Training in progress, step 572, 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 5327744
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:83da110c38b42c8f26332874e8bb627d6ead7522e50fc51e39079d4f9c35796a
|
| 3 |
size 5327744
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2857850
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b3646603ee02c3a1dc6142029b48d761a5d129039bb1a5936f191072a5323f6
|
| 3 |
size 2857850
|
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 sha256:691d0261780df62255146217a0fa66594f0de41a9521d0f54f49d2368cb05292
|
| 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 sha256:63ce17dd2c32e1042039dfe648c482c9ff0032ac68df46007019bf1f153ddc3e
|
| 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 +3042,1015 @@
|
|
| 3042 |
"eval_samples_per_second": 110.64,
|
| 3043 |
"eval_steps_per_second": 55.55,
|
| 3044 |
"step": 429
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3045 |
}
|
| 3046 |
],
|
| 3047 |
"logging_steps": 1,
|
|
@@ -3056,12 +4065,12 @@
|
|
| 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": 1.0008745080891999,
|
| 5 |
"eval_steps": 143,
|
| 6 |
+
"global_step": 572,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3042 |
"eval_samples_per_second": 110.64,
|
| 3043 |
"eval_steps_per_second": 55.55,
|
| 3044 |
"step": 429
|
| 3045 |
+
},
|
| 3046 |
+
{
|
| 3047 |
+
"epoch": 0.7520769567118496,
|
| 3048 |
+
"grad_norm": 6.074078559875488,
|
| 3049 |
+
"learning_rate": 2.9884712134280324e-05,
|
| 3050 |
+
"loss": 6.6575,
|
| 3051 |
+
"step": 430
|
| 3052 |
+
},
|
| 3053 |
+
{
|
| 3054 |
+
"epoch": 0.7538259728902492,
|
| 3055 |
+
"grad_norm": 6.187041759490967,
|
| 3056 |
+
"learning_rate": 2.948723496423379e-05,
|
| 3057 |
+
"loss": 7.9171,
|
| 3058 |
+
"step": 431
|
| 3059 |
+
},
|
| 3060 |
+
{
|
| 3061 |
+
"epoch": 0.7555749890686488,
|
| 3062 |
+
"grad_norm": 6.312707901000977,
|
| 3063 |
+
"learning_rate": 2.909196119613218e-05,
|
| 3064 |
+
"loss": 8.3438,
|
| 3065 |
+
"step": 432
|
| 3066 |
+
},
|
| 3067 |
+
{
|
| 3068 |
+
"epoch": 0.7573240052470486,
|
| 3069 |
+
"grad_norm": 5.885359287261963,
|
| 3070 |
+
"learning_rate": 2.8698903181597127e-05,
|
| 3071 |
+
"loss": 7.2738,
|
| 3072 |
+
"step": 433
|
| 3073 |
+
},
|
| 3074 |
+
{
|
| 3075 |
+
"epoch": 0.7590730214254482,
|
| 3076 |
+
"grad_norm": 6.485308647155762,
|
| 3077 |
+
"learning_rate": 2.8308073203011663e-05,
|
| 3078 |
+
"loss": 8.3997,
|
| 3079 |
+
"step": 434
|
| 3080 |
+
},
|
| 3081 |
+
{
|
| 3082 |
+
"epoch": 0.7608220376038478,
|
| 3083 |
+
"grad_norm": 5.899681091308594,
|
| 3084 |
+
"learning_rate": 2.7919483473136676e-05,
|
| 3085 |
+
"loss": 6.9346,
|
| 3086 |
+
"step": 435
|
| 3087 |
+
},
|
| 3088 |
+
{
|
| 3089 |
+
"epoch": 0.7625710537822474,
|
| 3090 |
+
"grad_norm": 6.48416805267334,
|
| 3091 |
+
"learning_rate": 2.753314613472906e-05,
|
| 3092 |
+
"loss": 8.0168,
|
| 3093 |
+
"step": 436
|
| 3094 |
+
},
|
| 3095 |
+
{
|
| 3096 |
+
"epoch": 0.7643200699606472,
|
| 3097 |
+
"grad_norm": 6.464148998260498,
|
| 3098 |
+
"learning_rate": 2.7149073260162416e-05,
|
| 3099 |
+
"loss": 7.6221,
|
| 3100 |
+
"step": 437
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"epoch": 0.7660690861390468,
|
| 3104 |
+
"grad_norm": 6.158417224884033,
|
| 3105 |
+
"learning_rate": 2.6767276851049816e-05,
|
| 3106 |
+
"loss": 8.3056,
|
| 3107 |
+
"step": 438
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"epoch": 0.7678181023174464,
|
| 3111 |
+
"grad_norm": 6.386941432952881,
|
| 3112 |
+
"learning_rate": 2.6387768837868597e-05,
|
| 3113 |
+
"loss": 7.7409,
|
| 3114 |
+
"step": 439
|
| 3115 |
+
},
|
| 3116 |
+
{
|
| 3117 |
+
"epoch": 0.7695671184958461,
|
| 3118 |
+
"grad_norm": 5.983461380004883,
|
| 3119 |
+
"learning_rate": 2.6010561079587813e-05,
|
| 3120 |
+
"loss": 8.2641,
|
| 3121 |
+
"step": 440
|
| 3122 |
+
},
|
| 3123 |
+
{
|
| 3124 |
+
"epoch": 0.7713161346742458,
|
| 3125 |
+
"grad_norm": 6.103724479675293,
|
| 3126 |
+
"learning_rate": 2.5635665363297422e-05,
|
| 3127 |
+
"loss": 7.4958,
|
| 3128 |
+
"step": 441
|
| 3129 |
+
},
|
| 3130 |
+
{
|
| 3131 |
+
"epoch": 0.7730651508526454,
|
| 3132 |
+
"grad_norm": 6.250931739807129,
|
| 3133 |
+
"learning_rate": 2.5263093403840142e-05,
|
| 3134 |
+
"loss": 8.1695,
|
| 3135 |
+
"step": 442
