Training in progress, step 1200, checkpoint
Browse files- last-checkpoint/model-00001-of-00004.safetensors +1 -1
- last-checkpoint/model-00002-of-00004.safetensors +1 -1
- last-checkpoint/model-00003-of-00004.safetensors +1 -1
- last-checkpoint/model-00004-of-00004.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +2103 -3
last-checkpoint/model-00001-of-00004.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4903351912
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:92dccb83c8c0a6f6345b737ce1db36c36d4d7992ca73e80b721d5b7f47b6b21f
|
3 |
size 4903351912
|
last-checkpoint/model-00002-of-00004.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4947570872
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7a2b24a229bd97979ae3a693e35a5956dc189e037a5b710e3036e7c6e73da46b
|
3 |
size 4947570872
|
last-checkpoint/model-00003-of-00004.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4962221464
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:392a4f12ee63cb0bc19fbef78019f437dd5c5304c182b08c02183eb4d2583232
|
3 |
size 4962221464
|
last-checkpoint/model-00004-of-00004.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3670322200
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0fdc80a8b0724b4471aa4d47e7c62e9887fbc389532013fa3f9588fb8779b16d
|
3 |
size 3670322200
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2216
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3452834f82acfa4bd3619f349e0974484197e8f142671485b30b15e286bebb6e
|
3 |
size 2216
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14244
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2b66e3cc7c452b707ddac5caf0aa17618afb9bc1a0333600a22c4afb353f3165
|
3 |
size 14244
|
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:aeed2188106e1a1def583b8f2acfd47fefac1069d1052695eda50a98d43b58c5
|
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":
|
5 |
"eval_steps": 500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -6307,6 +6307,2106 @@
|
|
6307 |
"learning_rate": 2.6725495145537665e-05,
|
6308 |
"loss": 1.1544,
|
6309 |
"step": 900
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6310 |
}
|
6311 |
],
|
6312 |
"logging_steps": 1,
|
@@ -6326,7 +8426,7 @@
|
|
6326 |
"attributes": {}
|
6327 |
}
|
6328 |
},
|
6329 |
-
"total_flos":
|
6330 |
"train_batch_size": 8,
|
6331 |
"trial_name": null,
|
6332 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.1235955056179776,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 1200,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
6307 |
"learning_rate": 2.6725495145537665e-05,
|
6308 |
"loss": 1.1544,
|
6309 |
"step": 900
|
6310 |
+
},
|
6311 |
+
{
|
6312 |
+
"epoch": 0.8436329588014981,
|
6313 |
+
"grad_norm": 0.0,
|
6314 |
+
"learning_rate": 2.669632358016719e-05,
|
6315 |
+
"loss": 1.18,
|
6316 |
+
"step": 901
|
6317 |
+
},
|
6318 |
+
{
|
6319 |
+
"epoch": 0.8445692883895131,
|
6320 |
+
"grad_norm": 0.0,
|
6321 |
+
"learning_rate": 2.6667135961213043e-05,
|
6322 |
+
"loss": 1.1473,
|
6323 |
+
"step": 902
|
6324 |
+
},
|
6325 |
+
{
|
6326 |
+
"epoch": 0.8455056179775281,
|
6327 |
+
"grad_norm": 0.0,
|
6328 |
+
"learning_rate": 2.6637932358648812e-05,
|
6329 |
+
"loss": 1.1294,
|
6330 |
+
"step": 903
|
6331 |
+
},
|
6332 |
+
{
|
6333 |
+
"epoch": 0.846441947565543,
|
6334 |
+
"grad_norm": 0.0,
|
6335 |
+
"learning_rate": 2.6608712842486438e-05,
|
6336 |
+
"loss": 1.1652,
|
6337 |
+
"step": 904
|
6338 |
+
},
|
6339 |
+
{
|
6340 |
+
"epoch": 0.8473782771535581,
|
6341 |
+
"grad_norm": 0.0,
|
6342 |
+
"learning_rate": 2.6579477482775985e-05,
|
6343 |
+
"loss": 1.1908,
|
6344 |
+
"step": 905
|
6345 |
+
},
|
6346 |
+
{
|
6347 |
+
"epoch": 0.848314606741573,
|
6348 |
+
"grad_norm": 0.0,
|
6349 |
+
"learning_rate": 2.65502263496055e-05,
|
6350 |
+
"loss": 1.1349,
|
6351 |
+
"step": 906
|
6352 |
+
},
|
6353 |
+
{
|
6354 |
+
"epoch": 0.849250936329588,
|
6355 |
+
"grad_norm": 0.0,
|
6356 |
+
"learning_rate": 2.6520959513100857e-05,
|
6357 |
+
"loss": 1.1694,
|
6358 |
+
"step": 907
|
6359 |
+
},
|
6360 |
+
{
|
6361 |
+
"epoch": 0.850187265917603,
|
6362 |
+
"grad_norm": 0.0,
|
6363 |
+
"learning_rate": 2.6491677043425573e-05,
|
6364 |
+
"loss": 1.1598,
|
6365 |
+
"step": 908
|
6366 |
+
},
|
6367 |
+
{
|
6368 |
+
"epoch": 0.851123595505618,
|
6369 |
+
"grad_norm": 0.0,
|
6370 |
+
"learning_rate": 2.646237901078065e-05,
|
6371 |
+
"loss": 1.1238,
|
6372 |
+
"step": 909
|
6373 |
+
},
|
6374 |
+
{
|
6375 |
+
"epoch": 0.8520599250936329,
|
6376 |
+
"grad_norm": 0.0,
|
6377 |
+
"learning_rate": 2.643306548540438e-05,
|
6378 |
+
"loss": 1.1607,
|
6379 |
+
"step": 910
|
6380 |
+
},
|
6381 |
+
{
|
6382 |
+
"epoch": 0.852996254681648,
|
6383 |
+
"grad_norm": 0.0,
|
6384 |
+
"learning_rate": 2.6403736537572222e-05,
|
6385 |
+
"loss": 1.1524,
|
6386 |
+
"step": 911
|
6387 |
+
},
|
6388 |
+
{
|
6389 |
+
"epoch": 0.8539325842696629,
|
6390 |
+
"grad_norm": 0.0,
|
6391 |
+
"learning_rate": 2.637439223759659e-05,
|
6392 |
+
"loss": 1.1592,
|
6393 |
+
"step": 912
|
6394 |
+
},
|
6395 |
+
{
|
6396 |
+
"epoch": 0.8548689138576779,
|
6397 |
+
"grad_norm": 0.0,
|
6398 |
+
"learning_rate": 2.6345032655826708e-05,
|
6399 |
+
"loss": 1.1102,
|
6400 |
+
"step": 913
|
6401 |
+
},
|
6402 |
+
{
|
6403 |
+
"epoch": 0.8558052434456929,
|
6404 |
+
"grad_norm": 0.0,
|
6405 |
+
"learning_rate": 2.631565786264844e-05,
|
6406 |
+
"loss": 1.1421,
|
6407 |
+
"step": 914
|
6408 |
+
},
|
6409 |
+
{
|
6410 |
+
"epoch": 0.8567415730337079,
|
6411 |
+
"grad_norm": 0.0,
|
6412 |
+
"learning_rate": 2.6286267928484112e-05,
|
6413 |
+
"loss": 1.1219,
|
6414 |
+
"step": 915
|
6415 |
+
},
|
6416 |
+
{
|
6417 |
+
"epoch": 0.8576779026217228,
|
6418 |
+
"grad_norm": 0.0,
|
6419 |
+
"learning_rate": 2.6256862923792356e-05,
|
6420 |
+
"loss": 1.165,
|
6421 |
+
"step": 916
|
6422 |
+
},
|
6423 |
+
{
|
6424 |
+
"epoch": 0.8586142322097379,
|
6425 |
+
"grad_norm": 0.0,
|
6426 |
+
"learning_rate": 2.6227442919067923e-05,
|
6427 |
+
"loss": 1.1583,
|
6428 |
+
"step": 917
|
6429 |
+
},
|
6430 |
+
{
|
6431 |
+
"epoch": 0.8595505617977528,
|
6432 |
+
"grad_norm": 0.0,
|
6433 |
+
"learning_rate": 2.6198007984841543e-05,
|
6434 |
+
"loss": 1.1407,
|
6435 |
+
"step": 918
|
6436 |
+
},
|
6437 |
+
{
|
6438 |
+
"epoch": 0.8604868913857678,
|
6439 |
+
"grad_norm": 0.0,
|
6440 |
+
"learning_rate": 2.6168558191679708e-05,
|
6441 |
+
"loss": 1.1676,
|
6442 |
+
"step": 919
|
6443 |
+
},
|
6444 |
+
{
|
6445 |
+
"epoch": 0.8614232209737828,
|
6446 |
+
"grad_norm": 0.0,
|
6447 |
+
"learning_rate": 2.6139093610184554e-05,
|
6448 |
+
"loss": 1.1316,
|
6449 |
+
"step": 920
|
6450 |
+
},
|
6451 |
+
{
|
6452 |
+
"epoch": 0.8623595505617978,
|
6453 |
+
"grad_norm": 0.0,
|
6454 |
+
"learning_rate": 2.6109614310993676e-05,
|
6455 |
+
"loss": 1.1515,
|
6456 |
+
"step": 921
|
6457 |
+
},
|
6458 |
+
{
|
6459 |
+
"epoch": 0.8632958801498127,
|
6460 |
+
"grad_norm": 0.0,
|
6461 |
+
"learning_rate": 2.608012036477994e-05,
|
6462 |
+
"loss": 1.0938,
|
6463 |
+
"step": 922
|
6464 |
+
},
|
6465 |
+
{
|
6466 |
+
"epoch": 0.8642322097378277,
|
6467 |
+
"grad_norm": 0.0,
|
6468 |
+
"learning_rate": 2.6050611842251315e-05,
|
6469 |
+
"loss": 1.0931,
|
6470 |
+
"step": 923
|
6471 |
+
},
|
6472 |
+
{
|
6473 |
+
"epoch": 0.8651685393258427,
|
6474 |
+
"grad_norm": 0.0,
|
6475 |
+
"learning_rate": 2.6021088814150747e-05,
|
6476 |
+
"loss": 1.1491,
|
6477 |
+
"step": 924
|
6478 |
+
},
|
6479 |
+
{
|
6480 |
+
"epoch": 0.8661048689138576,
|
6481 |
+
"grad_norm": 0.0,
|
6482 |
+
"learning_rate": 2.5991551351255923e-05,
|
6483 |
+
"loss": 1.1658,
|
6484 |
+
"step": 925
|
6485 |
+
},
|
6486 |
+
{
|
6487 |
+
"epoch": 0.8670411985018727,
|
6488 |
+
"grad_norm": 0.0,
|
6489 |
+
"learning_rate": 2.5961999524379162e-05,
|
6490 |
+
"loss": 1.1628,
|
6491 |
+
"step": 926
|
6492 |
+
},
|
6493 |
+
{
|
6494 |
+
