Commit
•
7eb9800
1
Parent(s):
53f3594
Model save
Browse files- README.md +3 -3
- all_results.json +10 -10
- eval_results.json +6 -6
- model.safetensors +1 -1
- tokenizer_config.json +1 -1
- train_results.json +5 -6
- trainer_state.json +2454 -54
README.md
CHANGED
|
@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
This model was trained from scratch on the None dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
-
- Loss:
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
@@ -34,14 +34,14 @@ More information needed
|
|
| 34 |
### Training hyperparameters
|
| 35 |
|
| 36 |
The following hyperparameters were used during training:
|
| 37 |
-
- learning_rate: 5e-
|
| 38 |
- train_batch_size: 2
|
| 39 |
- eval_batch_size: 2
|
| 40 |
- seed: 0
|
| 41 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 42 |
- lr_scheduler_type: cosine
|
| 43 |
- lr_scheduler_warmup_ratio: 0.2
|
| 44 |
-
- num_epochs:
|
| 45 |
|
| 46 |
### Training results
|
| 47 |
|
|
|
|
| 15 |
|
| 16 |
This model was trained from scratch on the None dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 11.5514
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 34 |
### Training hyperparameters
|
| 35 |
|
| 36 |
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-07
|
| 38 |
- train_batch_size: 2
|
| 39 |
- eval_batch_size: 2
|
| 40 |
- seed: 0
|
| 41 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 42 |
- lr_scheduler_type: cosine
|
| 43 |
- lr_scheduler_warmup_ratio: 0.2
|
| 44 |
+
- num_epochs: 3
|
| 45 |
|
| 46 |
### Training results
|
| 47 |
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss":
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples":
|
| 6 |
-
"eval_samples_per_second": 12.
|
| 7 |
-
"eval_steps_per_second": 6.
|
| 8 |
"total_flos": 8536543131303936.0,
|
| 9 |
-
"train_loss":
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second": 2.
|
| 12 |
-
"train_steps_per_second": 1.
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 11.551432609558105,
|
| 4 |
+
"eval_runtime": 50.9651,
|
| 5 |
+
"eval_samples": 615,
|
| 6 |
+
"eval_samples_per_second": 12.067,
|
| 7 |
+
"eval_steps_per_second": 6.043,
|
| 8 |
"total_flos": 8536543131303936.0,
|
| 9 |
+
"train_loss": 12.036643933186408,
|
| 10 |
+
"train_runtime": 6501.8144,
|
| 11 |
+
"train_samples_per_second": 2.552,
|
| 12 |
+
"train_steps_per_second": 1.276
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_loss":
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples":
|
| 6 |
-
"eval_samples_per_second": 12.
|
| 7 |
-
"eval_steps_per_second": 6.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_loss": 11.551432609558105,
|
| 4 |
+
"eval_runtime": 50.9651,
|
| 5 |
+
"eval_samples": 615,
|
| 6 |
+
"eval_samples_per_second": 12.067,
|
| 7 |
+
"eval_steps_per_second": 6.043
|
| 8 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4018095352
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b2f21cfaa81376695a587fc99b7aa0b43466336b0fc712a35eb98abf758182c
|
| 3 |
size 4018095352
|
tokenizer_config.json
CHANGED
|
@@ -122,7 +122,7 @@
|
|
| 122 |
"legacy": false,
|
| 123 |
"model_max_length": 4096,
|
| 124 |
"pad_token": "<unk>",
|
| 125 |
-
"padding_side": "
|
| 126 |
"sp_model_kwargs": {},
|
| 127 |
"tokenizer_class": "LlamaTokenizer",
|
| 128 |
"unk_token": "<unk>",
|
|
|
|
| 122 |
"legacy": false,
|
| 123 |
"model_max_length": 4096,
|
| 124 |
"pad_token": "<unk>",
|
| 125 |
+
"padding_side": "left",
|
| 126 |
"sp_model_kwargs": {},
|
| 127 |
"tokenizer_class": "LlamaTokenizer",
|
| 128 |
"unk_token": "<unk>",
|
train_results.json
CHANGED
|
@@ -1,8 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"
|
| 4 |
-
"
|
| 5 |
-
"
|
| 6 |
-
"
|
| 7 |
-
"train_steps_per_second": 1.281
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"train_loss": 12.036643933186408,
|
| 4 |
+
"train_runtime": 6501.8144,
|
| 5 |
+
"train_samples_per_second": 2.552,
|
| 6 |
+
"train_steps_per_second": 1.276
|
|
|
|
| 7 |
}
|
trainer_state.json
CHANGED
|
@@ -1,113 +1,2513 @@
|
|
| 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,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"loss": 13.
|
| 15 |
"step": 20
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
-
"learning_rate":
|
| 20 |
-
"loss":
|
| 21 |
"step": 40
|
| 22 |
},
|
| 23 |
{
|
| 24 |
-
"epoch": 0.
|
| 25 |
-
"learning_rate":
|
| 26 |
-
"loss":
|
| 27 |
"step": 60
|
| 28 |
},
|
| 29 |
{
|
| 30 |
-
"epoch": 0.
|
| 31 |
-
"learning_rate":
|
| 32 |
-
"loss":
|
| 33 |
"step": 80
|
| 34 |
},
|
| 35 |
{
|
| 36 |
-
"epoch": 0.
|
| 37 |
-
"learning_rate":
|
| 38 |
-
"loss":
|
| 39 |
"step": 100
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch": 0.
|
| 43 |
-
"learning_rate":
|
| 44 |
-
"loss":
|
| 45 |
"step": 120
|
| 46 |
},
|
| 47 |
{
|
| 48 |
-
"epoch": 0.
|
| 49 |
-
"learning_rate":
|
| 50 |
-
"loss":
|
| 51 |
"step": 140
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate":
|
| 56 |
-
"loss":
|
| 57 |
"step": 160
|
| 58 |
},
|
| 59 |
{
|
| 60 |
-
"epoch": 0.
|
| 61 |
-
"learning_rate":
|
| 62 |
-
"loss":
|
| 63 |
"step": 180
|
| 64 |
},
|
| 65 |
{
|
| 66 |
-
"epoch": 0.
|
| 67 |
-
"learning_rate":
|
| 68 |
-
"loss":
|
| 69 |
"step": 200
|
| 70 |
},
|
| 71 |
{
|
| 72 |
-
"epoch": 0.
|
| 73 |
-
"learning_rate":
|
| 74 |
-
"loss":
|
| 75 |
"step": 220
|
| 76 |
},
|
| 77 |
{
|
| 78 |
-
"epoch": 0.
|
| 79 |
-
"learning_rate":
|
| 80 |
-
"loss":
|
| 81 |
"step": 240
|
| 82 |
},
|
| 83 |
{
|
| 84 |
-
"epoch": 0.
|
| 85 |
-
"learning_rate":
|
| 86 |
-
"loss":
|
| 87 |
"step": 260
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch": 0.
|
| 91 |
-
"learning_rate":
|
| 92 |
-
"loss":
|
| 93 |
"step": 280
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch":
|
| 97 |
-
"
|
| 98 |
-
"
|
| 99 |
-
"
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 103 |
}
|
| 104 |
],
|
| 105 |
"logging_steps": 20,
|
| 106 |
-
"max_steps":
|
| 107 |
"num_input_tokens_seen": 0,
|
| 108 |
-
"num_train_epochs":
|
| 109 |
"save_steps": 100,
|
| 110 |
-
"total_flos":
|
| 111 |
"train_batch_size": 2,
|
| 112 |
"trial_name": null,
|
| 113 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 8295,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"learning_rate": 6.027727546714888e-09,
|
| 14 |
+
"loss": 13.638,
|
| 15 |
"step": 20
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 1.2055455093429776e-08,
|
| 20 |
+
"loss": 13.7323,
|
| 21 |
"step": 40
|
| 22 |
},
|
| 23 |
{
|
| 24 |
+
"epoch": 0.02,
|
| 25 |
+
"learning_rate": 1.8083182640144666e-08,
|
| 26 |
+
"loss": 13.5875,
|
| 27 |
"step": 60
|
| 28 |
},
|
| 29 |
{
|
| 30 |
+
"epoch": 0.03,
|
| 31 |
+
"learning_rate": 2.4110910186859552e-08,
|
| 32 |
+
"loss": 13.6351,
|
| 33 |
"step": 80
|
| 34 |
},
|
| 35 |
{
|
| 36 |
+
"epoch": 0.04,
|
| 37 |
+
"learning_rate": 3.013863773357444e-08,
|
| 38 |
+
"loss": 13.634,
|
| 39 |
"step": 100
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 0.04,
|
| 43 |
+
"learning_rate": 3.616636528028933e-08,
|
| 44 |
+
"loss": 13.6684,
|
| 45 |
"step": 120
|
| 46 |
},
|
| 47 |
{
|
| 48 |
+
"epoch": 0.05,
|
| 49 |
+
"learning_rate": 4.2194092827004215e-08,
|
| 50 |
+
"loss": 13.6572,
|
| 51 |
"step": 140
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.06,
|
| 55 |
+
"learning_rate": 4.8221820373719104e-08,
|
| 56 |
+
"loss": 13.6528,
|
| 57 |
"step": 160
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.07,
|
| 61 |
+
"learning_rate": 5.4249547920433994e-08,
|
| 62 |
+
"loss": 13.7057,
|
| 63 |
"step": 180
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.07,
|
| 67 |
+
"learning_rate": 6.027727546714888e-08,
|
| 68 |
+
"loss": 13.6378,
|
| 69 |
"step": 200
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.08,
|
| 73 |
+
"learning_rate": 6.630500301386377e-08,
|
| 74 |
+
"loss": 13.7237,
|
| 75 |
"step": 220
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.09,
|
| 79 |
+
"learning_rate": 7.233273056057866e-08,
|
| 80 |
+
"loss": 13.7124,
|
| 81 |
"step": 240
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 0.09,
|
| 85 |
+
"learning_rate": 7.836045810729356e-08,
|
| 86 |
+
"loss": 13.6658,
|
| 87 |
"step": 260
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.1,
|
| 91 |
+
"learning_rate": 8.438818565400843e-08,
|
| 92 |
+
"loss": 13.6358,
|
| 93 |
"step": 280
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.11,
|
| 97 |
+
"learning_rate": 9.041591320072333e-08,
|
| 98 |
+
"loss": 13.6806,
|
| 99 |
+
"step": 300
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.12,
|
| 103 |
+
"learning_rate": 9.644364074743821e-08,
|
| 104 |
+
"loss": 13.7341,
|
| 105 |
+
"step": 320
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.12,
|
| 109 |
+
"learning_rate": 1.0247136829415309e-07,
|
| 110 |
+
"loss": 13.6129,
|
| 111 |
+
"step": 340
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.13,
|
| 115 |
+
"learning_rate": 1.0849909584086799e-07,
|
| 116 |
+
"loss": 13.6649,
|
| 117 |
+
"step": 360
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.14,
|
| 121 |
+
"learning_rate": 1.1452682338758287e-07,
|
| 122 |
+
"loss": 13.6915,
|
| 123 |
+
"step": 380
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.14,
|
| 127 |
+
"learning_rate": 1.2055455093429777e-07,
|
| 128 |
+
"loss": 13.7191,
|
| 129 |
+
"step": 400
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.15,
|
| 133 |
+
"learning_rate": 1.2658227848101266e-07,
|
| 134 |
+
"loss": 13.5871,
|
| 135 |
+
"step": 420
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.16,
|
| 139 |
+