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 0.774814167031045,
|
| 3139 |
+
"grad_norm": 6.299392223358154,
|
| 3140 |
+
"learning_rate": 2.4892856843445322e-05,
|
| 3141 |
+
"loss": 8.1588,
|
| 3142 |
+
"step": 443
|
| 3143 |
+
},
|
| 3144 |
+
{
|
| 3145 |
+
"epoch": 0.7765631832094447,
|
| 3146 |
+
"grad_norm": 7.157833099365234,
|
| 3147 |
+
"learning_rate": 2.4524967251365026e-05,
|
| 3148 |
+
"loss": 8.4203,
|
| 3149 |
+
"step": 444
|
| 3150 |
+
},
|
| 3151 |
+
{
|
| 3152 |
+
"epoch": 0.7783121993878444,
|
| 3153 |
+
"grad_norm": 6.5352349281311035,
|
| 3154 |
+
"learning_rate": 2.4159436123512735e-05,
|
| 3155 |
+
"loss": 8.4377,
|
| 3156 |
+
"step": 445
|
| 3157 |
+
},
|
| 3158 |
+
{
|
| 3159 |
+
"epoch": 0.780061215566244,
|
| 3160 |
+
"grad_norm": 5.913573265075684,
|
| 3161 |
+
"learning_rate": 2.3796274882103964e-05,
|
| 3162 |
+
"loss": 6.8979,
|
| 3163 |
+
"step": 446
|
| 3164 |
+
},
|
| 3165 |
+
{
|
| 3166 |
+
"epoch": 0.7818102317446436,
|
| 3167 |
+
"grad_norm": 6.062626838684082,
|
| 3168 |
+
"learning_rate": 2.3435494875299314e-05,
|
| 3169 |
+
"loss": 9.2065,
|
| 3170 |
+
"step": 447
|
| 3171 |
+
},
|
| 3172 |
+
{
|
| 3173 |
+
"epoch": 0.7835592479230433,
|
| 3174 |
+
"grad_norm": 6.175731658935547,
|
| 3175 |
+
"learning_rate": 2.3077107376850005e-05,
|
| 3176 |
+
"loss": 7.3399,
|
| 3177 |
+
"step": 448
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 0.7853082641014429,
|
| 3181 |
+
"grad_norm": 5.536900997161865,
|
| 3182 |
+
"learning_rate": 2.2721123585745507e-05,
|
| 3183 |
+
"loss": 7.0864,
|
| 3184 |
+
"step": 449
|
| 3185 |
+
},
|
| 3186 |
+
{
|
| 3187 |
+
"epoch": 0.7870572802798426,
|
| 3188 |
+
"grad_norm": 6.569960117340088,
|
| 3189 |
+
"learning_rate": 2.2367554625863497e-05,
|
| 3190 |
+
"loss": 7.0612,
|
| 3191 |
+
"step": 450
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 0.7888062964582422,
|
| 3195 |
+
"grad_norm": 6.351140022277832,
|
| 3196 |
+
"learning_rate": 2.2016411545622495e-05,
|
| 3197 |
+
"loss": 7.9918,
|
| 3198 |
+
"step": 451
|
| 3199 |
+
},
|
| 3200 |
+
{
|
| 3201 |
+
"epoch": 0.7905553126366419,
|
| 3202 |
+
"grad_norm": 6.007497787475586,
|
| 3203 |
+
"learning_rate": 2.166770531763633e-05,
|
| 3204 |
+
"loss": 8.2044,
|
| 3205 |
+
"step": 452
|
| 3206 |
+
},
|
| 3207 |
+
{
|
| 3208 |
+
"epoch": 0.7923043288150415,
|
| 3209 |
+
"grad_norm": 6.685032844543457,
|
| 3210 |
+
"learning_rate": 2.132144683837155e-05,
|
| 3211 |
+
"loss": 8.174,
|
| 3212 |
+
"step": 453
|
| 3213 |
+
},
|
| 3214 |
+
{
|
| 3215 |
+
"epoch": 0.7940533449934412,
|
| 3216 |
+
"grad_norm": 5.885077953338623,
|
| 3217 |
+
"learning_rate": 2.0977646927806683e-05,
|
| 3218 |
+
"loss": 7.6945,
|
| 3219 |
+
"step": 454
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 0.7958023611718409,
|
| 3223 |
+
"grad_norm": 6.740251064300537,
|
| 3224 |
+
"learning_rate": 2.0636316329094317e-05,
|
| 3225 |
+
"loss": 7.4873,
|
| 3226 |
+
"step": 455
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 0.7975513773502405,
|
| 3230 |
+
"grad_norm": 6.132180690765381,
|
| 3231 |
+
"learning_rate": 2.0297465708225238e-05,
|
| 3232 |
+
"loss": 8.0784,
|
| 3233 |
+
"step": 456
|
| 3234 |
+
},
|
| 3235 |
+
{
|
| 3236 |
+
"epoch": 0.7993003935286401,
|
| 3237 |
+
"grad_norm": 5.6954779624938965,
|
| 3238 |
+
"learning_rate": 1.9961105653695266e-05,
|
| 3239 |
+
"loss": 8.4652,
|
| 3240 |
+
"step": 457
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 0.8010494097070398,
|
| 3244 |
+
"grad_norm": 6.492379665374756,
|
| 3245 |
+
"learning_rate": 1.962724667617436e-05,
|
| 3246 |
+
"loss": 8.178,
|
| 3247 |
+
"step": 458
|
| 3248 |
+
},
|
| 3249 |
+
{
|
| 3250 |
+
"epoch": 0.8027984258854395,
|
| 3251 |
+
"grad_norm": 6.245090007781982,
|
| 3252 |
+
"learning_rate": 1.929589920817806e-05,
|
| 3253 |
+
"loss": 6.97,
|
| 3254 |
+
"step": 459
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 0.8045474420638391,
|
| 3258 |
+
"grad_norm": 5.776219367980957,
|
| 3259 |
+
"learning_rate": 1.896707360374167e-05,
|
| 3260 |
+
"loss": 8.1005,
|
| 3261 |
+
"step": 460
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 0.8062964582422387,
|
| 3265 |
+
"grad_norm": 6.456238746643066,
|
| 3266 |
+
"learning_rate": 1.8640780138096513e-05,
|
| 3267 |
+
"loss": 7.4033,
|
| 3268 |
+
"step": 461
|