"epoch": 0.8679775280898876,
|
6495 |
+
"grad_norm": 0.0,
|
6496 |
+
"learning_rate": 2.593243340436721e-05,
|
6497 |
+
"loss": 1.1777,
|
6498 |
+
"step": 927
|
6499 |
+
},
|
6500 |
+
{
|
6501 |
+
"epoch": 0.8689138576779026,
|
6502 |
+
"grad_norm": 0.0,
|
6503 |
+
"learning_rate": 2.590285306210107e-05,
|
6504 |
+
"loss": 1.1687,
|
6505 |
+
"step": 928
|
6506 |
+
},
|
6507 |
+
{
|
6508 |
+
"epoch": 0.8698501872659176,
|
6509 |
+
"grad_norm": 0.0,
|
6510 |
+
"learning_rate": 2.5873258568495853e-05,
|
6511 |
+
"loss": 1.2063,
|
6512 |
+
"step": 929
|
6513 |
+
},
|
6514 |
+
{
|
6515 |
+
"epoch": 0.8707865168539326,
|
6516 |
+
"grad_norm": 0.0,
|
6517 |
+
"learning_rate": 2.584364999450059e-05,
|
6518 |
+
"loss": 1.159,
|
6519 |
+
"step": 930
|
6520 |
+
},
|
6521 |
+
{
|
6522 |
+
"epoch": 0.8717228464419475,
|
6523 |
+
"grad_norm": 0.0,
|
6524 |
+
"learning_rate": 2.5814027411098074e-05,
|
6525 |
+
"loss": 1.1388,
|
6526 |
+
"step": 931
|
6527 |
+
},
|
6528 |
+
{
|
6529 |
+
"epoch": 0.8726591760299626,
|
6530 |
+
"grad_norm": 0.0,
|
6531 |
+
"learning_rate": 2.5784390889304668e-05,
|
6532 |
+
"loss": 1.1621,
|
6533 |
+
"step": 932
|
6534 |
+
},
|
6535 |
+
{
|
6536 |
+
"epoch": 0.8735955056179775,
|
6537 |
+
"grad_norm": 0.0,
|
6538 |
+
"learning_rate": 2.5754740500170172e-05,
|
6539 |
+
"loss": 1.177,
|
6540 |
+
"step": 933
|
6541 |
+
},
|
6542 |
+
{
|
6543 |
+
"epoch": 0.8745318352059925,
|
6544 |
+
"grad_norm": 0.0,
|
6545 |
+
"learning_rate": 2.5725076314777623e-05,
|
6546 |
+
"loss": 1.1238,
|
6547 |
+
"step": 934
|
6548 |
+
},
|
6549 |
+
{
|
6550 |
+
"epoch": 0.8754681647940075,
|
6551 |
+
"grad_norm": 0.0,
|
6552 |
+
"learning_rate": 2.5695398404243116e-05,
|
6553 |
+
"loss": 1.1821,
|
6554 |
+
"step": 935
|
6555 |
+
},
|
6556 |
+
{
|
6557 |
+
"epoch": 0.8764044943820225,
|
6558 |
+
"grad_norm": 0.0,
|
6559 |
+
"learning_rate": 2.566570683971568e-05,
|
6560 |
+
"loss": 1.1406,
|
6561 |
+
"step": 936
|
6562 |
+
},
|
6563 |
+
{
|
6564 |
+
"epoch": 0.8773408239700374,
|
6565 |
+
"grad_norm": 0.0,
|
6566 |
+
"learning_rate": 2.5636001692377056e-05,
|
6567 |
+
"loss": 1.1174,
|
6568 |
+
"step": 937
|
6569 |
+
},
|
6570 |
+
{
|
6571 |
+
"epoch": 0.8782771535580525,
|
6572 |
+
"grad_norm": 0.0,
|
6573 |
+
"learning_rate": 2.5606283033441558e-05,
|
6574 |
+
"loss": 1.1988,
|
6575 |
+
"step": 938
|
6576 |
+
},
|
6577 |
+
{
|
6578 |
+
"epoch": 0.8792134831460674,
|
6579 |
+
"grad_norm": 0.0,
|
6580 |
+
"learning_rate": 2.5576550934155888e-05,
|
6581 |
+
"loss": 1.1353,
|
6582 |
+
"step": 939
|
6583 |
+
},
|
6584 |
+
{
|
6585 |
+
"epoch": 0.8801498127340824,
|
6586 |
+
"grad_norm": 0.0,
|
6587 |
+
"learning_rate": 2.5546805465798964e-05,
|
6588 |
+
"loss": 1.1899,
|
6589 |
+
"step": 940
|
6590 |
+
},
|
6591 |
+
{
|
6592 |
+
"epoch": 0.8810861423220974,
|
6593 |
+
"grad_norm": 0.0,
|
6594 |
+
"learning_rate": 2.551704669968177e-05,
|
6595 |
+
"loss": 1.1453,
|
6596 |
+
"step": 941
|
6597 |
+
},
|
6598 |
+
{
|
6599 |
+
"epoch": 0.8820224719101124,
|
6600 |
+
"grad_norm": 0.0,
|
6601 |
+
"learning_rate": 2.5487274707147158e-05,
|
6602 |
+
"loss": 1.1262,
|
6603 |
+
"step": 942
|
6604 |
+
},
|
6605 |
+
{
|
6606 |
+
"epoch": 0.8829588014981273,
|
6607 |
+
"grad_norm": 0.0,
|
6608 |
+
"learning_rate": 2.545748955956969e-05,
|
6609 |
+
"loss": 1.1586,
|
6610 |
+
"step": 943
|
6611 |
+
},
|
6612 |
+
{
|
6613 |
+
"epoch": 0.8838951310861424,
|
6614 |
+
"grad_norm": 0.0,
|
6615 |
+
"learning_rate": 2.5427691328355472e-05,
|
6616 |
+
"loss": 1.1651,
|
6617 |
+
"step": 944
|
6618 |
+
},
|
6619 |
+
{
|
6620 |
+
"epoch": 0.8848314606741573,
|
6621 |
+
"grad_norm": 0.0,
|
6622 |
+
"learning_rate": 2.5397880084941963e-05,
|
6623 |
+
"loss": 1.1174,
|
6624 |
+
"step": 945
|
6625 |
+
},
|
6626 |
+
{
|
6627 |
+
"epoch": 0.8857677902621723,
|
6628 |
+
"grad_norm": 0.0,
|
6629 |
+
"learning_rate": 2.5368055900797837e-05,
|
6630 |
+
"loss": 1.1689,
|
6631 |
+
"step": 946
|
6632 |
+
},
|
6633 |
+
{
|
6634 |
+
"epoch": 0.8867041198501873,
|
6635 |
+
"grad_norm": 0.0,
|
6636 |
+
"learning_rate": 2.5338218847422772e-05,
|
6637 |
+
"loss": 1.1399,
|
6638 |
+
"step": 947
|
6639 |
+
},
|
6640 |
+
{
|
6641 |
+
"epoch": 0.8876404494382022,
|
6642 |
+
"grad_norm": 0.0,
|
6643 |
+
"learning_rate": 2.5308368996347297e-05,
|
6644 |
+
"loss": 1.124,
|
6645 |
+
"step": 948
|
6646 |
+
},
|
6647 |
+
{
|
6648 |
+
"epoch": 0.8885767790262172,
|
6649 |
+
"grad_norm": 0.0,
|
6650 |
+
"learning_rate": 2.5278506419132663e-05,
|
6651 |
+
"loss": 1.1064,
|
6652 |
+
"step": 949
|
6653 |
+
},
|
6654 |
+
{
|
6655 |
+
"epoch": 0.8895131086142322,
|
6656 |
+
"grad_norm": 0.0,
|
6657 |
+
"learning_rate": 2.5248631187370568e-05,
|
6658 |
+
"loss": 1.1786,
|
6659 |
+
"step": 950
|
6660 |
+
},
|
6661 |
+
{
|
6662 |
+
"epoch": 0.8904494382022472,
|
6663 |
+
"grad_norm": 0.0,
|
6664 |
+
"learning_rate": 2.521874337268309e-05,
|
6665 |
+
"loss": 1.1104,
|
6666 |
+
"step": 951
|
6667 |
+
},
|
6668 |
+
{
|
6669 |
+
"epoch": 0.8913857677902621,
|
6670 |
+
"grad_norm": 0.0,
|
6671 |
+
"learning_rate": 2.5188843046722472e-05,
|
6672 |
+
"loss": 1.1299,
|
6673 |
+
"step": 952
|
6674 |
+
},
|
6675 |
+
{
|
6676 |
+
"epoch": 0.8923220973782772,
|
6677 |
+
"grad_norm": 0.0,
|
6678 |
+
"learning_rate": 2.5158930281170912e-05,
|
6679 |
+
"loss": 1.1297,
|
6680 |
+
"step": 953
|
6681 |
+
},
|
6682 |
+
{
|
6683 |
+
"epoch": 0.8932584269662921,
|
6684 |
+
"grad_norm": 0.0,
|
6685 |
+
"learning_rate": 2.5129005147740486e-05,
|
6686 |
+
"loss": 1.1213,
|
6687 |
+
"step": 954
|
6688 |
+
},
|
6689 |
+
{
|
6690 |
+
"epoch": 0.8941947565543071,
|
6691 |
+
"grad_norm": 0.0,
|
6692 |
+
"learning_rate": 2.5099067718172888e-05,
|
6693 |
+
"loss": 1.1517,
|
6694 |
+
"step": 955
|
6695 |
+
},
|
6696 |
+
{
|
6697 |
+
"epoch": 0.8951310861423221,
|
6698 |
+
"grad_norm": 0.0,
|
6699 |
+
"learning_rate": 2.5069118064239293e-05,
|
6700 |
+
"loss": 1.204,
|
6701 |
+
"step": 956
|
6702 |
+
},
|
6703 |
+
{
|
6704 |
+
"epoch": 0.8960674157303371,
|
6705 |
+
"grad_norm": 0.0,
|
6706 |
+
"learning_rate": 2.5039156257740178e-05,
|
6707 |
+
"loss": 1.1408,
|
6708 |
+
"step": 957
|
6709 |
+
},
|
6710 |
+
{
|
6711 |
+
"epoch": 0.897003745318352,
|
6712 |
+
"grad_norm": 0.0,
|
6713 |
+
"learning_rate": 2.5009182370505173e-05,
|
6714 |
+
"loss": 1.1192,
|
6715 |
+
"step": 958
|
6716 |
+
},
|
6717 |
+
{
|
6718 |
+
"epoch": 0.8979400749063671,
|
6719 |
+
"grad_norm": 0.0,
|
6720 |
+
"learning_rate": 2.4979196474392846e-05,
|
6721 |
+
"loss": 1.1376,
|
6722 |
+
"step": 959
|
6723 |
+
},
|
6724 |
+
{
|
6725 |
+
"epoch": 0.898876404494382,
|
6726 |
+
"grad_norm": 0.0,
|
6727 |
+
"learning_rate": 2.4949198641290582e-05,
|
6728 |
+
"loss": 1.1704,
|
6729 |
+
"step": 960
|
6730 |
+
},
|
6731 |
+
{
|
6732 |
+
"epoch": 0.899812734082397,
|
6733 |
+
"grad_norm": 0.0,
|
6734 |
+
"learning_rate": 2.491918894311436e-05,
|
6735 |
+
"loss": 1.1669,
|
6736 |
+
"step": 961
|
6737 |
+
},
|
6738 |
+
{
|
6739 |
+
"epoch": 0.900749063670412,
|
6740 |
+
"grad_norm": 0.0,
|
6741 |
+
"learning_rate": 2.4889167451808613e-05,
|
6742 |
+
"loss": 1.134,
|
6743 |
+
"step": 962
|
6744 |
+
},
|
6745 |
+
{
|
6746 |
+
"epoch": 0.901685393258427,
|
6747 |
+
"grad_norm": 0.0,
|
6748 |
+
"learning_rate": 2.4859134239346046e-05,
|
6749 |
+
"loss": 1.125,
|
6750 |
+
"step": 963
|
6751 |
+
},
|
6752 |
+
{
|
6753 |
+
"epoch": 0.9026217228464419,
|
6754 |
+
"grad_norm": 0.0,
|
6755 |
+
"learning_rate": 2.4829089377727464e-05,
|
6756 |
+
"loss": 1.116,
|
6757 |
+
"step": 964
|
6758 |
+
},
|
6759 |
+
{
|
6760 |
+
"epoch": 0.903558052434457,
|
6761 |
+
"grad_norm": 0.0,
|
6762 |
+
"learning_rate": 2.4799032938981604e-05,
|
6763 |
+
"loss": 1.1616,
|
6764 |
+
"step": 965
|
6765 |
+
},
|
6766 |
+
{
|
6767 |
+
"epoch": 0.9044943820224719,
|
6768 |
+
"grad_norm": 0.0,
|
6769 |
+