"learning_rate": 1.3261000602772753e-07,
|
| 140 |
+
"loss": 13.6615,
|
| 141 |
+
"step": 440
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.17,
|
| 145 |
+
"learning_rate": 1.3863773357444243e-07,
|
| 146 |
+
"loss": 13.5988,
|
| 147 |
+
"step": 460
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.17,
|
| 151 |
+
"learning_rate": 1.4466546112115733e-07,
|
| 152 |
+
"loss": 13.5958,
|
| 153 |
+
"step": 480
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.18,
|
| 157 |
+
"learning_rate": 1.506931886678722e-07,
|
| 158 |
+
"loss": 13.6678,
|
| 159 |
+
"step": 500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.19,
|
| 163 |
+
"learning_rate": 1.5672091621458712e-07,
|
| 164 |
+
"loss": 13.6995,
|
| 165 |
+
"step": 520
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.2,
|
| 169 |
+
"learning_rate": 1.62748643761302e-07,
|
| 170 |
+
"loss": 13.6663,
|
| 171 |
+
"step": 540
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.2,
|
| 175 |
+
"learning_rate": 1.6877637130801686e-07,
|
| 176 |
+
"loss": 13.659,
|
| 177 |
+
"step": 560
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.21,
|
| 181 |
+
"learning_rate": 1.7480409885473178e-07,
|
| 182 |
+
"loss": 13.554,
|
| 183 |
+
"step": 580
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.22,
|
| 187 |
+
"learning_rate": 1.8083182640144665e-07,
|
| 188 |
+
"loss": 13.5805,
|
| 189 |
+
"step": 600
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.22,
|
| 193 |
+
"learning_rate": 1.8685955394816152e-07,
|
| 194 |
+
"loss": 13.6546,
|
| 195 |
+
"step": 620
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.23,
|
| 199 |
+
"learning_rate": 1.9288728149487642e-07,
|
| 200 |
+
"loss": 13.5849,
|
| 201 |
+
"step": 640
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.24,
|
| 205 |
+
"learning_rate": 1.9891500904159131e-07,
|
| 206 |
+
"loss": 13.611,
|
| 207 |
+
"step": 660
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.25,
|
| 211 |
+
"learning_rate": 2.0494273658830618e-07,
|
| 212 |
+
"loss": 13.5593,
|
| 213 |
+
"step": 680
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.25,
|
| 217 |
+
"learning_rate": 2.1097046413502108e-07,
|
| 218 |
+
"loss": 13.6788,
|
| 219 |
+
"step": 700
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.26,
|
| 223 |
+
"learning_rate": 2.1699819168173598e-07,
|
| 224 |
+
"loss": 13.591,
|
| 225 |
+
"step": 720
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.27,
|
| 229 |
+
"learning_rate": 2.2302591922845087e-07,
|
| 230 |
+
"loss": 13.5295,
|
| 231 |
+
"step": 740
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.27,
|
| 235 |
+
"learning_rate": 2.2905364677516574e-07,
|
| 236 |
+
"loss": 13.4517,
|
| 237 |
+
"step": 760
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.28,
|
| 241 |
+
"learning_rate": 2.3508137432188064e-07,
|
| 242 |
+
"loss": 13.4802,
|
| 243 |
+
"step": 780
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.29,
|
| 247 |
+
"learning_rate": 2.4110910186859554e-07,
|
| 248 |
+
"loss": 13.4712,
|
| 249 |
+
"step": 800
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.3,
|
| 253 |
+
"learning_rate": 2.4713682941531043e-07,
|
| 254 |
+
"loss": 13.5632,
|
| 255 |
+
"step": 820
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.3,
|
| 259 |
+
"learning_rate": 2.5316455696202533e-07,
|
| 260 |
+
"loss": 13.6037,
|
| 261 |
+
"step": 840
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.31,
|
| 265 |
+
"learning_rate": 2.591922845087402e-07,
|
| 266 |
+
"loss": 13.537,
|
| 267 |
+
"step": 860
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.32,
|
| 271 |
+
"learning_rate": 2.6522001205545507e-07,
|
| 272 |
+
"loss": 13.555,
|
| 273 |
+
"step": 880
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.33,
|
| 277 |
+
"learning_rate": 2.7124773960216996e-07,
|
| 278 |
+
"loss": 13.5306,
|
| 279 |
+
"step": 900
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.33,
|
| 283 |
+
"learning_rate": 2.7727546714888486e-07,
|
| 284 |
+
"loss": 13.5647,
|
| 285 |
+
"step": 920
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.34,
|
| 289 |
+
"learning_rate": 2.8330319469559976e-07,
|
| 290 |
+
"loss": 13.6317,
|
| 291 |
+
"step": 940
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.35,
|
| 295 |
+
"learning_rate": 2.8933092224231465e-07,
|
| 296 |
+
"loss": 13.4934,
|
| 297 |
+
"step": 960
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.35,
|
| 301 |
+
"learning_rate": 2.9535864978902955e-07,
|
| 302 |
+
"loss": 13.6209,
|
| 303 |
+
"step": 980
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.36,
|
| 307 |
+
"learning_rate": 3.013863773357444e-07,
|
| 308 |
+
"loss": 13.5188,
|
| 309 |
+
"step": 1000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.37,
|
| 313 |
+
"learning_rate": 3.074141048824593e-07,
|
| 314 |
+
"loss": 13.5084,
|
| 315 |
+
"step": 1020
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.38,
|
| 319 |
+
"learning_rate": 3.1344183242917424e-07,
|
| 320 |
+
"loss": 13.4541,
|
| 321 |
+
"step": 1040
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.38,
|
| 325 |
+
"learning_rate": 3.194695599758891e-07,
|
| 326 |
+
"loss": 13.404,
|
| 327 |
+
"step": 1060
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.39,
|
| 331 |
+
"learning_rate": 3.25497287522604e-07,
|
| 332 |
+
"loss": 13.4051,
|
| 333 |
+
"step": 1080
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.4,
|
| 337 |
+
"learning_rate": 3.315250150693188e-07,
|
| 338 |
+
"loss": 13.4567,
|
| 339 |
+
"step": 1100
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.41,
|
| 343 |
+
"learning_rate": 3.375527426160337e-07,
|
| 344 |
+
"loss": 13.4018,
|
| 345 |
+
"step": 1120
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.41,
|
| 349 |
+
"learning_rate": 3.435804701627486e-07,
|
| 350 |
+
"loss": 13.3846,
|
| 351 |
+
"step": 1140
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.42,
|
| 355 |
+
"learning_rate": 3.4960819770946356e-07,
|
| 356 |
+
"loss": 13.3325,
|
| 357 |
+
"step": 1160
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.43,
|
| 361 |
+
"learning_rate": 3.556359252561784e-07,
|
| 362 |
+
"loss": 13.3005,
|
| 363 |
+
"step": 1180
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.43,
|
| 367 |
+
"learning_rate": 3.616636528028933e-07,
|
| 368 |
+
"loss": 13.3281,
|
| 369 |
+
"step": 1200
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.44,
|
| 373 |
+
"learning_rate": 3.6769138034960815e-07,
|
| 374 |
+
"loss": 13.258,
|
| 375 |
+
"step": 1220
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.45,
|
| 379 |
+
"learning_rate": 3.7371910789632304e-07,
|
| 380 |
+
"loss": 13.2784,
|
| 381 |
+
"step": 1240
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.46,
|
| 385 |
+
"learning_rate": 3.79746835443038e-07,
|
| 386 |
+
"loss": 13.2793,
|
| 387 |
+
"step": 1260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.46,
|
| 391 |
+
"learning_rate": 3.8577456298975284e-07,
|
| 392 |
+
"loss": 13.2464,
|
| 393 |
+
"step": 1280
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.47,
|
| 397 |
+
"learning_rate": 3.9180229053646773e-07,
|
| 398 |
+
"loss": 13.3347,
|
| 399 |
+
"step": 1300
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.48,
|
| 403 |
+
"learning_rate": 3.9783001808318263e-07,
|
| 404 |
+
"loss": 13.3884,
|
| 405 |
+
"step": 1320
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.48,
|
| 409 |
+
"learning_rate": 4.0385774562989747e-07,
|
| 410 |
+
"loss": 13.2666,
|
| 411 |
+
"step": 1340
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.49,
|
| 415 |
+
"learning_rate": 4.0988547317661237e-07,
|
| 416 |
+
"loss": 13.2176,
|
| 417 |
+
"step": 1360
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.5,
|
| 421 |
+
"learning_rate": 4.159132007233273e-07,
|
| 422 |
+
"loss": 13.233,
|
| 423 |
+
"step": 1380
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.51,
|
| 427 |
+
"learning_rate": 4.2194092827004216e-07,
|
| 428 |
+
"loss": 13.2696,
|
| 429 |
+
"step": 1400
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.51,
|
| 433 |
+
"learning_rate": 4.2796865581675706e-07,
|
| 434 |
+
"loss": 13.1597,
|
| 435 |
+
"step": 1420
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.52,
|
| 439 |
+
"learning_rate": 4.3399638336347195e-07,
|
| 440 |
+
"loss": 13.1763,
|
| 441 |
+
"step": 1440
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.53,
|
| 445 |
+
"learning_rate": 4.400241109101868e-07,
|
| 446 |
+
"loss": 13.2415,
|
| 447 |
+
"step": 1460
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.54,
|
| 451 |
+
"learning_rate": 4.4605183845690175e-07,
|
| 452 |
+
"loss": 13.2653,
|
| 453 |
+
"step": 1480
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.54,
|
| 457 |
+
"learning_rate": 4.5207956600361664e-07,
|
| 458 |
+
"loss": 13.1193,
|
| 459 |
+
"step": 1500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.55,
|
| 463 |
+
"learning_rate": 4.581072935503315e-07,
|
| 464 |
+
"loss": 13.1928,
|
| 465 |
+
"step": 1520
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.56,
|
| 469 |
+
"learning_rate": 4.641350210970464e-07,
|
| 470 |
+
"loss": 13.1678,
|
| 471 |
+
"step": 1540
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.56,
|
| 475 |
+
"learning_rate": 4.701627486437613e-07,
|
| 476 |
+
"loss": 13.1757,
|
| 477 |
+
"step": 1560
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.57,
|
| 481 |
+
"learning_rate": 4.761904761904761e-07,
|
| 482 |
+
"loss": 13.1773,
|
| 483 |
+
"step": 1580
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.58,
