Training in progress, step 570, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 81576
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51e98f67765a2eb2368c2f2e215e5056e6877b63a99d2c749fa289541026b705
|
| 3 |
size 81576
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 173094
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2cfe11ad55a96cad8bdac1448ebc4e55b90395474681435aab23be456282687f
|
| 3 |
size 173094
|
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:c2b1b4d4fc5af089dfffe964a368b7be036a5a5e60d94d3627c20a7cd7efe1e0
|
| 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 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e3efe9a7dcecfc294170eaf9d37d88929a06a263232a2f5ff76c24580b1ffbbf
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 143,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3042,6 +3042,993 @@
|
|
| 3042 |
"eval_samples_per_second": 317.257,
|
| 3043 |
"eval_steps_per_second": 158.629,
|
| 3044 |
"step": 429
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3045 |
}
|
| 3046 |
],
|
| 3047 |
"logging_steps": 1,
|
|
@@ -3056,12 +4043,12 @@
|
|
| 3056 |
"should_evaluate": false,
|
| 3057 |
"should_log": false,
|
| 3058 |
"should_save": true,
|
| 3059 |
-
"should_training_stop":
|
| 3060 |
},
|
| 3061 |
"attributes": {}
|
| 3062 |
}
|
| 3063 |
},
|
| 3064 |
-
"total_flos":
|
| 3065 |
"train_batch_size": 2,
|
| 3066 |
"trial_name": null,
|
| 3067 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 143,
|
| 6 |
+
"global_step": 570,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3042 |
"eval_samples_per_second": 317.257,
|
| 3043 |
"eval_steps_per_second": 158.629,
|
| 3044 |
"step": 429
|
| 3045 |
+
},
|
| 3046 |
+
{
|
| 3047 |
+
"epoch": 0.7543859649122807,
|
| 3048 |
+
"grad_norm": 0.29503875970840454,
|
| 3049 |
+
"learning_rate": 2.9289321881345254e-05,
|
| 3050 |
+
"loss": 10.691,
|
| 3051 |
+
"step": 430
|
| 3052 |
+
},
|
| 3053 |
+
{
|
| 3054 |
+
"epoch": 0.756140350877193,
|
| 3055 |
+
"grad_norm": 0.2280651479959488,
|
| 3056 |
+
"learning_rate": 2.8893750684111975e-05,
|
| 3057 |
+
"loss": 10.7043,
|
| 3058 |
+
"step": 431
|
| 3059 |
+
},
|
| 3060 |
+
{
|
| 3061 |
+
"epoch": 0.7578947368421053,
|
| 3062 |
+
"grad_norm": 0.2564852237701416,
|
| 3063 |
+
"learning_rate": 2.850041733353247e-05,
|
| 3064 |
+
"loss": 10.6683,
|
| 3065 |
+
"step": 432
|
| 3066 |
+
},
|
| 3067 |
+
{
|
| 3068 |
+
"epoch": 0.7596491228070176,
|
| 3069 |
+
"grad_norm": 0.19316798448562622,
|
| 3070 |
+
"learning_rate": 2.8109334208542993e-05,
|
| 3071 |
+
"loss": 10.6773,
|
| 3072 |
+
"step": 433
|
| 3073 |
+
},
|
| 3074 |
+
{
|
| 3075 |
+
"epoch": 0.7614035087719299,
|
| 3076 |
+
"grad_norm": 0.18743948638439178,
|
| 3077 |
+
"learning_rate": 2.7720513617260856e-05,
|
| 3078 |
+
"loss": 10.6676,
|
| 3079 |
+
"step": 434
|
| 3080 |
+
},
|
| 3081 |
+
{
|
| 3082 |
+
"epoch": 0.7631578947368421,
|
| 3083 |
+
"grad_norm": 0.21178874373435974,
|
| 3084 |
+
"learning_rate": 2.7333967796597315e-05,
|
| 3085 |
+
"loss": 10.7064,
|
| 3086 |
+
"step": 435
|
| 3087 |
+
},
|
| 3088 |
+
{
|
| 3089 |
+
"epoch": 0.7649122807017544,
|
| 3090 |
+
"grad_norm": 0.19119100272655487,
|
| 3091 |
+
"learning_rate": 2.694970891187225e-05,
|
| 3092 |
+
"loss": 10.6512,
|
| 3093 |
+
"step": 436
|
| 3094 |
+
},
|
| 3095 |
+
{
|
| 3096 |
+
"epoch": 0.7666666666666667,
|
| 3097 |
+
"grad_norm": 0.25586599111557007,
|
| 3098 |
+
"learning_rate": 2.6567749056431467e-05,
|
| 3099 |
+
"loss": 10.6905,
|
| 3100 |
+
"step": 437
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"epoch": 0.7684210526315789,
|
| 3104 |
+
"grad_norm": 0.17666222155094147,
|
| 3105 |
+
"learning_rate": 2.6188100251265945e-05,
|
| 3106 |
+
"loss": 10.6544,
|
| 3107 |
+
"step": 438
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"epoch": 0.7701754385964912,
|
| 3111 |
+
"grad_norm": 0.1568339616060257,
|
| 3112 |
+
"learning_rate": 2.5810774444633644e-05,
|
| 3113 |
+
"loss": 10.6804,
|
| 3114 |
+
"step": 439
|
| 3115 |
+
},
|
| 3116 |
+
{
|
| 3117 |
+
"epoch": 0.7719298245614035,
|
| 3118 |
+
"grad_norm": 0.2483753114938736,
|
| 3119 |
+
"learning_rate": 2.5435783511683443e-05,
|
| 3120 |
+
"loss": 10.6641,
|
| 3121 |
+
"step": 440
|
| 3122 |
+
},
|
| 3123 |
+
{
|
| 3124 |
+
"epoch": 0.7736842105263158,
|