| 3269 |
+
},
|
| 3270 |
+
{
|
| 3271 |
+
"epoch": 0.8080454744206383,
|
| 3272 |
+
"grad_norm": 6.484393119812012,
|
| 3273 |
+
"learning_rate": 1.8317029007349085e-05,
|
| 3274 |
+
"loss": 8.597,
|
| 3275 |
+
"step": 462
|
| 3276 |
+
},
|
| 3277 |
+
{
|
| 3278 |
+
"epoch": 0.8097944905990381,
|
| 3279 |
+
"grad_norm": 6.799935817718506,
|
| 3280 |
+
"learning_rate": 1.799583032816219e-05,
|
| 3281 |
+
"loss": 6.3503,
|
| 3282 |
+
"step": 463
|
| 3283 |
+
},
|
| 3284 |
+
{
|
| 3285 |
+
"epoch": 0.8115435067774377,
|
| 3286 |
+
"grad_norm": 5.950386047363281,
|
| 3287 |
+
"learning_rate": 1.7677194137439035e-05,
|
| 3288 |
+
"loss": 7.3482,
|
| 3289 |
+
"step": 464
|
| 3290 |
+
},
|
| 3291 |
+
{
|
| 3292 |
+
"epoch": 0.8132925229558373,
|
| 3293 |
+
"grad_norm": 6.03711462020874,
|
| 3294 |
+
"learning_rate": 1.7361130392009407e-05,
|
| 3295 |
+
"loss": 8.2516,
|
| 3296 |
+
"step": 465
|
| 3297 |
+
},
|
| 3298 |
+
{
|
| 3299 |
+
"epoch": 0.8150415391342369,
|
| 3300 |
+
"grad_norm": 6.34808874130249,
|
| 3301 |
+
"learning_rate": 1.7047648968318698e-05,
|
| 3302 |
+
"loss": 8.1899,
|
| 3303 |
+
"step": 466
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 0.8167905553126367,
|
| 3307 |
+
"grad_norm": 5.498723030090332,
|
| 3308 |
+
"learning_rate": 1.6736759662119183e-05,
|
| 3309 |
+
"loss": 8.3873,
|
| 3310 |
+
"step": 467
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"epoch": 0.8185395714910363,
|
| 3314 |
+
"grad_norm": 6.472886562347412,
|
| 3315 |
+
"learning_rate": 1.642847218816398e-05,
|
| 3316 |
+
"loss": 8.2182,
|
| 3317 |
+
"step": 468
|
| 3318 |
+
},
|
| 3319 |
+
{
|
| 3320 |
+
"epoch": 0.8202885876694359,
|
| 3321 |
+
"grad_norm": 7.618271827697754,
|
| 3322 |
+
"learning_rate": 1.6122796179903354e-05,
|
| 3323 |
+
"loss": 8.3238,
|
| 3324 |
+
"step": 469
|
| 3325 |
+
},
|
| 3326 |
+
{
|
| 3327 |
+
"epoch": 0.8220376038478356,
|
| 3328 |
+
"grad_norm": 6.079611778259277,
|
| 3329 |
+
"learning_rate": 1.58197411891839e-05,
|
| 3330 |
+
"loss": 7.5876,
|
| 3331 |
+
"step": 470
|
| 3332 |
+
},
|
| 3333 |
+
{
|
| 3334 |
+
"epoch": 0.8237866200262353,
|
| 3335 |
+
"grad_norm": 6.017045497894287,
|
| 3336 |
+
"learning_rate": 1.5519316685949903e-05,
|
| 3337 |
+
"loss": 7.4218,
|
| 3338 |
+
"step": 471
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 0.8255356362046349,
|
| 3342 |
+
"grad_norm": 5.659122943878174,
|
| 3343 |
+
"learning_rate": 1.5221532057947419e-05,
|
| 3344 |
+
"loss": 8.3443,
|
| 3345 |
+
"step": 472
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 0.8272846523830345,
|
| 3349 |
+
"grad_norm": 7.566473484039307,
|
| 3350 |
+
"learning_rate": 1.4926396610431059e-05,
|
| 3351 |
+
"loss": 7.7792,
|
| 3352 |
+
"step": 473
|
| 3353 |
+
},
|
| 3354 |
+
{
|
| 3355 |
+
"epoch": 0.8290336685614342,
|
| 3356 |
+
"grad_norm": 6.2454729080200195,
|
| 3357 |
+
"learning_rate": 1.4633919565873033e-05,
|
| 3358 |
+
"loss": 8.3766,
|
| 3359 |
+
"step": 474
|
| 3360 |
+
},
|
| 3361 |
+
{
|
| 3362 |
+
"epoch": 0.8307826847398339,
|
| 3363 |
+
"grad_norm": 6.340610980987549,
|
| 3364 |
+
"learning_rate": 1.4344110063675142e-05,
|
| 3365 |
+
"loss": 7.4127,
|
| 3366 |
+
"step": 475
|
| 3367 |
+
},
|
| 3368 |
+
{
|
| 3369 |
+
"epoch": 0.8325317009182335,
|
| 3370 |
+
"grad_norm": 5.981843948364258,
|
| 3371 |
+
"learning_rate": 1.4056977159883012e-05,
|
| 3372 |
+
"loss": 7.7706,
|
| 3373 |
+
"step": 476
|
| 3374 |
+
},
|
| 3375 |
+
{
|
| 3376 |
+
"epoch": 0.8342807170966331,
|
| 3377 |
+
"grad_norm": 6.074410438537598,
|
| 3378 |
+
"learning_rate": 1.3772529826903269e-05,
|
| 3379 |
+
"loss": 7.0402,
|
| 3380 |
+
"step": 477
|
| 3381 |
+
},
|
| 3382 |
+
{
|
| 3383 |
+
"epoch": 0.8360297332750328,
|
| 3384 |
+
"grad_norm": 6.144327163696289,
|
| 3385 |
+
"learning_rate": 1.3490776953223105e-05,
|
| 3386 |
+
"loss": 7.6445,
|
| 3387 |
+
"step": 478
|
| 3388 |
+
},
|
| 3389 |
+
{
|
| 3390 |
+
"epoch": 0.8377787494534324,
|
| 3391 |
+
"grad_norm": 5.650998115539551,
|
| 3392 |
+
"learning_rate": 1.321172734313244e-05,
|
| 3393 |
+
"loss": 8.1351,
|
| 3394 |
+
"step": 479
|
| 3395 |
+
},
|
| 3396 |
+
{
|
| 3397 |
+
"epoch": 0.8395277656318321,
|
| 3398 |
+
"grad_norm": 6.38914680480957,
|
| 3399 |
+
"learning_rate": 1.2935389716448976e-05,
|
| 3400 |
+
"loss": 8.9728,
|
| 3401 |