"learning_rate": 2.4768964995164947e-05,
|
6770 |
+
"loss": 1.1526,
|
6771 |
+
"step": 966
|
6772 |
+
},
|
6773 |
+
{
|
6774 |
+
"epoch": 0.9054307116104869,
|
6775 |
+
"grad_norm": 0.0,
|
6776 |
+
"learning_rate": 2.473888561836157e-05,
|
6777 |
+
"loss": 1.1605,
|
6778 |
+
"step": 967
|
6779 |
+
},
|
6780 |
+
{
|
6781 |
+
"epoch": 0.9063670411985019,
|
6782 |
+
"grad_norm": 0.0,
|
6783 |
+
"learning_rate": 2.470879488068294e-05,
|
6784 |
+
"loss": 1.1418,
|
6785 |
+
"step": 968
|
6786 |
+
},
|
6787 |
+
{
|
6788 |
+
"epoch": 0.9073033707865169,
|
6789 |
+
"grad_norm": 0.0,
|
6790 |
+
"learning_rate": 2.4678692854267784e-05,
|
6791 |
+
"loss": 1.1297,
|
6792 |
+
"step": 969
|
6793 |
+
},
|
6794 |
+
{
|
6795 |
+
"epoch": 0.9082397003745318,
|
6796 |
+
"grad_norm": 0.0,
|
6797 |
+
"learning_rate": 2.4648579611281868e-05,
|
6798 |
+
"loss": 1.1487,
|
6799 |
+
"step": 970
|
6800 |
+
},
|
6801 |
+
{
|
6802 |
+
"epoch": 0.9091760299625468,
|
6803 |
+
"grad_norm": 0.0,
|
6804 |
+
"learning_rate": 2.4618455223917868e-05,
|
6805 |
+
"loss": 1.1858,
|
6806 |
+
"step": 971
|
6807 |
+
},
|
6808 |
+
{
|
6809 |
+
"epoch": 0.9101123595505618,
|
6810 |
+
"grad_norm": 0.0,
|
6811 |
+
"learning_rate": 2.4588319764395164e-05,
|
6812 |
+
"loss": 1.151,
|
6813 |
+
"step": 972
|
6814 |
+
},
|
6815 |
+
{
|
6816 |
+
"epoch": 0.9110486891385767,
|
6817 |
+
"grad_norm": 0.0,
|
6818 |
+
"learning_rate": 2.45581733049597e-05,
|
6819 |
+
"loss": 1.1049,
|
6820 |
+
"step": 973
|
6821 |
+
},
|
6822 |
+
{
|
6823 |
+
"epoch": 0.9119850187265918,
|
6824 |
+
"grad_norm": 0.0,
|
6825 |
+
"learning_rate": 2.4528015917883754e-05,
|
6826 |
+
"loss": 1.1616,
|
6827 |
+
"step": 974
|
6828 |
+
},
|
6829 |
+
{
|
6830 |
+
"epoch": 0.9129213483146067,
|
6831 |
+
"grad_norm": 0.0,
|
6832 |
+
"learning_rate": 2.4497847675465834e-05,
|
6833 |
+
"loss": 1.1636,
|
6834 |
+
"step": 975
|
6835 |
+
},
|
6836 |
+
{
|
6837 |
+
"epoch": 0.9138576779026217,
|
6838 |
+
"grad_norm": 0.0,
|
6839 |
+
"learning_rate": 2.4467668650030474e-05,
|
6840 |
+
"loss": 1.1109,
|
6841 |
+
"step": 976
|
6842 |
+
},
|
6843 |
+
{
|
6844 |
+
"epoch": 0.9147940074906367,
|
6845 |
+
"grad_norm": 0.0,
|
6846 |
+
"learning_rate": 2.4437478913928026e-05,
|
6847 |
+
"loss": 1.1602,
|
6848 |
+
"step": 977
|
6849 |
+
},
|
6850 |
+
{
|
6851 |
+
"epoch": 0.9157303370786517,
|
6852 |
+
"grad_norm": 0.0,
|
6853 |
+
"learning_rate": 2.4407278539534555e-05,
|
6854 |
+
"loss": 1.1346,
|
6855 |
+
"step": 978
|
6856 |
+
},
|
6857 |
+
{
|
6858 |
+
"epoch": 0.9166666666666666,
|
6859 |
+
"grad_norm": 0.0,
|
6860 |
+
"learning_rate": 2.4377067599251616e-05,
|
6861 |
+
"loss": 1.1642,
|
6862 |
+
"step": 979
|
6863 |
+
},
|
6864 |
+
{
|
6865 |
+
"epoch": 0.9176029962546817,
|
6866 |
+
"grad_norm": 0.0,
|
6867 |
+
"learning_rate": 2.4346846165506085e-05,
|
6868 |
+
"loss": 1.1231,
|
6869 |
+
"step": 980
|
6870 |
+
},
|
6871 |
+
{
|
6872 |
+
"epoch": 0.9185393258426966,
|
6873 |
+
"grad_norm": 0.0,
|
6874 |
+
"learning_rate": 2.4316614310750007e-05,
|
6875 |
+
"loss": 1.1812,
|
6876 |
+
"step": 981
|
6877 |
+
},
|
6878 |
+
{
|
6879 |
+
"epoch": 0.9194756554307116,
|
6880 |
+
"grad_norm": 0.0,
|
6881 |
+
"learning_rate": 2.428637210746041e-05,
|
6882 |
+
"loss": 1.1728,
|
6883 |
+
"step": 982
|
6884 |
+
},
|
6885 |
+
{
|
6886 |
+
"epoch": 0.9204119850187266,
|
6887 |
+
"grad_norm": 0.0,
|
6888 |
+
"learning_rate": 2.425611962813913e-05,
|
6889 |
+
"loss": 1.1271,
|
6890 |
+
"step": 983
|
6891 |
+
},
|
6892 |
+
{
|
6893 |
+
"epoch": 0.9213483146067416,
|
6894 |
+
"grad_norm": 0.0,
|
6895 |
+
"learning_rate": 2.4225856945312633e-05,
|
6896 |
+
"loss": 1.107,
|
6897 |
+
"step": 984
|
6898 |
+
},
|
6899 |
+
{
|
6900 |
+
"epoch": 0.9222846441947565,
|
6901 |
+
"grad_norm": 0.0,
|
6902 |
+
"learning_rate": 2.4195584131531845e-05,
|
6903 |
+
"loss": 1.1474,
|
6904 |
+
"step": 985
|
6905 |
+
},
|
6906 |
+
{
|
6907 |
+
"epoch": 0.9232209737827716,
|
6908 |
+
"grad_norm": 0.0,
|
6909 |
+
"learning_rate": 2.4165301259371995e-05,
|
6910 |
+
"loss": 1.154,
|
6911 |
+
"step": 986
|
6912 |
+
},
|
6913 |
+
{
|
6914 |
+
"epoch": 0.9241573033707865,
|
6915 |
+
"grad_norm": 0.0,
|
6916 |
+
"learning_rate": 2.4135008401432404e-05,
|
6917 |
+
"loss": 1.1346,
|
6918 |
+
"step": 987
|
6919 |
+
},
|
6920 |
+
{
|
6921 |
+
"epoch": 0.9250936329588015,
|
6922 |
+
"grad_norm": 0.0,
|
6923 |
+
"learning_rate": 2.4104705630336353e-05,
|
6924 |
+
"loss": 1.1752,
|
6925 |
+
"step": 988
|
6926 |
+
},
|
6927 |
+
{
|
6928 |
+
"epoch": 0.9260299625468165,
|
6929 |
+
"grad_norm": 0.0,
|
6930 |
+
"learning_rate": 2.4074393018730882e-05,
|
6931 |
+
"loss": 1.1118,
|
6932 |
+
"step": 989
|
6933 |
+
},
|
6934 |
+
{
|
6935 |
+
"epoch": 0.9269662921348315,
|
6936 |
+
"grad_norm": 0.0,
|
6937 |
+
"learning_rate": 2.404407063928661e-05,
|
6938 |
+
"loss": 1.1517,
|
6939 |
+
"step": 990
|
6940 |
+
},
|
6941 |
+
{
|
6942 |
+
"epoch": 0.9279026217228464,
|
6943 |
+
"grad_norm": 0.0,
|
6944 |
+
"learning_rate": 2.401373856469759e-05,
|
6945 |
+
"loss": 1.1422,
|
6946 |
+
"step": 991
|
6947 |
+
},
|
6948 |
+
{
|
6949 |
+
"epoch": 0.9288389513108615,
|
6950 |
+
"grad_norm": 0.0,
|
6951 |
+
"learning_rate": 2.398339686768111e-05,
|
6952 |
+
"loss": 1.1641,
|
6953 |
+
"step": 992
|
6954 |
+
},
|
6955 |
+
{
|
6956 |
+
"epoch": 0.9297752808988764,
|
6957 |
+
"grad_norm": 0.0,
|
6958 |
+
"learning_rate": 2.395304562097753e-05,
|
6959 |
+
"loss": 1.1659,
|
6960 |
+
"step": 993
|
6961 |
+
},
|
6962 |
+
{
|
6963 |
+
"epoch": 0.9307116104868914,
|
6964 |
+
"grad_norm": 0.0,
|
6965 |
+
"learning_rate": 2.3922684897350093e-05,
|
6966 |
+
"loss": 1.1311,
|
6967 |
+
"step": 994
|
6968 |
+
},
|
6969 |
+
{
|
6970 |
+
"epoch": 0.9316479400749064,
|
6971 |
+
"grad_norm": 0.0,
|
6972 |
+
"learning_rate": 2.3892314769584778e-05,
|
6973 |
+
"loss": 1.1237,
|
6974 |
+
"step": 995
|
6975 |
+
},
|
6976 |
+
{
|
6977 |
+
"epoch": 0.9325842696629213,
|
6978 |
+
"grad_norm": 0.0,
|
6979 |
+
"learning_rate": 2.3861935310490103e-05,
|
6980 |
+
"loss": 1.1237,
|
6981 |
+
"step": 996
|
6982 |
+
},
|
6983 |
+
{
|
6984 |
+
"epoch": 0.9335205992509363,
|
6985 |
+
"grad_norm": 0.0,
|
6986 |
+
"learning_rate": 2.3831546592896954e-05,
|
6987 |
+
"loss": 1.156,
|
6988 |
+
"step": 997
|
6989 |
+
},
|
6990 |
+
{
|
6991 |
+
"epoch": 0.9344569288389513,
|
6992 |
+
"grad_norm": 0.0,
|
6993 |
+
"learning_rate": 2.3801148689658406e-05,
|
6994 |
+
"loss": 1.1329,
|
6995 |
+
"step": 998
|
6996 |
+
},
|
6997 |
+
{
|
6998 |
+
"epoch": 0.9353932584269663,
|
6999 |
+
"grad_norm": 0.0,
|
7000 |
+
"learning_rate": 2.3770741673649577e-05,
|
7001 |
+
"loss": 1.1102,
|
7002 |
+
"step": 999
|
7003 |
+
},
|
7004 |
+
{
|
7005 |
+
"epoch": 0.9363295880149812,
|
7006 |
+
"grad_norm": 0.0,
|
7007 |
+
"learning_rate": 2.3740325617767406e-05,
|
7008 |
+
"loss": 1.1609,
|
7009 |
+
"step": 1000
|
7010 |
+
},
|
7011 |
+
{
|
7012 |
+
"epoch": 0.9372659176029963,
|
7013 |
+
"grad_norm": 0.0,
|
7014 |
+
"learning_rate": 2.3709900594930533e-05,
|
7015 |
+
"loss": 1.1311,
|
7016 |
+
"step": 1001
|
7017 |
+
},
|
7018 |
+
{
|
7019 |
+
"epoch": 0.9382022471910112,
|
7020 |
+
"grad_norm": 0.0,
|
7021 |
+
"learning_rate": 2.367946667807906e-05,
|
7022 |
+
"loss": 1.1693,
|
7023 |
+
"step": 1002
|
7024 |
+
},
|
7025 |
+
{
|
7026 |
+
"epoch": 0.9391385767790262,
|
7027 |
+
"grad_norm": 0.0,
|
7028 |
+
"learning_rate": 2.3649023940174435e-05,
|
7029 |
+
"loss": 1.1453,
|
7030 |
+
"step": 1003
|
7031 |
+
},
|
7032 |
+
{
|
7033 |
+
"epoch": 0.9400749063670412,
|
7034 |
+
"grad_norm": 0.0,
|
7035 |
+
"learning_rate": 2.3618572454199264e-05,
|
7036 |
+
"loss": 1.128,
|
7037 |
+
"step": 1004
|
7038 |
+
},
|
7039 |
+
{
|
7040 |
+
"epoch": 0.9410112359550562,
|
7041 |
+
"grad_norm": 0.0,
|
7042 |
+
"learning_rate": 2.358811229315709e-05,
|
7043 |
+
"loss": 1.1582,
|
7044 |
+
"step": 1005