| 3125 |
+
"grad_norm": 0.22110183537006378,
|
| 3126 |
+
"learning_rate": 2.506313925408127e-05,
|
| 3127 |
+
"loss": 10.6885,
|
| 3128 |
+
"step": 441
|
| 3129 |
+
},
|
| 3130 |
+
{
|
| 3131 |
+
"epoch": 0.775438596491228,
|
| 3132 |
+
"grad_norm": 0.1530926376581192,
|
| 3133 |
+
"learning_rate": 2.4692853399638917e-05,
|
| 3134 |
+
"loss": 10.6533,
|
| 3135 |
+
"step": 442
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 0.7771929824561403,
|
| 3139 |
+
"grad_norm": 0.23491761088371277,
|
| 3140 |
+
"learning_rate": 2.4324937601944685e-05,
|
| 3141 |
+
"loss": 10.6488,
|
| 3142 |
+
"step": 443
|
| 3143 |
+
},
|
| 3144 |
+
{
|
| 3145 |
+
"epoch": 0.7789473684210526,
|
| 3146 |
+
"grad_norm": 0.21756233274936676,
|
| 3147 |
+
"learning_rate": 2.3959403439996907e-05,
|
| 3148 |
+
"loss": 10.6742,
|
| 3149 |
+
"step": 444
|
| 3150 |
+
},
|
| 3151 |
+
{
|
| 3152 |
+
"epoch": 0.7807017543859649,
|
| 3153 |
+
"grad_norm": 0.18125437200069427,
|
| 3154 |
+
"learning_rate": 2.3596262417839255e-05,
|
| 3155 |
+
"loss": 10.6726,
|
| 3156 |
+
"step": 445
|
| 3157 |
+
},
|
| 3158 |
+
{
|
| 3159 |
+
"epoch": 0.7824561403508772,
|
| 3160 |
+
"grad_norm": 0.23887260258197784,
|
| 3161 |
+
"learning_rate": 2.323552596419889e-05,
|
| 3162 |
+
"loss": 10.6596,
|
| 3163 |
+
"step": 446
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 0.7842105263157895,
|
| 3167 |
+
"grad_norm": 0.18337543308734894,
|
| 3168 |
+
"learning_rate": 2.2877205432126657e-05,
|
| 3169 |
+
"loss": 10.6511,
|
| 3170 |
+
"step": 447
|
| 3171 |
+
},
|
| 3172 |
+
{
|
| 3173 |
+
"epoch": 0.7859649122807018,
|
| 3174 |
+
"grad_norm": 0.2415776550769806,
|
| 3175 |
+
"learning_rate": 2.2521312098639914e-05,
|
| 3176 |
+
"loss": 10.6739,
|
| 3177 |
+
"step": 448
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 0.787719298245614,
|
| 3181 |
+
"grad_norm": 0.18577896058559418,
|
| 3182 |
+
"learning_rate": 2.2167857164367422e-05,
|
| 3183 |
+
"loss": 10.6954,
|
| 3184 |
+
"step": 449
|
| 3185 |
+
},
|
| 3186 |
+
{
|
| 3187 |
+
"epoch": 0.7894736842105263,
|
| 3188 |
+
"grad_norm": 0.2739926278591156,
|
| 3189 |
+
"learning_rate": 2.181685175319702e-05,
|
| 3190 |
+
"loss": 10.6477,
|
| 3191 |
+
"step": 450
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 0.7912280701754386,
|
| 3195 |
+
"grad_norm": 0.2336394488811493,
|
| 3196 |
+
"learning_rate": 2.146830691192553e-05,
|
| 3197 |
+
"loss": 10.6947,
|
| 3198 |
+
"step": 451
|
| 3199 |
+
},
|
| 3200 |
+
{
|
| 3201 |
+
"epoch": 0.7929824561403509,
|
| 3202 |
+
"grad_norm": 0.22242297232151031,
|
| 3203 |
+
"learning_rate": 2.11222336099109e-05,
|
| 3204 |
+
"loss": 10.6602,
|
| 3205 |
+
"step": 452
|
| 3206 |
+
},
|
| 3207 |
+
{
|
| 3208 |
+
"epoch": 0.7947368421052632,
|
| 3209 |
+
"grad_norm": 0.18758748471736908,
|
| 3210 |
+
"learning_rate": 2.0778642738727272e-05,
|
| 3211 |
+
"loss": 10.6843,
|
| 3212 |
+
"step": 453
|
| 3213 |
+
},
|
| 3214 |
+
{
|
| 3215 |
+
"epoch": 0.7964912280701755,
|
| 3216 |
+
"grad_norm": 0.1851445883512497,
|
| 3217 |
+
"learning_rate": 2.043754511182191e-05,
|
| 3218 |
+
"loss": 10.6776,
|
| 3219 |
+
"step": 454
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 0.7982456140350878,
|
| 3223 |
+
"grad_norm": 0.1795864701271057,
|
| 3224 |
+
"learning_rate": 2.009895146417512e-05,
|
| 3225 |
+
"loss": 10.6947,
|
| 3226 |
+
"step": 455
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 0.8,
|
| 3230 |
+
"grad_norm": 0.21275727450847626,
|
| 3231 |
+
"learning_rate": 1.9762872451962212e-05,
|
| 3232 |
+
"loss": 10.69,
|
| 3233 |
+
"step": 456
|
| 3234 |
+
},
|
| 3235 |
+
{
|
| 3236 |
+
"epoch": 0.8017543859649123,
|
| 3237 |
+
"grad_norm": 0.1958458423614502,
|
| 3238 |
+
"learning_rate": 1.9429318652218276e-05,
|
| 3239 |
+
"loss": 10.6889,
|
| 3240 |
+
"step": 457
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 0.8035087719298246,
|
| 3244 |
+
"grad_norm": 0.3333258032798767,
|
| 3245 |
+
"learning_rate": 1.9098300562505266e-05,
|
| 3246 |
+
"loss": 10.652,
|
| 3247 |
+
"step": 458
|
| 3248 |
+
},
|
| 3249 |
+
{
|
| 3250 |
+
"epoch": 0.8052631578947368,
|
| 3251 |
+
"grad_norm": 0.22521115839481354,
|
| 3252 |
+
"learning_rate": 1.8769828600581496e-05,
|
| 3253 |
+
"loss": 10.6767,
|
| 3254 |
+
"step": 459
