5950
|
| 4215 |
+
},
|
| 4216 |
+
{
|
| 4217 |
+
"epoch": 0.26450535612251125,
|
| 4218 |
+
"grad_norm": 92.14616394042969,
|
| 4219 |
+
"learning_rate": 9.989667766691239e-06,
|
| 4220 |
+
"loss": 11.6492,
|
| 4221 |
+
"step": 5960
|
| 4222 |
+
},
|
| 4223 |
+
{
|
| 4224 |
+
"epoch": 0.26494915705560274,
|
| 4225 |
+
"grad_norm": 95.84307861328125,
|
| 4226 |
+
"learning_rate": 9.98965043072931e-06,
|
| 4227 |
+
"loss": 12.0779,
|
| 4228 |
+
"step": 5970
|
| 4229 |
+
},
|
| 4230 |
+
{
|
| 4231 |
+
"epoch": 0.26539295798869417,
|
| 4232 |
+
"grad_norm": 102.63638305664062,
|
| 4233 |
+
"learning_rate": 9.989633094767383e-06,
|
| 4234 |
+
"loss": 12.4218,
|
| 4235 |
+
"step": 5980
|
| 4236 |
+
},
|
| 4237 |
+
{
|
| 4238 |
+
"epoch": 0.26583675892178565,
|
| 4239 |
+
"grad_norm": 110.38811492919922,
|
| 4240 |
+
"learning_rate": 9.989615758805456e-06,
|
| 4241 |
+
"loss": 12.0949,
|
| 4242 |
+
"step": 5990
|
| 4243 |
+
},
|
| 4244 |
+
{
|
| 4245 |
+
"epoch": 0.2662805598548771,
|
| 4246 |
+
"grad_norm": 88.99332427978516,
|
| 4247 |
+
"learning_rate": 9.989598422843528e-06,
|
| 4248 |
+
"loss": 12.2146,
|
| 4249 |
+
"step": 6000
|
| 4250 |
+
},
|
| 4251 |
+
{
|
| 4252 |
+
"epoch": 0.2662805598548771,
|
| 4253 |
+
"eval_loss": 0.3712849020957947,
|
| 4254 |
+
"eval_runtime": 674.6735,
|
| 4255 |
+
"eval_samples_per_second": 1799.968,
|
| 4256 |
+
"eval_steps_per_second": 56.249,
|
| 4257 |
+
"step": 6000
|
| 4258 |
+
},
|
| 4259 |
+
{
|
| 4260 |
+
"epoch": 0.2667243607879686,
|
| 4261 |
+
"grad_norm": 106.79357147216797,
|
| 4262 |
+
"learning_rate": 9.9895810868816e-06,
|
| 4263 |
+
"loss": 11.816,
|
| 4264 |
+
"step": 6010
|
| 4265 |
+
},
|
| 4266 |
+
{
|
| 4267 |
+
"epoch": 0.26716816172106,
|
| 4268 |
+
"grad_norm": 101.64129638671875,
|
| 4269 |
+
"learning_rate": 9.989563750919674e-06,
|
| 4270 |
+
"loss": 11.551,
|
| 4271 |
+
"step": 6020
|
| 4272 |
+
},
|
| 4273 |
+
{
|
| 4274 |
+
"epoch": 0.2676119626541515,
|
| 4275 |
+
"grad_norm": 124.5982437133789,
|
| 4276 |
+
"learning_rate": 9.989546414957745e-06,
|
| 4277 |
+
"loss": 11.2434,
|
| 4278 |
+
"step": 6030
|
| 4279 |
+
},
|
| 4280 |
+
{
|
| 4281 |
+
"epoch": 0.2680557635872429,
|
| 4282 |
+
"grad_norm": 86.55455780029297,
|
| 4283 |
+
"learning_rate": 9.989529078995818e-06,
|
| 4284 |
+
"loss": 11.6358,
|
| 4285 |
+
"step": 6040
|
| 4286 |
+
},
|
| 4287 |
+
{
|
| 4288 |
+
"epoch": 0.2684995645203344,
|
| 4289 |
+
"grad_norm": 104.90837860107422,
|
| 4290 |
+
"learning_rate": 9.989511743033891e-06,
|
| 4291 |
+
"loss": 11.7908,
|
| 4292 |
+
"step": 6050
|
| 4293 |
+
},
|
| 4294 |
+
{
|
| 4295 |
+
"epoch": 0.26894336545342584,
|
| 4296 |
+
"grad_norm": 91.85465240478516,
|
| 4297 |
+
"learning_rate": 9.989494407071964e-06,
|
| 4298 |
+
"loss": 11.675,
|
| 4299 |
+
"step": 6060
|
| 4300 |
+
},
|
| 4301 |
+
{
|
| 4302 |
+
"epoch": 0.26938716638651733,
|
| 4303 |
+
"grad_norm": 110.98849487304688,
|
| 4304 |
+
"learning_rate": 9.989477071110036e-06,
|
| 4305 |
+
"loss": 11.9682,
|
| 4306 |
+
"step": 6070
|
| 4307 |
+
},
|
| 4308 |
+
{
|
| 4309 |
+
"epoch": 0.2698309673196088,
|
| 4310 |
+
"grad_norm": 117.15697479248047,
|
| 4311 |
+
"learning_rate": 9.989459735148109e-06,
|
| 4312 |
+
"loss": 12.6112,
|
| 4313 |
+
"step": 6080
|
| 4314 |
+
},
|
| 4315 |
+
{
|
| 4316 |
+
"epoch": 0.27027476825270025,
|
| 4317 |
+
"grad_norm": 116.58313751220703,
|
| 4318 |
+
"learning_rate": 9.989442399186182e-06,
|
| 4319 |
+
"loss": 11.8428,
|
| 4320 |
+
"step": 6090
|
| 4321 |
+
},
|
| 4322 |
+
{
|
| 4323 |
+
"epoch": 0.27071856918579174,
|
| 4324 |
+
"grad_norm": 141.74298095703125,
|
| 4325 |
+
"learning_rate": 9.989425063224253e-06,
|
| 4326 |
+
"loss": 11.678,
|
| 4327 |
+
"step": 6100
|
| 4328 |
+
},
|
| 4329 |
+
{
|
| 4330 |
+
"epoch": 0.27116237011888317,
|
| 4331 |
+
"grad_norm": 117.08960723876953,
|
| 4332 |
+
"learning_rate": 9.989407727262326e-06,
|
| 4333 |
+
"loss": 11.6677,
|
| 4334 |
+
"step": 6110
|
| 4335 |
+
},
|
| 4336 |
+
{
|
| 4337 |
+
"epoch": 0.27160617105197465,
|
| 4338 |
+
"grad_norm": 102.23706817626953,
|
| 4339 |
+
"learning_rate": 9.9893903913004e-06,
|
| 4340 |
+
"loss": 11.5424,
|
| 4341 |
+
"step": 6120
|
| 4342 |
+
},
|
| 4343 |
+
{
|
| 4344 |
+
"epoch": 0.2720499719850661,
|
| 4345 |
+
"grad_norm": 100.09423065185547,
|
| 4346 |
+
"learning_rate": 9.98937305533847e-06,
|
| 4347 |
+
"loss": 12.0797,
|
| 4348 |
+
"step": 6130
|
| 4349 |
+
},
|
| 4350 |
+
{
|
| 4351 |
+
"epoch": 0.2724937729181576,
|
| 4352 |
+
"grad_norm": 113.59011840820312,
|
| 4353 |