|
7045 |
+
},
|
7046 |
+
{
|
7047 |
+
"epoch": 0.9419475655430711,
|
7048 |
+
"grad_norm": 0.0,
|
7049 |
+
"learning_rate": 2.3557643530072284e-05,
|
7050 |
+
"loss": 1.1259,
|
7051 |
+
"step": 1006
|
7052 |
+
},
|
7053 |
+
{
|
7054 |
+
"epoch": 0.9428838951310862,
|
7055 |
+
"grad_norm": 0.0,
|
7056 |
+
"learning_rate": 2.3527166237989825e-05,
|
7057 |
+
"loss": 1.1497,
|
7058 |
+
"step": 1007
|
7059 |
+
},
|
7060 |
+
{
|
7061 |
+
"epoch": 0.9438202247191011,
|
7062 |
+
"grad_norm": 0.0,
|
7063 |
+
"learning_rate": 2.3496680489975136e-05,
|
7064 |
+
"loss": 1.1312,
|
7065 |
+
"step": 1008
|
7066 |
+
},
|
7067 |
+
{
|
7068 |
+
"epoch": 0.9447565543071161,
|
7069 |
+
"grad_norm": 0.0,
|
7070 |
+
"learning_rate": 2.346618635911393e-05,
|
7071 |
+
"loss": 1.0997,
|
7072 |
+
"step": 1009
|
7073 |
+
},
|
7074 |
+
{
|
7075 |
+
"epoch": 0.9456928838951311,
|
7076 |
+
"grad_norm": 0.0,
|
7077 |
+
"learning_rate": 2.3435683918512e-05,
|
7078 |
+
"loss": 1.1551,
|
7079 |
+
"step": 1010
|
7080 |
+
},
|
7081 |
+
{
|
7082 |
+
"epoch": 0.9466292134831461,
|
7083 |
+
"grad_norm": 0.0,
|
7084 |
+
"learning_rate": 2.340517324129507e-05,
|
7085 |
+
"loss": 1.063,
|
7086 |
+
"step": 1011
|
7087 |
+
},
|
7088 |
+
{
|
7089 |
+
"epoch": 0.947565543071161,
|
7090 |
+
"grad_norm": 0.0,
|
7091 |
+
"learning_rate": 2.3374654400608595e-05,
|
7092 |
+
"loss": 1.1224,
|
7093 |
+
"step": 1012
|
7094 |
+
},
|
7095 |
+
{
|
7096 |
+
"epoch": 0.9485018726591761,
|
7097 |
+
"grad_norm": 0.0,
|
7098 |
+
"learning_rate": 2.3344127469617624e-05,
|
7099 |
+
"loss": 1.1681,
|
7100 |
+
"step": 1013
|
7101 |
+
},
|
7102 |
+
{
|
7103 |
+
"epoch": 0.949438202247191,
|
7104 |
+
"grad_norm": 0.0,
|
7105 |
+
"learning_rate": 2.3313592521506583e-05,
|
7106 |
+
"loss": 1.1698,
|
7107 |
+
"step": 1014
|
7108 |
+
},
|
7109 |
+
{
|
7110 |
+
"epoch": 0.950374531835206,
|
7111 |
+
"grad_norm": 0.0,
|
7112 |
+
"learning_rate": 2.328304962947913e-05,
|
7113 |
+
"loss": 1.1309,
|
7114 |
+
"step": 1015
|
7115 |
+
},
|
7116 |
+
{
|
7117 |
+
"epoch": 0.951310861423221,
|
7118 |
+
"grad_norm": 0.0,
|
7119 |
+
"learning_rate": 2.3252498866757952e-05,
|
7120 |
+
"loss": 1.1284,
|
7121 |
+
"step": 1016
|
7122 |
+
},
|
7123 |
+
{
|
7124 |
+
"epoch": 0.952247191011236,
|
7125 |
+
"grad_norm": 0.0,
|
7126 |
+
"learning_rate": 2.3221940306584627e-05,
|
7127 |
+
"loss": 1.1681,
|
7128 |
+
"step": 1017
|
7129 |
+
},
|
7130 |
+
{
|
7131 |
+
"epoch": 0.9531835205992509,
|
7132 |
+
"grad_norm": 0.0,
|
7133 |
+
"learning_rate": 2.3191374022219407e-05,
|
7134 |
+
"loss": 1.1858,
|
7135 |
+
"step": 1018
|
7136 |
+
},
|
7137 |
+
{
|
7138 |
+
"epoch": 0.954119850187266,
|
7139 |
+
"grad_norm": 0.0,
|
7140 |
+
"learning_rate": 2.3160800086941075e-05,
|
7141 |
+
"loss": 1.1038,
|
7142 |
+
"step": 1019
|
7143 |
+
},
|
7144 |
+
{
|
7145 |
+
"epoch": 0.9550561797752809,
|
7146 |
+
"grad_norm": 0.0,
|
7147 |
+
"learning_rate": 2.313021857404674e-05,
|
7148 |
+
"loss": 1.1651,
|
7149 |
+
"step": 1020
|
7150 |
+
},
|
7151 |
+
{
|
7152 |
+
"epoch": 0.9559925093632958,
|
7153 |
+
"grad_norm": 0.0,
|
7154 |
+
"learning_rate": 2.30996295568517e-05,
|
7155 |
+
"loss": 1.1355,
|
7156 |
+
"step": 1021
|
7157 |
+
},
|
7158 |
+
{
|
7159 |
+
"epoch": 0.9569288389513109,
|
7160 |
+
"grad_norm": 0.0,
|
7161 |
+
"learning_rate": 2.306903310868922e-05,
|
7162 |
+
"loss": 1.1433,
|
7163 |
+
"step": 1022
|
7164 |
+
},
|
7165 |
+
{
|
7166 |
+
"epoch": 0.9578651685393258,
|
7167 |
+
"grad_norm": 0.0,
|
7168 |
+
"learning_rate": 2.3038429302910402e-05,
|
7169 |
+
"loss": 1.1333,
|
7170 |
+
"step": 1023
|
7171 |
+
},
|
7172 |
+
{
|
7173 |
+
"epoch": 0.9588014981273408,
|
7174 |
+
"grad_norm": 0.0,
|
7175 |
+
"learning_rate": 2.3007818212883967e-05,
|
7176 |
+
"loss": 1.1077,
|
7177 |
+
"step": 1024
|
7178 |
+
},
|
7179 |
+
{
|
7180 |
+
"epoch": 0.9597378277153558,
|
7181 |
+
"grad_norm": 0.0,
|
7182 |
+
"learning_rate": 2.297719991199611e-05,
|
7183 |
+
"loss": 1.0909,
|
7184 |
+
"step": 1025
|
7185 |
+
},
|
7186 |
+
{
|
7187 |
+
"epoch": 0.9606741573033708,
|
7188 |
+
"grad_norm": 0.0,
|
7189 |
+
"learning_rate": 2.2946574473650318e-05,
|
7190 |
+
"loss": 1.1333,
|
7191 |
+
"step": 1026
|
7192 |
+
},
|
7193 |
+
{
|
7194 |
+
"epoch": 0.9616104868913857,
|
7195 |
+
"grad_norm": 0.0,
|
7196 |
+
"learning_rate": 2.2915941971267172e-05,
|
7197 |
+
"loss": 1.1264,
|
7198 |
+
"step": 1027
|
7199 |
+
},
|
7200 |
+
{
|
7201 |
+
"epoch": 0.9625468164794008,
|
7202 |
+
"grad_norm": 0.0,
|
7203 |
+
"learning_rate": 2.288530247828421e-05,
|
7204 |
+
"loss": 1.1528,
|
7205 |
+
"step": 1028
|
7206 |
+
},
|
7207 |
+
{
|
7208 |
+
"epoch": 0.9634831460674157,
|
7209 |
+
"grad_norm": 0.0,
|
7210 |
+
"learning_rate": 2.2854656068155706e-05,
|
7211 |
+
"loss": 1.1345,
|
7212 |
+
"step": 1029
|
7213 |
+
},
|
7214 |
+
{
|
7215 |
+
"epoch": 0.9644194756554307,
|
7216 |
+
"grad_norm": 0.0,
|
7217 |
+
"learning_rate": 2.282400281435253e-05,
|
7218 |
+
"loss": 1.1975,
|
7219 |
+
"step": 1030
|
7220 |
+
},
|
7221 |
+
{
|
7222 |
+
"epoch": 0.9653558052434457,
|
7223 |
+
"grad_norm": 0.0,
|
7224 |
+
"learning_rate": 2.2793342790361966e-05,
|
7225 |
+
"loss": 1.1709,
|
7226 |
+
"step": 1031
|
7227 |
+
},
|
7228 |
+
{
|
7229 |
+
"epoch": 0.9662921348314607,
|
7230 |
+
"grad_norm": 0.0,
|
7231 |
+
"learning_rate": 2.2762676069687517e-05,
|
7232 |
+
"loss": 1.1301,
|
7233 |
+
"step": 1032
|
7234 |
+
},
|
7235 |
+
{
|
7236 |
+
"epoch": 0.9672284644194756,
|
7237 |
+
"grad_norm": 0.0,
|
7238 |
+
"learning_rate": 2.273200272584874e-05,
|
7239 |
+
"loss": 1.1683,
|
7240 |
+
"step": 1033
|
7241 |
+
},
|
7242 |
+
{
|
7243 |
+
"epoch": 0.9681647940074907,
|
7244 |
+
"grad_norm": 0.0,
|
7245 |
+
"learning_rate": 2.2701322832381072e-05,
|
7246 |
+
"loss": 1.0972,
|
7247 |
+
"step": 1034
|
7248 |
+
},
|
7249 |
+
{
|
7250 |
+
"epoch": 0.9691011235955056,
|
7251 |
+
"grad_norm": 0.0,
|
7252 |
+
"learning_rate": 2.2670636462835666e-05,
|
7253 |
+
"loss": 1.1972,
|
7254 |
+
"step": 1035
|
7255 |
+
},
|
7256 |
+
{
|
7257 |
+
"epoch": 0.9700374531835206,
|
7258 |
+
"grad_norm": 0.0,
|
7259 |
+
"learning_rate": 2.2639943690779168e-05,
|
7260 |
+
"loss": 1.1733,
|
7261 |
+
"step": 1036
|
7262 |
+
},
|
7263 |
+
{
|
7264 |
+
"epoch": 0.9709737827715356,
|
7265 |
+
"grad_norm": 0.0,
|
7266 |
+
"learning_rate": 2.2609244589793606e-05,
|
7267 |
+
"loss": 1.1569,
|
7268 |
+
"step": 1037
|
7269 |
+
},
|
7270 |
+
{
|
7271 |
+
"epoch": 0.9719101123595506,
|
7272 |
+
"grad_norm": 0.0,
|
7273 |
+
"learning_rate": 2.2578539233476176e-05,
|
7274 |
+
"loss": 1.1468,
|
7275 |
+
"step": 1038
|
7276 |
+
},
|
7277 |
+
{
|
7278 |
+
"epoch": 0.9728464419475655,
|
7279 |
+
"grad_norm": 0.0,
|
7280 |
+
"learning_rate": 2.2547827695439046e-05,
|
7281 |
+
"loss": 1.1192,
|
7282 |
+
"step": 1039
|
7283 |
+
},
|
7284 |
+
{
|
7285 |
+
"epoch": 0.9737827715355806,
|
7286 |
+
"grad_norm": 0.0,
|
7287 |
+
"learning_rate": 2.2517110049309222e-05,
|
7288 |
+
"loss": 1.2081,
|
7289 |
+
"step": 1040
|
7290 |
+
},
|
7291 |
+
{
|
7292 |
+
"epoch": 0.9747191011235955,
|
7293 |
+
"grad_norm": 0.0,
|
7294 |
+
"learning_rate": 2.2486386368728367e-05,
|
7295 |
+
"loss": 1.1871,
|
7296 |
+
"step": 1041
|
7297 |
+
},
|
7298 |
+
{
|
7299 |
+
"epoch": 0.9756554307116105,
|
7300 |
+
"grad_norm": 0.0,
|
7301 |
+
"learning_rate": 2.2455656727352578e-05,
|
7302 |
+
"loss": 1.1106,
|
7303 |
+
"step": 1042
|
7304 |
+
},
|
7305 |
+
{
|
7306 |
+
"epoch": 0.9765917602996255,
|
7307 |
+
"grad_norm": 0.0,
|
7308 |
+
"learning_rate": 2.242492119885227e-05,
|
7309 |
+
"loss": 1.1256,
|
7310 |
+
"step": 1043
|
7311 |
+
},
|
7312 |
+
{
|
7313 |
+
"epoch": 0.9775280898876404,
|
7314 |
+
"grad_norm": 0.0,
|
7315 |
+
"learning_rate": 2.239417985691196e-05,
|
7316 |
+
"loss": 1.1416,
|
7317 |
+
"step": 1044
|
7318 |
+
},
|
7319 |
+
{
|
7320 |
+