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 0.8070175438596491,
|
| 3258 |
+
"grad_norm": 0.22420458495616913,
|
| 3259 |
+
"learning_rate": 1.8443913104073983e-05,
|
| 3260 |
+
"loss": 10.6827,
|
| 3261 |
+
"step": 460
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 0.8087719298245614,
|
| 3265 |
+
"grad_norm": 0.28353458642959595,
|
| 3266 |
+
"learning_rate": 1.8120564330152923e-05,
|
| 3267 |
+
"loss": 10.6371,
|
| 3268 |
+
"step": 461
|
| 3269 |
+
},
|
| 3270 |
+
{
|
| 3271 |
+
"epoch": 0.8105263157894737,
|
| 3272 |
+
"grad_norm": 0.27955901622772217,
|
| 3273 |
+
"learning_rate": 1.7799792455209018e-05,
|
| 3274 |
+
"loss": 10.6391,
|
| 3275 |
+
"step": 462
|
| 3276 |
+
},
|
| 3277 |
+
{
|
| 3278 |
+
"epoch": 0.8122807017543859,
|
| 3279 |
+
"grad_norm": 0.24275584518909454,
|
| 3280 |
+
"learning_rate": 1.7481607574533044e-05,
|
| 3281 |
+
"loss": 10.669,
|
| 3282 |
+
"step": 463
|
| 3283 |
+
},
|
| 3284 |
+
{
|
| 3285 |
+
"epoch": 0.8140350877192982,
|
| 3286 |
+
"grad_norm": 0.21551206707954407,
|
| 3287 |
+
"learning_rate": 1.716601970199836e-05,
|
| 3288 |
+
"loss": 10.6878,
|
| 3289 |
+
"step": 464
|
| 3290 |
+
},
|
| 3291 |
+
{
|
| 3292 |
+
"epoch": 0.8157894736842105,
|
| 3293 |
+
"grad_norm": 0.20878919959068298,
|
| 3294 |
+
"learning_rate": 1.6853038769745467e-05,
|
| 3295 |
+
"loss": 10.6937,
|
| 3296 |
+
"step": 465
|
| 3297 |
+
},
|
| 3298 |
+
{
|
| 3299 |
+
"epoch": 0.8175438596491228,
|
| 3300 |
+
"grad_norm": 0.21660566329956055,
|
| 3301 |
+
"learning_rate": 1.6542674627869737e-05,
|
| 3302 |
+
"loss": 10.6919,
|
| 3303 |
+
"step": 466
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 0.8192982456140351,
|
| 3307 |
+
"grad_norm": 0.2458791881799698,
|
| 3308 |
+
"learning_rate": 1.6234937044111152e-05,
|
| 3309 |
+
"loss": 10.6738,
|
| 3310 |
+
"step": 467
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"epoch": 0.8210526315789474,
|
| 3314 |
+
"grad_norm": 0.22026145458221436,
|
| 3315 |
+
"learning_rate": 1.5929835703546993e-05,
|
| 3316 |
+
"loss": 10.6797,
|
| 3317 |
+
"step": 468
|
| 3318 |
+
},
|
| 3319 |
+
{
|
| 3320 |
+
"epoch": 0.8228070175438597,
|
| 3321 |
+
"grad_norm": 0.21497632563114166,
|
| 3322 |
+
"learning_rate": 1.5627380208287114e-05,
|
| 3323 |
+
"loss": 10.6807,
|
| 3324 |
+
"step": 469
|
| 3325 |
+
},
|
| 3326 |
+
{
|
| 3327 |
+
"epoch": 0.8245614035087719,
|
| 3328 |
+
"grad_norm": 0.24622896313667297,
|
| 3329 |
+
"learning_rate": 1.5327580077171587e-05,
|
| 3330 |
+
"loss": 10.6612,
|
| 3331 |
+
"step": 470
|
| 3332 |
+
},
|
| 3333 |
+
{
|
| 3334 |
+
"epoch": 0.8263157894736842,
|
| 3335 |
+
"grad_norm": 0.2639427185058594,
|
| 3336 |
+
"learning_rate": 1.5030444745471294e-05,
|
| 3337 |
+
"loss": 10.6531,
|
| 3338 |
+
"step": 471
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 0.8280701754385965,
|
| 3342 |
+
"grad_norm": 0.21809430420398712,
|
| 3343 |
+
"learning_rate": 1.4735983564590783e-05,
|
| 3344 |
+
"loss": 10.6821,
|
| 3345 |
+
"step": 472
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 0.8298245614035088,
|
| 3349 |
+
"grad_norm": 0.2837026119232178,
|
| 3350 |
+
"learning_rate": 1.4444205801774202e-05,
|
| 3351 |
+
"loss": 10.6852,
|
| 3352 |
+
"step": 473
|
| 3353 |
+
},
|
| 3354 |
+
{
|
| 3355 |
+
"epoch": 0.8315789473684211,
|
| 3356 |
+
"grad_norm": 0.2315063625574112,
|
| 3357 |
+
"learning_rate": 1.415512063981339e-05,
|
| 3358 |
+
"loss": 10.6723,
|
| 3359 |
+
"step": 474
|
| 3360 |
+
},
|
| 3361 |
+
{
|
| 3362 |
+
"epoch": 0.8333333333333334,
|
| 3363 |
+
"grad_norm": 0.2142920196056366,
|
| 3364 |
+
"learning_rate": 1.3868737176759106e-05,
|
| 3365 |
+
"loss": 10.6347,
|
| 3366 |
+
"step": 475
|
| 3367 |
+
},
|
| 3368 |
+
{
|
| 3369 |
+
"epoch": 0.8350877192982457,
|
| 3370 |
+
"grad_norm": 0.22398516535758972,
|
| 3371 |
+
"learning_rate": 1.3585064425634542e-05,
|
| 3372 |
+
"loss": 10.6954,
|
| 3373 |
+
"step": 476
|
| 3374 |
+
},
|
| 3375 |
+
{
|
| 3376 |
+
"epoch": 0.8368421052631579,
|
| 3377 |
+
"grad_norm": 0.18427734076976776,
|
| 3378 |
+
"learning_rate": 1.330411131415169e-05,
|
| 3379 |
+
"loss": 10.6796,
|
| 3380 |
+
"step": 477
|
| 3381 |
+
},
|
| 3382 |
+
{
|
| 3383 |
+
"epoch": 0.8385964912280702,
|
| 3384 |
+
"grad_norm": 0.20222590863704681,
|
| 3385 |
+
"learning_rate": 1.3025886684430467e-05,
|