|
| 487 |
+
"learning_rate": 4.822182037371911e-07,
|
| 488 |
+
"loss": 13.1095,
|
| 489 |
+
"step": 1600
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.59,
|
| 493 |
+
"learning_rate": 4.88245931283906e-07,
|
| 494 |
+
"loss": 13.146,
|
| 495 |
+
"step": 1620
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.59,
|
| 499 |
+
"learning_rate": 4.942736588306209e-07,
|
| 500 |
+
"loss": 13.1242,
|
| 501 |
+
"step": 1640
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.6,
|
| 505 |
+
"learning_rate": 4.999999719845892e-07,
|
| 506 |
+
"loss": 13.0289,
|
| 507 |
+
"step": 1660
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.61,
|
| 511 |
+
"learning_rate": 4.99987645305352e-07,
|
| 512 |
+
"loss": 13.1282,
|
| 513 |
+
"step": 1680
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.61,
|
| 517 |
+
"learning_rate": 4.999529075720485e-07,
|
| 518 |
+
"loss": 13.1208,
|
| 519 |
+
"step": 1700
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.62,
|
| 523 |
+
"learning_rate": 4.998957618988693e-07,
|
| 524 |
+
"loss": 13.1201,
|
| 525 |
+
"step": 1720
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.63,
|
| 529 |
+
"learning_rate": 4.998162134088466e-07,
|
| 530 |
+
"loss": 12.9442,
|
| 531 |
+
"step": 1740
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.64,
|
| 535 |
+
"learning_rate": 4.997142692333949e-07,
|
| 536 |
+
"loss": 13.114,
|
| 537 |
+
"step": 1760
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.64,
|
| 541 |
+
"learning_rate": 4.995899385116716e-07,
|
| 542 |
+
"loss": 13.0605,
|
| 543 |
+
"step": 1780
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.65,
|
| 547 |
+
"learning_rate": 4.994432323897575e-07,
|
| 548 |
+
"loss": 13.0292,
|
| 549 |
+
"step": 1800
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.66,
|
| 553 |
+
"learning_rate": 4.992741640196581e-07,
|
| 554 |
+
"loss": 13.0242,
|
| 555 |
+
"step": 1820
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.67,
|
| 559 |
+
"learning_rate": 4.990827485581237e-07,
|
| 560 |
+
"loss": 13.0583,
|
| 561 |
+
"step": 1840
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.67,
|
| 565 |
+
"learning_rate": 4.988690031652916e-07,
|
| 566 |
+
"loss": 13.0788,
|
| 567 |
+
"step": 1860
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.68,
|
| 571 |
+
"learning_rate": 4.986329470031468e-07,
|
| 572 |
+
"loss": 13.0321,
|
| 573 |
+
"step": 1880
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.69,
|
| 577 |
+
"learning_rate": 4.983746012338049e-07,
|
| 578 |
+
"loss": 12.9585,
|
| 579 |
+
"step": 1900
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.69,
|
| 583 |
+
"learning_rate": 4.980939890176143e-07,
|
| 584 |
+
"loss": 13.0254,
|
| 585 |
+
"step": 1920
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.7,
|
| 589 |
+
"learning_rate": 4.977911355110807e-07,
|
| 590 |
+
"loss": 13.0148,
|
| 591 |
+
"step": 1940
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.71,
|
| 595 |
+
"learning_rate": 4.974660678646106e-07,
|
| 596 |
+
"loss": 12.8612,
|
| 597 |
+
"step": 1960
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.72,
|
| 601 |
+
"learning_rate": 4.971188152200791e-07,
|
| 602 |
+
"loss": 12.8905,
|
| 603 |
+
"step": 1980
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.72,
|
| 607 |
+
"learning_rate": 4.967494087082155e-07,
|
| 608 |
+
"loss": 12.7506,
|
| 609 |
+
"step": 2000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.73,
|
| 613 |
+
"learning_rate": 4.963578814458136e-07,
|
| 614 |
+
"loss": 12.7439,
|
| 615 |
+
"step": 2020
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.74,
|
| 619 |
+
"learning_rate": 4.959442685327627e-07,
|
| 620 |
+
"loss": 12.5732,
|
| 621 |
+
"step": 2040
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.75,
|
| 625 |
+
"learning_rate": 4.955086070489006e-07,
|
| 626 |
+
"loss": 12.4223,
|
| 627 |
+
"step": 2060
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.75,
|
| 631 |
+
"learning_rate": 4.950509360506898e-07,
|
| 632 |
+
"loss": 12.4088,
|
| 633 |
+
"step": 2080
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.76,
|
| 637 |
+
"learning_rate": 4.945712965677158e-07,
|
| 638 |
+
"loss": 12.4423,
|
| 639 |
+
"step": 2100
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.77,
|
| 643 |
+
"learning_rate": 4.94069731599009e-07,
|
| 644 |
+
"loss": 12.2641,
|
| 645 |
+
"step": 2120
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.77,
|
| 649 |
+
"learning_rate": 4.935462861091903e-07,
|
| 650 |
+
"loss": 12.3842,
|
| 651 |
+
"step": 2140
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.78,
|
| 655 |
+
"learning_rate": 4.93001007024439e-07,
|
| 656 |
+
"loss": 12.1959,
|
| 657 |
+
"step": 2160
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.79,
|
| 661 |
+
"learning_rate": 4.924339432282875e-07,
|
| 662 |
+
"loss": 12.2833,
|
| 663 |
+
"step": 2180
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.8,
|
| 667 |
+
"learning_rate": 4.918451455572374e-07,
|
| 668 |
+
"loss": 12.146,
|
| 669 |
+
"step": 2200
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.8,
|
| 673 |
+
"learning_rate": 4.912346667962032e-07,
|
| 674 |
+
"loss": 12.1976,
|
| 675 |
+
"step": 2220
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.81,
|
| 679 |
+
"learning_rate": 4.906025616737794e-07,
|
| 680 |
+
"loss": 12.0598,
|
| 681 |
+
"step": 2240
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.82,
|
| 685 |
+
"learning_rate": 4.899488868573349e-07,
|
| 686 |
+
"loss": 12.0777,
|
| 687 |
+
"step": 2260
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.82,
|
| 691 |
+
"learning_rate": 4.892737009479322e-07,
|
| 692 |
+
"loss": 12.0493,
|
| 693 |
+
"step": 2280
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.83,
|
| 697 |
+
"learning_rate": 4.885770644750743e-07,
|
| 698 |
+
"loss": 12.0849,
|
| 699 |
+
"step": 2300
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.84,
|
| 703 |
+
"learning_rate": 4.878590398912782e-07,
|
| 704 |
+
"loss": 11.9523,
|
| 705 |
+
"step": 2320
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.85,
|
| 709 |
+
"learning_rate": 4.871196915664761e-07,
|
| 710 |
+
"loss": 11.976,
|
| 711 |
+
"step": 2340
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.85,
|
| 715 |
+
"learning_rate": 4.863590857822445e-07,
|
| 716 |
+
"loss": 11.8329,
|
| 717 |
+
"step": 2360
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.86,
|
| 721 |
+
"learning_rate": 4.855772907258626e-07,
|
| 722 |
+
"loss": 11.9747,
|
| 723 |
+
"step": 2380
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.87,
|
| 727 |
+
"learning_rate": 4.847743764841993e-07,
|
| 728 |
+
"loss": 11.9154,
|
| 729 |
+
"step": 2400
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.88,
|
| 733 |
+
"learning_rate": 4.839504150374295e-07,
|
| 734 |
+
"loss": 11.8741,
|
| 735 |
+
"step": 2420
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.88,
|
| 739 |
+
"learning_rate": 4.831054802525819e-07,
|
| 740 |
+
"loss": 11.914,
|
| 741 |
+
"step": 2440
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.89,
|
| 745 |
+
"learning_rate": 4.822396478769162e-07,
|
| 746 |
+
"loss": 11.7922,
|
| 747 |
+
"step": 2460
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.9,
|
| 751 |
+
"learning_rate": 4.813529955311333e-07,
|
| 752 |
+
"loss": 11.693,
|
| 753 |
+
"step": 2480
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.9,
|
| 757 |
+
"learning_rate": 4.804456027024159e-07,
|
| 758 |
+
"loss": 11.8023,
|
| 759 |
+
"step": 2500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.91,
|
| 763 |
+
"learning_rate": 4.795175507373028e-07,
|
| 764 |
+
"loss": 11.8223,
|
| 765 |
+
"step": 2520
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.92,
|
| 769 |
+
"learning_rate": 4.785689228343971e-07,
|
| 770 |
+
"loss": 11.6569,
|
| 771 |
+
"step": 2540
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.93,
|
| 775 |
+
"learning_rate": 4.77599804036906e-07,
|
| 776 |
+
"loss": 11.6978,
|
| 777 |
+
"step": 2560
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.93,
|
| 781 |
+
"learning_rate": 4.766102812250183e-07,
|
| 782 |
+
"loss": 11.7304,
|
| 783 |
+
"step": 2580
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.94,
|
| 787 |
+
"learning_rate": 4.756004431081149e-07,
|
| 788 |
+
"loss": 11.6343,
|
| 789 |
+
"step": 2600
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.95,
|
| 793 |
+
"learning_rate": 4.745703802168161e-07,
|
| 794 |
+
"loss": 11.7166,
|
| 795 |
+
"step": 2620
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.95,
|
| 799 |
+
"learning_rate": 4.7352018489486606e-07,
|
| 800 |
+
"loss": 11.5801,
|
| 801 |
+
"step": 2640
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.96,
|
| 805 |
+
"learning_rate": 4.724499512908541e-07,
|
| 806 |
+
"loss": 11.5505,
|
| 807 |
+
"step": 2660
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.97,
|
| 811 |
+
"learning_rate": 4.7135977534977425e-07,
|
| 812 |
+
"loss": 11.573,
|
| 813 |
+
"step": 2680
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.98,
|
| 817 |
+
"learning_rate": 4.702497548044243e-07,
|
| 818 |
+
"loss": 11.7114,
|
| 819 |
+
"step": 2700
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.98,
|
| 823 |
+
"learning_rate": 4.6911998916664375e-07,
|
| 824 |
+
"loss": 11.6111,
|
| 825 |
+
"step": 2720
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.99,
|
| 829 |
+
"learning_rate": 4.6797057971839306e-07,
|
| 830 |
+
"loss": 11.6444,
|
| 831 |
+
"step": 2740