"epoch": 0.9784644194756554,
|
7321 |
+
"grad_norm": 0.0,
|
7322 |
+
"learning_rate": 2.2363432775230108e-05,
|
7323 |
+
"loss": 1.1216,
|
7324 |
+
"step": 1045
|
7325 |
+
},
|
7326 |
+
{
|
7327 |
+
"epoch": 0.9794007490636704,
|
7328 |
+
"grad_norm": 0.0,
|
7329 |
+
"learning_rate": 2.2332680027518917e-05,
|
7330 |
+
"loss": 1.1514,
|
7331 |
+
"step": 1046
|
7332 |
+
},
|
7333 |
+
{
|
7334 |
+
"epoch": 0.9803370786516854,
|
7335 |
+
"grad_norm": 0.0,
|
7336 |
+
"learning_rate": 2.2301921687504193e-05,
|
7337 |
+
"loss": 1.1427,
|
7338 |
+
"step": 1047
|
7339 |
+
},
|
7340 |
+
{
|
7341 |
+
"epoch": 0.9812734082397003,
|
7342 |
+
"grad_norm": 0.0,
|
7343 |
+
"learning_rate": 2.2271157828925145e-05,
|
7344 |
+
"loss": 1.1516,
|
7345 |
+
"step": 1048
|
7346 |
+
},
|
7347 |
+
{
|
7348 |
+
"epoch": 0.9822097378277154,
|
7349 |
+
"grad_norm": 0.0,
|
7350 |
+
"learning_rate": 2.2240388525534208e-05,
|
7351 |
+
"loss": 1.1911,
|
7352 |
+
"step": 1049
|
7353 |
+
},
|
7354 |
+
{
|
7355 |
+
"epoch": 0.9831460674157303,
|
7356 |
+
"grad_norm": 0.0,
|
7357 |
+
"learning_rate": 2.220961385109687e-05,
|
7358 |
+
"loss": 1.1308,
|
7359 |
+
"step": 1050
|
7360 |
+
},
|
7361 |
+
{
|
7362 |
+
"epoch": 0.9840823970037453,
|
7363 |
+
"grad_norm": 0.0,
|
7364 |
+
"learning_rate": 2.21788338793915e-05,
|
7365 |
+
"loss": 1.1448,
|
7366 |
+
"step": 1051
|
7367 |
+
},
|
7368 |
+
{
|
7369 |
+
"epoch": 0.9850187265917603,
|
7370 |
+
"grad_norm": 0.0,
|
7371 |
+
"learning_rate": 2.214804868420916e-05,
|
7372 |
+
"loss": 1.1489,
|
7373 |
+
"step": 1052
|
7374 |
+
},
|
7375 |
+
{
|
7376 |
+
"epoch": 0.9859550561797753,
|
7377 |
+
"grad_norm": 0.0,
|
7378 |
+
"learning_rate": 2.211725833935343e-05,
|
7379 |
+
"loss": 1.1567,
|
7380 |
+
"step": 1053
|
7381 |
+
},
|
7382 |
+
{
|
7383 |
+
"epoch": 0.9868913857677902,
|
7384 |
+
"grad_norm": 0.0,
|
7385 |
+
"learning_rate": 2.2086462918640264e-05,
|
7386 |
+
"loss": 1.1834,
|
7387 |
+
"step": 1054
|
7388 |
+
},
|
7389 |
+
{
|
7390 |
+
"epoch": 0.9878277153558053,
|
7391 |
+
"grad_norm": 0.0,
|
7392 |
+
"learning_rate": 2.2055662495897745e-05,
|
7393 |
+
"loss": 1.1503,
|
7394 |
+
"step": 1055
|
7395 |
+
},
|
7396 |
+
{
|
7397 |
+
"epoch": 0.9887640449438202,
|
7398 |
+
"grad_norm": 0.0,
|
7399 |
+
"learning_rate": 2.2024857144965978e-05,
|
7400 |
+
"loss": 1.1624,
|
7401 |
+
"step": 1056
|
7402 |
+
},
|
7403 |
+
{
|
7404 |
+
"epoch": 0.9897003745318352,
|
7405 |
+
"grad_norm": 0.0,
|
7406 |
+
"learning_rate": 2.1994046939696877e-05,
|
7407 |
+
"loss": 1.1295,
|
7408 |
+
"step": 1057
|
7409 |
+
},
|
7410 |
+
{
|
7411 |
+
"epoch": 0.9906367041198502,
|
7412 |
+
"grad_norm": 0.0,
|
7413 |
+
"learning_rate": 2.196323195395397e-05,
|
7414 |
+
"loss": 1.1411,
|
7415 |
+
"step": 1058
|
7416 |
+
},
|
7417 |
+
{
|
7418 |
+
"epoch": 0.9915730337078652,
|
7419 |
+
"grad_norm": 0.0,
|
7420 |
+
"learning_rate": 2.193241226161228e-05,
|
7421 |
+
"loss": 1.1263,
|
7422 |
+
"step": 1059
|
7423 |
+
},
|
7424 |
+
{
|
7425 |
+
"epoch": 0.9925093632958801,
|
7426 |
+
"grad_norm": 0.0,
|
7427 |
+
"learning_rate": 2.190158793655809e-05,
|
7428 |
+
"loss": 1.1258,
|
7429 |
+
"step": 1060
|
7430 |
+
},
|
7431 |
+
{
|
7432 |
+
"epoch": 0.9934456928838952,
|
7433 |
+
"grad_norm": 0.0,
|
7434 |
+
"learning_rate": 2.1870759052688795e-05,
|
7435 |
+
"loss": 1.1264,
|
7436 |
+
"step": 1061
|
7437 |
+
},
|
7438 |
+
{
|
7439 |
+
"epoch": 0.9943820224719101,
|
7440 |
+
"grad_norm": 0.0,
|
7441 |
+
"learning_rate": 2.1839925683912734e-05,
|
7442 |
+
"loss": 1.1199,
|
7443 |
+
"step": 1062
|
7444 |
+
},
|
7445 |
+
{
|
7446 |
+
"epoch": 0.9953183520599251,
|
7447 |
+
"grad_norm": 0.0,
|
7448 |
+
"learning_rate": 2.1809087904148974e-05,
|
7449 |
+
"loss": 1.1546,
|
7450 |
+
"step": 1063
|
7451 |
+
},
|
7452 |
+
{
|
7453 |
+
"epoch": 0.9962546816479401,
|
7454 |
+
"grad_norm": 0.0,
|
7455 |
+
"learning_rate": 2.1778245787327168e-05,
|
7456 |
+
"loss": 1.1243,
|
7457 |
+
"step": 1064
|
7458 |
+
},
|
7459 |
+
{
|
7460 |
+
"epoch": 0.9971910112359551,
|
7461 |
+
"grad_norm": 0.0,
|
7462 |
+
"learning_rate": 2.1747399407387366e-05,
|
7463 |
+
"loss": 1.1177,
|
7464 |
+
"step": 1065
|
7465 |
+
},
|
7466 |
+
{
|
7467 |
+
"epoch": 0.99812734082397,
|
7468 |
+
"grad_norm": 0.0,
|
7469 |
+
"learning_rate": 2.171654883827985e-05,
|
7470 |
+
"loss": 1.1337,
|
7471 |
+
"step": 1066
|
7472 |
+
},
|
7473 |
+
{
|
7474 |
+
"epoch": 0.9990636704119851,
|
7475 |
+
"grad_norm": 0.0,
|
7476 |
+
"learning_rate": 2.1685694153964923e-05,
|
7477 |
+
"loss": 1.0604,
|
7478 |
+
"step": 1067
|
7479 |
+
},
|
7480 |
+
{
|
7481 |
+
"epoch": 1.0,
|
7482 |
+
"grad_norm": 0.0,
|
7483 |
+
"learning_rate": 2.165483542841276e-05,
|
7484 |
+
"loss": 1.1105,
|
7485 |
+
"step": 1068
|
7486 |
+
},
|
7487 |
+
{
|
7488 |
+
"epoch": 1.000936329588015,
|
7489 |
+
"grad_norm": 0.0,
|
7490 |
+
"learning_rate": 2.1623972735603237e-05,
|
7491 |
+
"loss": 1.1807,
|
7492 |
+
"step": 1069
|
7493 |
+
},
|
7494 |
+
{
|
7495 |
+
"epoch": 1.0018726591760299,
|
7496 |
+
"grad_norm": 0.0,
|
7497 |
+
"learning_rate": 2.1593106149525735e-05,
|
7498 |
+
"loss": 1.1901,
|
7499 |
+
"step": 1070
|
7500 |
+
},
|
7501 |
+
{
|
7502 |
+
"epoch": 1.002808988764045,
|
7503 |
+
"grad_norm": 0.0,
|
7504 |
+
"learning_rate": 2.156223574417896e-05,
|
7505 |
+
"loss": 1.1817,
|
7506 |
+
"step": 1071
|
7507 |
+
},
|
7508 |
+
{
|
7509 |
+
"epoch": 1.00374531835206,
|
7510 |
+
"grad_norm": 0.0,
|
7511 |
+
"learning_rate": 2.1531361593570788e-05,
|
7512 |
+
"loss": 1.1032,
|
7513 |
+
"step": 1072
|
7514 |
+
},
|
7515 |
+
{
|
7516 |
+
"epoch": 1.0046816479400749,
|
7517 |
+
"grad_norm": 0.0,
|
7518 |
+
"learning_rate": 2.1500483771718056e-05,
|
7519 |
+
"loss": 1.1515,
|
7520 |
+
"step": 1073
|
7521 |
+
},
|
7522 |
+
{
|
7523 |
+
"epoch": 1.0056179775280898,
|
7524 |
+
"grad_norm": 0.0,
|
7525 |
+
"learning_rate": 2.1469602352646417e-05,
|
7526 |
+
"loss": 1.1682,
|
7527 |
+
"step": 1074
|
7528 |
+
},
|
7529 |
+
{
|
7530 |
+
"epoch": 1.006554307116105,
|
7531 |
+
"grad_norm": 0.0,
|
7532 |
+
"learning_rate": 2.1438717410390138e-05,
|
7533 |
+
"loss": 1.1747,
|
7534 |
+
"step": 1075
|
7535 |
+
},
|
7536 |
+
{
|
7537 |
+
"epoch": 1.0074906367041199,
|
7538 |
+
"grad_norm": 0.0,
|
7539 |
+
"learning_rate": 2.1407829018991957e-05,
|
7540 |
+
"loss": 1.1021,
|
7541 |
+
"step": 1076
|
7542 |
+
},
|
7543 |
+
{
|
7544 |
+
"epoch": 1.0084269662921348,
|
7545 |
+
"grad_norm": 0.0,
|
7546 |
+
"learning_rate": 2.1376937252502846e-05,
|
7547 |
+
"loss": 1.1787,
|
7548 |
+
"step": 1077
|
7549 |
+
},
|
7550 |
+
{
|
7551 |
+
"epoch": 1.0093632958801497,
|
7552 |
+
"grad_norm": 0.0,
|
7553 |
+
"learning_rate": 2.1346042184981886e-05,
|
7554 |
+
"loss": 1.0454,
|
7555 |
+
"step": 1078
|
7556 |
+
},
|
7557 |
+
{
|
7558 |
+
"epoch": 1.0102996254681649,
|
7559 |
+
"grad_norm": 0.0,
|
7560 |
+
"learning_rate": 2.1315143890496074e-05,
|
7561 |
+
"loss": 1.1214,
|
7562 |
+
"step": 1079
|
7563 |
+
},
|
7564 |
+
{
|
7565 |
+
"epoch": 1.0112359550561798,
|
7566 |
+
"grad_norm": 0.0,
|
7567 |
+
"learning_rate": 2.1284242443120133e-05,
|
7568 |
+
"loss": 1.1246,
|
7569 |
+
"step": 1080
|
7570 |
+
},
|
7571 |
+
{
|
7572 |
+
"epoch": 1.0121722846441947,
|
7573 |
+
"grad_norm": 0.0,
|
7574 |
+
"learning_rate": 2.125333791693635e-05,
|
7575 |
+
"loss": 1.1144,
|
7576 |
+
"step": 1081
|
7577 |
+
},
|
7578 |
+
{
|
7579 |
+
"epoch": 1.0131086142322097,
|
7580 |
+
"grad_norm": 0.0,
|
7581 |
+
"learning_rate": 2.1222430386034406e-05,
|
7582 |
+
"loss": 1.1229,
|
7583 |
+
"step": 1082
|
7584 |
+
},
|
7585 |
+
{
|
7586 |
+
"epoch": 1.0140449438202248,
|
7587 |
+
"grad_norm": 0.0,
|
7588 |
+
"learning_rate": 2.1191519924511167e-05,
|
7589 |
+
"loss": 1.1704,
|
7590 |
+
"step": 1083
|
7591 |
+
},
|
7592 |
+
{
|
7593 |
+
"epoch": 1.0149812734082397,
|
7594 |
+
"grad_norm": 0.0,
|
7595 |