+
"learning_rate": 9.989355719376544e-06,
|
| 4354 |
+
"loss": 12.0007,
|
| 4355 |
+
"step": 6140
|
| 4356 |
+
},
|
| 4357 |
+
{
|
| 4358 |
+
"epoch": 0.272937573851249,
|
| 4359 |
+
"grad_norm": 94.99832153320312,
|
| 4360 |
+
"learning_rate": 9.989338383414617e-06,
|
| 4361 |
+
"loss": 11.4787,
|
| 4362 |
+
"step": 6150
|
| 4363 |
+
},
|
| 4364 |
+
{
|
| 4365 |
+
"epoch": 0.2733813747843405,
|
| 4366 |
+
"grad_norm": 135.75721740722656,
|
| 4367 |
+
"learning_rate": 9.989321047452688e-06,
|
| 4368 |
+
"loss": 11.9264,
|
| 4369 |
+
"step": 6160
|
| 4370 |
+
},
|
| 4371 |
+
{
|
| 4372 |
+
"epoch": 0.2738251757174319,
|
| 4373 |
+
"grad_norm": 112.27375793457031,
|
| 4374 |
+
"learning_rate": 9.989303711490761e-06,
|
| 4375 |
+
"loss": 11.4877,
|
| 4376 |
+
"step": 6170
|
| 4377 |
+
},
|
| 4378 |
+
{
|
| 4379 |
+
"epoch": 0.2742689766505234,
|
| 4380 |
+
"grad_norm": 115.73841094970703,
|
| 4381 |
+
"learning_rate": 9.989286375528834e-06,
|
| 4382 |
+
"loss": 11.6287,
|
| 4383 |
+
"step": 6180
|
| 4384 |
+
},
|
| 4385 |
+
{
|
| 4386 |
+
"epoch": 0.27471277758361484,
|
| 4387 |
+
"grad_norm": 108.56095886230469,
|
| 4388 |
+
"learning_rate": 9.989269039566908e-06,
|
| 4389 |
+
"loss": 12.2162,
|
| 4390 |
+
"step": 6190
|
| 4391 |
+
},
|
| 4392 |
+
{
|
| 4393 |
+
"epoch": 0.27515657851670633,
|
| 4394 |
+
"grad_norm": 106.11347961425781,
|
| 4395 |
+
"learning_rate": 9.989251703604979e-06,
|
| 4396 |
+
"loss": 12.4926,
|
| 4397 |
+
"step": 6200
|
| 4398 |
+
},
|
| 4399 |
+
{
|
| 4400 |
+
"epoch": 0.2756003794497978,
|
| 4401 |
+
"grad_norm": 93.0553207397461,
|
| 4402 |
+
"learning_rate": 9.989234367643052e-06,
|
| 4403 |
+
"loss": 11.5661,
|
| 4404 |
+
"step": 6210
|
| 4405 |
+
},
|
| 4406 |
+
{
|
| 4407 |
+
"epoch": 0.27604418038288925,
|
| 4408 |
+
"grad_norm": 114.06632232666016,
|
| 4409 |
+
"learning_rate": 9.989217031681125e-06,
|
| 4410 |
+
"loss": 11.9102,
|
| 4411 |
+
"step": 6220
|
| 4412 |
+
},
|
| 4413 |
+
{
|
| 4414 |
+
"epoch": 0.27648798131598074,
|
| 4415 |
+
"grad_norm": 105.9919204711914,
|
| 4416 |
+
"learning_rate": 9.989199695719196e-06,
|
| 4417 |
+
"loss": 11.4711,
|
| 4418 |
+
"step": 6230
|
| 4419 |
+
},
|
| 4420 |
+
{
|
| 4421 |
+
"epoch": 0.27693178224907217,
|
| 4422 |
+
"grad_norm": 84.44627380371094,
|
| 4423 |
+
"learning_rate": 9.98918235975727e-06,
|
| 4424 |
+
"loss": 11.6304,
|
| 4425 |
+
"step": 6240
|
| 4426 |
+
},
|
| 4427 |
+
{
|
| 4428 |
+
"epoch": 0.27737558318216365,
|
| 4429 |
+
"grad_norm": 96.19264221191406,
|
| 4430 |
+
"learning_rate": 9.989165023795343e-06,
|
| 4431 |
+
"loss": 11.414,
|
| 4432 |
+
"step": 6250
|
| 4433 |
+
},
|
| 4434 |
+
{
|
| 4435 |
+
"epoch": 0.2778193841152551,
|
| 4436 |
+
"grad_norm": 127.19998931884766,
|
| 4437 |
+
"learning_rate": 9.989147687833414e-06,
|
| 4438 |
+
"loss": 11.9031,
|
| 4439 |
+
"step": 6260
|
| 4440 |
+
},
|
| 4441 |
+
{
|
| 4442 |
+
"epoch": 0.2782631850483466,
|
| 4443 |
+
"grad_norm": 108.23567199707031,
|
| 4444 |
+
"learning_rate": 9.989130351871487e-06,
|
| 4445 |
+
"loss": 11.9478,
|
| 4446 |
+
"step": 6270
|
| 4447 |
+
},
|
| 4448 |
+
{
|
| 4449 |
+
"epoch": 0.278706985981438,
|
| 4450 |
+
"grad_norm": 96.11833190917969,
|
| 4451 |
+
"learning_rate": 9.98911301590956e-06,
|
| 4452 |
+
"loss": 12.053,
|
| 4453 |
+
"step": 6280
|
| 4454 |
+
},
|
| 4455 |
+
{
|
| 4456 |
+
"epoch": 0.2791507869145295,
|
| 4457 |
+
"grad_norm": 89.6744613647461,
|
| 4458 |
+
"learning_rate": 9.989095679947633e-06,
|
| 4459 |
+
"loss": 11.7384,
|
| 4460 |
+
"step": 6290
|
| 4461 |
+
},
|
| 4462 |
+
{
|
| 4463 |
+
"epoch": 0.2795945878476209,
|
| 4464 |
+
"grad_norm": 88.66259002685547,
|
| 4465 |
+
"learning_rate": 9.989078343985705e-06,
|
| 4466 |
+
"loss": 11.6472,
|
| 4467 |
+
"step": 6300
|
| 4468 |
+
},
|
| 4469 |
+
{
|
| 4470 |
+
"epoch": 0.2800383887807124,
|
| 4471 |
+
"grad_norm": 112.74403381347656,
|
| 4472 |
+
"learning_rate": 9.989061008023778e-06,
|
| 4473 |
+
"loss": 11.8507,
|
| 4474 |
+
"step": 6310
|
| 4475 |
+
},
|
| 4476 |
+
{
|
| 4477 |
+
"epoch": 0.2804821897138039,
|
| 4478 |
+
"grad_norm": 100.31635284423828,
|
| 4479 |
+
"learning_rate": 9.98904367206185e-06,
|
| 4480 |
+
"loss": 11.3632,
|
| 4481 |
+
"step": 6320
|
| 4482 |
+
},
|
| 4483 |
+
{
|
| 4484 |
+
"epoch": 0.28092599064689533,
|
| 4485 |
+
"grad_norm": 92.06365966796875,
|
| 4486 |
+
"learning_rate": 9.989026336099922e-06,
|
| 4487 |
+
"loss": 11.9611,
|
| 4488 |
+
"step": 6330
|
| 4489 |
+
},
|
| 4490 |
+
{
|
| 4491 |
+
"epoch": 0.2813697915799868,
|
| 4492 |
+