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 1.0,
|
| 835 |
+
"learning_rate": 4.6680162950267356e-07,
|
| 836 |
+
"loss": 11.5578,
|
| 837 |
+
"step": 2760
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 1.01,
|
| 841 |
+
"learning_rate": 4.656132433142901e-07,
|
| 842 |
+
"loss": 11.5535,
|
| 843 |
+
"step": 2780
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 1.01,
|
| 847 |
+
"learning_rate": 4.6440552769045604e-07,
|
| 848 |
+
"loss": 11.6784,
|
| 849 |
+
"step": 2800
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 1.02,
|
| 853 |
+
"learning_rate": 4.631785909012426e-07,
|
| 854 |
+
"loss": 11.5629,
|
| 855 |
+
"step": 2820
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 1.03,
|
| 859 |
+
"learning_rate": 4.619325429398724e-07,
|
| 860 |
+
"loss": 11.5986,
|
| 861 |
+
"step": 2840
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 1.03,
|
| 865 |
+
"learning_rate": 4.606674955128586e-07,
|
| 866 |
+
"loss": 11.5513,
|
| 867 |
+
"step": 2860
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 1.04,
|
| 871 |
+
"learning_rate": 4.593835620299911e-07,
|
| 872 |
+
"loss": 11.4675,
|
| 873 |
+
"step": 2880
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 1.05,
|
| 877 |
+
"learning_rate": 4.580808575941688e-07,
|
| 878 |
+
"loss": 11.5263,
|
| 879 |
+
"step": 2900
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 1.06,
|
| 883 |
+
"learning_rate": 4.5675949899108133e-07,
|
| 884 |
+
"loss": 11.6017,
|
| 885 |
+
"step": 2920
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 1.06,
|
| 889 |
+
"learning_rate": 4.554196046787392e-07,
|
| 890 |
+
"loss": 11.6048,
|
| 891 |
+
"step": 2940
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 1.07,
|
| 895 |
+
"learning_rate": 4.5406129477685406e-07,
|
| 896 |
+
"loss": 11.4808,
|
| 897 |
+
"step": 2960
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 1.08,
|
| 901 |
+
"learning_rate": 4.526846910560702e-07,
|
| 902 |
+
"loss": 11.5947,
|
| 903 |
+
"step": 2980
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 1.08,
|
| 907 |
+
"learning_rate": 4.512899169270481e-07,
|
| 908 |
+
"loss": 11.5756,
|
| 909 |
+
"step": 3000
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.09,
|
| 913 |
+
"learning_rate": 4.4987709742940077e-07,
|
| 914 |
+
"loss": 11.5105,
|
| 915 |
+
"step": 3020
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 1.1,
|
| 919 |
+
"learning_rate": 4.4844635922048405e-07,
|
| 920 |
+
"loss": 11.5413,
|
| 921 |
+
"step": 3040
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 1.11,
|
| 925 |
+
"learning_rate": 4.46997830564042e-07,
|
| 926 |
+
"loss": 11.4868,
|
| 927 |
+
"step": 3060
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 1.11,
|
| 931 |
+
"learning_rate": 4.4553164131870813e-07,
|
| 932 |
+
"loss": 11.557,
|
| 933 |
+
"step": 3080
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 1.12,
|
| 937 |
+
"learning_rate": 4.440479229263641e-07,
|
| 938 |
+
"loss": 11.5823,
|
| 939 |
+
"step": 3100
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.13,
|
| 943 |
+
"learning_rate": 4.4254680840035554e-07,
|
| 944 |
+
"loss": 11.4501,
|
| 945 |
+
"step": 3120
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 1.14,
|
| 949 |
+
"learning_rate": 4.410284323135682e-07,
|
| 950 |
+
"loss": 11.5275,
|
| 951 |
+
"step": 3140
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 1.14,
|
| 955 |
+
"learning_rate": 4.394929307863632e-07,
|
| 956 |
+
"loss": 11.5249,
|
| 957 |
+
"step": 3160
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 1.15,
|
| 961 |
+
"learning_rate": 4.3794044147437437e-07,
|
| 962 |
+
"loss": 11.479,
|
| 963 |
+
"step": 3180
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 1.16,
|
| 967 |
+
"learning_rate": 4.3637110355616735e-07,
|
| 968 |
+
"loss": 11.5742,
|
| 969 |
+
"step": 3200
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 1.16,
|
| 973 |
+
"learning_rate": 4.347850577207626e-07,
|
| 974 |
+
"loss": 11.5581,
|
| 975 |
+
"step": 3220
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 1.17,
|
| 979 |
+
"learning_rate": 4.3318244615502254e-07,
|
| 980 |
+
"loss": 11.4662,
|
| 981 |
+
"step": 3240
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 1.18,
|
| 985 |
+
"learning_rate": 4.3156341253090524e-07,
|
| 986 |
+
"loss": 11.5468,
|
| 987 |
+
"step": 3260
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 1.19,
|
| 991 |
+
"learning_rate": 4.2992810199258373e-07,
|
| 992 |
+
"loss": 11.5638,
|
| 993 |
+
"step": 3280
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 1.19,
|
| 997 |
+
"learning_rate": 4.2827666114343463e-07,
|
| 998 |
+
"loss": 11.5076,
|
| 999 |
+
"step": 3300
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 1.2,
|
| 1003 |
+
"learning_rate": 4.26609238032895e-07,
|
| 1004 |
+
"loss": 11.5381,
|
| 1005 |
+
"step": 3320
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 1.21,
|
| 1009 |
+
"learning_rate": 4.2492598214319e-07,
|
| 1010 |
+
"loss": 11.5322,
|
| 1011 |
+
"step": 3340
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 1.22,
|
| 1015 |
+
"learning_rate": 4.232270443759319e-07,
|
| 1016 |
+
"loss": 11.5848,
|
| 1017 |
+
"step": 3360
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 1.22,
|
| 1021 |
+
"learning_rate": 4.215125770385923e-07,
|
| 1022 |
+
"loss": 11.5532,
|
| 1023 |
+
"step": 3380
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 1.23,
|
| 1027 |
+
"learning_rate": 4.197827338308475e-07,
|
| 1028 |
+
"loss": 11.5282,
|
| 1029 |
+
"step": 3400
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 1.24,
|
| 1033 |
+
"learning_rate": 4.1803766983080006e-07,
|
| 1034 |
+
"loss": 11.5016,
|
| 1035 |
+
"step": 3420
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 1.24,
|
| 1039 |
+
"learning_rate": 4.162775414810757e-07,
|
| 1040 |
+
"loss": 11.5488,
|
| 1041 |
+
"step": 3440
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 1.25,
|
| 1045 |
+
"learning_rate": 4.145025065747988e-07,
|
| 1046 |
+
"loss": 11.4712,
|
| 1047 |
+
"step": 3460
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 1.26,
|
| 1051 |
+
"learning_rate": 4.1271272424144645e-07,
|
| 1052 |
+
"loss": 11.6967,
|
| 1053 |
+
"step": 3480
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 1.27,
|
| 1057 |
+
"learning_rate": 4.1090835493258235e-07,
|
| 1058 |
+
"loss": 11.5829,
|
| 1059 |
+
"step": 3500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 1.27,
|
| 1063 |
+
"learning_rate": 4.0908956040747315e-07,
|
| 1064 |
+
"loss": 11.5436,
|
| 1065 |
+
"step": 3520
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 1.28,
|
| 1069 |
+
"learning_rate": 4.0725650371858646e-07,
|
| 1070 |
+
"loss": 11.6792,
|
| 1071 |
+
"step": 3540
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 1.29,
|
| 1075 |
+
"learning_rate": 4.054093491969734e-07,
|
| 1076 |
+
"loss": 11.5023,
|
| 1077 |
+
"step": 3560
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 1.29,
|
| 1081 |
+
"learning_rate": 4.03548262437537e-07,
|
| 1082 |
+
"loss": 11.4841,
|
| 1083 |
+
"step": 3580
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 1.3,
|
| 1087 |
+
"learning_rate": 4.0167341028418655e-07,
|
| 1088 |
+
"loss": 11.5712,
|
| 1089 |
+
"step": 3600
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 1.31,
|
| 1093 |
+
"learning_rate": 3.997849608148799e-07,
|
| 1094 |
+
"loss": 11.566,
|
| 1095 |
+
"step": 3620
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.32,
|
| 1099 |
+
"learning_rate": 3.978830833265563e-07,
|
| 1100 |
+
"loss": 11.5569,
|
| 1101 |
+
"step": 3640
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 1.32,
|
| 1105 |
+
"learning_rate": 3.9596794831995863e-07,
|
| 1106 |
+
"loss": 11.5758,
|
| 1107 |
+
"step": 3660
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 1.33,
|
| 1111 |
+
"learning_rate": 3.940397274843481e-07,
|
| 1112 |
+
"loss": 11.4845,
|
| 1113 |
+
"step": 3680
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 1.34,
|
| 1117 |
+
"learning_rate": 3.9209859368211304e-07,
|
| 1118 |
+
"loss": 11.5205,
|
| 1119 |
+
"step": 3700
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 1.35,
|
| 1123 |
+
"learning_rate": 3.9014472093327164e-07,
|
| 1124 |
+
"loss": 11.5794,
|
| 1125 |
+
"step": 3720
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 1.35,
|
| 1129 |
+
"learning_rate": 3.8817828439987154e-07,
|
| 1130 |
+
"loss": 11.5827,
|
| 1131 |
+
"step": 3740
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 1.36,
|
| 1135 |
+
"learning_rate": 3.8619946037028644e-07,
|
| 1136 |
+
"loss": 11.5148,
|
| 1137 |
+
"step": 3760
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.37,
|
| 1141 |
+
"learning_rate": 3.842084262434125e-07,
|
| 1142 |
+
"loss": 11.6089,
|
| 1143 |
+
"step": 3780
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 1.37,
|
| 1147 |
+
"learning_rate": 3.822053605127645e-07,
|
| 1148 |
+
"loss": 11.6853,
|
| 1149 |
+
"step": 3800
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 1.38,
|
| 1153 |
+
"learning_rate": 3.801904427504743e-07,
|
| 1154 |
+
"loss": 11.5037,
|
| 1155 |
+
"step": 3820
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 1.39,
|
| 1159 |
+
"learning_rate": 3.781638535911922e-07,
|
| 1160 |
+
"loss": 11.4557,
|
| 1161 |
+
"step": 3840
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 1.4,
|
| 1165 |
+
"learning_rate": 3.7612577471589355e-07,
|
| 1166 |
+
"loss": 11.5509,
|
| 1167 |
+
"step": 3860
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 1.4,
|
| 1171 |
+
"learning_rate": 3.7407638883559117e-07,
|