+
"step": 480
|
| 3402 |
+
},
|
| 3403 |
+
{
|
| 3404 |
+
"epoch": 0.8412767818102318,
|
| 3405 |
+
"grad_norm": 6.161402225494385,
|
| 3406 |
+
"learning_rate": 1.2661772708245535e-05,
|
| 3407 |
+
"loss": 8.9174,
|
| 3408 |
+
"step": 481
|
| 3409 |
+
},
|
| 3410 |
+
{
|
| 3411 |
+
"epoch": 0.8430257979886314,
|
| 3412 |
+
"grad_norm": 5.9002685546875,
|
| 3413 |
+
"learning_rate": 1.23908848685804e-05,
|
| 3414 |
+
"loss": 6.2995,
|
| 3415 |
+
"step": 482
|
| 3416 |
+
},
|
| 3417 |
+
{
|
| 3418 |
+
"epoch": 0.844774814167031,
|
| 3419 |
+
"grad_norm": 5.791549205780029,
|
| 3420 |
+
"learning_rate": 1.2122734662229984e-05,
|
| 3421 |
+
"loss": 6.7385,
|
| 3422 |
+
"step": 483
|
| 3423 |
+
},
|
| 3424 |
+
{
|
| 3425 |
+
"epoch": 0.8465238303454307,
|
| 3426 |
+
"grad_norm": 7.234724521636963,
|
| 3427 |
+
"learning_rate": 1.1857330468424466e-05,
|
| 3428 |
+
"loss": 6.838,
|
| 3429 |
+
"step": 484
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 0.8482728465238304,
|
| 3433 |
+
"grad_norm": 5.775229454040527,
|
| 3434 |
+
"learning_rate": 1.1594680580585814e-05,
|
| 3435 |
+
"loss": 8.2034,
|
| 3436 |
+
"step": 485
|
| 3437 |
+
},
|
| 3438 |
+
{
|
| 3439 |
+
"epoch": 0.85002186270223,
|
| 3440 |
+
"grad_norm": 6.084395885467529,
|
| 3441 |
+
"learning_rate": 1.133479320606874e-05,
|
| 3442 |
+
"loss": 8.3378,
|
| 3443 |
+
"step": 486
|
| 3444 |
+
},
|
| 3445 |
+
{
|
| 3446 |
+
"epoch": 0.8517708788806296,
|
| 3447 |
+
"grad_norm": 6.647040843963623,
|
| 3448 |
+
"learning_rate": 1.1077676465904208e-05,
|
| 3449 |
+
"loss": 8.3666,
|
| 3450 |
+
"step": 487
|
| 3451 |
+
},
|
| 3452 |
+
{
|
| 3453 |
+
"epoch": 0.8535198950590293,
|
| 3454 |
+
"grad_norm": 6.034310340881348,
|
| 3455 |
+
"learning_rate": 1.082333839454559e-05,
|
| 3456 |
+
"loss": 8.4993,
|
| 3457 |
+
"step": 488
|
| 3458 |
+
},
|
| 3459 |
+
{
|
| 3460 |
+
"epoch": 0.855268911237429,
|
| 3461 |
+
"grad_norm": 6.046447277069092,
|
| 3462 |
+
"learning_rate": 1.0571786939617712e-05,
|
| 3463 |
+
"loss": 6.8808,
|
| 3464 |
+
"step": 489
|
| 3465 |
+
},
|
| 3466 |
+
{
|
| 3467 |
+
"epoch": 0.8570179274158286,
|
| 3468 |
+
"grad_norm": 7.073398590087891,
|
| 3469 |
+
"learning_rate": 1.0323029961668462e-05,
|
| 3470 |
+
"loss": 8.4804,
|
| 3471 |
+
"step": 490
|
| 3472 |
+
},
|
| 3473 |
+
{
|
| 3474 |
+
"epoch": 0.8587669435942282,
|
| 3475 |
+
"grad_norm": 7.022246837615967,
|
| 3476 |
+
"learning_rate": 1.0077075233923116e-05,
|
| 3477 |
+
"loss": 8.4708,
|
| 3478 |
+
"step": 491
|
| 3479 |
+
},
|
| 3480 |
+
{
|
| 3481 |
+
"epoch": 0.8605159597726278,
|
| 3482 |
+
"grad_norm": 6.369495868682861,
|
| 3483 |
+
"learning_rate": 9.833930442041506e-06,
|
| 3484 |
+
"loss": 8.9675,
|
| 3485 |
+
"step": 492
|
| 3486 |
+
},
|
| 3487 |
+
{
|
| 3488 |
+
"epoch": 0.8622649759510276,
|
| 3489 |
+
"grad_norm": 7.497471809387207,
|
| 3490 |
+
"learning_rate": 9.593603183877841e-06,
|
| 3491 |
+
"loss": 9.2559,
|
| 3492 |
+
"step": 493
|
| 3493 |
+
},
|
| 3494 |
+
{
|
| 3495 |
+
"epoch": 0.8640139921294272,
|
| 3496 |
+
"grad_norm": 6.291429042816162,
|
| 3497 |
+
"learning_rate": 9.35610096924323e-06,
|
| 3498 |
+
"loss": 8.7969,
|
| 3499 |
+
"step": 494
|
| 3500 |
+
},
|
| 3501 |
+
{
|
| 3502 |
+
"epoch": 0.8657630083078268,
|
| 3503 |
+
"grad_norm": 5.909082412719727,
|
| 3504 |
+
"learning_rate": 9.121431219671095e-06,
|
| 3505 |
+
"loss": 9.1292,
|
| 3506 |
+
"step": 495
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 0.8675120244862266,
|
| 3510 |
+
"grad_norm": 6.053621292114258,
|
| 3511 |
+
"learning_rate": 8.889601268185232e-06,
|
| 3512 |
+
"loss": 8.1841,
|
| 3513 |
+
"step": 496
|
| 3514 |
+
},
|
| 3515 |
+
{
|
| 3516 |
+
"epoch": 0.8692610406646262,
|
| 3517 |
+
"grad_norm": 7.89780855178833,
|
| 3518 |
+
"learning_rate": 8.660618359070604e-06,
|
| 3519 |
+
"loss": 8.2031,
|
| 3520 |
+
"step": 497
|
| 3521 |
+
},
|
| 3522 |
+
{
|
| 3523 |
+
"epoch": 0.8710100568430258,
|
| 3524 |
+
"grad_norm": 5.878389835357666,
|
| 3525 |
+
"learning_rate": 8.434489647647092e-06,
|
| 3526 |
+
"loss": 7.5146,
|
| 3527 |
+
"step": 498
|
| 3528 |
+
},
|
| 3529 |
+
{
|
| 3530 |
+
"epoch": 0.8727590730214254,
|
| 3531 |
+
"grad_norm": 6.576045513153076,
|
| 3532 |
+
"learning_rate": 8.211222200045788e-06,
|
| 3533 |
+
"loss": 6.9011,
|
| 3534 |
+