"grad_norm": 106.90202331542969,
|
| 4493 |
+
"learning_rate": 9.989009000137995e-06,
|
| 4494 |
+
"loss": 11.9601,
|
| 4495 |
+
"step": 6340
|
| 4496 |
+
},
|
| 4497 |
+
{
|
| 4498 |
+
"epoch": 0.28181359251307825,
|
| 4499 |
+
"grad_norm": 93.49620056152344,
|
| 4500 |
+
"learning_rate": 9.988991664176068e-06,
|
| 4501 |
+
"loss": 12.0237,
|
| 4502 |
+
"step": 6350
|
| 4503 |
+
},
|
| 4504 |
+
{
|
| 4505 |
+
"epoch": 0.28225739344616974,
|
| 4506 |
+
"grad_norm": 101.77459716796875,
|
| 4507 |
+
"learning_rate": 9.98897432821414e-06,
|
| 4508 |
+
"loss": 11.6238,
|
| 4509 |
+
"step": 6360
|
| 4510 |
+
},
|
| 4511 |
+
{
|
| 4512 |
+
"epoch": 0.28270119437926117,
|
| 4513 |
+
"grad_norm": 94.77214813232422,
|
| 4514 |
+
"learning_rate": 9.988956992252213e-06,
|
| 4515 |
+
"loss": 11.6693,
|
| 4516 |
+
"step": 6370
|
| 4517 |
+
},
|
| 4518 |
+
{
|
| 4519 |
+
"epoch": 0.28314499531235265,
|
| 4520 |
+
"grad_norm": 110.98194885253906,
|
| 4521 |
+
"learning_rate": 9.988939656290286e-06,
|
| 4522 |
+
"loss": 11.8191,
|
| 4523 |
+
"step": 6380
|
| 4524 |
+
},
|
| 4525 |
+
{
|
| 4526 |
+
"epoch": 0.2835887962454441,
|
| 4527 |
+
"grad_norm": 96.43860626220703,
|
| 4528 |
+
"learning_rate": 9.988922320328357e-06,
|
| 4529 |
+
"loss": 11.2866,
|
| 4530 |
+
"step": 6390
|
| 4531 |
+
},
|
| 4532 |
+
{
|
| 4533 |
+
"epoch": 0.2840325971785356,
|
| 4534 |
+
"grad_norm": 125.56005096435547,
|
| 4535 |
+
"learning_rate": 9.98890498436643e-06,
|
| 4536 |
+
"loss": 12.1676,
|
| 4537 |
+
"step": 6400
|
| 4538 |
+
},
|
| 4539 |
+
{
|
| 4540 |
+
"epoch": 0.284476398111627,
|
| 4541 |
+
"grad_norm": 122.20761108398438,
|
| 4542 |
+
"learning_rate": 9.988887648404503e-06,
|
| 4543 |
+
"loss": 12.474,
|
| 4544 |
+
"step": 6410
|
| 4545 |
+
},
|
| 4546 |
+
{
|
| 4547 |
+
"epoch": 0.2849201990447185,
|
| 4548 |
+
"grad_norm": 101.35332489013672,
|
| 4549 |
+
"learning_rate": 9.988870312442576e-06,
|
| 4550 |
+
"loss": 11.733,
|
| 4551 |
+
"step": 6420
|
| 4552 |
+
},
|
| 4553 |
+
{
|
| 4554 |
+
"epoch": 0.28536399997781,
|
| 4555 |
+
"grad_norm": 97.87244415283203,
|
| 4556 |
+
"learning_rate": 9.988852976480648e-06,
|
| 4557 |
+
"loss": 11.9138,
|
| 4558 |
+
"step": 6430
|
| 4559 |
+
},
|
| 4560 |
+
{
|
| 4561 |
+
"epoch": 0.2858078009109014,
|
| 4562 |
+
"grad_norm": 106.80064392089844,
|
| 4563 |
+
"learning_rate": 9.988835640518721e-06,
|
| 4564 |
+
"loss": 12.4942,
|
| 4565 |
+
"step": 6440
|
| 4566 |
+
},
|
| 4567 |
+
{
|
| 4568 |
+
"epoch": 0.2862516018439929,
|
| 4569 |
+
"grad_norm": 106.28848266601562,
|
| 4570 |
+
"learning_rate": 9.988818304556794e-06,
|
| 4571 |
+
"loss": 11.9158,
|
| 4572 |
+
"step": 6450
|
| 4573 |
+
},
|
| 4574 |
+
{
|
| 4575 |
+
"epoch": 0.28669540277708433,
|
| 4576 |
+
"grad_norm": 100.0877914428711,
|
| 4577 |
+
"learning_rate": 9.988800968594865e-06,
|
| 4578 |
+
"loss": 11.2765,
|
| 4579 |
+
"step": 6460
|
| 4580 |
+
},
|
| 4581 |
+
{
|
| 4582 |
+
"epoch": 0.2871392037101758,
|
| 4583 |
+
"grad_norm": 100.36260223388672,
|
| 4584 |
+
"learning_rate": 9.988783632632938e-06,
|
| 4585 |
+
"loss": 11.7645,
|
| 4586 |
+
"step": 6470
|
| 4587 |
+
},
|
| 4588 |
+
{
|
| 4589 |
+
"epoch": 0.28758300464326725,
|
| 4590 |
+
"grad_norm": 102.455078125,
|
| 4591 |
+
"learning_rate": 9.988766296671012e-06,
|
| 4592 |
+
"loss": 11.5561,
|
| 4593 |
+
"step": 6480
|
| 4594 |
+
},
|
| 4595 |
+
{
|
| 4596 |
+
"epoch": 0.28802680557635874,
|
| 4597 |
+
"grad_norm": 91.07699584960938,
|
| 4598 |
+
"learning_rate": 9.988748960709083e-06,
|
| 4599 |
+
"loss": 11.8894,
|
| 4600 |
+
"step": 6490
|
| 4601 |
+
},
|
| 4602 |
+
{
|
| 4603 |
+
"epoch": 0.28847060650945017,
|
| 4604 |
+
"grad_norm": 96.68830108642578,
|
| 4605 |
+
"learning_rate": 9.988731624747156e-06,
|
| 4606 |
+
"loss": 11.7245,
|
| 4607 |
+
"step": 6500
|
| 4608 |
+
},
|
| 4609 |
+
{
|
| 4610 |
+
"epoch": 0.28891440744254165,
|
| 4611 |
+
"grad_norm": 86.85298919677734,
|
| 4612 |
+
"learning_rate": 9.988714288785229e-06,
|
| 4613 |
+
"loss": 11.685,
|
| 4614 |
+
"step": 6510
|
| 4615 |
+
},
|
| 4616 |
+
{
|
| 4617 |
+
"epoch": 0.2893582083756331,
|
| 4618 |
+
"grad_norm": 100.70494842529297,
|
| 4619 |
+
"learning_rate": 9.9886969528233e-06,
|
| 4620 |
+
"loss": 11.6402,
|
| 4621 |
+
"step": 6520
|
| 4622 |
+
},
|
| 4623 |
+
{
|
| 4624 |
+
"epoch": 0.2898020093087246,
|
| 4625 |
+
"grad_norm": 101.03349304199219,
|
| 4626 |
+
"learning_rate": 9.988679616861374e-06,
|
| 4627 |
+
"loss": 12.2626,
|
| 4628 |
+
"step": 6530
|
| 4629 |
+
},
|
| 4630 |
+
{
|
| 4631 |
+