| 1172 |
+
"loss": 11.5338,
|
| 1173 |
+
"step": 3880
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 1.41,
|
| 1177 |
+
"learning_rate": 3.720158796749556e-07,
|
| 1178 |
+
"loss": 11.5365,
|
| 1179 |
+
"step": 3900
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.42,
|
| 1183 |
+
"learning_rate": 3.699444319558444e-07,
|
| 1184 |
+
"loss": 11.5654,
|
| 1185 |
+
"step": 3920
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.42,
|
| 1189 |
+
"learning_rate": 3.6786223138074195e-07,
|
| 1190 |
+
"loss": 11.5881,
|
| 1191 |
+
"step": 3940
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 1.43,
|
| 1195 |
+
"learning_rate": 3.657694646161119e-07,
|
| 1196 |
+
"loss": 11.6041,
|
| 1197 |
+
"step": 3960
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 1.44,
|
| 1201 |
+
"learning_rate": 3.6366631927566223e-07,
|
| 1202 |
+
"loss": 11.6121,
|
| 1203 |
+
"step": 3980
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 1.45,
|
| 1207 |
+
"learning_rate": 3.6155298390352637e-07,
|
| 1208 |
+
"loss": 11.5077,
|
| 1209 |
+
"step": 4000
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 1.45,
|
| 1213 |
+
"learning_rate": 3.594296479573602e-07,
|
| 1214 |
+
"loss": 11.6038,
|
| 1215 |
+
"step": 4020
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 1.46,
|
| 1219 |
+
"learning_rate": 3.5729650179135743e-07,
|
| 1220 |
+
"loss": 11.5351,
|
| 1221 |
+
"step": 4040
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.47,
|
| 1225 |
+
"learning_rate": 3.5515373663918504e-07,
|
| 1226 |
+
"loss": 11.6607,
|
| 1227 |
+
"step": 4060
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.48,
|
| 1231 |
+
"learning_rate": 3.530015445968388e-07,
|
| 1232 |
+
"loss": 11.5155,
|
| 1233 |
+
"step": 4080
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 1.48,
|
| 1237 |
+
"learning_rate": 3.508401186054224e-07,
|
| 1238 |
+
"loss": 11.5073,
|
| 1239 |
+
"step": 4100
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 1.49,
|
| 1243 |
+
"learning_rate": 3.4866965243385093e-07,
|
| 1244 |
+
"loss": 11.5617,
|
| 1245 |
+
"step": 4120
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 1.5,
|
| 1249 |
+
"learning_rate": 3.4649034066147894e-07,
|
| 1250 |
+
"loss": 11.4584,
|
| 1251 |
+
"step": 4140
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 1.5,
|
| 1255 |
+
"learning_rate": 3.4430237866065735e-07,
|
| 1256 |
+
"loss": 11.5588,
|
| 1257 |
+
"step": 4160
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.51,
|
| 1261 |
+
"learning_rate": 3.4210596257921844e-07,
|
| 1262 |
+
"loss": 11.4431,
|
| 1263 |
+
"step": 4180
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.52,
|
| 1267 |
+
"learning_rate": 3.399012893228912e-07,
|
| 1268 |
+
"loss": 11.5465,
|
| 1269 |
+
"step": 4200
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.53,
|
| 1273 |
+
"learning_rate": 3.376885565376491e-07,
|
| 1274 |
+
"loss": 11.5845,
|
| 1275 |
+
"step": 4220
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 1.53,
|
| 1279 |
+
"learning_rate": 3.3546796259199185e-07,
|
| 1280 |
+
"loss": 11.6001,
|
| 1281 |
+
"step": 4240
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 1.54,
|
| 1285 |
+
"learning_rate": 3.3323970655916115e-07,
|
| 1286 |
+
"loss": 11.5957,
|
| 1287 |
+
"step": 4260
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 1.55,
|
| 1291 |
+
"learning_rate": 3.3100398819929473e-07,
|
| 1292 |
+
"loss": 11.5674,
|
| 1293 |
+
"step": 4280
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 1.56,
|
| 1297 |
+
"learning_rate": 3.2876100794151755e-07,
|
| 1298 |
+
"loss": 11.554,
|
| 1299 |
+
"step": 4300
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 1.56,
|
| 1303 |
+
"learning_rate": 3.2651096686597423e-07,
|
| 1304 |
+
"loss": 11.5188,
|
| 1305 |
+
"step": 4320
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.57,
|
| 1309 |
+
"learning_rate": 3.2425406668580185e-07,
|
| 1310 |
+
"loss": 11.5757,
|
| 1311 |
+
"step": 4340
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 1.58,
|
| 1315 |
+
"learning_rate": 3.2199050972904706e-07,
|
| 1316 |
+
"loss": 11.4447,
|
| 1317 |
+
"step": 4360
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 1.58,
|
| 1321 |
+
"learning_rate": 3.197204989205276e-07,
|
| 1322 |
+
"loss": 11.5552,
|
| 1323 |
+
"step": 4380
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 1.59,
|
| 1327 |
+
"learning_rate": 3.174442377636401e-07,
|
| 1328 |
+
"loss": 11.5855,
|
| 1329 |
+
"step": 4400
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 1.6,
|
| 1333 |
+
"learning_rate": 3.151619303221166e-07,
|
| 1334 |
+
"loss": 11.5776,
|
| 1335 |
+
"step": 4420
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 1.61,
|
| 1339 |
+
"learning_rate": 3.1287378120173045e-07,
|
| 1340 |
+
"loss": 11.4827,
|
| 1341 |
+
"step": 4440
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 1.61,
|
| 1345 |
+
"learning_rate": 3.1057999553195337e-07,
|
| 1346 |
+
"loss": 11.7251,
|
| 1347 |
+
"step": 4460
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.62,
|
| 1351 |
+
"learning_rate": 3.082807789475662e-07,
|
| 1352 |
+
"loss": 11.6293,
|
| 1353 |
+
"step": 4480
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 1.63,
|
| 1357 |
+
"learning_rate": 3.059763375702241e-07,
|
| 1358 |
+
"loss": 11.494,
|
| 1359 |
+
"step": 4500
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 1.63,
|
| 1363 |
+
"learning_rate": 3.0366687798997756e-07,
|
| 1364 |
+
"loss": 11.5105,
|
| 1365 |
+
"step": 4520
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 1.64,
|
| 1369 |
+
"learning_rate": 3.0135260724675236e-07,
|
| 1370 |
+
"loss": 11.4956,
|
| 1371 |
+
"step": 4540
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 1.65,
|
| 1375 |
+
"learning_rate": 2.990337328117886e-07,
|
| 1376 |
+
"loss": 11.5945,
|
| 1377 |
+
"step": 4560
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 1.66,
|
| 1381 |
+
"learning_rate": 2.9671046256904075e-07,
|
| 1382 |
+
"loss": 11.5549,
|
| 1383 |
+
"step": 4580
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 1.66,
|
| 1387 |
+
"learning_rate": 2.9438300479654196e-07,
|
| 1388 |
+
"loss": 11.5096,
|
| 1389 |
+
"step": 4600
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.67,
|
| 1393 |
+
"learning_rate": 2.9205156814773143e-07,
|
| 1394 |
+
"loss": 11.6499,
|
| 1395 |
+
"step": 4620
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.68,
|
| 1399 |
+
"learning_rate": 2.8971636163274924e-07,
|
| 1400 |
+
"loss": 11.5242,
|
| 1401 |
+
"step": 4640
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 1.69,
|
| 1405 |
+
"learning_rate": 2.8737759459969925e-07,
|
| 1406 |
+
"loss": 11.6741,
|
| 1407 |
+
"step": 4660
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 1.69,
|
| 1411 |
+
"learning_rate": 2.850354767158804e-07,
|
| 1412 |
+
"loss": 11.5848,
|
| 1413 |
+
"step": 4680
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 1.7,
|
| 1417 |
+
"learning_rate": 2.826902179489913e-07,
|
| 1418 |
+
"loss": 11.3855,
|
| 1419 |
+
"step": 4700
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 1.71,
|
| 1423 |
+
"learning_rate": 2.803420285483062e-07,
|
| 1424 |
+
"loss": 11.6143,
|
| 1425 |
+
"step": 4720
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 1.71,
|
| 1429 |
+
"learning_rate": 2.7799111902582693e-07,
|
| 1430 |
+
"loss": 11.5048,
|
| 1431 |
+
"step": 4740
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.72,
|
| 1435 |
+
"learning_rate": 2.7563770013741036e-07,
|
| 1436 |
+
"loss": 11.6516,
|
| 1437 |
+
"step": 4760
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.73,
|
| 1441 |
+
"learning_rate": 2.7328198286387473e-07,
|
| 1442 |
+
"loss": 11.496,
|
| 1443 |
+
"step": 4780
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 1.74,
|
| 1447 |
+
"learning_rate": 2.7092417839208537e-07,
|
| 1448 |
+
"loss": 11.5108,
|
| 1449 |
+
"step": 4800
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 1.74,
|
| 1453 |
+
"learning_rate": 2.685644980960221e-07,
|
| 1454 |
+
"loss": 11.6525,
|
| 1455 |
+
"step": 4820
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 1.75,
|
| 1459 |
+
"learning_rate": 2.662031535178298e-07,
|
| 1460 |
+
"loss": 11.6017,
|
| 1461 |
+
"step": 4840
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 1.76,
|
| 1465 |
+
"learning_rate": 2.638403563488542e-07,
|
| 1466 |
+
"loss": 11.5589,
|
| 1467 |
+
"step": 4860
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 1.76,
|
| 1471 |
+
"learning_rate": 2.614763184106635e-07,
|
| 1472 |
+
"loss": 11.4828,
|
| 1473 |
+
"step": 4880
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.77,
|
| 1477 |
+
"learning_rate": 2.591112516360594e-07,
|
| 1478 |
+
"loss": 11.5234,
|
| 1479 |
+
"step": 4900
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.78,
|
| 1483 |
+
"learning_rate": 2.567453680500774e-07,
|
| 1484 |
+
"loss": 11.4855,
|
| 1485 |
+
"step": 4920
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 1.79,
|
| 1489 |
+
"learning_rate": 2.5437887975097887e-07,
|
| 1490 |
+
"loss": 11.4654,
|
| 1491 |
+
"step": 4940
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 1.79,
|
| 1495 |
+
"learning_rate": 2.5201199889123715e-07,
|
| 1496 |
+
"loss": 11.5591,
|
| 1497 |
+
"step": 4960
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 1.8,
|
| 1501 |
+
"learning_rate": 2.4964493765851795e-07,
|
| 1502 |
+
"loss": 11.5175,
|
| 1503 |
+
"step": 4980
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 1.81,
|
| 1507 |
+
"learning_rate": 2.4727790825665703e-07,
|
| 1508 |
+