"step": 499
|
| 3535 |
+
},
|
| 3536 |
+
{
|
| 3537 |
+
"epoch": 0.8745080891998251,
|
| 3538 |
+
"grad_norm": 6.503655433654785,
|
| 3539 |
+
"learning_rate": 7.990822992988267e-06,
|
| 3540 |
+
"loss": 7.2335,
|
| 3541 |
+
"step": 500
|
| 3542 |
+
},
|
| 3543 |
+
{
|
| 3544 |
+
"epoch": 0.8762571053782248,
|
| 3545 |
+
"grad_norm": 6.5767998695373535,
|
| 3546 |
+
"learning_rate": 7.773298913568505e-06,
|
| 3547 |
+
"loss": 8.625,
|
| 3548 |
+
"step": 501
|
| 3549 |
+
},
|
| 3550 |
+
{
|
| 3551 |
+
"epoch": 0.8780061215566244,
|
| 3552 |
+
"grad_norm": 6.215456485748291,
|
| 3553 |
+
"learning_rate": 7.558656759037797e-06,
|
| 3554 |
+
"loss": 6.6789,
|
| 3555 |
+
"step": 502
|
| 3556 |
+
},
|
| 3557 |
+
{
|
| 3558 |
+
"epoch": 0.879755137735024,
|
| 3559 |
+
"grad_norm": 6.190433502197266,
|
| 3560 |
+
"learning_rate": 7.346903236592162e-06,
|
| 3561 |
+
"loss": 7.0194,
|
| 3562 |
+
"step": 503
|
| 3563 |
+
},
|
| 3564 |
+
{
|
| 3565 |
+
"epoch": 0.8815041539134237,
|
| 3566 |
+
"grad_norm": 6.697807788848877,
|
| 3567 |
+
"learning_rate": 7.13804496316296e-06,
|
| 3568 |
+
"loss": 7.5084,
|
| 3569 |
+
"step": 504
|
| 3570 |
+
},
|
| 3571 |
+
{
|
| 3572 |
+
"epoch": 0.8832531700918234,
|
| 3573 |
+
"grad_norm": 6.033225059509277,
|
| 3574 |
+
"learning_rate": 6.9320884652099406e-06,
|
| 3575 |
+
"loss": 7.5227,
|
| 3576 |
+
"step": 505
|
| 3577 |
+
},
|
| 3578 |
+
{
|
| 3579 |
+
"epoch": 0.885002186270223,
|
| 3580 |
+
"grad_norm": 6.665908336639404,
|
| 3581 |
+
"learning_rate": 6.729040178517454e-06,
|
| 3582 |
+
"loss": 7.7323,
|
| 3583 |
+
"step": 506
|
| 3584 |
+
},
|
| 3585 |
+
{
|
| 3586 |
+
"epoch": 0.8867512024486226,
|
| 3587 |
+
"grad_norm": 6.476802825927734,
|
| 3588 |
+
"learning_rate": 6.528906447993288e-06,
|
| 3589 |
+
"loss": 7.7431,
|
| 3590 |
+
"step": 507
|
| 3591 |
+
},
|
| 3592 |
+
{
|
| 3593 |
+
"epoch": 0.8885002186270223,
|
| 3594 |
+
"grad_norm": 5.98831844329834,
|
| 3595 |
+
"learning_rate": 6.331693527470306e-06,
|
| 3596 |
+
"loss": 7.3852,
|
| 3597 |
+
"step": 508
|
| 3598 |
+
},
|
| 3599 |
+
{
|
| 3600 |
+
"epoch": 0.890249234805422,
|
| 3601 |
+
"grad_norm": 6.885638236999512,
|
| 3602 |
+
"learning_rate": 6.137407579511212e-06,
|
| 3603 |
+
"loss": 7.667,
|
| 3604 |
+
"step": 509
|
| 3605 |
+
},
|
| 3606 |
+
{
|
| 3607 |
+
"epoch": 0.8919982509838216,
|
| 3608 |
+
"grad_norm": 7.12498140335083,
|
| 3609 |
+
"learning_rate": 5.946054675215784e-06,
|
| 3610 |
+
"loss": 8.1353,
|
| 3611 |
+
"step": 510
|
| 3612 |
+
},
|
| 3613 |
+
{
|
| 3614 |
+
"epoch": 0.8937472671622213,
|
| 3615 |
+
"grad_norm": 6.516235828399658,
|
| 3616 |
+
"learning_rate": 5.757640794031361e-06,
|
| 3617 |
+
"loss": 7.4876,
|
| 3618 |
+
"step": 511
|
| 3619 |
+
},
|
| 3620 |
+
{
|
| 3621 |
+
"epoch": 0.8954962833406209,
|
| 3622 |
+
"grad_norm": 6.076444149017334,
|
| 3623 |
+
"learning_rate": 5.572171823565797e-06,
|
| 3624 |
+
"loss": 9.0379,
|
| 3625 |
+
"step": 512
|
| 3626 |
+
},
|
| 3627 |
+
{
|
| 3628 |
+
"epoch": 0.8972452995190205,
|
| 3629 |
+
"grad_norm": 5.914377689361572,
|
| 3630 |
+
"learning_rate": 5.389653559403629e-06,
|
| 3631 |
+
"loss": 8.1439,
|
| 3632 |
+
"step": 513
|
| 3633 |
+
},
|
| 3634 |
+
{
|
| 3635 |
+
"epoch": 0.8989943156974202,
|
| 3636 |
+
"grad_norm": 6.191368579864502,
|
| 3637 |
+
"learning_rate": 5.210091704924946e-06,
|
| 3638 |
+
"loss": 8.6825,
|
| 3639 |
+
"step": 514
|
| 3640 |
+
},
|
| 3641 |
+
{
|
| 3642 |
+
"epoch": 0.9007433318758199,
|
| 3643 |
+
"grad_norm": 6.621984958648682,
|
| 3644 |
+
"learning_rate": 5.033491871127105e-06,
|
| 3645 |
+
"loss": 8.3616,
|
| 3646 |
+
"step": 515
|
| 3647 |
+
},
|
| 3648 |
+
{
|
| 3649 |
+
"epoch": 0.9024923480542195,
|
| 3650 |
+
"grad_norm": 6.113316059112549,
|
| 3651 |
+
"learning_rate": 4.859859576449444e-06,
|
| 3652 |
+
"loss": 7.6225,
|
| 3653 |
+
"step": 516
|
| 3654 |
+
},
|
| 3655 |
+
{
|
| 3656 |
+
"epoch": 0.9042413642326191,
|
| 3657 |
+
"grad_norm": 6.2419915199279785,
|
| 3658 |
+
"learning_rate": 4.689200246600867e-06,
|
| 3659 |
+
"loss": 7.8226,
|
| 3660 |
+
"step": 517
|
| 3661 |
+
},
|
| 3662 |
+
{
|
| 3663 |
+
"epoch": 0.9059903804110188,
|
| 3664 |
+
"grad_norm": 6.267433166503906,
|
| 3665 |
+
"learning_rate": 4.521519214390257e-06,
|
| 3666 |
+
"loss": 7.5588,
|
| 3667 |
+
"step": 518