"epoch": 0.29024581024181606,
|
| 4632 |
+
"grad_norm": 91.4577865600586,
|
| 4633 |
+
"learning_rate": 9.988662280899447e-06,
|
| 4634 |
+
"loss": 11.8911,
|
| 4635 |
+
"step": 6540
|
| 4636 |
+
},
|
| 4637 |
+
{
|
| 4638 |
+
"epoch": 0.2906896111749075,
|
| 4639 |
+
"grad_norm": 97.14340209960938,
|
| 4640 |
+
"learning_rate": 9.98864494493752e-06,
|
| 4641 |
+
"loss": 11.6866,
|
| 4642 |
+
"step": 6550
|
| 4643 |
+
},
|
| 4644 |
+
{
|
| 4645 |
+
"epoch": 0.291133412107999,
|
| 4646 |
+
"grad_norm": 94.9050064086914,
|
| 4647 |
+
"learning_rate": 9.988627608975591e-06,
|
| 4648 |
+
"loss": 11.713,
|
| 4649 |
+
"step": 6560
|
| 4650 |
+
},
|
| 4651 |
+
{
|
| 4652 |
+
"epoch": 0.2915772130410904,
|
| 4653 |
+
"grad_norm": 94.01126861572266,
|
| 4654 |
+
"learning_rate": 9.988610273013664e-06,
|
| 4655 |
+
"loss": 11.6159,
|
| 4656 |
+
"step": 6570
|
| 4657 |
+
},
|
| 4658 |
+
{
|
| 4659 |
+
"epoch": 0.2920210139741819,
|
| 4660 |
+
"grad_norm": 89.02586364746094,
|
| 4661 |
+
"learning_rate": 9.988592937051737e-06,
|
| 4662 |
+
"loss": 11.6739,
|
| 4663 |
+
"step": 6580
|
| 4664 |
+
},
|
| 4665 |
+
{
|
| 4666 |
+
"epoch": 0.29246481490727333,
|
| 4667 |
+
"grad_norm": 126.11679077148438,
|
| 4668 |
+
"learning_rate": 9.988575601089809e-06,
|
| 4669 |
+
"loss": 11.7169,
|
| 4670 |
+
"step": 6590
|
| 4671 |
+
},
|
| 4672 |
+
{
|
| 4673 |
+
"epoch": 0.2929086158403648,
|
| 4674 |
+
"grad_norm": 91.34590148925781,
|
| 4675 |
+
"learning_rate": 9.988558265127882e-06,
|
| 4676 |
+
"loss": 11.5195,
|
| 4677 |
+
"step": 6600
|
| 4678 |
+
},
|
| 4679 |
+
{
|
| 4680 |
+
"epoch": 0.29335241677345625,
|
| 4681 |
+
"grad_norm": 108.34649658203125,
|
| 4682 |
+
"learning_rate": 9.988540929165955e-06,
|
| 4683 |
+
"loss": 11.6583,
|
| 4684 |
+
"step": 6610
|
| 4685 |
+
},
|
| 4686 |
+
{
|
| 4687 |
+
"epoch": 0.29379621770654774,
|
| 4688 |
+
"grad_norm": 111.23458099365234,
|
| 4689 |
+
"learning_rate": 9.988523593204026e-06,
|
| 4690 |
+
"loss": 11.5678,
|
| 4691 |
+
"step": 6620
|
| 4692 |
+
},
|
| 4693 |
+
{
|
| 4694 |
+
"epoch": 0.29424001863963917,
|
| 4695 |
+
"grad_norm": 92.88959503173828,
|
| 4696 |
+
"learning_rate": 9.9885062572421e-06,
|
| 4697 |
+
"loss": 12.0539,
|
| 4698 |
+
"step": 6630
|
| 4699 |
+
},
|
| 4700 |
+
{
|
| 4701 |
+
"epoch": 0.29468381957273065,
|
| 4702 |
+
"grad_norm": 119.5103530883789,
|
| 4703 |
+
"learning_rate": 9.988488921280172e-06,
|
| 4704 |
+
"loss": 11.7499,
|
| 4705 |
+
"step": 6640
|
| 4706 |
+
},
|
| 4707 |
+
{
|
| 4708 |
+
"epoch": 0.2951276205058221,
|
| 4709 |
+
"grad_norm": 107.3758544921875,
|
| 4710 |
+
"learning_rate": 9.988471585318244e-06,
|
| 4711 |
+
"loss": 11.3844,
|
| 4712 |
+
"step": 6650
|
| 4713 |
+
},
|
| 4714 |
+
{
|
| 4715 |
+
"epoch": 0.2955714214389136,
|
| 4716 |
+
"grad_norm": 103.09275817871094,
|
| 4717 |
+
"learning_rate": 9.988454249356317e-06,
|
| 4718 |
+
"loss": 11.3347,
|
| 4719 |
+
"step": 6660
|
| 4720 |
+
},
|
| 4721 |
+
{
|
| 4722 |
+
"epoch": 0.29601522237200506,
|
| 4723 |
+
"grad_norm": 104.75767517089844,
|
| 4724 |
+
"learning_rate": 9.98843691339439e-06,
|
| 4725 |
+
"loss": 12.1328,
|
| 4726 |
+
"step": 6670
|
| 4727 |
+
},
|
| 4728 |
+
{
|
| 4729 |
+
"epoch": 0.2964590233050965,
|
| 4730 |
+
"grad_norm": 105.1933364868164,
|
| 4731 |
+
"learning_rate": 9.988419577432463e-06,
|
| 4732 |
+
"loss": 12.1966,
|
| 4733 |
+
"step": 6680
|
| 4734 |
+
},
|
| 4735 |
+
{
|
| 4736 |
+
"epoch": 0.296902824238188,
|
| 4737 |
+
"grad_norm": 95.23583221435547,
|
| 4738 |
+
"learning_rate": 9.988402241470534e-06,
|
| 4739 |
+
"loss": 11.925,
|
| 4740 |
+
"step": 6690
|
| 4741 |
+
},
|
| 4742 |
+
{
|
| 4743 |
+
"epoch": 0.2973466251712794,
|
| 4744 |
+
"grad_norm": 93.23466491699219,
|
| 4745 |
+
"learning_rate": 9.988384905508607e-06,
|
| 4746 |
+
"loss": 11.9893,
|
| 4747 |
+
"step": 6700
|
| 4748 |
+
},
|
| 4749 |
+
{
|
| 4750 |
+
"epoch": 0.2977904261043709,
|
| 4751 |
+
"grad_norm": 96.76582336425781,
|
| 4752 |
+
"learning_rate": 9.98836756954668e-06,
|
| 4753 |
+
"loss": 11.4291,
|
| 4754 |
+
"step": 6710
|
| 4755 |
+
},
|
| 4756 |
+
{
|
| 4757 |
+
"epoch": 0.29823422703746233,
|
| 4758 |
+
"grad_norm": 95.39541625976562,
|
| 4759 |
+
"learning_rate": 9.988350233584752e-06,
|
| 4760 |
+
"loss": 11.6534,
|
| 4761 |
+
"step": 6720
|
| 4762 |
+
},
|
| 4763 |
+
{
|
| 4764 |
+
"epoch": 0.2986780279705538,
|
| 4765 |
+
"grad_norm": 111.02050018310547,
|
| 4766 |
+
"learning_rate": 9.988332897622825e-06,
|
| 4767 |
+
"loss": 11.215,
|
| 4768 |
+
"step": 6730
|
| 4769 |
+
},
|