| 3386 |
+
"loss": 10.6647,
|
| 3387 |
+
"step": 478
|
| 3388 |
+
},
|
| 3389 |
+
{
|
| 3390 |
+
"epoch": 0.8403508771929824,
|
| 3391 |
+
"grad_norm": 0.19547833502292633,
|
| 3392 |
+
"learning_rate": 1.2750399292720283e-05,
|
| 3393 |
+
"loss": 10.6801,
|
| 3394 |
+
"step": 479
|
| 3395 |
+
},
|
| 3396 |
+
{
|
| 3397 |
+
"epoch": 0.8421052631578947,
|
| 3398 |
+
"grad_norm": 0.23069126904010773,
|
| 3399 |
+
"learning_rate": 1.2477657809124631e-05,
|
| 3400 |
+
"loss": 10.6665,
|
| 3401 |
+
"step": 480
|
| 3402 |
+
},
|
| 3403 |
+
{
|
| 3404 |
+
"epoch": 0.843859649122807,
|
| 3405 |
+
"grad_norm": 0.17966169118881226,
|
| 3406 |
+
"learning_rate": 1.2207670817328066e-05,
|
| 3407 |
+
"loss": 10.6755,
|
| 3408 |
+
"step": 481
|
| 3409 |
+
},
|
| 3410 |
+
{
|
| 3411 |
+
"epoch": 0.8456140350877193,
|
| 3412 |
+
"grad_norm": 0.2698460817337036,
|
| 3413 |
+
"learning_rate": 1.19404468143262e-05,
|
| 3414 |
+
"loss": 10.6617,
|
| 3415 |
+
"step": 482
|
| 3416 |
+
},
|
| 3417 |
+
{
|
| 3418 |
+
"epoch": 0.8473684210526315,
|
| 3419 |
+
"grad_norm": 0.1589714139699936,
|
| 3420 |
+
"learning_rate": 1.1675994210158181e-05,
|
| 3421 |
+
"loss": 10.6625,
|
| 3422 |
+
"step": 483
|
| 3423 |
+
},
|
| 3424 |
+
{
|
| 3425 |
+
"epoch": 0.8491228070175438,
|
| 3426 |
+
"grad_norm": 0.23473794758319855,
|
| 3427 |
+
"learning_rate": 1.141432132764202e-05,
|
| 3428 |
+
"loss": 10.6698,
|
| 3429 |
+
"step": 484
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 0.8508771929824561,
|
| 3433 |
+
"grad_norm": 0.25006183981895447,
|
| 3434 |
+
"learning_rate": 1.1155436402112785e-05,
|
| 3435 |
+
"loss": 10.6668,
|
| 3436 |
+
"step": 485
|
| 3437 |
+
},
|
| 3438 |
+
{
|
| 3439 |
+
"epoch": 0.8526315789473684,
|
| 3440 |
+
"grad_norm": 0.200173020362854,
|
| 3441 |
+
"learning_rate": 1.0899347581163221e-05,
|
| 3442 |
+
"loss": 10.6611,
|
| 3443 |
+
"step": 486
|
| 3444 |
+
},
|
| 3445 |
+
{
|
| 3446 |
+
"epoch": 0.8543859649122807,
|
| 3447 |
+
"grad_norm": 0.18156000971794128,
|
| 3448 |
+
"learning_rate": 1.0646062924387512e-05,
|
| 3449 |
+
"loss": 10.6593,
|
| 3450 |
+
"step": 487
|
| 3451 |
+
},
|
| 3452 |
+
{
|
| 3453 |
+
"epoch": 0.856140350877193,
|
| 3454 |
+
"grad_norm": 0.2163092941045761,
|
| 3455 |
+
"learning_rate": 1.0395590403127486e-05,
|
| 3456 |
+
"loss": 10.6704,
|
| 3457 |
+
"step": 488
|
| 3458 |
+
},
|
| 3459 |
+
{
|
| 3460 |
+
"epoch": 0.8578947368421053,
|
| 3461 |
+
"grad_norm": 0.29484280943870544,
|
| 3462 |
+
"learning_rate": 1.0147937900221883e-05,
|
| 3463 |
+
"loss": 10.6809,
|
| 3464 |
+
"step": 489
|
| 3465 |
+
},
|
| 3466 |
+
{
|
| 3467 |
+
"epoch": 0.8596491228070176,
|
| 3468 |
+
"grad_norm": 0.2818056046962738,
|
| 3469 |
+
"learning_rate": 9.903113209758096e-06,
|
| 3470 |
+
"loss": 10.6291,
|
| 3471 |
+
"step": 490
|
| 3472 |
+
},
|
| 3473 |
+
{
|
| 3474 |
+
"epoch": 0.8614035087719298,
|
| 3475 |
+
"grad_norm": 0.19995582103729248,
|
| 3476 |
+
"learning_rate": 9.661124036827063e-06,
|
| 3477 |
+
"loss": 10.6645,
|
| 3478 |
+
"step": 491
|
| 3479 |
+
},
|
| 3480 |
+
{
|
| 3481 |
+
"epoch": 0.8631578947368421,
|
| 3482 |
+
"grad_norm": 0.25295835733413696,
|
| 3483 |
+
"learning_rate": 9.421977997280596e-06,
|
| 3484 |
+
"loss": 10.6527,
|
| 3485 |
+
"step": 492
|
| 3486 |
+
},
|
| 3487 |
+
{
|
| 3488 |
+
"epoch": 0.8649122807017544,
|
| 3489 |
+
"grad_norm": 0.1981017291545868,
|
| 3490 |
+
"learning_rate": 9.185682617491863e-06,
|
| 3491 |
+
"loss": 10.6562,
|
| 3492 |
+
"step": 493
|
| 3493 |
+
},
|
| 3494 |
+
{
|
| 3495 |
+
"epoch": 0.8666666666666667,
|
| 3496 |
+
"grad_norm": 0.16144883632659912,
|
| 3497 |
+
"learning_rate": 8.952245334118414e-06,
|
| 3498 |
+
"loss": 10.6626,
|
| 3499 |
+
"step": 494
|
| 3500 |
+
},
|
| 3501 |
+
{
|
| 3502 |
+
"epoch": 0.868421052631579,
|
| 3503 |
+
"grad_norm": 0.23267914354801178,
|
| 3504 |
+
"learning_rate": 8.72167349386811e-06,
|
| 3505 |
+
"loss": 10.6941,
|
| 3506 |
+
"step": 495
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 0.8701754385964913,
|
| 3510 |
+
"grad_norm": 0.20202142000198364,
|
| 3511 |
+
"learning_rate": 8.493974353268019e-06,
|
| 3512 |
+
"loss": 10.6771,
|
| 3513 |
+
"step": 496
|
| 3514 |
+
},
|
| 3515 |
+
{
|
| 3516 |
+
"epoch": 0.8719298245614036,
|
| 3517 |
+
"grad_norm": 0.34231188893318176,
|