+
"learning_rate": 2.1160606606470523e-05,
|
7596 |
+
"loss": 1.1265,
|
7597 |
+
"step": 1084
|
7598 |
+
},
|
7599 |
+
{
|
7600 |
+
"epoch": 1.0159176029962547,
|
7601 |
+
"grad_norm": 0.0,
|
7602 |
+
"learning_rate": 2.1129690506023234e-05,
|
7603 |
+
"loss": 1.1264,
|
7604 |
+
"step": 1085
|
7605 |
+
},
|
7606 |
+
{
|
7607 |
+
"epoch": 1.0168539325842696,
|
7608 |
+
"grad_norm": 0.0,
|
7609 |
+
"learning_rate": 2.1098771697286703e-05,
|
7610 |
+
"loss": 1.1549,
|
7611 |
+
"step": 1086
|
7612 |
+
},
|
7613 |
+
{
|
7614 |
+
"epoch": 1.0177902621722847,
|
7615 |
+
"grad_norm": 0.0,
|
7616 |
+
"learning_rate": 2.1067850254384853e-05,
|
7617 |
+
"loss": 1.1881,
|
7618 |
+
"step": 1087
|
7619 |
+
},
|
7620 |
+
{
|
7621 |
+
"epoch": 1.0187265917602997,
|
7622 |
+
"grad_norm": 0.0,
|
7623 |
+
"learning_rate": 2.10369262514479e-05,
|
7624 |
+
"loss": 1.1419,
|
7625 |
+
"step": 1088
|
7626 |
+
},
|
7627 |
+
{
|
7628 |
+
"epoch": 1.0196629213483146,
|
7629 |
+
"grad_norm": 0.0,
|
7630 |
+
"learning_rate": 2.1005999762612208e-05,
|
7631 |
+
"loss": 1.156,
|
7632 |
+
"step": 1089
|
7633 |
+
},
|
7634 |
+
{
|
7635 |
+
"epoch": 1.0205992509363295,
|
7636 |
+
"grad_norm": 0.0,
|
7637 |
+
"learning_rate": 2.09750708620201e-05,
|
7638 |
+
"loss": 1.1403,
|
7639 |
+
"step": 1090
|
7640 |
+
},
|
7641 |
+
{
|
7642 |
+
"epoch": 1.0215355805243447,
|
7643 |
+
"grad_norm": 0.0,
|
7644 |
+
"learning_rate": 2.0944139623819676e-05,
|
7645 |
+
"loss": 1.1485,
|
7646 |
+
"step": 1091
|
7647 |
+
},
|
7648 |
+
{
|
7649 |
+
"epoch": 1.0224719101123596,
|
7650 |
+
"grad_norm": 0.0,
|
7651 |
+
"learning_rate": 2.0913206122164643e-05,
|
7652 |
+
"loss": 1.1764,
|
7653 |
+
"step": 1092
|
7654 |
+
},
|
7655 |
+
{
|
7656 |
+
"epoch": 1.0234082397003745,
|
7657 |
+
"grad_norm": 0.0,
|
7658 |
+
"learning_rate": 2.088227043121414e-05,
|
7659 |
+
"loss": 1.1304,
|
7660 |
+
"step": 1093
|
7661 |
+
},
|
7662 |
+
{
|
7663 |
+
"epoch": 1.0243445692883895,
|
7664 |
+
"grad_norm": 0.0,
|
7665 |
+
"learning_rate": 2.0851332625132554e-05,
|
7666 |
+
"loss": 1.1896,
|
7667 |
+
"step": 1094
|
7668 |
+
},
|
7669 |
+
{
|
7670 |
+
"epoch": 1.0252808988764044,
|
7671 |
+
"grad_norm": 0.0,
|
7672 |
+
"learning_rate": 2.0820392778089323e-05,
|
7673 |
+
"loss": 1.1414,
|
7674 |
+
"step": 1095
|
7675 |
+
},
|
7676 |
+
{
|
7677 |
+
"epoch": 1.0262172284644195,
|
7678 |
+
"grad_norm": 0.0,
|
7679 |
+
"learning_rate": 2.0789450964258805e-05,
|
7680 |
+
"loss": 1.1761,
|
7681 |
+
"step": 1096
|
7682 |
+
},
|
7683 |
+
{
|
7684 |
+
"epoch": 1.0271535580524345,
|
7685 |
+
"grad_norm": 0.0,
|
7686 |
+
"learning_rate": 2.075850725782005e-05,
|
7687 |
+
"loss": 1.1246,
|
7688 |
+
"step": 1097
|
7689 |
+
},
|
7690 |
+
{
|
7691 |
+
"epoch": 1.0280898876404494,
|
7692 |
+
"grad_norm": 0.0,
|
7693 |
+
"learning_rate": 2.0727561732956672e-05,
|
7694 |
+
"loss": 1.0788,
|
7695 |
+
"step": 1098
|
7696 |
+
},
|
7697 |
+
{
|
7698 |
+
"epoch": 1.0290262172284643,
|
7699 |
+
"grad_norm": 0.0,
|
7700 |
+
"learning_rate": 2.0696614463856624e-05,
|
7701 |
+
"loss": 1.1586,
|
7702 |
+
"step": 1099
|
7703 |
+
},
|
7704 |
+
{
|
7705 |
+
"epoch": 1.0299625468164795,
|
7706 |
+
"grad_norm": 0.0,
|
7707 |
+
"learning_rate": 2.066566552471204e-05,
|
7708 |
+
"loss": 1.1517,
|
7709 |
+
"step": 1100
|
7710 |
+
},
|
7711 |
+
{
|
7712 |
+
"epoch": 1.0308988764044944,
|
7713 |
+
"grad_norm": 0.0,
|
7714 |
+
"learning_rate": 2.0634714989719065e-05,
|
7715 |
+
"loss": 1.1935,
|
7716 |
+
"step": 1101
|
7717 |
+
},
|
7718 |
+
{
|
7719 |
+
"epoch": 1.0318352059925093,
|
7720 |
+
"grad_norm": 0.0,
|
7721 |
+
"learning_rate": 2.0603762933077675e-05,
|
7722 |
+
"loss": 1.1288,
|
7723 |
+
"step": 1102
|
7724 |
+
},
|
7725 |
+
{
|
7726 |
+
"epoch": 1.0327715355805243,
|
7727 |
+
"grad_norm": 0.0,
|
7728 |
+
"learning_rate": 2.0572809428991473e-05,
|
7729 |
+
"loss": 1.1996,
|
7730 |
+
"step": 1103
|
7731 |
+
},
|
7732 |
+
{
|
7733 |
+
"epoch": 1.0337078651685394,
|
7734 |
+
"grad_norm": 0.0,
|
7735 |
+
"learning_rate": 2.054185455166756e-05,
|
7736 |
+
"loss": 1.1287,
|
7737 |
+
"step": 1104
|
7738 |
+
},
|
7739 |
+
{
|
7740 |
+
"epoch": 1.0346441947565543,
|
7741 |
+
"grad_norm": 0.0,
|
7742 |
+
"learning_rate": 2.051089837531631e-05,
|
7743 |
+
"loss": 1.1478,
|
7744 |
+
"step": 1105
|
7745 |
+
},
|
7746 |
+
{
|
7747 |
+
"epoch": 1.0355805243445693,
|
7748 |
+
"grad_norm": 0.0,
|
7749 |
+
"learning_rate": 2.0479940974151217e-05,
|
7750 |
+
"loss": 1.1181,
|
7751 |
+
"step": 1106
|
7752 |
+
},
|
7753 |
+
{
|
7754 |
+
"epoch": 1.0365168539325842,
|
7755 |
+
"grad_norm": 0.0,
|
7756 |
+
"learning_rate": 2.044898242238871e-05,
|
7757 |
+
"loss": 1.1039,
|
7758 |
+
"step": 1107
|
7759 |
+
},
|
7760 |
+
{
|
7761 |
+
"epoch": 1.0374531835205993,
|
7762 |
+
"grad_norm": 0.0,
|
7763 |
+
"learning_rate": 2.041802279424798e-05,
|
7764 |
+
"loss": 1.1194,
|
7765 |
+
"step": 1108
|
7766 |
+
},
|
7767 |
+
{
|
7768 |
+
"epoch": 1.0383895131086143,
|
7769 |
+
"grad_norm": 0.0,
|
7770 |
+
"learning_rate": 2.0387062163950792e-05,
|
7771 |
+
"loss": 1.1967,
|
7772 |
+
"step": 1109
|
7773 |
+
},
|
7774 |
+
{
|
7775 |
+
"epoch": 1.0393258426966292,
|
7776 |
+
"grad_norm": 0.0,
|
7777 |
+
"learning_rate": 2.0356100605721324e-05,
|
7778 |
+
"loss": 1.1687,
|
7779 |
+
"step": 1110
|
7780 |
+
},
|
7781 |
+
{
|
7782 |
+
"epoch": 1.0402621722846441,
|
7783 |
+
"grad_norm": 0.0,
|
7784 |
+
"learning_rate": 2.0325138193785966e-05,
|
7785 |
+
"loss": 1.1686,
|
7786 |
+
"step": 1111
|
7787 |
+
},
|
7788 |
+
{
|
7789 |
+
"epoch": 1.0411985018726593,
|
7790 |
+
"grad_norm": 0.0,
|
7791 |
+
"learning_rate": 2.0294175002373157e-05,
|
7792 |
+
"loss": 1.1532,
|
7793 |
+
"step": 1112
|
7794 |
+
},
|
7795 |
+
{
|
7796 |
+
"epoch": 1.0421348314606742,
|
7797 |
+
"grad_norm": 0.0,
|
7798 |
+
"learning_rate": 2.0263211105713218e-05,
|
7799 |
+
"loss": 1.1625,
|
7800 |
+
"step": 1113
|
7801 |
+
},
|
7802 |
+
{
|
7803 |
+
"epoch": 1.0430711610486891,
|
7804 |
+
"grad_norm": 0.0,
|
7805 |
+
"learning_rate": 2.0232246578038145e-05,
|
7806 |
+
"loss": 1.1232,
|
7807 |
+
"step": 1114
|
7808 |
+
},
|
7809 |
+
{
|
7810 |
+
"epoch": 1.044007490636704,
|
7811 |
+
"grad_norm": 0.0,
|
7812 |
+
"learning_rate": 2.0201281493581457e-05,
|
7813 |
+
"loss": 1.1184,
|
7814 |
+
"step": 1115
|
7815 |
+
},
|
7816 |
+
{
|
7817 |
+
"epoch": 1.0449438202247192,
|
7818 |
+
"grad_norm": 0.0,
|
7819 |
+
"learning_rate": 2.0170315926578e-05,
|
7820 |
+
"loss": 1.1284,
|
7821 |
+
"step": 1116
|
7822 |
+
},
|
7823 |
+
{
|
7824 |
+
"epoch": 1.0458801498127341,
|
7825 |
+
"grad_norm": 0.0,
|
7826 |
+
"learning_rate": 2.0139349951263783e-05,
|
7827 |
+
"loss": 1.1369,
|
7828 |
+
"step": 1117
|
7829 |
+
},
|
7830 |
+
{
|
7831 |
+
"epoch": 1.046816479400749,
|
7832 |
+
"grad_norm": 0.0,
|
7833 |
+
"learning_rate": 2.010838364187579e-05,
|
7834 |
+
"loss": 1.1234,
|
7835 |
+
"step": 1118
|
7836 |
+
},
|
7837 |
+
{
|
7838 |
+
"epoch": 1.047752808988764,
|
7839 |
+
"grad_norm": 0.0,
|
7840 |
+
"learning_rate": 2.0077417072651807e-05,
|
7841 |
+
"loss": 1.1382,
|
7842 |
+
"step": 1119
|
7843 |
+
},
|
7844 |
+
{
|
7845 |
+
"epoch": 1.048689138576779,
|
7846 |
+
"grad_norm": 0.0,
|
7847 |
+
"learning_rate": 2.004645031783024e-05,
|
7848 |
+
"loss": 1.1416,
|
7849 |
+
"step": 1120
|
7850 |
+
},
|
7851 |
+
{
|
7852 |
+
"epoch": 1.049625468164794,
|
7853 |
+
"grad_norm": 0.0,
|
7854 |
+
"learning_rate": 2.001548345164995e-05,
|
7855 |
+
"loss": 1.1364,
|
7856 |
+
"step": 1121
|
7857 |
+
},
|
7858 |
+
{
|
7859 |
+
"epoch": 1.050561797752809,
|
7860 |
+
"grad_norm": 0.0,
|
7861 |
+
"learning_rate": 1.9984516548350053e-05,
|
7862 |
+
"loss": 1.1506,
|
7863 |
+
"step": 1122
|
7864 |
+
},
|
7865 |
+
{
|
7866 |
+
"epoch": 1.051498127340824,
|
7867 |
+
"grad_norm": 0.0,
|
7868 |
+
"learning_rate": 1.9953549682169767e-05,
|