| 4770 |
+
{
|
| 4771 |
+
"epoch": 0.29912182890364525,
|
| 4772 |
+
"grad_norm": 93.41612243652344,
|
| 4773 |
+
"learning_rate": 9.988315561660898e-06,
|
| 4774 |
+
"loss": 11.8045,
|
| 4775 |
+
"step": 6740
|
| 4776 |
+
},
|
| 4777 |
+
{
|
| 4778 |
+
"epoch": 0.29956562983673674,
|
| 4779 |
+
"grad_norm": 103.25040435791016,
|
| 4780 |
+
"learning_rate": 9.98829822569897e-06,
|
| 4781 |
+
"loss": 11.7352,
|
| 4782 |
+
"step": 6750
|
| 4783 |
+
},
|
| 4784 |
+
{
|
| 4785 |
+
"epoch": 0.30000943076982817,
|
| 4786 |
+
"grad_norm": 110.95781707763672,
|
| 4787 |
+
"learning_rate": 9.988280889737043e-06,
|
| 4788 |
+
"loss": 11.9544,
|
| 4789 |
+
"step": 6760
|
| 4790 |
+
},
|
| 4791 |
+
{
|
| 4792 |
+
"epoch": 0.30045323170291965,
|
| 4793 |
+
"grad_norm": 90.73185729980469,
|
| 4794 |
+
"learning_rate": 9.988263553775116e-06,
|
| 4795 |
+
"loss": 11.5593,
|
| 4796 |
+
"step": 6770
|
| 4797 |
+
},
|
| 4798 |
+
{
|
| 4799 |
+
"epoch": 0.30089703263601114,
|
| 4800 |
+
"grad_norm": 90.01508331298828,
|
| 4801 |
+
"learning_rate": 9.988246217813187e-06,
|
| 4802 |
+
"loss": 11.1305,
|
| 4803 |
+
"step": 6780
|
| 4804 |
+
},
|
| 4805 |
+
{
|
| 4806 |
+
"epoch": 0.3013408335691026,
|
| 4807 |
+
"grad_norm": 99.30635070800781,
|
| 4808 |
+
"learning_rate": 9.98822888185126e-06,
|
| 4809 |
+
"loss": 11.1892,
|
| 4810 |
+
"step": 6790
|
| 4811 |
+
},
|
| 4812 |
+
{
|
| 4813 |
+
"epoch": 0.30178463450219406,
|
| 4814 |
+
"grad_norm": 112.32921600341797,
|
| 4815 |
+
"learning_rate": 9.988211545889333e-06,
|
| 4816 |
+
"loss": 11.9221,
|
| 4817 |
+
"step": 6800
|
| 4818 |
+
},
|
| 4819 |
+
{
|
| 4820 |
+
"epoch": 0.3022284354352855,
|
| 4821 |
+
"grad_norm": 97.16471099853516,
|
| 4822 |
+
"learning_rate": 9.988194209927405e-06,
|
| 4823 |
+
"loss": 11.8916,
|
| 4824 |
+
"step": 6810
|
| 4825 |
+
},
|
| 4826 |
+
{
|
| 4827 |
+
"epoch": 0.302672236368377,
|
| 4828 |
+
"grad_norm": 93.2406234741211,
|
| 4829 |
+
"learning_rate": 9.988176873965478e-06,
|
| 4830 |
+
"loss": 11.4169,
|
| 4831 |
+
"step": 6820
|
| 4832 |
+
},
|
| 4833 |
+
{
|
| 4834 |
+
"epoch": 0.3031160373014684,
|
| 4835 |
+
"grad_norm": 111.53374481201172,
|
| 4836 |
+
"learning_rate": 9.98815953800355e-06,
|
| 4837 |
+
"loss": 11.1438,
|
| 4838 |
+
"step": 6830
|
| 4839 |
+
},
|
| 4840 |
+
{
|
| 4841 |
+
"epoch": 0.3035598382345599,
|
| 4842 |
+
"grad_norm": 94.99114990234375,
|
| 4843 |
+
"learning_rate": 9.988142202041622e-06,
|
| 4844 |
+
"loss": 12.0838,
|
| 4845 |
+
"step": 6840
|
| 4846 |
+
},
|
| 4847 |
+
{
|
| 4848 |
+
"epoch": 0.30400363916765133,
|
| 4849 |
+
"grad_norm": 117.2374496459961,
|
| 4850 |
+
"learning_rate": 9.988124866079695e-06,
|
| 4851 |
+
"loss": 11.4959,
|
| 4852 |
+
"step": 6850
|
| 4853 |
+
},
|
| 4854 |
+
{
|
| 4855 |
+
"epoch": 0.3044474401007428,
|
| 4856 |
+
"grad_norm": 97.72772216796875,
|
| 4857 |
+
"learning_rate": 9.988107530117768e-06,
|
| 4858 |
+
"loss": 11.8559,
|
| 4859 |
+
"step": 6860
|
| 4860 |
+
},
|
| 4861 |
+
{
|
| 4862 |
+
"epoch": 0.30489124103383425,
|
| 4863 |
+
"grad_norm": 105.31409454345703,
|
| 4864 |
+
"learning_rate": 9.98809019415584e-06,
|
| 4865 |
+
"loss": 11.537,
|
| 4866 |
+
"step": 6870
|
| 4867 |
+
},
|
| 4868 |
+
{
|
| 4869 |
+
"epoch": 0.30533504196692574,
|
| 4870 |
+
"grad_norm": 125.29707336425781,
|
| 4871 |
+
"learning_rate": 9.988072858193913e-06,
|
| 4872 |
+
"loss": 11.4942,
|
| 4873 |
+
"step": 6880
|
| 4874 |
+
},
|
| 4875 |
+
{
|
| 4876 |
+
"epoch": 0.3057788429000172,
|
| 4877 |
+
"grad_norm": 101.15106201171875,
|
| 4878 |
+
"learning_rate": 9.988055522231986e-06,
|
| 4879 |
+
"loss": 11.3884,
|
| 4880 |
+
"step": 6890
|
| 4881 |
+
},
|
| 4882 |
+
{
|
| 4883 |
+
"epoch": 0.30622264383310865,
|
| 4884 |
+
"grad_norm": 97.56317138671875,
|
| 4885 |
+
"learning_rate": 9.988038186270059e-06,
|
| 4886 |
+
"loss": 11.5489,
|
| 4887 |
+
"step": 6900
|
| 4888 |
+
},
|
| 4889 |
+
{
|
| 4890 |
+
"epoch": 0.30666644476620014,
|
| 4891 |
+
"grad_norm": 112.57002258300781,
|
| 4892 |
+
"learning_rate": 9.98802085030813e-06,
|
| 4893 |
+
"loss": 11.3843,
|
| 4894 |
+
"step": 6910
|
| 4895 |
+
},
|
| 4896 |
+
{
|
| 4897 |
+
"epoch": 0.3071102456992916,
|
| 4898 |
+
"grad_norm": 100.37068176269531,
|
| 4899 |
+
"learning_rate": 9.988003514346203e-06,
|
| 4900 |
+
"loss": 11.4439,
|
| 4901 |
+
"step": 6920
|
| 4902 |
+
},
|
| 4903 |
+
{
|
| 4904 |
+
"epoch": 0.30755404663238306,
|
| 4905 |
+
"grad_norm": 119.5300521850586,
|
| 4906 |
+
"learning_rate": 9.987986178384276e-06,
|
| 4907 |
+
"loss": 11.4964,