| 3518 |
+
"learning_rate": 8.269155078435931e-06,
|
| 3519 |
+
"loss": 10.7252,
|
| 3520 |
+
"step": 497
|
| 3521 |
+
},
|
| 3522 |
+
{
|
| 3523 |
+
"epoch": 0.8736842105263158,
|
| 3524 |
+
"grad_norm": 0.17752091586589813,
|
| 3525 |
+
"learning_rate": 8.047222744854943e-06,
|
| 3526 |
+
"loss": 10.6724,
|
| 3527 |
+
"step": 498
|
| 3528 |
+
},
|
| 3529 |
+
{
|
| 3530 |
+
"epoch": 0.875438596491228,
|
| 3531 |
+
"grad_norm": 0.26747509837150574,
|
| 3532 |
+
"learning_rate": 7.828184337150613e-06,
|
| 3533 |
+
"loss": 10.7056,
|
| 3534 |
+
"step": 499
|
| 3535 |
+
},
|
| 3536 |
+
{
|
| 3537 |
+
"epoch": 0.8771929824561403,
|
| 3538 |
+
"grad_norm": 0.26659560203552246,
|
| 3539 |
+
"learning_rate": 7.612046748871327e-06,
|
| 3540 |
+
"loss": 10.7047,
|
| 3541 |
+
"step": 500
|
| 3542 |
+
},
|
| 3543 |
+
{
|
| 3544 |
+
"epoch": 0.8789473684210526,
|
| 3545 |
+
"grad_norm": 0.19917453825473785,
|
| 3546 |
+
"learning_rate": 7.398816782271223e-06,
|
| 3547 |
+
"loss": 10.704,
|
| 3548 |
+
"step": 501
|
| 3549 |
+
},
|
| 3550 |
+
{
|
| 3551 |
+
"epoch": 0.8807017543859649,
|
| 3552 |
+
"grad_norm": 0.23498232662677765,
|
| 3553 |
+
"learning_rate": 7.1885011480961164e-06,
|
| 3554 |
+
"loss": 10.6675,
|
| 3555 |
+
"step": 502
|
| 3556 |
+
},
|
| 3557 |
+
{
|
| 3558 |
+
"epoch": 0.8824561403508772,
|
| 3559 |
+
"grad_norm": 0.21307386457920074,
|
| 3560 |
+
"learning_rate": 6.981106465372389e-06,
|
| 3561 |
+
"loss": 10.6562,
|
| 3562 |
+
"step": 503
|
| 3563 |
+
},
|
| 3564 |
+
{
|
| 3565 |
+
"epoch": 0.8842105263157894,
|
| 3566 |
+
"grad_norm": 0.3048202097415924,
|
| 3567 |
+
"learning_rate": 6.776639261198581e-06,
|
| 3568 |
+
"loss": 10.6504,
|
| 3569 |
+
"step": 504
|
| 3570 |
+
},
|
| 3571 |
+
{
|
| 3572 |
+
"epoch": 0.8859649122807017,
|
| 3573 |
+
"grad_norm": 0.221920445561409,
|
| 3574 |
+
"learning_rate": 6.5751059705400295e-06,
|
| 3575 |
+
"loss": 10.6859,
|
| 3576 |
+
"step": 505
|
| 3577 |
+
},
|
| 3578 |
+
{
|
| 3579 |
+
"epoch": 0.887719298245614,
|
| 3580 |
+
"grad_norm": 0.19181109964847565,
|
| 3581 |
+
"learning_rate": 6.37651293602628e-06,
|
| 3582 |
+
"loss": 10.6674,
|
| 3583 |
+
"step": 506
|
| 3584 |
+
},
|
| 3585 |
+
{
|
| 3586 |
+
"epoch": 0.8894736842105263,
|
| 3587 |
+
"grad_norm": 0.23929765820503235,
|
| 3588 |
+
"learning_rate": 6.180866407751595e-06,
|
| 3589 |
+
"loss": 10.6818,
|
| 3590 |
+
"step": 507
|
| 3591 |
+
},
|
| 3592 |
+
{
|
| 3593 |
+
"epoch": 0.8912280701754386,
|
| 3594 |
+
"grad_norm": 0.28008386492729187,
|
| 3595 |
+
"learning_rate": 5.988172543078097e-06,
|
| 3596 |
+
"loss": 10.7264,
|
| 3597 |
+
"step": 508
|
| 3598 |
+
},
|
| 3599 |
+
{
|
| 3600 |
+
"epoch": 0.8929824561403509,
|
| 3601 |
+
"grad_norm": 0.31902721524238586,
|
| 3602 |
+
"learning_rate": 5.7984374064421035e-06,
|
| 3603 |
+
"loss": 10.6314,
|
| 3604 |
+
"step": 509
|
| 3605 |
+
},
|
| 3606 |
+
{
|
| 3607 |
+
"epoch": 0.8947368421052632,
|
| 3608 |
+
"grad_norm": 0.22529898583889008,
|
| 3609 |
+
"learning_rate": 5.611666969163243e-06,
|
| 3610 |
+
"loss": 10.6476,
|
| 3611 |
+
"step": 510
|
| 3612 |
+
},
|
| 3613 |
+
{
|
| 3614 |
+
"epoch": 0.8964912280701754,
|
| 3615 |
+
"grad_norm": 0.18052135407924652,
|
| 3616 |
+
"learning_rate": 5.427867109256457e-06,
|
| 3617 |
+
"loss": 10.6507,
|
| 3618 |
+
"step": 511
|
| 3619 |
+
},
|
| 3620 |
+
{
|
| 3621 |
+
"epoch": 0.8982456140350877,
|
| 3622 |
+
"grad_norm": 0.20663322508335114,
|
| 3623 |
+
"learning_rate": 5.247043611247127e-06,
|
| 3624 |
+
"loss": 10.6917,
|
| 3625 |
+
"step": 512
|
| 3626 |
+
},
|
| 3627 |
+
{
|
| 3628 |
+
"epoch": 0.9,
|
| 3629 |
+
"grad_norm": 0.19402727484703064,
|
| 3630 |
+
"learning_rate": 5.0692021659888735e-06,
|
| 3631 |
+
"loss": 10.6435,
|
| 3632 |
+
"step": 513
|
| 3633 |
+
},
|
| 3634 |
+
{
|
| 3635 |
+
"epoch": 0.9017543859649123,
|
| 3636 |
+
"grad_norm": 0.18608134984970093,
|
| 3637 |
+
"learning_rate": 4.8943483704846475e-06,
|
| 3638 |
+
"loss": 10.6597,
|
| 3639 |
+
"step": 514
|
| 3640 |
+
},
|
| 3641 |
+
{
|
| 3642 |
+
"epoch": 0.9035087719298246,
|
| 3643 |
+
"grad_norm": 0.23060499131679535,
|
| 3644 |
+
"learning_rate": 4.722487727710368e-06,
|
| 3645 |
+
"loss": 10.673,
|
| 3646 |
+
"step": 515
|
| 3647 |
+
},
|
| 3648 |
+
{
|
| 3649 |
+
"epoch": 0.9052631578947369,
|
| 3650 |