7869 |
+
"loss": 1.1834,
|
7870 |
+
"step": 1123
|
7871 |
+
},
|
7872 |
+
{
|
7873 |
+
"epoch": 1.0524344569288389,
|
7874 |
+
"grad_norm": 0.0,
|
7875 |
+
"learning_rate": 1.9922582927348203e-05,
|
7876 |
+
"loss": 1.1532,
|
7877 |
+
"step": 1124
|
7878 |
+
},
|
7879 |
+
{
|
7880 |
+
"epoch": 1.053370786516854,
|
7881 |
+
"grad_norm": 0.0,
|
7882 |
+
"learning_rate": 1.9891616358124217e-05,
|
7883 |
+
"loss": 1.1567,
|
7884 |
+
"step": 1125
|
7885 |
+
},
|
7886 |
+
{
|
7887 |
+
"epoch": 1.054307116104869,
|
7888 |
+
"grad_norm": 0.0,
|
7889 |
+
"learning_rate": 1.9860650048736224e-05,
|
7890 |
+
"loss": 1.1665,
|
7891 |
+
"step": 1126
|
7892 |
+
},
|
7893 |
+
{
|
7894 |
+
"epoch": 1.0552434456928839,
|
7895 |
+
"grad_norm": 0.0,
|
7896 |
+
"learning_rate": 1.982968407342201e-05,
|
7897 |
+
"loss": 1.1224,
|
7898 |
+
"step": 1127
|
7899 |
+
},
|
7900 |
+
{
|
7901 |
+
"epoch": 1.0561797752808988,
|
7902 |
+
"grad_norm": 0.0,
|
7903 |
+
"learning_rate": 1.9798718506418546e-05,
|
7904 |
+
"loss": 1.1162,
|
7905 |
+
"step": 1128
|
7906 |
+
},
|
7907 |
+
{
|
7908 |
+
"epoch": 1.057116104868914,
|
7909 |
+
"grad_norm": 0.0,
|
7910 |
+
"learning_rate": 1.976775342196186e-05,
|
7911 |
+
"loss": 1.1807,
|
7912 |
+
"step": 1129
|
7913 |
+
},
|
7914 |
+
{
|
7915 |
+
"epoch": 1.0580524344569289,
|
7916 |
+
"grad_norm": 0.0,
|
7917 |
+
"learning_rate": 1.973678889428679e-05,
|
7918 |
+
"loss": 1.126,
|
7919 |
+
"step": 1130
|
7920 |
+
},
|
7921 |
+
{
|
7922 |
+
"epoch": 1.0589887640449438,
|
7923 |
+
"grad_norm": 0.0,
|
7924 |
+
"learning_rate": 1.9705824997626843e-05,
|
7925 |
+
"loss": 1.1369,
|
7926 |
+
"step": 1131
|
7927 |
+
},
|
7928 |
+
{
|
7929 |
+
"epoch": 1.0599250936329587,
|
7930 |
+
"grad_norm": 0.0,
|
7931 |
+
"learning_rate": 1.9674861806214044e-05,
|
7932 |
+
"loss": 1.2066,
|
7933 |
+
"step": 1132
|
7934 |
+
},
|
7935 |
+
{
|
7936 |
+
"epoch": 1.0608614232209739,
|
7937 |
+
"grad_norm": 0.0,
|
7938 |
+
"learning_rate": 1.9643899394278686e-05,
|
7939 |
+
"loss": 1.1466,
|
7940 |
+
"step": 1133
|
7941 |
+
},
|
7942 |
+
{
|
7943 |
+
"epoch": 1.0617977528089888,
|
7944 |
+
"grad_norm": 0.0,
|
7945 |
+
"learning_rate": 1.961293783604921e-05,
|
7946 |
+
"loss": 1.1395,
|
7947 |
+
"step": 1134
|
7948 |
+
},
|
7949 |
+
{
|
7950 |
+
"epoch": 1.0627340823970037,
|
7951 |
+
"grad_norm": 0.0,
|
7952 |
+
"learning_rate": 1.9581977205752028e-05,
|
7953 |
+
"loss": 1.1214,
|
7954 |
+
"step": 1135
|
7955 |
+
},
|
7956 |
+
{
|
7957 |
+
"epoch": 1.0636704119850187,
|
7958 |
+
"grad_norm": 0.0,
|
7959 |
+
"learning_rate": 1.95510175776113e-05,
|
7960 |
+
"loss": 1.1134,
|
7961 |
+
"step": 1136
|
7962 |
+
},
|
7963 |
+
{
|
7964 |
+
"epoch": 1.0646067415730338,
|
7965 |
+
"grad_norm": 0.0,
|
7966 |
+
"learning_rate": 1.9520059025848787e-05,
|
7967 |
+
"loss": 1.1354,
|
7968 |
+
"step": 1137
|
7969 |
+
},
|
7970 |
+
{
|
7971 |
+
"epoch": 1.0655430711610487,
|
7972 |
+
"grad_norm": 0.0,
|
7973 |
+
"learning_rate": 1.9489101624683697e-05,
|
7974 |
+
"loss": 1.1322,
|
7975 |
+
"step": 1138
|
7976 |
+
},
|
7977 |
+
{
|
7978 |
+
"epoch": 1.0664794007490637,
|
7979 |
+
"grad_norm": 0.0,
|
7980 |
+
"learning_rate": 1.945814544833245e-05,
|
7981 |
+
"loss": 1.1701,
|
7982 |
+
"step": 1139
|
7983 |
+
},
|
7984 |
+
{
|
7985 |
+
"epoch": 1.0674157303370786,
|
7986 |
+
"grad_norm": 0.0,
|
7987 |
+
"learning_rate": 1.942719057100853e-05,
|
7988 |
+
"loss": 1.1884,
|
7989 |
+
"step": 1140
|
7990 |
+
},
|
7991 |
+
{
|
7992 |
+
"epoch": 1.0683520599250937,
|
7993 |
+
"grad_norm": 0.0,
|
7994 |
+
"learning_rate": 1.9396237066922335e-05,
|
7995 |
+
"loss": 1.1579,
|
7996 |
+
"step": 1141
|
7997 |
+
},
|
7998 |
+
{
|
7999 |
+
"epoch": 1.0692883895131087,
|
8000 |
+
"grad_norm": 0.0,
|
8001 |
+
"learning_rate": 1.9365285010280942e-05,
|
8002 |
+
"loss": 1.1618,
|
8003 |
+
"step": 1142
|
8004 |
+
},
|
8005 |
+
{
|
8006 |
+
"epoch": 1.0702247191011236,
|
8007 |
+
"grad_norm": 0.0,
|
8008 |
+
"learning_rate": 1.9334334475287963e-05,
|
8009 |
+
"loss": 1.0985,
|
8010 |
+
"step": 1143
|
8011 |
+
},
|
8012 |
+
{
|
8013 |
+
"epoch": 1.0711610486891385,
|
8014 |
+
"grad_norm": 0.0,
|
8015 |
+
"learning_rate": 1.9303385536143383e-05,
|
8016 |
+
"loss": 1.084,
|
8017 |
+
"step": 1144
|
8018 |
+
},
|
8019 |
+
{
|
8020 |
+
"epoch": 1.0720973782771535,
|
8021 |
+
"grad_norm": 0.0,
|
8022 |
+
"learning_rate": 1.9272438267043338e-05,
|
8023 |
+
"loss": 1.152,
|
8024 |
+
"step": 1145
|
8025 |
+
},
|
8026 |
+
{
|
8027 |
+
"epoch": 1.0730337078651686,
|
8028 |
+
"grad_norm": 0.0,
|
8029 |
+
"learning_rate": 1.9241492742179952e-05,
|
8030 |
+
"loss": 1.1679,
|
8031 |
+
"step": 1146
|
8032 |
+
},
|
8033 |
+
{
|
8034 |
+
"epoch": 1.0739700374531835,
|
8035 |
+
"grad_norm": 0.0,
|
8036 |
+
"learning_rate": 1.9210549035741205e-05,
|
8037 |
+
"loss": 1.149,
|
8038 |
+
"step": 1147
|
8039 |
+
},
|
8040 |
+
{
|
8041 |
+
"epoch": 1.0749063670411985,
|
8042 |
+
"grad_norm": 0.0,
|
8043 |
+
"learning_rate": 1.9179607221910687e-05,
|
8044 |
+
"loss": 1.1561,
|
8045 |
+
"step": 1148
|
8046 |
+
},
|
8047 |
+
{
|
8048 |
+
"epoch": 1.0758426966292134,
|
8049 |
+
"grad_norm": 0.0,
|
8050 |
+
"learning_rate": 1.9148667374867453e-05,
|
8051 |
+
"loss": 1.1496,
|
8052 |
+
"step": 1149
|
8053 |
+
},
|
8054 |
+
{
|
8055 |
+
"epoch": 1.0767790262172285,
|
8056 |
+
"grad_norm": 0.0,
|
8057 |
+
"learning_rate": 1.9117729568785864e-05,
|
8058 |
+
"loss": 1.1308,
|
8059 |
+
"step": 1150
|
8060 |
+
},
|
8061 |
+
{
|
8062 |
+
"epoch": 1.0777153558052435,
|
8063 |
+
"grad_norm": 0.0,
|
8064 |
+
"learning_rate": 1.9086793877835364e-05,
|
8065 |
+
"loss": 1.1308,
|
8066 |
+
"step": 1151
|
8067 |
+
},
|
8068 |
+
{
|
8069 |
+
"epoch": 1.0786516853932584,
|
8070 |
+
"grad_norm": 0.0,
|
8071 |
+
"learning_rate": 1.9055860376180334e-05,
|
8072 |
+
"loss": 1.1575,
|
8073 |
+
"step": 1152
|
8074 |
+
},
|
8075 |
+
{
|
8076 |
+
"epoch": 1.0795880149812733,
|
8077 |
+
"grad_norm": 0.0,
|
8078 |
+
"learning_rate": 1.902492913797991e-05,
|
8079 |
+
"loss": 1.0963,
|
8080 |
+
"step": 1153
|
8081 |
+
},
|
8082 |
+
{
|
8083 |
+
"epoch": 1.0805243445692885,
|
8084 |
+
"grad_norm": 0.0,
|
8085 |
+
"learning_rate": 1.899400023738779e-05,
|
8086 |
+
"loss": 1.1545,
|
8087 |
+
"step": 1154
|
8088 |
+
},
|
8089 |
+
{
|
8090 |
+
"epoch": 1.0814606741573034,
|
8091 |
+
"grad_norm": 0.0,
|
8092 |
+
"learning_rate": 1.8963073748552104e-05,
|
8093 |
+
"loss": 1.1085,
|
8094 |
+
"step": 1155
|
8095 |
+
},
|
8096 |
+
{
|
8097 |
+
"epoch": 1.0823970037453183,
|
8098 |
+
"grad_norm": 0.0,
|
8099 |
+
"learning_rate": 1.8932149745615153e-05,
|
8100 |
+
"loss": 1.1315,
|
8101 |
+
"step": 1156
|
8102 |
+
},
|
8103 |
+
{
|
8104 |
+
"epoch": 1.0833333333333333,
|
8105 |
+
"grad_norm": 0.0,
|
8106 |
+
"learning_rate": 1.8901228302713297e-05,
|
8107 |
+
"loss": 1.157,
|
8108 |
+
"step": 1157
|
8109 |
+
},
|
8110 |
+
{
|
8111 |
+
"epoch": 1.0842696629213484,
|
8112 |
+
"grad_norm": 0.0,
|
8113 |
+
"learning_rate": 1.887030949397677e-05,
|
8114 |
+
"loss": 1.1458,
|
8115 |
+
"step": 1158
|
8116 |
+
},
|
8117 |
+
{
|
8118 |
+
"epoch": 1.0852059925093633,
|
8119 |
+
"grad_norm": 0.0,
|
8120 |
+
"learning_rate": 1.883939339352948e-05,
|
8121 |
+
"loss": 1.1097,
|
8122 |
+
"step": 1159
|
8123 |
+
},
|
8124 |
+
{
|
8125 |
+
"epoch": 1.0861423220973783,
|
8126 |
+
"grad_norm": 0.0,
|
8127 |
+
"learning_rate": 1.8808480075488837e-05,
|
8128 |
+
"loss": 1.1668,
|
8129 |
+
"step": 1160
|
8130 |
+
},
|
8131 |
+
{
|
8132 |
+
"epoch": 1.0870786516853932,
|
8133 |
+
"grad_norm": 0.0,
|
8134 |
+
"learning_rate": 1.8777569613965597e-05,
|
8135 |
+
"loss": 1.1657,
|
8136 |
+
"step": 1161
|
8137 |
+
},
|
8138 |
+
{
|
8139 |
+
"epoch": 1.0880149812734083,
|
8140 |
+
"grad_norm": 0.0,
|
8141 |
+
"learning_rate": 1.8746662083063656e-05,
|
8142 |
+
"loss": 1.1379,
|
8143 |
+