|
| 4908 |
+
"step": 6930
|
| 4909 |
+
},
|
| 4910 |
+
{
|
| 4911 |
+
"epoch": 0.3079978475654745,
|
| 4912 |
+
"grad_norm": 88.61663055419922,
|
| 4913 |
+
"learning_rate": 9.987968842422348e-06,
|
| 4914 |
+
"loss": 11.6089,
|
| 4915 |
+
"step": 6940
|
| 4916 |
+
},
|
| 4917 |
+
{
|
| 4918 |
+
"epoch": 0.308441648498566,
|
| 4919 |
+
"grad_norm": 101.71082305908203,
|
| 4920 |
+
"learning_rate": 9.98795150646042e-06,
|
| 4921 |
+
"loss": 11.5646,
|
| 4922 |
+
"step": 6950
|
| 4923 |
+
},
|
| 4924 |
+
{
|
| 4925 |
+
"epoch": 0.3088854494316574,
|
| 4926 |
+
"grad_norm": 111.21810913085938,
|
| 4927 |
+
"learning_rate": 9.987934170498494e-06,
|
| 4928 |
+
"loss": 11.2611,
|
| 4929 |
+
"step": 6960
|
| 4930 |
+
},
|
| 4931 |
+
{
|
| 4932 |
+
"epoch": 0.3093292503647489,
|
| 4933 |
+
"grad_norm": 110.25248718261719,
|
| 4934 |
+
"learning_rate": 9.987916834536565e-06,
|
| 4935 |
+
"loss": 11.5723,
|
| 4936 |
+
"step": 6970
|
| 4937 |
+
},
|
| 4938 |
+
{
|
| 4939 |
+
"epoch": 0.30977305129784033,
|
| 4940 |
+
"grad_norm": 104.40763092041016,
|
| 4941 |
+
"learning_rate": 9.987899498574638e-06,
|
| 4942 |
+
"loss": 12.0642,
|
| 4943 |
+
"step": 6980
|
| 4944 |
+
},
|
| 4945 |
+
{
|
| 4946 |
+
"epoch": 0.3102168522309318,
|
| 4947 |
+
"grad_norm": 89.2273941040039,
|
| 4948 |
+
"learning_rate": 9.987882162612711e-06,
|
| 4949 |
+
"loss": 11.4713,
|
| 4950 |
+
"step": 6990
|
| 4951 |
+
},
|
| 4952 |
+
{
|
| 4953 |
+
"epoch": 0.3106606531640233,
|
| 4954 |
+
"grad_norm": 85.18675994873047,
|
| 4955 |
+
"learning_rate": 9.987864826650783e-06,
|
| 4956 |
+
"loss": 11.8243,
|
| 4957 |
+
"step": 7000
|
| 4958 |
+
},
|
| 4959 |
+
{
|
| 4960 |
+
"epoch": 0.3106606531640233,
|
| 4961 |
+
"eval_loss": 0.36315852403640747,
|
| 4962 |
+
"eval_runtime": 674.7882,
|
| 4963 |
+
"eval_samples_per_second": 1799.663,
|
| 4964 |
+
"eval_steps_per_second": 56.24,
|
| 4965 |
+
"step": 7000
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.31110445409711474,
|
| 4969 |
+
"grad_norm": 123.10943603515625,
|
| 4970 |
+
"learning_rate": 9.987847490688856e-06,
|
| 4971 |
+
"loss": 11.6931,
|
| 4972 |
+
"step": 7010
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.3115482550302062,
|
| 4976 |
+
"grad_norm": 130.8447723388672,
|
| 4977 |
+
"learning_rate": 9.987830154726929e-06,
|
| 4978 |
+
"loss": 11.5164,
|
| 4979 |
+
"step": 7020
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.31199205596329765,
|
| 4983 |
+
"grad_norm": 101.97557830810547,
|
| 4984 |
+
"learning_rate": 9.987812818765e-06,
|
| 4985 |
+
"loss": 11.8152,
|
| 4986 |
+
"step": 7030
|
| 4987 |
+
},
|
| 4988 |
+
{
|
| 4989 |
+
"epoch": 0.31243585689638914,
|
| 4990 |
+
"grad_norm": 93.44794464111328,
|
| 4991 |
+
"learning_rate": 9.987795482803073e-06,
|
| 4992 |
+
"loss": 11.6298,
|
| 4993 |
+
"step": 7040
|
| 4994 |
+
},
|
| 4995 |
+
{
|
| 4996 |
+
"epoch": 0.3128796578294806,
|
| 4997 |
+
"grad_norm": 97.57392120361328,
|
| 4998 |
+
"learning_rate": 9.987778146841147e-06,
|
| 4999 |
+
"loss": 11.3152,
|
| 5000 |
+
"step": 7050
|
| 5001 |
+
},
|
| 5002 |
+
{
|
| 5003 |
+
"epoch": 0.31332345876257206,
|
| 5004 |
+
"grad_norm": 102.27783203125,
|
| 5005 |
+
"learning_rate": 9.987760810879218e-06,
|
| 5006 |
+
"loss": 10.9668,
|
| 5007 |
+
"step": 7060
|
| 5008 |
+
},
|
| 5009 |
+
{
|
| 5010 |
+
"epoch": 0.3137672596956635,
|
| 5011 |
+
"grad_norm": 108.09846496582031,
|
| 5012 |
+
"learning_rate": 9.987743474917291e-06,
|
| 5013 |
+
"loss": 11.4038,
|
| 5014 |
+
"step": 7070
|
| 5015 |
+
},
|
| 5016 |
+
{
|
| 5017 |
+
"epoch": 0.314211060628755,
|
| 5018 |
+
"grad_norm": 90.89788055419922,
|
| 5019 |
+
"learning_rate": 9.987726138955364e-06,
|
| 5020 |
+
"loss": 11.3717,
|
| 5021 |
+
"step": 7080
|
| 5022 |
+
},
|
| 5023 |
+
{
|
| 5024 |
+
"epoch": 0.3146548615618464,
|
| 5025 |
+
"grad_norm": 115.7634506225586,
|
| 5026 |
+
"learning_rate": 9.987708802993435e-06,
|
| 5027 |
+
"loss": 11.682,
|
| 5028 |
+
"step": 7090
|
| 5029 |
+
},
|
| 5030 |
+
{
|
| 5031 |
+
"epoch": 0.3150986624949379,
|
| 5032 |
+
"grad_norm": 109.29789733886719,
|
| 5033 |
+
"learning_rate": 9.987691467031509e-06,
|
| 5034 |
+
"loss": 11.7889,
|
| 5035 |
+
"step": 7100
|
| 5036 |
+
},
|
| 5037 |
+
{
|
| 5038 |
+
"epoch": 0.3155424634280294,
|
| 5039 |
+
"grad_norm": 95.27775573730469,
|
| 5040 |
+
"learning_rate": 9.987674131069582e-06,
|
| 5041 |
+
"loss": 11.2461,
|
| 5042 |
+
"step": 7110
|
| 5043 |
+
},
|
| 5044 |
+
{
|
| 5045 |
+
"epoch": 0.3159862643611208,
|
| 5046 |
+
"grad_norm": 110.1352310180664,