"loss": 11.5616,
|
| 1509 |
+
"step": 5000
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 1.82,
|
| 1513 |
+
"learning_rate": 2.4491112288663675e-07,
|
| 1514 |
+
"loss": 11.6138,
|
| 1515 |
+
"step": 5020
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.82,
|
| 1519 |
+
"learning_rate": 2.4254479372756236e-07,
|
| 1520 |
+
"loss": 11.5176,
|
| 1521 |
+
"step": 5040
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.83,
|
| 1525 |
+
"learning_rate": 2.4017913291764045e-07,
|
| 1526 |
+
"loss": 11.5173,
|
| 1527 |
+
"step": 5060
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 1.84,
|
| 1531 |
+
"learning_rate": 2.3781435253516076e-07,
|
| 1532 |
+
"loss": 11.4549,
|
| 1533 |
+
"step": 5080
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 1.84,
|
| 1537 |
+
"learning_rate": 2.354506645794842e-07,
|
| 1538 |
+
"loss": 11.5263,
|
| 1539 |
+
"step": 5100
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 1.85,
|
| 1543 |
+
"learning_rate": 2.330882809520369e-07,
|
| 1544 |
+
"loss": 11.5152,
|
| 1545 |
+
"step": 5120
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 1.86,
|
| 1549 |
+
"learning_rate": 2.3072741343731383e-07,
|
| 1550 |
+
"loss": 11.4778,
|
| 1551 |
+
"step": 5140
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 1.87,
|
| 1555 |
+
"learning_rate": 2.2836827368389245e-07,
|
| 1556 |
+
"loss": 11.4824,
|
| 1557 |
+
"step": 5160
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.87,
|
| 1561 |
+
"learning_rate": 2.2601107318545874e-07,
|
| 1562 |
+
"loss": 11.5656,
|
| 1563 |
+
"step": 5180
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.88,
|
| 1567 |
+
"learning_rate": 2.2365602326184714e-07,
|
| 1568 |
+
"loss": 11.5009,
|
| 1569 |
+
"step": 5200
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 1.89,
|
| 1573 |
+
"learning_rate": 2.21303335040096e-07,
|
| 1574 |
+
"loss": 11.5992,
|
| 1575 |
+
"step": 5220
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 1.9,
|
| 1579 |
+
"learning_rate": 2.1895321943552024e-07,
|
| 1580 |
+
"loss": 11.6539,
|
| 1581 |
+
"step": 5240
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 1.9,
|
| 1585 |
+
"learning_rate": 2.1660588713280316e-07,
|
| 1586 |
+
"loss": 11.5716,
|
| 1587 |
+
"step": 5260
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 1.91,
|
| 1591 |
+
"learning_rate": 2.142615485671088e-07,
|
| 1592 |
+
"loss": 11.6155,
|
| 1593 |
+
"step": 5280
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 1.92,
|
| 1597 |
+
"learning_rate": 2.119204139052167e-07,
|
| 1598 |
+
"loss": 11.6557,
|
| 1599 |
+
"step": 5300
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.92,
|
| 1603 |
+
"learning_rate": 2.0958269302668075e-07,
|
| 1604 |
+
"loss": 11.6282,
|
| 1605 |
+
"step": 5320
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 1.93,
|
| 1609 |
+
"learning_rate": 2.0724859550501393e-07,
|
| 1610 |
+
"loss": 11.3897,
|
| 1611 |
+
"step": 5340
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 1.94,
|
| 1615 |
+
"learning_rate": 2.0491833058890002e-07,
|
| 1616 |
+
"loss": 11.6027,
|
| 1617 |
+
"step": 5360
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 1.95,
|
| 1621 |
+
"learning_rate": 2.0259210718343495e-07,
|
| 1622 |
+
"loss": 11.6817,
|
| 1623 |
+
"step": 5380
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 1.95,
|
| 1627 |
+
"learning_rate": 2.002701338313987e-07,
|
| 1628 |
+
"loss": 11.6283,
|
| 1629 |
+
"step": 5400
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 1.96,
|
| 1633 |
+
"learning_rate": 1.9795261869455967e-07,
|
| 1634 |
+
"loss": 11.6081,
|
| 1635 |
+
"step": 5420
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 1.97,
|
| 1639 |
+
"learning_rate": 1.9563976953501333e-07,
|
| 1640 |
+
"loss": 11.6218,
|
| 1641 |
+
"step": 5440
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.97,
|
| 1645 |
+
"learning_rate": 1.933317936965566e-07,
|
| 1646 |
+
"loss": 11.6083,
|
| 1647 |
+
"step": 5460
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.98,
|
| 1651 |
+
"learning_rate": 1.9102889808609993e-07,
|
| 1652 |
+
"loss": 11.5382,
|
| 1653 |
+
"step": 5480
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 1.99,
|
| 1657 |
+
"learning_rate": 1.8873128915511782e-07,
|
| 1658 |
+
"loss": 11.7335,
|
| 1659 |
+
"step": 5500
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 2.0,
|
| 1663 |
+
"learning_rate": 1.8643917288114146e-07,
|
| 1664 |
+
"loss": 11.7195,
|
| 1665 |
+
"step": 5520
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 2.0,
|
| 1669 |
+
"learning_rate": 1.8415275474929265e-07,
|
| 1670 |
+
"loss": 11.6062,
|
| 1671 |
+
"step": 5540
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 2.01,
|
| 1675 |
+
"learning_rate": 1.8187223973386243e-07,
|
| 1676 |
+
"loss": 11.4377,
|
| 1677 |
+
"step": 5560
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 2.02,
|
| 1681 |
+
"learning_rate": 1.795978322799354e-07,
|
| 1682 |
+
"loss": 11.4762,
|
| 1683 |
+
"step": 5580
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 2.03,
|
| 1687 |
+
"learning_rate": 1.773297362850617e-07,
|
| 1688 |
+
"loss": 11.7563,
|
| 1689 |
+
"step": 5600
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 2.03,
|
| 1693 |
+
"learning_rate": 1.7506815508097754e-07,
|
| 1694 |
+
"loss": 11.5128,
|
| 1695 |
+
"step": 5620
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 2.04,
|
| 1699 |
+
"learning_rate": 1.728132914153771e-07,
|
| 1700 |
+
"loss": 11.604,
|
| 1701 |
+
"step": 5640
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 2.05,
|
| 1705 |
+
"learning_rate": 1.7056534743373617e-07,
|
| 1706 |
+
"loss": 11.4812,
|
| 1707 |
+
"step": 5660
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 2.05,
|
| 1711 |
+
"learning_rate": 1.6832452466119039e-07,
|
| 1712 |
+
"loss": 11.529,
|
| 1713 |
+
"step": 5680
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 2.06,
|
| 1717 |
+
"learning_rate": 1.6609102398446852e-07,
|
| 1718 |
+
"loss": 11.5573,
|
| 1719 |
+
"step": 5700
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 2.07,
|
| 1723 |
+
"learning_rate": 1.6386504563388343e-07,
|
| 1724 |
+
"loss": 11.5652,
|
| 1725 |
+
"step": 5720
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 2.08,
|
| 1729 |
+
"learning_rate": 1.6164678916538154e-07,
|
| 1730 |
+
"loss": 11.594,
|
| 1731 |
+
"step": 5740
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 2.08,
|
| 1735 |
+
"learning_rate": 1.5943645344265312e-07,
|
| 1736 |
+
"loss": 11.5252,
|
| 1737 |
+
"step": 5760
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 2.09,
|
| 1741 |
+
"learning_rate": 1.572342366193042e-07,
|
| 1742 |
+
"loss": 11.5946,
|
| 1743 |
+
"step": 5780
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 2.1,
|
| 1747 |
+
"learning_rate": 1.5504033612109246e-07,
|
| 1748 |
+
"loss": 11.5068,
|
| 1749 |
+
"step": 5800
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 2.1,
|
| 1753 |
+
"learning_rate": 1.528549486282284e-07,
|
| 1754 |
+
"loss": 11.5782,
|
| 1755 |
+
"step": 5820
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 2.11,
|
| 1759 |
+
"learning_rate": 1.5067827005774297e-07,
|
| 1760 |
+
"loss": 11.5785,
|
| 1761 |
+
"step": 5840
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 2.12,
|
| 1765 |
+
"learning_rate": 1.4851049554592403e-07,
|
| 1766 |
+
"loss": 11.4974,
|
| 1767 |
+
"step": 5860
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 2.13,
|
| 1771 |
+
"learning_rate": 1.4635181943082284e-07,
|
| 1772 |
+
"loss": 11.636,
|
| 1773 |
+
"step": 5880
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 2.13,
|
| 1777 |
+
"learning_rate": 1.442024352348314e-07,
|
| 1778 |
+
"loss": 11.5021,
|
| 1779 |
+
"step": 5900
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 2.14,
|
| 1783 |
+
"learning_rate": 1.42062535647334e-07,
|
| 1784 |
+
"loss": 11.4928,
|
| 1785 |
+
"step": 5920
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 2.15,
|
| 1789 |
+
"learning_rate": 1.3993231250743243e-07,
|
| 1790 |
+
"loss": 11.5453,
|
| 1791 |
+
"step": 5940
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 2.16,
|
| 1795 |
+
"learning_rate": 1.3781195678674817e-07,
|
| 1796 |
+
"loss": 11.5554,
|
| 1797 |
+
"step": 5960
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 2.16,
|
| 1801 |
+
"learning_rate": 1.3570165857230176e-07,
|
| 1802 |
+
"loss": 11.4993,
|
| 1803 |
+
"step": 5980
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 2.17,
|
| 1807 |
+
"learning_rate": 1.3360160704947221e-07,
|
| 1808 |
+
"loss": 11.5695,
|
| 1809 |
+
"step": 6000
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 2.18,
|
| 1813 |
+
"learning_rate": 1.3151199048503625e-07,
|
| 1814 |
+
"loss": 11.5464,
|
| 1815 |
+
"step": 6020
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 2.18,
|
| 1819 |
+
"learning_rate": 1.2943299621029076e-07,
|
| 1820 |
+
"loss": 11.5437,
|
| 1821 |
+
"step": 6040
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 2.19,
|
| 1825 |
+
"learning_rate": 1.273648106042587e-07,
|
| 1826 |
+
"loss": 11.6056,
|
| 1827 |
+
"step": 6060
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 2.2,
|
| 1831 |
+
"learning_rate": 1.2530761907698062e-07,
|
| 1832 |
+
"loss": 11.5232,
|
| 1833 |
+
"step": 6080
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 2.21,
|
| 1837 |
+
"learning_rate": 1.2326160605289272e-07,
|
| 1838 |
+
"loss": 11.5292,
|
| 1839 |
+
"step": 6100
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 2.21,
|
| 1843 |
+
"learning_rate": 1.2122695495429347e-07,
|
| 1844 |
+