|
| 3668 |
+
},
|
| 3669 |
+
{
|
| 3670 |
+
"epoch": 0.9077393965894185,
|
| 3671 |
+
"grad_norm": 7.470646381378174,
|
| 3672 |
+
"learning_rate": 4.356821719559812e-06,
|
| 3673 |
+
"loss": 7.5397,
|
| 3674 |
+
"step": 519
|
| 3675 |
+
},
|
| 3676 |
+
{
|
| 3677 |
+
"epoch": 0.9094884127678181,
|
| 3678 |
+
"grad_norm": 5.909048080444336,
|
| 3679 |
+
"learning_rate": 4.195112908621402e-06,
|
| 3680 |
+
"loss": 8.8447,
|
| 3681 |
+
"step": 520
|
| 3682 |
+
},
|
| 3683 |
+
{
|
| 3684 |
+
"epoch": 0.9112374289462177,
|
| 3685 |
+
"grad_norm": 6.860408782958984,
|
| 3686 |
+
"learning_rate": 4.03639783469566e-06,
|
| 3687 |
+
"loss": 8.8253,
|
| 3688 |
+
"step": 521
|
| 3689 |
+
},
|
| 3690 |
+
{
|
| 3691 |
+
"epoch": 0.9129864451246174,
|
| 3692 |
+
"grad_norm": 7.22110652923584,
|
| 3693 |
+
"learning_rate": 3.880681457354118e-06,
|
| 3694 |
+
"loss": 7.9479,
|
| 3695 |
+
"step": 522
|
| 3696 |
+
},
|
| 3697 |
+
{
|
| 3698 |
+
"epoch": 0.9147354613030171,
|
| 3699 |
+
"grad_norm": 5.659202575683594,
|
| 3700 |
+
"learning_rate": 3.727968642464241e-06,
|
| 3701 |
+
"loss": 7.4659,
|
| 3702 |
+
"step": 523
|
| 3703 |
+
},
|
| 3704 |
+
{
|
| 3705 |
+
"epoch": 0.9164844774814167,
|
| 3706 |
+
"grad_norm": 7.78839111328125,
|
| 3707 |
+
"learning_rate": 3.578264162037348e-06,
|
| 3708 |
+
"loss": 7.8924,
|
| 3709 |
+
"step": 524
|
| 3710 |
+
},
|
| 3711 |
+
{
|
| 3712 |
+
"epoch": 0.9182334936598163,
|
| 3713 |
+
"grad_norm": 6.77100133895874,
|
| 3714 |
+
"learning_rate": 3.4315726940795433e-06,
|
| 3715 |
+
"loss": 8.6822,
|
| 3716 |
+
"step": 525
|
| 3717 |
+
},
|
| 3718 |
+
{
|
| 3719 |
+
"epoch": 0.9199825098382161,
|
| 3720 |
+
"grad_norm": 7.059048175811768,
|
| 3721 |
+
"learning_rate": 3.2878988224454344e-06,
|
| 3722 |
+
"loss": 8.3176,
|
| 3723 |
+
"step": 526
|
| 3724 |
+
},
|
| 3725 |
+
{
|
| 3726 |
+
"epoch": 0.9217315260166157,
|
| 3727 |
+
"grad_norm": 5.931784152984619,
|
| 3728 |
+
"learning_rate": 3.1472470366950334e-06,
|
| 3729 |
+
"loss": 7.5494,
|
| 3730 |
+
"step": 527
|
| 3731 |
+
},
|
| 3732 |
+
{
|
| 3733 |
+
"epoch": 0.9234805421950153,
|
| 3734 |
+
"grad_norm": 6.578647613525391,
|
| 3735 |
+
"learning_rate": 3.0096217319533382e-06,
|
| 3736 |
+
"loss": 8.5582,
|
| 3737 |
+
"step": 528
|
| 3738 |
+
},
|
| 3739 |
+
{
|
| 3740 |
+
"epoch": 0.9252295583734149,
|
| 3741 |
+
"grad_norm": 5.951269626617432,
|
| 3742 |
+
"learning_rate": 2.875027208773118e-06,
|
| 3743 |
+
"loss": 7.3472,
|
| 3744 |
+
"step": 529
|
| 3745 |
+
},
|
| 3746 |
+
{
|
| 3747 |
+
"epoch": 0.9269785745518146,
|
| 3748 |
+
"grad_norm": 6.003902912139893,
|
| 3749 |
+
"learning_rate": 2.7434676730003884e-06,
|
| 3750 |
+
"loss": 8.754,
|
| 3751 |
+
"step": 530
|
| 3752 |
+
},
|
| 3753 |
+
{
|
| 3754 |
+
"epoch": 0.9287275907302143,
|
| 3755 |
+
"grad_norm": 6.373345851898193,
|
| 3756 |
+
"learning_rate": 2.614947235643106e-06,
|
| 3757 |
+
"loss": 7.5758,
|
| 3758 |
+
"step": 531
|
| 3759 |
+
},
|
| 3760 |
+
{
|
| 3761 |
+
"epoch": 0.9304766069086139,
|
| 3762 |
+
"grad_norm": 6.780086517333984,
|
| 3763 |
+
"learning_rate": 2.4894699127426367e-06,
|
| 3764 |
+
"loss": 9.3402,
|
| 3765 |
+
"step": 532
|
| 3766 |
+
},
|
| 3767 |
+
{
|
| 3768 |
+
"epoch": 0.9322256230870135,
|
| 3769 |
+
"grad_norm": 6.614811897277832,
|
| 3770 |
+
"learning_rate": 2.367039625248302e-06,
|
| 3771 |
+
"loss": 7.8778,
|
| 3772 |
+
"step": 533
|
| 3773 |
+
},
|
| 3774 |
+
{
|
| 3775 |
+
"epoch": 0.9339746392654132,
|
| 3776 |
+
"grad_norm": 6.098567008972168,
|
| 3777 |
+
"learning_rate": 2.2476601988947966e-06,
|
| 3778 |
+
"loss": 7.928,
|
| 3779 |
+
"step": 534
|
| 3780 |
+
},
|
| 3781 |
+
{
|
| 3782 |
+
"epoch": 0.9357236554438129,
|
| 3783 |
+
"grad_norm": 5.995659828186035,
|
| 3784 |
+
"learning_rate": 2.1313353640827206e-06,
|
| 3785 |
+
"loss": 9.0382,
|
| 3786 |
+
"step": 535
|
| 3787 |
+
},
|
| 3788 |
+
{
|
| 3789 |
+
"epoch": 0.9374726716222125,
|
| 3790 |
+
"grad_norm": 7.359274387359619,
|
| 3791 |
+
"learning_rate": 2.0180687557619816e-06,
|
| 3792 |
+
"loss": 8.395,
|
| 3793 |
+
"step": 536
|
| 3794 |
+
},
|
| 3795 |
+
{
|
| 3796 |
+
"epoch": 0.9392216878006121,
|
| 3797 |
+
"grad_norm": 6.481335639953613,
|
| 3798 |
+
"learning_rate": 1.907863913318153e-06,
|
| 3799 |
+
"loss": 8.3956,
|
| 3800 |
+
"step": 537