|
| 5047 |
+
"learning_rate": 9.987656795107655e-06,
|
| 5048 |
+
"loss": 11.3557,
|
| 5049 |
+
"step": 7120
|
| 5050 |
+
},
|
| 5051 |
+
{
|
| 5052 |
+
"epoch": 0.3164300652942123,
|
| 5053 |
+
"grad_norm": 82.80999755859375,
|
| 5054 |
+
"learning_rate": 9.987639459145726e-06,
|
| 5055 |
+
"loss": 11.4224,
|
| 5056 |
+
"step": 7130
|
| 5057 |
+
},
|
| 5058 |
+
{
|
| 5059 |
+
"epoch": 0.31687386622730374,
|
| 5060 |
+
"grad_norm": 94.51629638671875,
|
| 5061 |
+
"learning_rate": 9.987622123183799e-06,
|
| 5062 |
+
"loss": 11.6793,
|
| 5063 |
+
"step": 7140
|
| 5064 |
+
},
|
| 5065 |
+
{
|
| 5066 |
+
"epoch": 0.3173176671603952,
|
| 5067 |
+
"grad_norm": 106.12570190429688,
|
| 5068 |
+
"learning_rate": 9.987604787221872e-06,
|
| 5069 |
+
"loss": 11.3187,
|
| 5070 |
+
"step": 7150
|
| 5071 |
+
},
|
| 5072 |
+
{
|
| 5073 |
+
"epoch": 0.31776146809348665,
|
| 5074 |
+
"grad_norm": 110.01351928710938,
|
| 5075 |
+
"learning_rate": 9.987587451259944e-06,
|
| 5076 |
+
"loss": 11.3371,
|
| 5077 |
+
"step": 7160
|
| 5078 |
+
},
|
| 5079 |
+
{
|
| 5080 |
+
"epoch": 0.31820526902657814,
|
| 5081 |
+
"grad_norm": 97.4388427734375,
|
| 5082 |
+
"learning_rate": 9.987570115298017e-06,
|
| 5083 |
+
"loss": 11.5326,
|
| 5084 |
+
"step": 7170
|
| 5085 |
+
},
|
| 5086 |
+
{
|
| 5087 |
+
"epoch": 0.3186490699596696,
|
| 5088 |
+
"grad_norm": 80.90787506103516,
|
| 5089 |
+
"learning_rate": 9.98755277933609e-06,
|
| 5090 |
+
"loss": 11.7102,
|
| 5091 |
+
"step": 7180
|
| 5092 |
+
},
|
| 5093 |
+
{
|
| 5094 |
+
"epoch": 0.31909287089276106,
|
| 5095 |
+
"grad_norm": 102.96366882324219,
|
| 5096 |
+
"learning_rate": 9.987535443374161e-06,
|
| 5097 |
+
"loss": 11.9188,
|
| 5098 |
+
"step": 7190
|
| 5099 |
+
},
|
| 5100 |
+
{
|
| 5101 |
+
"epoch": 0.3195366718258525,
|
| 5102 |
+
"grad_norm": 101.80105590820312,
|
| 5103 |
+
"learning_rate": 9.987518107412234e-06,
|
| 5104 |
+
"loss": 12.0429,
|
| 5105 |
+
"step": 7200
|
| 5106 |
+
},
|
| 5107 |
+
{
|
| 5108 |
+
"epoch": 0.319980472758944,
|
| 5109 |
+
"grad_norm": 90.94136047363281,
|
| 5110 |
+
"learning_rate": 9.987500771450307e-06,
|
| 5111 |
+
"loss": 10.825,
|
| 5112 |
+
"step": 7210
|
| 5113 |
+
},
|
| 5114 |
+
{
|
| 5115 |
+
"epoch": 0.3204242736920354,
|
| 5116 |
+
"grad_norm": 109.56967163085938,
|
| 5117 |
+
"learning_rate": 9.987483435488379e-06,
|
| 5118 |
+
"loss": 11.5279,
|
| 5119 |
+
"step": 7220
|
| 5120 |
+
},
|
| 5121 |
+
{
|
| 5122 |
+
"epoch": 0.3208680746251269,
|
| 5123 |
+
"grad_norm": 95.71635437011719,
|
| 5124 |
+
"learning_rate": 9.987466099526452e-06,
|
| 5125 |
+
"loss": 11.6178,
|
| 5126 |
+
"step": 7230
|
| 5127 |
+
},
|
| 5128 |
+
{
|
| 5129 |
+
"epoch": 0.3213118755582184,
|
| 5130 |
+
"grad_norm": 93.75989532470703,
|
| 5131 |
+
"learning_rate": 9.987448763564525e-06,
|
| 5132 |
+
"loss": 11.4102,
|
| 5133 |
+
"step": 7240
|
| 5134 |
+
},
|
| 5135 |
+
{
|
| 5136 |
+
"epoch": 0.3217556764913098,
|
| 5137 |
+
"grad_norm": 100.95425415039062,
|
| 5138 |
+
"learning_rate": 9.987431427602596e-06,
|
| 5139 |
+
"loss": 11.7016,
|
| 5140 |
+
"step": 7250
|
| 5141 |
+
},
|
| 5142 |
+
{
|
| 5143 |
+
"epoch": 0.3221994774244013,
|
| 5144 |
+
"grad_norm": 97.52702331542969,
|
| 5145 |
+
"learning_rate": 9.98741409164067e-06,
|
| 5146 |
+
"loss": 11.8841,
|
| 5147 |
+
"step": 7260
|
| 5148 |
+
},
|
| 5149 |
+
{
|
| 5150 |
+
"epoch": 0.32264327835749274,
|
| 5151 |
+
"grad_norm": 91.30542755126953,
|
| 5152 |
+
"learning_rate": 9.987396755678742e-06,
|
| 5153 |
+
"loss": 11.4776,
|
| 5154 |
+
"step": 7270
|
| 5155 |
+
},
|
| 5156 |
+
{
|
| 5157 |
+
"epoch": 0.3230870792905842,
|
| 5158 |
+
"grad_norm": 106.42786407470703,
|
| 5159 |
+
"learning_rate": 9.987379419716814e-06,
|
| 5160 |
+
"loss": 11.5718,
|
| 5161 |
+
"step": 7280
|
| 5162 |
+
},
|
| 5163 |
+
{
|
| 5164 |
+
"epoch": 0.32353088022367565,
|
| 5165 |
+
"grad_norm": 93.21934509277344,
|
| 5166 |
+
"learning_rate": 9.987362083754887e-06,
|
| 5167 |
+
"loss": 11.1913,
|
| 5168 |
+
"step": 7290
|
| 5169 |
+
},
|
| 5170 |
+
{
|
| 5171 |
+
"epoch": 0.32397468115676714,
|
| 5172 |
+
"grad_norm": 102.5538101196289,
|
| 5173 |
+
"learning_rate": 9.98734474779296e-06,
|
| 5174 |
+
"loss": 11.5974,
|
| 5175 |
+
"step": 7300
|
| 5176 |
+
},
|
| 5177 |
+
{
|
| 5178 |
+
"epoch": 0.3244184820898586,
|
| 5179 |
+
"grad_norm": 96.53839874267578,
|
| 5180 |
+
"learning_rate": 9.987327411831033e-06,
|
| 5181 |
+
"loss": 11.6911,
|
| 5182 |
+
"step": 7310
|
| 5183 |
+
},
|
| 5184 |
+
{
|
| 5185 |
+
"epoch": 0.32486228302295006,
|