"loss": 11.4655,
|
| 1845 |
+
"step": 6120
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 2.22,
|
| 1849 |
+
"learning_rate": 1.1920384818490018e-07,
|
| 1850 |
+
"loss": 11.5054,
|
| 1851 |
+
"step": 6140
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 2.23,
|
| 1855 |
+
"learning_rate": 1.1719246711349664e-07,
|
| 1856 |
+
"loss": 11.484,
|
| 1857 |
+
"step": 6160
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 2.24,
|
| 1861 |
+
"learning_rate": 1.151929920576737e-07,
|
| 1862 |
+
"loss": 11.4898,
|
| 1863 |
+
"step": 6180
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 2.24,
|
| 1867 |
+
"learning_rate": 1.1320560226766396e-07,
|
| 1868 |
+
"loss": 11.5315,
|
| 1869 |
+
"step": 6200
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 2.25,
|
| 1873 |
+
"learning_rate": 1.112304759102723e-07,
|
| 1874 |
+
"loss": 11.5258,
|
| 1875 |
+
"step": 6220
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 2.26,
|
| 1879 |
+
"learning_rate": 1.0926779005290365e-07,
|
| 1880 |
+
"loss": 11.4383,
|
| 1881 |
+
"step": 6240
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 2.26,
|
| 1885 |
+
"learning_rate": 1.0731772064768876e-07,
|
| 1886 |
+
"loss": 11.5047,
|
| 1887 |
+
"step": 6260
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 2.27,
|
| 1891 |
+
"learning_rate": 1.0538044251571057e-07,
|
| 1892 |
+
"loss": 11.5129,
|
| 1893 |
+
"step": 6280
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 2.28,
|
| 1897 |
+
"learning_rate": 1.0345612933133166e-07,
|
| 1898 |
+
"loss": 11.516,
|
| 1899 |
+
"step": 6300
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 2.29,
|
| 1903 |
+
"learning_rate": 1.0154495360662463e-07,
|
| 1904 |
+
"loss": 11.589,
|
| 1905 |
+
"step": 6320
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 2.29,
|
| 1909 |
+
"learning_rate": 9.964708667590655e-08,
|
| 1910 |
+
"loss": 11.5469,
|
| 1911 |
+
"step": 6340
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 2.3,
|
| 1915 |
+
"learning_rate": 9.776269868037928e-08,
|
| 1916 |
+
"loss": 11.5621,
|
| 1917 |
+
"step": 6360
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 2.31,
|
| 1921 |
+
"learning_rate": 9.589195855287623e-08,
|
| 1922 |
+
"loss": 11.5689,
|
| 1923 |
+
"step": 6380
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 2.31,
|
| 1927 |
+
"learning_rate": 9.403503400271798e-08,
|
| 1928 |
+
"loss": 11.4723,
|
| 1929 |
+
"step": 6400
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 2.32,
|
| 1933 |
+
"learning_rate": 9.219209150067725e-08,
|
| 1934 |
+
"loss": 11.5355,
|
| 1935 |
+
"step": 6420
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 2.33,
|
| 1939 |
+
"learning_rate": 9.036329626405506e-08,
|
| 1940 |
+
"loss": 11.402,
|
| 1941 |
+
"step": 6440
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 2.34,
|
| 1945 |
+
"learning_rate": 8.854881224186933e-08,
|
| 1946 |
+
"loss": 11.5974,
|
| 1947 |
+
"step": 6460
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 2.34,
|
| 1951 |
+
"learning_rate": 8.67488021001569e-08,
|
| 1952 |
+
"loss": 11.6007,
|
| 1953 |
+
"step": 6480
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 2.35,
|
| 1957 |
+
"learning_rate": 8.496342720739078e-08,
|
| 1958 |
+
"loss": 11.4786,
|
| 1959 |
+
"step": 6500
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 2.36,
|
| 1963 |
+
"learning_rate": 8.319284762001371e-08,
|
| 1964 |
+
"loss": 11.557,
|
| 1965 |
+
"step": 6520
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 2.37,
|
| 1969 |
+
"learning_rate": 8.143722206808959e-08,
|
| 1970 |
+
"loss": 11.4715,
|
| 1971 |
+
"step": 6540
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 2.37,
|
| 1975 |
+
"learning_rate": 7.969670794107294e-08,
|
| 1976 |
+
"loss": 11.5074,
|
| 1977 |
+
"step": 6560
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 2.38,
|
| 1981 |
+
"learning_rate": 7.797146127369966e-08,
|
| 1982 |
+
"loss": 11.533,
|
| 1983 |
+
"step": 6580
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 2.39,
|
| 1987 |
+
"learning_rate": 7.626163673199848e-08,
|
| 1988 |
+
"loss": 11.4245,
|
| 1989 |
+
"step": 6600
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 2.39,
|
| 1993 |
+
"learning_rate": 7.456738759942549e-08,
|
| 1994 |
+
"loss": 11.6053,
|
| 1995 |
+
"step": 6620
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 2.4,
|
| 1999 |
+
"learning_rate": 7.288886576312242e-08,
|
| 2000 |
+
"loss": 11.6554,
|
| 2001 |
+
"step": 6640
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 2.41,
|
| 2005 |
+
"learning_rate": 7.122622170030016e-08,
|
| 2006 |
+
"loss": 11.604,
|
| 2007 |
+
"step": 6660
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 2.42,
|
| 2011 |
+
"learning_rate": 6.957960446474873e-08,
|
| 2012 |
+
"loss": 11.5528,
|
| 2013 |
+
"step": 6680
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 2.42,
|
| 2017 |
+
"learning_rate": 6.794916167347473e-08,
|
| 2018 |
+
"loss": 11.5293,
|
| 2019 |
+
"step": 6700
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 2.43,
|
| 2023 |
+
"learning_rate": 6.633503949346775e-08,
|
| 2024 |
+
"loss": 11.4711,
|
| 2025 |
+
"step": 6720
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 2.44,
|
| 2029 |
+
"learning_rate": 6.473738262859668e-08,
|
| 2030 |
+
"loss": 11.5405,
|
| 2031 |
+
"step": 6740
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 2.44,
|
| 2035 |
+
"learning_rate": 6.315633430663708e-08,
|
| 2036 |
+
"loss": 11.4586,
|
| 2037 |
+
"step": 6760
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 2.45,
|
| 2041 |
+
"learning_rate": 6.159203626643137e-08,
|
| 2042 |
+
"loss": 11.547,
|
| 2043 |
+
"step": 6780
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 2.46,
|
| 2047 |
+
"learning_rate": 6.004462874518157e-08,
|
| 2048 |
+
"loss": 11.5252,
|
| 2049 |
+
"step": 6800
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 2.47,
|
| 2053 |
+
"learning_rate": 5.8514250465877826e-08,
|
| 2054 |
+
"loss": 11.5623,
|
| 2055 |
+
"step": 6820
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 2.47,
|
| 2059 |
+
"learning_rate": 5.700103862486158e-08,
|
| 2060 |
+
"loss": 11.5316,
|
| 2061 |
+
"step": 6840
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 2.48,
|
| 2065 |
+
"learning_rate": 5.550512887952641e-08,
|
| 2066 |
+
"loss": 11.5136,
|
| 2067 |
+
"step": 6860
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 2.49,
|
| 2071 |
+
"learning_rate": 5.4026655336156384e-08,
|
| 2072 |
+
"loss": 11.6511,
|
| 2073 |
+
"step": 6880
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 2.5,
|
| 2077 |
+
"learning_rate": 5.2565750537903716e-08,
|
| 2078 |
+
"loss": 11.5241,
|
| 2079 |
+
"step": 6900
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 2.5,
|
| 2083 |
+
"learning_rate": 5.112254545290634e-08,
|
| 2084 |
+
"loss": 11.5878,
|
| 2085 |
+
"step": 6920
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 2.51,
|
| 2089 |
+
"learning_rate": 4.96971694625469e-08,
|
| 2090 |
+
"loss": 11.4855,
|
| 2091 |
+
"step": 6940
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 2.52,
|
| 2095 |
+
"learning_rate": 4.8289750349853834e-08,
|
| 2096 |
+
"loss": 11.5632,
|
| 2097 |
+
"step": 6960
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 2.52,
|
| 2101 |
+
"learning_rate": 4.6900414288045726e-08,
|
| 2102 |
+
"loss": 11.6167,
|
| 2103 |
+
"step": 6980
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 2.53,
|
| 2107 |
+
"learning_rate": 4.552928582922022e-08,
|
| 2108 |
+
"loss": 11.5751,
|
| 2109 |
+
"step": 7000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 2.54,
|
| 2113 |
+
"learning_rate": 4.4176487893187956e-08,
|
| 2114 |
+
"loss": 11.5228,
|
| 2115 |
+
"step": 7020
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 2.55,
|
| 2119 |
+
"learning_rate": 4.284214175645301e-08,
|
| 2120 |
+
"loss": 11.624,
|
| 2121 |
+
"step": 7040
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 2.55,
|
| 2125 |
+
"learning_rate": 4.152636704134055e-08,
|
| 2126 |
+
"loss": 11.4761,
|
| 2127 |
+
"step": 7060
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 2.56,
|
| 2131 |
+
"learning_rate": 4.022928170527315e-08,
|
| 2132 |
+
"loss": 11.579,
|
| 2133 |
+
"step": 7080
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 2.57,
|
| 2137 |
+
"learning_rate": 3.8951002030195636e-08,
|
| 2138 |
+
"loss": 11.5007,
|
| 2139 |
+
"step": 7100
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 2.58,
|
| 2143 |
+
"learning_rate": 3.7691642612150826e-08,
|
| 2144 |
+
"loss": 11.6567,
|
| 2145 |
+
"step": 7120
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 2.58,
|
| 2149 |
+
"learning_rate": 3.6451316351006153e-08,
|
| 2150 |
+
"loss": 11.517,
|
| 2151 |
+
"step": 7140
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 2.59,
|
| 2155 |
+
"learning_rate": 3.5230134440332236e-08,
|
| 2156 |
+
"loss": 11.5614,
|
| 2157 |
+
"step": 7160
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 2.6,
|
| 2161 |
+
"learning_rate": 3.402820635743464e-08,
|
| 2162 |
+
"loss": 11.5687,
|
| 2163 |
+
"step": 7180
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 2.6,
|
| 2167 |
+
"learning_rate": 3.284563985353925e-08,
|
| 2168 |
+
"loss": 11.4573,
|
| 2169 |
+
"step": 7200
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 2.61,
|
| 2173 |
+
"learning_rate": 3.168254094413264e-08,
|
| 2174 |
+
"loss": 11.5399,
|
| 2175 |
+
"step": 7220
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 2.62,
|
| 2179 |
+
"learning_rate": 3.053901389945798e-08,
|
| 2180 |
+
"loss": 11.5103,