|
| 3801 |
+
},
|
| 3802 |
+
{
|
| 3803 |
+
"epoch": 0.9409707039790118,
|
| 3804 |
+
"grad_norm": 5.920598983764648,
|
| 3805 |
+
"learning_rate": 1.8007242804619628e-06,
|
| 3806 |
+
"loss": 7.8273,
|
| 3807 |
+
"step": 538
|
| 3808 |
+
},
|
| 3809 |
+
{
|
| 3810 |
+
"epoch": 0.9427197201574115,
|
| 3811 |
+
"grad_norm": 6.889771461486816,
|
| 3812 |
+
"learning_rate": 1.696653205121612e-06,
|
| 3813 |
+
"loss": 7.9301,
|
| 3814 |
+
"step": 539
|
| 3815 |
+
},
|
| 3816 |
+
{
|
| 3817 |
+
"epoch": 0.9444687363358111,
|
| 3818 |
+
"grad_norm": 7.978548049926758,
|
| 3819 |
+
"learning_rate": 1.595653939338204e-06,
|
| 3820 |
+
"loss": 8.1209,
|
| 3821 |
+
"step": 540
|
| 3822 |
+
},
|
| 3823 |
+
{
|
| 3824 |
+
"epoch": 0.9462177525142108,
|
| 3825 |
+
"grad_norm": 6.091990947723389,
|
| 3826 |
+
"learning_rate": 1.4977296391641026e-06,
|
| 3827 |
+
"loss": 7.4813,
|
| 3828 |
+
"step": 541
|
| 3829 |
+
},
|
| 3830 |
+
{
|
| 3831 |
+
"epoch": 0.9479667686926104,
|
| 3832 |
+
"grad_norm": 5.8564372062683105,
|
| 3833 |
+
"learning_rate": 1.4028833645643113e-06,
|
| 3834 |
+
"loss": 8.299,
|
| 3835 |
+
"step": 542
|
| 3836 |
+
},
|
| 3837 |
+
{
|
| 3838 |
+
"epoch": 0.94971578487101,
|
| 3839 |
+
"grad_norm": 6.623847484588623,
|
| 3840 |
+
"learning_rate": 1.31111807932085e-06,
|
| 3841 |
+
"loss": 7.3619,
|
| 3842 |
+
"step": 543
|
| 3843 |
+
},
|
| 3844 |
+
{
|
| 3845 |
+
"epoch": 0.9514648010494097,
|
| 3846 |
+
"grad_norm": 6.130828380584717,
|
| 3847 |
+
"learning_rate": 1.222436650940173e-06,
|
| 3848 |
+
"loss": 8.2312,
|
| 3849 |
+
"step": 544
|
| 3850 |
+
},
|
| 3851 |
+
{
|
| 3852 |
+
"epoch": 0.9532138172278094,
|
| 3853 |
+
"grad_norm": 5.993081569671631,
|
| 3854 |
+
"learning_rate": 1.1368418505635302e-06,
|
| 3855 |
+
"loss": 9.4749,
|
| 3856 |
+
"step": 545
|
| 3857 |
+
},
|
| 3858 |
+
{
|
| 3859 |
+
"epoch": 0.954962833406209,
|
| 3860 |
+
"grad_norm": 6.714028835296631,
|
| 3861 |
+
"learning_rate": 1.0543363528803696e-06,
|
| 3862 |
+
"loss": 8.6106,
|
| 3863 |
+
"step": 546
|
| 3864 |
+
},
|
| 3865 |
+
{
|
| 3866 |
+
"epoch": 0.9567118495846086,
|
| 3867 |
+
"grad_norm": 6.759970664978027,
|
| 3868 |
+
"learning_rate": 9.749227360448143e-07,
|
| 3869 |
+
"loss": 8.9938,
|
| 3870 |
+
"step": 547
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 0.9584608657630083,
|
| 3874 |
+
"grad_norm": 6.770102500915527,
|
| 3875 |
+
"learning_rate": 8.986034815950172e-07,
|
| 3876 |
+
"loss": 9.3766,
|
| 3877 |
+
"step": 548
|
| 3878 |
+
},
|
| 3879 |
+
{
|
| 3880 |
+
"epoch": 0.960209881941408,
|
| 3881 |
+
"grad_norm": 5.988943576812744,
|
| 3882 |
+
"learning_rate": 8.253809743756668e-07,
|
| 3883 |
+
"loss": 8.0326,
|
| 3884 |
+
"step": 549
|
| 3885 |
+
},
|
| 3886 |
+
{
|
| 3887 |
+
"epoch": 0.9619588981198076,
|
| 3888 |
+
"grad_norm": 6.561304092407227,
|
| 3889 |
+
"learning_rate": 7.552575024634689e-07,
|
| 3890 |
+
"loss": 7.3145,
|
| 3891 |
+
"step": 550
|
| 3892 |
+
},
|
| 3893 |
+
{
|
| 3894 |
+
"epoch": 0.9637079142982072,
|
| 3895 |
+
"grad_norm": 6.857665061950684,
|
| 3896 |
+
"learning_rate": 6.882352570956485e-07,
|
| 3897 |
+
"loss": 8.4195,
|
| 3898 |
+
"step": 551
|
| 3899 |
+
},
|
| 3900 |
+
{
|
| 3901 |
+
"epoch": 0.9654569304766069,
|
| 3902 |
+
"grad_norm": 7.414341926574707,
|
| 3903 |
+
"learning_rate": 6.243163326014267e-07,
|
| 3904 |
+
"loss": 7.5319,
|
| 3905 |
+
"step": 552
|
| 3906 |
+
},
|
| 3907 |
+
{
|
| 3908 |
+
"epoch": 0.9672059466550066,
|
| 3909 |
+
"grad_norm": 7.155992031097412,
|
| 3910 |
+
"learning_rate": 5.635027263366399e-07,
|
| 3911 |
+
"loss": 7.9228,
|
| 3912 |
+
"step": 553
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 0.9689549628334062,
|
| 3916 |
+
"grad_norm": 5.8317551612854,
|
| 3917 |
+
"learning_rate": 5.057963386213116e-07,
|
| 3918 |
+
"loss": 8.3518,
|
| 3919 |
+
"step": 554
|
| 3920 |
+
},
|
| 3921 |
+
{
|
| 3922 |
+
"epoch": 0.9707039790118058,
|
| 3923 |
+
"grad_norm": 6.625657081604004,
|
| 3924 |
+
"learning_rate": 4.5119897268023347e-07,
|
| 3925 |
+
"loss": 8.2842,
|
| 3926 |
+
"step": 555
|
| 3927 |
+
},
|
| 3928 |
+
{
|
| 3929 |
+
"epoch": 0.9724529951902056,
|
| 3930 |
+
"grad_norm": 6.393413543701172,
|
| 3931 |
+
"learning_rate": 3.9971233458665493e-07,
|
| 3932 |
+
"loss": 8.9058,
|
| 3933 |
+
"step": 556