+
"grad_norm": 0.2104937583208084,
|
| 3651 |
+
"learning_rate": 4.553625646441928e-06,
|
| 3652 |
+
"loss": 10.6911,
|
| 3653 |
+
"step": 516
|
| 3654 |
+
},
|
| 3655 |
+
{
|
| 3656 |
+
"epoch": 0.9070175438596492,
|
| 3657 |
+
"grad_norm": 0.17810438573360443,
|
| 3658 |
+
"learning_rate": 4.3877674410848e-06,
|
| 3659 |
+
"loss": 10.6848,
|
| 3660 |
+
"step": 517
|
| 3661 |
+
},
|
| 3662 |
+
{
|
| 3663 |
+
"epoch": 0.9087719298245615,
|
| 3664 |
+
"grad_norm": 0.25825297832489014,
|
| 3665 |
+
"learning_rate": 4.224918331506955e-06,
|
| 3666 |
+
"loss": 10.6735,
|
| 3667 |
+
"step": 518
|
| 3668 |
+
},
|
| 3669 |
+
{
|
| 3670 |
+
"epoch": 0.9105263157894737,
|
| 3671 |
+
"grad_norm": 0.16333693265914917,
|
| 3672 |
+
"learning_rate": 4.065083442874418e-06,
|
| 3673 |
+
"loss": 10.6632,
|
| 3674 |
+
"step": 519
|
| 3675 |
+
},
|
| 3676 |
+
{
|
| 3677 |
+
"epoch": 0.9122807017543859,
|
| 3678 |
+
"grad_norm": 0.21222874522209167,
|
| 3679 |
+
"learning_rate": 3.908267805490051e-06,
|
| 3680 |
+
"loss": 10.6984,
|
| 3681 |
+
"step": 520
|
| 3682 |
+
},
|
| 3683 |
+
{
|
| 3684 |
+
"epoch": 0.9140350877192982,
|
| 3685 |
+
"grad_norm": 0.20313438773155212,
|
| 3686 |
+
"learning_rate": 3.7544763546352834e-06,
|
| 3687 |
+
"loss": 10.6889,
|
| 3688 |
+
"step": 521
|
| 3689 |
+
},
|
| 3690 |
+
{
|
| 3691 |
+
"epoch": 0.9157894736842105,
|
| 3692 |
+
"grad_norm": 0.201041117310524,
|
| 3693 |
+
"learning_rate": 3.6037139304146762e-06,
|
| 3694 |
+
"loss": 10.6284,
|
| 3695 |
+
"step": 522
|
| 3696 |
+
},
|
| 3697 |
+
{
|
| 3698 |
+
"epoch": 0.9175438596491228,
|
| 3699 |
+
"grad_norm": 0.2828705906867981,
|
| 3700 |
+
"learning_rate": 3.455985277603713e-06,
|
| 3701 |
+
"loss": 10.6829,
|
| 3702 |
+
"step": 523
|
| 3703 |
+
},
|
| 3704 |
+
{
|
| 3705 |
+
"epoch": 0.9192982456140351,
|
| 3706 |
+
"grad_norm": 0.2506016492843628,
|
| 3707 |
+
"learning_rate": 3.311295045499363e-06,
|
| 3708 |
+
"loss": 10.6884,
|
| 3709 |
+
"step": 524
|
| 3710 |
+
},
|
| 3711 |
+
{
|
| 3712 |
+
"epoch": 0.9210526315789473,
|
| 3713 |
+
"grad_norm": 0.42757734656333923,
|
| 3714 |
+
"learning_rate": 3.169647787773866e-06,
|
| 3715 |
+
"loss": 10.6356,
|
| 3716 |
+
"step": 525
|
| 3717 |
+
},
|
| 3718 |
+
{
|
| 3719 |
+
"epoch": 0.9228070175438596,
|
| 3720 |
+
"grad_norm": 0.2824893295764923,
|
| 3721 |
+
"learning_rate": 3.0310479623313127e-06,
|
| 3722 |
+
"loss": 10.7104,
|
| 3723 |
+
"step": 526
|
| 3724 |
+
},
|
| 3725 |
+
{
|
| 3726 |
+
"epoch": 0.9245614035087719,
|
| 3727 |
+
"grad_norm": 0.2309618890285492,
|
| 3728 |
+
"learning_rate": 2.8954999311674558e-06,
|
| 3729 |
+
"loss": 10.68,
|
| 3730 |
+
"step": 527
|
| 3731 |
+
},
|
| 3732 |
+
{
|
| 3733 |
+
"epoch": 0.9263157894736842,
|
| 3734 |
+
"grad_norm": 0.23892293870449066,
|
| 3735 |
+
"learning_rate": 2.7630079602323442e-06,
|
| 3736 |
+
"loss": 10.6244,
|
| 3737 |
+
"step": 528
|
| 3738 |
+
},
|
| 3739 |
+
{
|
| 3740 |
+
"epoch": 0.9280701754385965,
|
| 3741 |
+
"grad_norm": 0.21768365800380707,
|
| 3742 |
+
"learning_rate": 2.6335762192960743e-06,
|
| 3743 |
+
"loss": 10.6715,
|
| 3744 |
+
"step": 529
|
| 3745 |
+
},
|
| 3746 |
+
{
|
| 3747 |
+
"epoch": 0.9298245614035088,
|
| 3748 |
+
"grad_norm": 0.22950419783592224,
|
| 3749 |
+
"learning_rate": 2.5072087818176382e-06,
|
| 3750 |
+
"loss": 10.6401,
|
| 3751 |
+
"step": 530
|
| 3752 |
+
},
|
| 3753 |
+
{
|
| 3754 |
+
"epoch": 0.9315789473684211,
|
| 3755 |
+
"grad_norm": 0.20661982893943787,
|
| 3756 |
+
"learning_rate": 2.383909624816616e-06,
|
| 3757 |
+
"loss": 10.6522,
|
| 3758 |
+
"step": 531
|
| 3759 |
+
},
|
| 3760 |
+
{
|
| 3761 |
+
"epoch": 0.9333333333333333,
|
| 3762 |
+
"grad_norm": 0.18806463479995728,
|
| 3763 |
+
"learning_rate": 2.2636826287480873e-06,
|
| 3764 |
+
"loss": 10.6811,
|
| 3765 |
+
"step": 532
|
| 3766 |
+
},
|
| 3767 |
+
{
|
| 3768 |
+
"epoch": 0.9350877192982456,
|
| 3769 |
+
"grad_norm": 0.3133368194103241,
|
| 3770 |
+
"learning_rate": 2.1465315773804616e-06,
|
| 3771 |
+
"loss": 10.6325,
|
| 3772 |
+
"step": 533
|
| 3773 |
+
},
|
| 3774 |
+
{
|
| 3775 |
+
"epoch": 0.9368421052631579,
|
| 3776 |
+
"grad_norm": 0.23937764763832092,
|
| 3777 |
+
"learning_rate": 2.032460157676452e-06,
|
| 3778 |
+
"loss": 10.669,
|
| 3779 |
+
"step": 534
|
| 3780 |
+
},
|
| 3781 |
+
{