|
| 2181 |
+
"step": 7240
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 2.63,
|
| 2185 |
+
"learning_rate": 2.94151612351671e-08,
|
| 2186 |
+
"loss": 11.5896,
|
| 2187 |
+
"step": 7260
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 2.63,
|
| 2191 |
+
"learning_rate": 2.8311083703130346e-08,
|
| 2192 |
+
"loss": 11.4998,
|
| 2193 |
+
"step": 7280
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 2.64,
|
| 2197 |
+
"learning_rate": 2.722688028240419e-08,
|
| 2198 |
+
"loss": 11.5447,
|
| 2199 |
+
"step": 7300
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 2.65,
|
| 2203 |
+
"learning_rate": 2.616264817035793e-08,
|
| 2204 |
+
"loss": 11.6846,
|
| 2205 |
+
"step": 7320
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 2.65,
|
| 2209 |
+
"learning_rate": 2.5118482773960002e-08,
|
| 2210 |
+
"loss": 11.5462,
|
| 2211 |
+
"step": 7340
|
| 2212 |
+
},
|
| 2213 |
+
{
|
| 2214 |
+
"epoch": 2.66,
|
| 2215 |
+
"learning_rate": 2.4094477701225163e-08,
|
| 2216 |
+
"loss": 11.5512,
|
| 2217 |
+
"step": 7360
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 2.67,
|
| 2221 |
+
"learning_rate": 2.30907247528222e-08,
|
| 2222 |
+
"loss": 11.4978,
|
| 2223 |
+
"step": 7380
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 2.68,
|
| 2227 |
+
"learning_rate": 2.2107313913844432e-08,
|
| 2228 |
+
"loss": 11.4415,
|
| 2229 |
+
"step": 7400
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 2.68,
|
| 2233 |
+
"learning_rate": 2.114433334574267e-08,
|
| 2234 |
+
"loss": 11.5456,
|
| 2235 |
+
"step": 7420
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 2.69,
|
| 2239 |
+
"learning_rate": 2.0201869378421497e-08,
|
| 2240 |
+
"loss": 11.5887,
|
| 2241 |
+
"step": 7440
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 2.7,
|
| 2245 |
+
"learning_rate": 1.9280006502500056e-08,
|
| 2246 |
+
"loss": 11.6359,
|
| 2247 |
+
"step": 7460
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 2.71,
|
| 2251 |
+
"learning_rate": 1.8378827361737543e-08,
|
| 2252 |
+
"loss": 11.6285,
|
| 2253 |
+
"step": 7480
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 2.71,
|
| 2257 |
+
"learning_rate": 1.749841274562422e-08,
|
| 2258 |
+
"loss": 11.74,
|
| 2259 |
+
"step": 7500
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 2.72,
|
| 2263 |
+
"learning_rate": 1.6638841582138908e-08,
|
| 2264 |
+
"loss": 11.5993,
|
| 2265 |
+
"step": 7520
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 2.73,
|
| 2269 |
+
"learning_rate": 1.580019093067303e-08,
|
| 2270 |
+
"loss": 11.6595,
|
| 2271 |
+
"step": 7540
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 2.73,
|
| 2275 |
+
"learning_rate": 1.4982535975122474e-08,
|
| 2276 |
+
"loss": 11.6267,
|
| 2277 |
+
"step": 7560
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 2.74,
|
| 2281 |
+
"learning_rate": 1.4185950017147369e-08,
|
| 2282 |
+
"loss": 11.5287,
|
| 2283 |
+
"step": 7580
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 2.75,
|
| 2287 |
+
"learning_rate": 1.3410504469600708e-08,
|
| 2288 |
+
"loss": 11.6259,
|
| 2289 |
+
"step": 7600
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 2.76,
|
| 2293 |
+
"learning_rate": 1.2656268850126411e-08,
|
| 2294 |
+
"loss": 11.6673,
|
| 2295 |
+
"step": 7620
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 2.76,
|
| 2299 |
+
"learning_rate": 1.1923310774926948e-08,
|
| 2300 |
+
"loss": 11.5539,
|
| 2301 |
+
"step": 7640
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 2.77,
|
| 2305 |
+
"learning_rate": 1.1211695952701717e-08,
|
| 2306 |
+
"loss": 11.5689,
|
| 2307 |
+
"step": 7660
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 2.78,
|
| 2311 |
+
"learning_rate": 1.0521488178756532e-08,
|
| 2312 |
+
"loss": 11.471,
|
| 2313 |
+
"step": 7680
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 2.78,
|
| 2317 |
+
"learning_rate": 9.852749329284093e-09,
|
| 2318 |
+
"loss": 11.5166,
|
| 2319 |
+
"step": 7700
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 2.79,
|
| 2323 |
+
"learning_rate": 9.20553935581725e-09,
|
| 2324 |
+
"loss": 11.5288,
|
| 2325 |
+
"step": 7720
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 2.8,
|
| 2329 |
+
"learning_rate": 8.57991627985416e-09,
|
| 2330 |
+
"loss": 11.6492,
|
| 2331 |
+
"step": 7740
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 2.81,
|
| 2335 |
+
"learning_rate": 7.975936187656852e-09,
|
| 2336 |
+
"loss": 11.4599,
|
| 2337 |
+
"step": 7760
|
| 2338 |
+
},
|
| 2339 |
+
{
|
| 2340 |
+
"epoch": 2.81,
|
| 2341 |
+
"learning_rate": 7.393653225223145e-09,
|
| 2342 |
+
"loss": 11.5627,
|
| 2343 |
+
"step": 7780
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 2.82,
|
| 2347 |
+
"learning_rate": 6.833119593432607e-09,
|
| 2348 |
+
"loss": 11.6969,
|
| 2349 |
+
"step": 7800
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.83,
|
| 2353 |
+
"learning_rate": 6.294385543366642e-09,
|
| 2354 |
+
"loss": 11.651,
|
| 2355 |
+
"step": 7820
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 2.84,
|
| 2359 |
+
"learning_rate": 5.777499371803745e-09,
|
| 2360 |
+
"loss": 11.5684,
|
| 2361 |
+
"step": 7840
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 2.84,
|
| 2365 |
+
"learning_rate": 5.282507416889625e-09,
|
| 2366 |
+
"loss": 11.6632,
|
| 2367 |
+
"step": 7860
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 2.85,
|
| 2371 |
+
"learning_rate": 4.809454053983087e-09,
|
| 2372 |
+
"loss": 11.4329,
|
| 2373 |
+
"step": 7880
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 2.86,
|
| 2377 |
+
"learning_rate": 4.358381691677931e-09,
|
| 2378 |
+
"loss": 11.6679,
|
| 2379 |
+
"step": 7900
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 2.86,
|
| 2383 |
+
"learning_rate": 3.929330768000949e-09,
|
| 2384 |
+
"loss": 11.479,
|
| 2385 |
+
"step": 7920
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 2.87,
|
| 2389 |
+
"learning_rate": 3.5223397467867565e-09,
|
| 2390 |
+
"loss": 11.6362,
|
| 2391 |
+
"step": 7940
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.88,
|
| 2395 |
+
"learning_rate": 3.1374451142295867e-09,
|
| 2396 |
+
"loss": 11.7412,
|
| 2397 |
+
"step": 7960
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 2.89,
|
| 2401 |
+
"learning_rate": 2.774681375612292e-09,
|
| 2402 |
+
"loss": 11.5321,
|
| 2403 |
+
"step": 7980
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 2.89,
|
| 2407 |
+
"learning_rate": 2.4340810522131274e-09,
|
| 2408 |
+
"loss": 11.5223,
|
| 2409 |
+
"step": 8000
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 2.9,
|
| 2413 |
+
"learning_rate": 2.115674678390078e-09,
|
| 2414 |
+
"loss": 11.6288,
|
| 2415 |
+
"step": 8020
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 2.91,
|
| 2419 |
+
"learning_rate": 1.8194907988436093e-09,
|
| 2420 |
+
"loss": 11.5262,
|
| 2421 |
+
"step": 8040
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 2.92,
|
| 2425 |
+
"learning_rate": 1.5455559660577377e-09,
|
| 2426 |
+
"loss": 11.5915,
|
| 2427 |
+
"step": 8060
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 2.92,
|
| 2431 |
+
"learning_rate": 1.293894737919382e-09,
|
| 2432 |
+
"loss": 11.5543,
|
| 2433 |
+
"step": 8080
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.93,
|
| 2437 |
+
"learning_rate": 1.0645296755171229e-09,
|
| 2438 |
+
"loss": 11.5801,
|
| 2439 |
+
"step": 8100
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 2.94,
|
| 2443 |
+
"learning_rate": 8.574813411183213e-10,
|
| 2444 |
+
"loss": 11.6047,
|
| 2445 |
+
"step": 8120
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 2.94,
|
| 2449 |
+
"learning_rate": 6.727682963259818e-10,
|
| 2450 |
+
"loss": 11.5458,
|
| 2451 |
+
"step": 8140
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 2.95,
|
| 2455 |
+
"learning_rate": 5.10407100414556e-10,
|
| 2456 |
+
"loss": 11.5238,
|
| 2457 |
+
"step": 8160
|
| 2458 |
+
},
|
| 2459 |
+
{
|
| 2460 |
+
"epoch": 2.96,
|
| 2461 |
+
"learning_rate": 3.704123088455191e-10,
|
| 2462 |
+
"loss": 11.5343,
|
| 2463 |
+
"step": 8180
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 2.97,
|
| 2467 |
+
"learning_rate": 2.5279647196246867e-10,
|
| 2468 |
+
"loss": 11.5845,
|
| 2469 |
+
"step": 8200
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 2.97,
|
| 2473 |
+
"learning_rate": 1.5757013386599738e-10,
|
| 2474 |
+
"loss": 11.6215,
|
| 2475 |
+
"step": 8220
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.98,
|
| 2479 |
+
"learning_rate": 8.474183146844894e-11,
|
| 2480 |
+
"loss": 11.5309,
|
| 2481 |
+
"step": 8240
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 2.99,
|
| 2485 |
+
"learning_rate": 3.431809372853034e-11,
|
| 2486 |
+
"loss": 11.528,
|
| 2487 |
+
"step": 8260
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 2.99,
|
| 2491 |
+
"learning_rate": 6.3034410661133574e-12,
|
| 2492 |
+
"loss": 11.5548,
|
| 2493 |
+
"step": 8280
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 3.0,
|
| 2497 |
+
"step": 8295,
|
| 2498 |
+
"total_flos": 2.451174829221888e+17,
|
| 2499 |
+
"train_loss": 12.036643933186408,
|
| 2500 |
+
"train_runtime": 6501.8144,
|
| 2501 |
+
"train_samples_per_second": 2.552,
|
| 2502 |
+
"train_steps_per_second": 1.276
|
| 2503 |
}
|
| 2504 |
],
|
| 2505 |
"logging_steps": 20,
|
| 2506 |
+
"max_steps": 8295,
|
| 2507 |
"num_input_tokens_seen": 0,
|
| 2508 |
+
"num_train_epochs": 3,
|
| 2509 |
"save_steps": 100,
|
| 2510 |
+
"total_flos": 2.451174829221888e+17,
|
| 2511 |
"train_batch_size": 2,
|
| 2512 |
"trial_name": null,
|
| 2513 |
"trial_params": null
|