| 5186 |
+
"grad_norm": 94.2166519165039,
|
| 5187 |
+
"learning_rate": 9.987310075869104e-06,
|
| 5188 |
+
"loss": 11.4638,
|
| 5189 |
+
"step": 7320
|
| 5190 |
+
},
|
| 5191 |
+
{
|
| 5192 |
+
"epoch": 0.3253060839560415,
|
| 5193 |
+
"grad_norm": 100.33325958251953,
|
| 5194 |
+
"learning_rate": 9.987292739907177e-06,
|
| 5195 |
+
"loss": 11.3276,
|
| 5196 |
+
"step": 7330
|
| 5197 |
+
},
|
| 5198 |
+
{
|
| 5199 |
+
"epoch": 0.325749884889133,
|
| 5200 |
+
"grad_norm": 94.38825988769531,
|
| 5201 |
+
"learning_rate": 9.98727540394525e-06,
|
| 5202 |
+
"loss": 11.408,
|
| 5203 |
+
"step": 7340
|
| 5204 |
+
},
|
| 5205 |
+
{
|
| 5206 |
+
"epoch": 0.32619368582222447,
|
| 5207 |
+
"grad_norm": 106.76789855957031,
|
| 5208 |
+
"learning_rate": 9.987258067983322e-06,
|
| 5209 |
+
"loss": 12.0534,
|
| 5210 |
+
"step": 7350
|
| 5211 |
+
},
|
| 5212 |
+
{
|
| 5213 |
+
"epoch": 0.3266374867553159,
|
| 5214 |
+
"grad_norm": 92.51216888427734,
|
| 5215 |
+
"learning_rate": 9.987240732021395e-06,
|
| 5216 |
+
"loss": 11.0724,
|
| 5217 |
+
"step": 7360
|
| 5218 |
+
},
|
| 5219 |
+
{
|
| 5220 |
+
"epoch": 0.3270812876884074,
|
| 5221 |
+
"grad_norm": 90.04621124267578,
|
| 5222 |
+
"learning_rate": 9.987223396059468e-06,
|
| 5223 |
+
"loss": 11.682,
|
| 5224 |
+
"step": 7370
|
| 5225 |
+
},
|
| 5226 |
+
{
|
| 5227 |
+
"epoch": 0.3275250886214988,
|
| 5228 |
+
"grad_norm": 112.20062255859375,
|
| 5229 |
+
"learning_rate": 9.98720606009754e-06,
|
| 5230 |
+
"loss": 11.4545,
|
| 5231 |
+
"step": 7380
|
| 5232 |
+
},
|
| 5233 |
+
{
|
| 5234 |
+
"epoch": 0.3279688895545903,
|
| 5235 |
+
"grad_norm": 98.15010833740234,
|
| 5236 |
+
"learning_rate": 9.987188724135613e-06,
|
| 5237 |
+
"loss": 11.6855,
|
| 5238 |
+
"step": 7390
|
| 5239 |
+
},
|
| 5240 |
+
{
|
| 5241 |
+
"epoch": 0.32841269048768174,
|
| 5242 |
+
"grad_norm": 79.66431427001953,
|
| 5243 |
+
"learning_rate": 9.987171388173686e-06,
|
| 5244 |
+
"loss": 11.2667,
|
| 5245 |
+
"step": 7400
|
| 5246 |
+
},
|
| 5247 |
+
{
|
| 5248 |
+
"epoch": 0.3288564914207732,
|
| 5249 |
+
"grad_norm": 103.56822204589844,
|
| 5250 |
+
"learning_rate": 9.987154052211757e-06,
|
| 5251 |
+
"loss": 11.5097,
|
| 5252 |
+
"step": 7410
|
| 5253 |
+
},
|
| 5254 |
+
{
|
| 5255 |
+
"epoch": 0.32930029235386465,
|
| 5256 |
+
"grad_norm": 101.66146850585938,
|
| 5257 |
+
"learning_rate": 9.98713671624983e-06,
|
| 5258 |
+
"loss": 11.2951,
|
| 5259 |
+
"step": 7420
|
| 5260 |
+
},
|
| 5261 |
+
{
|
| 5262 |
+
"epoch": 0.32974409328695614,
|
| 5263 |
+
"grad_norm": 95.05819702148438,
|
| 5264 |
+
"learning_rate": 9.987119380287903e-06,
|
| 5265 |
+
"loss": 11.6909,
|
| 5266 |
+
"step": 7430
|
| 5267 |
+
},
|
| 5268 |
+
{
|
| 5269 |
+
"epoch": 0.3301878942200476,
|
| 5270 |
+
"grad_norm": 97.97879028320312,
|
| 5271 |
+
"learning_rate": 9.987102044325975e-06,
|
| 5272 |
+
"loss": 10.9043,
|
| 5273 |
+
"step": 7440
|
| 5274 |
+
},
|
| 5275 |
+
{
|
| 5276 |
+
"epoch": 0.33063169515313906,
|
| 5277 |
+
"grad_norm": 139.6067352294922,
|
| 5278 |
+
"learning_rate": 9.987084708364048e-06,
|
| 5279 |
+
"loss": 11.3615,
|
| 5280 |
+
"step": 7450
|
| 5281 |
+
},
|
| 5282 |
+
{
|
| 5283 |
+
"epoch": 0.33107549608623055,
|
| 5284 |
+
"grad_norm": 106.51782989501953,
|
| 5285 |
+
"learning_rate": 9.98706737240212e-06,
|
| 5286 |
+
"loss": 11.3788,
|
| 5287 |
+
"step": 7460
|
| 5288 |
+
},
|
| 5289 |
+
{
|
| 5290 |
+
"epoch": 0.331519297019322,
|
| 5291 |
+
"grad_norm": 89.64512634277344,
|
| 5292 |
+
"learning_rate": 9.987050036440192e-06,
|
| 5293 |
+
"loss": 11.7731,
|
| 5294 |
+
"step": 7470
|
| 5295 |
+
},
|
| 5296 |
+
{
|
| 5297 |
+
"epoch": 0.33196309795241347,
|
| 5298 |
+
"grad_norm": 90.49308776855469,
|
| 5299 |
+
"learning_rate": 9.987032700478265e-06,
|
| 5300 |
+
"loss": 11.8977,
|
| 5301 |
+
"step": 7480
|
| 5302 |
+
},
|
| 5303 |
+
{
|
| 5304 |
+
"epoch": 0.3324068988855049,
|
| 5305 |
+
"grad_norm": 103.32622528076172,
|
| 5306 |
+
"learning_rate": 9.987015364516338e-06,
|
| 5307 |
+
"loss": 11.5663,
|
| 5308 |
+
"step": 7490
|
| 5309 |
+
},
|
| 5310 |
+
{
|
| 5311 |
+
"epoch": 0.3328506998185964,
|
| 5312 |
+
"grad_norm": 88.08988189697266,
|
| 5313 |
+
"learning_rate": 9.98699802855441e-06,
|
| 5314 |
+
"loss": 11.8236,
|
| 5315 |
+
"step": 7500
|
| 5316 |
}
|
| 5317 |
],
|
| 5318 |
"logging_steps": 10,
|
|
|
|
| 5332 |
"attributes": {}
|
| 5333 |
}
|
| 5334 |
},
|
| 5335 |
+
"total_flos": 2.617292403769344e+18,
|
| 5336 |
"train_batch_size": 4,
|
| 5337 |
"trial_name": null,
|
| 5338 |
"trial_params": null
|