|
| 3782 |
+
"epoch": 0.9385964912280702,
|
| 3783 |
+
"grad_norm": 0.21383170783519745,
|
| 3784 |
+
"learning_rate": 1.921471959676957e-06,
|
| 3785 |
+
"loss": 10.6751,
|
| 3786 |
+
"step": 535
|
| 3787 |
+
},
|
| 3788 |
+
{
|
| 3789 |
+
"epoch": 0.9403508771929825,
|
| 3790 |
+
"grad_norm": 0.3915383219718933,
|
| 3791 |
+
"learning_rate": 1.81357047638816e-06,
|
| 3792 |
+
"loss": 10.6673,
|
| 3793 |
+
"step": 536
|
| 3794 |
+
},
|
| 3795 |
+
{
|
| 3796 |
+
"epoch": 0.9421052631578948,
|
| 3797 |
+
"grad_norm": 0.21916410326957703,
|
| 3798 |
+
"learning_rate": 1.7087591036715534e-06,
|
| 3799 |
+
"loss": 10.7053,
|
| 3800 |
+
"step": 537
|
| 3801 |
+
},
|
| 3802 |
+
{
|
| 3803 |
+
"epoch": 0.9438596491228071,
|
| 3804 |
+
"grad_norm": 0.26692306995391846,
|
| 3805 |
+
"learning_rate": 1.6070411401370334e-06,
|
| 3806 |
+
"loss": 10.6689,
|
| 3807 |
+
"step": 538
|
| 3808 |
+
},
|
| 3809 |
+
{
|
| 3810 |
+
"epoch": 0.9456140350877194,
|
| 3811 |
+
"grad_norm": 0.19968879222869873,
|
| 3812 |
+
"learning_rate": 1.5084197870391837e-06,
|
| 3813 |
+
"loss": 10.7208,
|
| 3814 |
+
"step": 539
|
| 3815 |
+
},
|
| 3816 |
+
{
|
| 3817 |
+
"epoch": 0.9473684210526315,
|
| 3818 |
+
"grad_norm": 0.21784716844558716,
|
| 3819 |
+
"learning_rate": 1.4128981481764115e-06,
|
| 3820 |
+
"loss": 10.6356,
|
| 3821 |
+
"step": 540
|
| 3822 |
+
},
|
| 3823 |
+
{
|
| 3824 |
+
"epoch": 0.9491228070175438,
|
| 3825 |
+
"grad_norm": 0.1975242644548416,
|
| 3826 |
+
"learning_rate": 1.3204792297933588e-06,
|
| 3827 |
+
"loss": 10.7024,
|
| 3828 |
+
"step": 541
|
| 3829 |
+
},
|
| 3830 |
+
{
|
| 3831 |
+
"epoch": 0.9508771929824561,
|
| 3832 |
+
"grad_norm": 0.2712066173553467,
|
| 3833 |
+
"learning_rate": 1.231165940486234e-06,
|
| 3834 |
+
"loss": 10.6415,
|
| 3835 |
+
"step": 542
|
| 3836 |
+
},
|
| 3837 |
+
{
|
| 3838 |
+
"epoch": 0.9526315789473684,
|
| 3839 |
+
"grad_norm": 0.19951169192790985,
|
| 3840 |
+
"learning_rate": 1.1449610911112741e-06,
|
| 3841 |
+
"loss": 10.6615,
|
| 3842 |
+
"step": 543
|
| 3843 |
+
},
|
| 3844 |
+
{
|
| 3845 |
+
"epoch": 0.9543859649122807,
|
| 3846 |
+
"grad_norm": 0.21636833250522614,
|
| 3847 |
+
"learning_rate": 1.0618673946963365e-06,
|
| 3848 |
+
"loss": 10.67,
|
| 3849 |
+
"step": 544
|
| 3850 |
+
},
|
| 3851 |
+
{
|
| 3852 |
+
"epoch": 0.956140350877193,
|
| 3853 |
+
"grad_norm": 0.24349869787693024,
|
| 3854 |
+
"learning_rate": 9.818874663554357e-07,
|
| 3855 |
+
"loss": 10.658,
|
| 3856 |
+
"step": 545
|
| 3857 |
+
},
|
| 3858 |
+
{
|
| 3859 |
+
"epoch": 0.9578947368421052,
|
| 3860 |
+
"grad_norm": 0.22371384501457214,
|
| 3861 |
+
"learning_rate": 9.0502382320653e-07,
|
| 3862 |
+
"loss": 10.6703,
|
| 3863 |
+
"step": 546
|
| 3864 |
+
},
|
| 3865 |
+
{
|
| 3866 |
+
"epoch": 0.9596491228070175,
|
| 3867 |
+
"grad_norm": 0.2083524912595749,
|
| 3868 |
+
"learning_rate": 8.31278884292186e-07,
|
| 3869 |
+
"loss": 10.6766,
|
| 3870 |
+
"step": 547
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 0.9614035087719298,
|
| 3874 |
+
"grad_norm": 0.2264506220817566,
|
| 3875 |
+
"learning_rate": 7.606549705035937e-07,
|
| 3876 |
+
"loss": 10.672,
|
| 3877 |
+
"step": 548
|
| 3878 |
+
},
|
| 3879 |
+
{
|
| 3880 |
+
"epoch": 0.9631578947368421,
|
| 3881 |
+
"grad_norm": 0.21471793949604034,
|
| 3882 |
+
"learning_rate": 6.931543045073708e-07,
|
| 3883 |
+
"loss": 10.7041,
|
| 3884 |
+
"step": 549
|
| 3885 |
+
},
|
| 3886 |
+
{
|
| 3887 |
+
"epoch": 0.9649122807017544,
|
| 3888 |
+
"grad_norm": 0.20497629046440125,
|
| 3889 |
+
"learning_rate": 6.287790106757396e-07,
|
| 3890 |
+
"loss": 10.6923,
|
| 3891 |
+
"step": 550
|
| 3892 |
+
},
|
| 3893 |
+
{
|
| 3894 |
+
"epoch": 0.9666666666666667,
|
| 3895 |
+
"grad_norm": 0.18738481402397156,
|
| 3896 |
+
"learning_rate": 5.675311150195928e-07,
|
| 3897 |
+
"loss": 10.6622,
|
| 3898 |
+
"step": 551
|
| 3899 |
+
},
|
| 3900 |
+
{
|
| 3901 |
+
"epoch": 0.968421052631579,
|
| 3902 |
+
"grad_norm": 0.24018672108650208,
|
| 3903 |
+
"learning_rate": 5.094125451247655e-07,
|
| 3904 |
+
"loss": 10.629,
|
| 3905 |
+
"step": 552
|
| 3906 |
+
},
|
| 3907 |
+
{
|
| 3908 |
+
"epoch": 0.9701754385964912,
|
| 3909 |
+
"grad_norm": 0.23290672898292542,
|
| 3910 |
+
"learning_rate": 4.544251300913405e-07,
|
| 3911 |
+
"loss": 10.6888,
|
| 3912 |