"step": 1162
|
8144 |
+
},
|
8145 |
+
{
|
8146 |
+
"epoch": 1.0889513108614233,
|
8147 |
+
"grad_norm": 0.0,
|
8148 |
+
"learning_rate": 1.8715757556879873e-05,
|
8149 |
+
"loss": 1.1593,
|
8150 |
+
"step": 1163
|
8151 |
+
},
|
8152 |
+
{
|
8153 |
+
"epoch": 1.0898876404494382,
|
8154 |
+
"grad_norm": 0.0,
|
8155 |
+
"learning_rate": 1.8684856109503932e-05,
|
8156 |
+
"loss": 1.1339,
|
8157 |
+
"step": 1164
|
8158 |
+
},
|
8159 |
+
{
|
8160 |
+
"epoch": 1.0908239700374531,
|
8161 |
+
"grad_norm": 0.0,
|
8162 |
+
"learning_rate": 1.865395781501812e-05,
|
8163 |
+
"loss": 1.0974,
|
8164 |
+
"step": 1165
|
8165 |
+
},
|
8166 |
+
{
|
8167 |
+
"epoch": 1.0917602996254683,
|
8168 |
+
"grad_norm": 0.0,
|
8169 |
+
"learning_rate": 1.8623062747497154e-05,
|
8170 |
+
"loss": 1.1288,
|
8171 |
+
"step": 1166
|
8172 |
+
},
|
8173 |
+
{
|
8174 |
+
"epoch": 1.0926966292134832,
|
8175 |
+
"grad_norm": 0.0,
|
8176 |
+
"learning_rate": 1.859217098100805e-05,
|
8177 |
+
"loss": 1.1327,
|
8178 |
+
"step": 1167
|
8179 |
+
},
|
8180 |
+
{
|
8181 |
+
"epoch": 1.0936329588014981,
|
8182 |
+
"grad_norm": 0.0,
|
8183 |
+
"learning_rate": 1.8561282589609865e-05,
|
8184 |
+
"loss": 1.1578,
|
8185 |
+
"step": 1168
|
8186 |
+
},
|
8187 |
+
{
|
8188 |
+
"epoch": 1.094569288389513,
|
8189 |
+
"grad_norm": 0.0,
|
8190 |
+
"learning_rate": 1.853039764735359e-05,
|
8191 |
+
"loss": 1.1472,
|
8192 |
+
"step": 1169
|
8193 |
+
},
|
8194 |
+
{
|
8195 |
+
"epoch": 1.095505617977528,
|
8196 |
+
"grad_norm": 0.0,
|
8197 |
+
"learning_rate": 1.8499516228281954e-05,
|
8198 |
+
"loss": 1.1581,
|
8199 |
+
"step": 1170
|
8200 |
+
},
|
8201 |
+
{
|
8202 |
+
"epoch": 1.0964419475655431,
|
8203 |
+
"grad_norm": 0.0,
|
8204 |
+
"learning_rate": 1.8468638406429226e-05,
|
8205 |
+
"loss": 1.1156,
|
8206 |
+
"step": 1171
|
8207 |
+
},
|
8208 |
+
{
|
8209 |
+
"epoch": 1.097378277153558,
|
8210 |
+
"grad_norm": 0.0,
|
8211 |
+
"learning_rate": 1.8437764255821043e-05,
|
8212 |
+
"loss": 1.1533,
|
8213 |
+
"step": 1172
|
8214 |
+
},
|
8215 |
+
{
|
8216 |
+
"epoch": 1.098314606741573,
|
8217 |
+
"grad_norm": 0.0,
|
8218 |
+
"learning_rate": 1.8406893850474272e-05,
|
8219 |
+
"loss": 1.2105,
|
8220 |
+
"step": 1173
|
8221 |
+
},
|
8222 |
+
{
|
8223 |
+
"epoch": 1.099250936329588,
|
8224 |
+
"grad_norm": 0.0,
|
8225 |
+
"learning_rate": 1.8376027264396766e-05,
|
8226 |
+
"loss": 1.1536,
|
8227 |
+
"step": 1174
|
8228 |
+
},
|
8229 |
+
{
|
8230 |
+
"epoch": 1.100187265917603,
|
8231 |
+
"grad_norm": 0.0,
|
8232 |
+
"learning_rate": 1.8345164571587245e-05,
|
8233 |
+
"loss": 1.102,
|
8234 |
+
"step": 1175
|
8235 |
+
},
|
8236 |
+
{
|
8237 |
+
"epoch": 1.101123595505618,
|
8238 |
+
"grad_norm": 0.0,
|
8239 |
+
"learning_rate": 1.8314305846035084e-05,
|
8240 |
+
"loss": 1.1133,
|
8241 |
+
"step": 1176
|
8242 |
+
},
|
8243 |
+
{
|
8244 |
+
"epoch": 1.102059925093633,
|
8245 |
+
"grad_norm": 0.0,
|
8246 |
+
"learning_rate": 1.828345116172016e-05,
|
8247 |
+
"loss": 1.1443,
|
8248 |
+
"step": 1177
|
8249 |
+
},
|
8250 |
+
{
|
8251 |
+
"epoch": 1.1029962546816479,
|
8252 |
+
"grad_norm": 0.0,
|
8253 |
+
"learning_rate": 1.8252600592612634e-05,
|
8254 |
+
"loss": 1.1128,
|
8255 |
+
"step": 1178
|
8256 |
+
},
|
8257 |
+
{
|
8258 |
+
"epoch": 1.103932584269663,
|
8259 |
+
"grad_norm": 0.0,
|
8260 |
+
"learning_rate": 1.822175421267284e-05,
|
8261 |
+
"loss": 1.1466,
|
8262 |
+
"step": 1179
|
8263 |
+
},
|
8264 |
+
{
|
8265 |
+
"epoch": 1.104868913857678,
|
8266 |
+
"grad_norm": 0.0,
|
8267 |
+
"learning_rate": 1.8190912095851033e-05,
|
8268 |
+
"loss": 1.1,
|
8269 |
+
"step": 1180
|
8270 |
+
},
|
8271 |
+
{
|
8272 |
+
"epoch": 1.1058052434456929,
|
8273 |
+
"grad_norm": 0.0,
|
8274 |
+
"learning_rate": 1.816007431608727e-05,
|
8275 |
+
"loss": 1.1727,
|
8276 |
+
"step": 1181
|
8277 |
+
},
|
8278 |
+
{
|
8279 |
+
"epoch": 1.1067415730337078,
|
8280 |
+
"grad_norm": 0.0,
|
8281 |
+
"learning_rate": 1.8129240947311208e-05,
|
8282 |
+
"loss": 1.1262,
|
8283 |
+
"step": 1182
|
8284 |
+
},
|
8285 |
+
{
|
8286 |
+
"epoch": 1.107677902621723,
|
8287 |
+
"grad_norm": 0.0,
|
8288 |
+
"learning_rate": 1.8098412063441923e-05,
|
8289 |
+
"loss": 1.14,
|
8290 |
+
"step": 1183
|
8291 |
+
},
|
8292 |
+
{
|
8293 |
+
"epoch": 1.1086142322097379,
|
8294 |
+
"grad_norm": 0.0,
|
8295 |
+
"learning_rate": 1.8067587738387725e-05,
|
8296 |
+
"loss": 1.1644,
|
8297 |
+
"step": 1184
|
8298 |
+
},
|
8299 |
+
{
|
8300 |
+
"epoch": 1.1095505617977528,
|
8301 |
+
"grad_norm": 0.0,
|
8302 |
+
"learning_rate": 1.8036768046046038e-05,
|
8303 |
+
"loss": 1.1326,
|
8304 |
+
"step": 1185
|
8305 |
+
},
|
8306 |
+
{
|
8307 |
+
"epoch": 1.1104868913857677,
|
8308 |
+
"grad_norm": 0.0,
|
8309 |
+
"learning_rate": 1.8005953060303133e-05,
|
8310 |
+
"loss": 1.1411,
|
8311 |
+
"step": 1186
|
8312 |
+
},
|
8313 |
+
{
|
8314 |
+
"epoch": 1.1114232209737827,
|
8315 |
+
"grad_norm": 0.0,
|
8316 |
+
"learning_rate": 1.797514285503402e-05,
|
8317 |
+
"loss": 1.1796,
|
8318 |
+
"step": 1187
|
8319 |
+
},
|
8320 |
+
{
|
8321 |
+
"epoch": 1.1123595505617978,
|
8322 |
+
"grad_norm": 0.0,
|
8323 |
+
"learning_rate": 1.7944337504102258e-05,
|
8324 |
+
"loss": 1.1085,
|
8325 |
+
"step": 1188
|
8326 |
+
},
|
8327 |
+
{
|
8328 |
+
"epoch": 1.1132958801498127,
|
8329 |
+
"grad_norm": 0.0,
|
8330 |
+
"learning_rate": 1.7913537081359746e-05,
|
8331 |
+
"loss": 1.1831,
|
8332 |
+
"step": 1189
|
8333 |
+
},
|
8334 |
+
{
|
8335 |
+
"epoch": 1.1142322097378277,
|
8336 |
+
"grad_norm": 0.0,
|
8337 |
+
"learning_rate": 1.7882741660646572e-05,
|
8338 |
+
"loss": 1.1521,
|
8339 |
+
"step": 1190
|
8340 |
+
},
|
8341 |
+
{
|
8342 |
+
"epoch": 1.1151685393258428,
|
8343 |
+
"grad_norm": 0.0,
|
8344 |
+
"learning_rate": 1.7851951315790848e-05,
|
8345 |
+
"loss": 1.1385,
|
8346 |
+
"step": 1191
|
8347 |
+
},
|
8348 |
+
{
|
8349 |
+
"epoch": 1.1161048689138577,
|
8350 |
+
"grad_norm": 0.0,
|
8351 |
+
"learning_rate": 1.782116612060851e-05,
|
8352 |
+
"loss": 1.1865,
|
8353 |
+
"step": 1192
|
8354 |
+
},
|
8355 |
+
{
|
8356 |
+
"epoch": 1.1170411985018727,
|
8357 |
+
"grad_norm": 0.0,
|
8358 |
+
"learning_rate": 1.7790386148903134e-05,
|
8359 |
+
"loss": 1.0923,
|
8360 |
+
"step": 1193
|
8361 |
+
},
|
8362 |
+
{
|
8363 |
+
"epoch": 1.1179775280898876,
|
8364 |
+
"grad_norm": 0.0,
|
8365 |
+
"learning_rate": 1.77596114744658e-05,
|
8366 |
+
"loss": 1.1301,
|
8367 |
+
"step": 1194
|
8368 |
+
},
|
8369 |
+
{
|
8370 |
+
"epoch": 1.1189138576779025,
|
8371 |
+
"grad_norm": 0.0,
|
8372 |
+
"learning_rate": 1.7728842171074865e-05,
|
8373 |
+
"loss": 1.117,
|
8374 |
+
"step": 1195
|
8375 |
+
},
|
8376 |
+
{
|
8377 |
+
"epoch": 1.1198501872659177,
|
8378 |
+
"grad_norm": 0.0,
|
8379 |
+
"learning_rate": 1.7698078312495813e-05,
|
8380 |
+
"loss": 1.1607,
|
8381 |
+
"step": 1196
|
8382 |
+
},
|
8383 |
+
{
|
8384 |
+
"epoch": 1.1207865168539326,
|
8385 |
+
"grad_norm": 0.0,
|
8386 |
+
"learning_rate": 1.766731997248109e-05,
|
8387 |
+
"loss": 1.1629,
|
8388 |
+
"step": 1197
|
8389 |
+
},
|
8390 |
+
{
|
8391 |
+
"epoch": 1.1217228464419475,
|
8392 |
+
"grad_norm": 0.0,
|
8393 |
+
"learning_rate": 1.7636567224769906e-05,
|
8394 |
+
"loss": 1.0561,
|
8395 |
+
"step": 1198
|
8396 |
+
},
|
8397 |
+
{
|
8398 |
+
"epoch": 1.1226591760299625,
|
8399 |
+
"grad_norm": 0.0,
|
8400 |
+
"learning_rate": 1.760582014308804e-05,
|
8401 |
+
"loss": 1.1449,
|
8402 |
+
"step": 1199
|
8403 |
+
},
|
8404 |
+
{
|
8405 |
+
"epoch": 1.1235955056179776,
|
8406 |
+
"grad_norm": 0.0,
|
8407 |
+
"learning_rate": 1.7575078801147736e-05,
|
8408 |
+
"loss": 1.1362,
|
8409 |
+
"step": 1200
|
8410 |
}
|
8411 |
],
|
8412 |
"logging_steps": 1,
|
|
|
8426 |
"attributes": {}
|
8427 |
}
|
8428 |
},
|
8429 |
+
"total_flos": 3.925805532929065e+18,
|
8430 |
"train_batch_size": 8,
|
8431 |
"trial_name": null,
|
8432 |
"trial_params": null
|