|
| 3934 |
+
},
|
| 3935 |
+
{
|
| 3936 |
+
"epoch": 0.9742020113686052,
|
| 3937 |
+
"grad_norm": 6.962483882904053,
|
| 3938 |
+
"learning_rate": 3.5133803320896994e-07,
|
| 3939 |
+
"loss": 7.8907,
|
| 3940 |
+
"step": 557
|
| 3941 |
+
},
|
| 3942 |
+
{
|
| 3943 |
+
"epoch": 0.9759510275470048,
|
| 3944 |
+
"grad_norm": 6.612078666687012,
|
| 3945 |
+
"learning_rate": 3.060775801604354e-07,
|
| 3946 |
+
"loss": 8.0385,
|
| 3947 |
+
"step": 558
|
| 3948 |
+
},
|
| 3949 |
+
{
|
| 3950 |
+
"epoch": 0.9777000437254044,
|
| 3951 |
+
"grad_norm": 5.987767696380615,
|
| 3952 |
+
"learning_rate": 2.639323897518975e-07,
|
| 3953 |
+
"loss": 8.6059,
|
| 3954 |
+
"step": 559
|
| 3955 |
+
},
|
| 3956 |
+
{
|
| 3957 |
+
"epoch": 0.9794490599038042,
|
| 3958 |
+
"grad_norm": 5.532016277313232,
|
| 3959 |
+
"learning_rate": 2.2490377894768267e-07,
|
| 3960 |
+
"loss": 7.1836,
|
| 3961 |
+
"step": 560
|
| 3962 |
+
},
|
| 3963 |
+
{
|
| 3964 |
+
"epoch": 0.9811980760822038,
|
| 3965 |
+
"grad_norm": 6.169713020324707,
|
| 3966 |
+
"learning_rate": 1.889929673243529e-07,
|
| 3967 |
+
"loss": 7.9154,
|
| 3968 |
+
"step": 561
|
| 3969 |
+
},
|
| 3970 |
+
{
|
| 3971 |
+
"epoch": 0.9829470922606034,
|
| 3972 |
+
"grad_norm": 6.610177040100098,
|
| 3973 |
+
"learning_rate": 1.562010770326916e-07,
|
| 3974 |
+
"loss": 7.9639,
|
| 3975 |
+
"step": 562
|
| 3976 |
+
},
|
| 3977 |
+
{
|
| 3978 |
+
"epoch": 0.984696108439003,
|
| 3979 |
+
"grad_norm": 5.848151683807373,
|
| 3980 |
+
"learning_rate": 1.2652913276250955e-07,
|
| 3981 |
+
"loss": 7.69,
|
| 3982 |
+
"step": 563
|
| 3983 |
+
},
|
| 3984 |
+
{
|
| 3985 |
+
"epoch": 0.9864451246174027,
|
| 3986 |
+
"grad_norm": 6.159895420074463,
|
| 3987 |
+
"learning_rate": 9.99780617107815e-08,
|
| 3988 |
+
"loss": 7.9916,
|
| 3989 |
+
"step": 564
|
| 3990 |
+
},
|
| 3991 |
+
{
|
| 3992 |
+
"epoch": 0.9881941407958024,
|
| 3993 |
+
"grad_norm": 6.261981964111328,
|
| 3994 |
+
"learning_rate": 7.654869355252504e-08,
|
| 3995 |
+
"loss": 8.4931,
|
| 3996 |
+
"step": 565
|
| 3997 |
+
},
|
| 3998 |
+
{
|
| 3999 |
+
"epoch": 0.989943156974202,
|
| 4000 |
+
"grad_norm": 6.078785419464111,
|
| 4001 |
+
"learning_rate": 5.6241760414987856e-08,
|
| 4002 |
+
"loss": 9.2733,
|
| 4003 |
+
"step": 566
|
| 4004 |
+
},
|
| 4005 |
+
{
|
| 4006 |
+
"epoch": 0.9916921731526016,
|
| 4007 |
+
"grad_norm": 6.701304912567139,
|
| 4008 |
+
"learning_rate": 3.905789685471062e-08,
|
| 4009 |
+
"loss": 8.1601,
|
| 4010 |
+
"step": 567
|
| 4011 |
+
},
|
| 4012 |
+
{
|
| 4013 |
+
"epoch": 0.9934411893310013,
|
| 4014 |
+
"grad_norm": 6.677995681762695,
|
| 4015 |
+
"learning_rate": 2.4997639837687213e-08,
|
| 4016 |
+
"loss": 7.0004,
|
| 4017 |
+
"step": 568
|
| 4018 |
+
},
|
| 4019 |
+
{
|
| 4020 |
+
"epoch": 0.995190205509401,
|
| 4021 |
+
"grad_norm": 6.922513484954834,
|
| 4022 |
+
"learning_rate": 1.4061428722633718e-08,
|
| 4023 |
+
"loss": 8.7729,
|
| 4024 |
+
"step": 569
|
| 4025 |
+
},
|
| 4026 |
+
{
|
| 4027 |
+
"epoch": 0.9969392216878006,
|
| 4028 |
+
"grad_norm": 6.470855236053467,
|
| 4029 |
+
"learning_rate": 6.2496052472549304e-09,
|
| 4030 |
+
"loss": 8.6932,
|
| 4031 |
+
"step": 570
|
| 4032 |
+
},
|
| 4033 |
+
{
|
| 4034 |
+
"epoch": 0.9986882378662003,
|
| 4035 |
+
"grad_norm": 6.063665390014648,
|
| 4036 |
+
"learning_rate": 1.5624135174974186e-09,
|
| 4037 |
+
"loss": 8.2856,
|
| 4038 |
+
"step": 571
|
| 4039 |
+
},
|
| 4040 |
+
{
|
| 4041 |
+
"epoch": 1.0008745080891999,
|
| 4042 |
+
"grad_norm": 6.530557155609131,
|
| 4043 |
+
"learning_rate": 0.0,
|
| 4044 |
+
"loss": 8.5752,
|
| 4045 |
+
"step": 572
|
| 4046 |
+
},
|
| 4047 |
+
{
|
| 4048 |
+
"epoch": 1.0008745080891999,
|
| 4049 |
+
"eval_loss": 1.9150428771972656,
|
| 4050 |
+
"eval_runtime": 2.1677,
|
| 4051 |
+
"eval_samples_per_second": 111.18,
|
| 4052 |
+
"eval_steps_per_second": 55.821,
|
| 4053 |
+
"step": 572
|
| 4054 |
}
|
| 4055 |
],
|
| 4056 |
"logging_steps": 1,
|
|
|
|
| 4065 |
"should_evaluate": false,
|
| 4066 |
"should_log": false,
|
| 4067 |
"should_save": true,
|
| 4068 |
+
"should_training_stop": true
|
| 4069 |
},
|
| 4070 |
"attributes": {}
|
| 4071 |
}
|
| 4072 |
},
|
| 4073 |
+
"total_flos": 2428656131506176.0,
|
| 4074 |
"train_batch_size": 2,
|
| 4075 |
"trial_name": null,
|
| 4076 |
"trial_params": null
|