+
"step": 553
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 0.9719298245614035,
|
| 3916 |
+
"grad_norm": 0.3501298129558563,
|
| 3917 |
+
"learning_rate": 4.025706004760932e-07,
|
| 3918 |
+
"loss": 10.6527,
|
| 3919 |
+
"step": 554
|
| 3920 |
+
},
|
| 3921 |
+
{
|
| 3922 |
+
"epoch": 0.9736842105263158,
|
| 3923 |
+
"grad_norm": 0.2251330465078354,
|
| 3924 |
+
"learning_rate": 3.5385058823809156e-07,
|
| 3925 |
+
"loss": 10.696,
|
| 3926 |
+
"step": 555
|
| 3927 |
+
},
|
| 3928 |
+
{
|
| 3929 |
+
"epoch": 0.9754385964912281,
|
| 3930 |
+
"grad_norm": 0.24375322461128235,
|
| 3931 |
+
"learning_rate": 3.0826662668720364e-07,
|
| 3932 |
+
"loss": 10.7063,
|
| 3933 |
+
"step": 556
|
| 3934 |
+
},
|
| 3935 |
+
{
|
| 3936 |
+
"epoch": 0.9771929824561404,
|
| 3937 |
+
"grad_norm": 0.2920641303062439,
|
| 3938 |
+
"learning_rate": 2.658201504359803e-07,
|
| 3939 |
+
"loss": 10.645,
|
| 3940 |
+
"step": 557
|
| 3941 |
+
},
|
| 3942 |
+
{
|
| 3943 |
+
"epoch": 0.9789473684210527,
|
| 3944 |
+
"grad_norm": 0.29870396852493286,
|
| 3945 |
+
"learning_rate": 2.265124953543918e-07,
|
| 3946 |
+
"loss": 10.6955,
|
| 3947 |
+
"step": 558
|
| 3948 |
+
},
|
| 3949 |
+
{
|
| 3950 |
+
"epoch": 0.980701754385965,
|
| 3951 |
+
"grad_norm": 0.2032945305109024,
|
| 3952 |
+
"learning_rate": 1.9034489852787218e-07,
|
| 3953 |
+
"loss": 10.7028,
|
| 3954 |
+
"step": 559
|
| 3955 |
+
},
|
| 3956 |
+
{
|
| 3957 |
+
"epoch": 0.9824561403508771,
|
| 3958 |
+
"grad_norm": 0.16992153227329254,
|
| 3959 |
+
"learning_rate": 1.5731849821833954e-07,
|
| 3960 |
+
"loss": 10.6516,
|
| 3961 |
+
"step": 560
|
| 3962 |
+
},
|
| 3963 |
+
{
|
| 3964 |
+
"epoch": 0.9842105263157894,
|
| 3965 |
+
"grad_norm": 0.19095556437969208,
|
| 3966 |
+
"learning_rate": 1.274343338283801e-07,
|
| 3967 |
+
"loss": 10.6713,
|
| 3968 |
+
"step": 561
|
| 3969 |
+
},
|
| 3970 |
+
{
|
| 3971 |
+
"epoch": 0.9859649122807017,
|
| 3972 |
+
"grad_norm": 0.2375078797340393,
|
| 3973 |
+
"learning_rate": 1.0069334586854107e-07,
|
| 3974 |
+
"loss": 10.6812,
|
| 3975 |
+
"step": 562
|
| 3976 |
+
},
|
| 3977 |
+
{
|
| 3978 |
+
"epoch": 0.987719298245614,
|
| 3979 |
+
"grad_norm": 0.1917848289012909,
|
| 3980 |
+
"learning_rate": 7.709637592770991e-08,
|
| 3981 |
+
"loss": 10.6814,
|
| 3982 |
+
"step": 563
|
| 3983 |
+
},
|
| 3984 |
+
{
|
| 3985 |
+
"epoch": 0.9894736842105263,
|
| 3986 |
+
"grad_norm": 0.2119506448507309,
|
| 3987 |
+
"learning_rate": 5.6644166646668826e-08,
|
| 3988 |
+
"loss": 10.7086,
|
| 3989 |
+
"step": 564
|
| 3990 |
+
},
|
| 3991 |
+
{
|
| 3992 |
+
"epoch": 0.9912280701754386,
|
| 3993 |
+
"grad_norm": 0.21006697416305542,
|
| 3994 |
+
"learning_rate": 3.933736169471347e-08,
|
| 3995 |
+
"loss": 10.6638,
|
| 3996 |
+
"step": 565
|
| 3997 |
+
},
|
| 3998 |
+
{
|
| 3999 |
+
"epoch": 0.9929824561403509,
|
| 4000 |
+
"grad_norm": 0.19843615591526031,
|
| 4001 |
+
"learning_rate": 2.5176505749346936e-08,
|
| 4002 |
+
"loss": 10.6587,
|
| 4003 |
+
"step": 566
|
| 4004 |
+
},
|
| 4005 |
+
{
|
| 4006 |
+
"epoch": 0.9947368421052631,
|
| 4007 |
+
"grad_norm": 0.292624831199646,
|
| 4008 |
+
"learning_rate": 1.4162044479182346e-08,
|
| 4009 |
+
"loss": 10.6584,
|
| 4010 |
+
"step": 567
|
| 4011 |
+
},
|
| 4012 |
+
{
|
| 4013 |
+
"epoch": 0.9964912280701754,
|
| 4014 |
+
"grad_norm": 0.14934279024600983,
|
| 4015 |
+
"learning_rate": 6.294324529942941e-09,
|
| 4016 |
+
"loss": 10.6876,
|
| 4017 |
+
"step": 568
|
| 4018 |
+
},
|
| 4019 |
+
{
|
| 4020 |
+
"epoch": 0.9982456140350877,
|
| 4021 |
+
"grad_norm": 0.24475471675395966,
|
| 4022 |
+
"learning_rate": 1.5735935134708613e-09,
|
| 4023 |
+
"loss": 10.6859,
|
| 4024 |
+
"step": 569
|
| 4025 |
+
},
|
| 4026 |
+
{
|
| 4027 |
+
"epoch": 1.0,
|
| 4028 |
+
"grad_norm": 0.17199726402759552,
|
| 4029 |
+
"learning_rate": 0.0,
|
| 4030 |
+
"loss": 10.7093,
|
| 4031 |
+
"step": 570
|
| 4032 |
}
|
| 4033 |
],
|
| 4034 |
"logging_steps": 1,
|
|
|
|
| 4043 |
"should_evaluate": false,
|
| 4044 |
"should_log": false,
|
| 4045 |
"should_save": true,
|
| 4046 |
+
"should_training_stop": true
|
| 4047 |
},
|
| 4048 |
"attributes": {}
|
| 4049 |
}
|
| 4050 |
},
|
| 4051 |
+
"total_flos": 56846496301056.0,
|
| 4052 |
"train_batch_size": 2,
|
| 4053 |
"trial_name": null,
|
| 4054 |
"trial_params": null
|