End of training
Browse files- README.md +28 -68
- config.json +1 -1
- pytorch_model.bin +1 -1
- special_tokens_map.json +102 -0
- trainer_state.json +945 -1345
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
-
base_model:
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
model-index:
|
|
@@ -13,11 +13,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 13 |
|
| 14 |
# text-normalization-ru-new
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
-
- Loss: 0.
|
| 19 |
- Mean Distance: 0
|
| 20 |
-
- Max Distance:
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
@@ -36,79 +36,39 @@ More information needed
|
|
| 36 |
### Training hyperparameters
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
-
- learning_rate: 0.
|
| 40 |
- train_batch_size: 30
|
| 41 |
- eval_batch_size: 30
|
| 42 |
- seed: 42
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_ratio: 0.1
|
| 46 |
-
- num_epochs:
|
| 47 |
|
| 48 |
### Training results
|
| 49 |
|
| 50 |
-
| Training Loss | Epoch | Step
|
| 51 |
-
|:-------------:|:-----:|:-----
|
| 52 |
-
| 0.
|
| 53 |
-
| 0.
|
| 54 |
-
| 0.
|
| 55 |
-
| 0.
|
| 56 |
-
| 0.
|
| 57 |
-
| 0.
|
| 58 |
-
| 0.
|
| 59 |
-
| 0.
|
| 60 |
-
| 0.
|
| 61 |
-
| 0.
|
| 62 |
-
| 0.
|
| 63 |
-
| 0.
|
| 64 |
-
| 0.
|
| 65 |
-
| 0.
|
| 66 |
-
| 0.
|
| 67 |
-
| 0.
|
| 68 |
-
| 0.
|
| 69 |
-
| 0.
|
| 70 |
-
| 0.
|
| 71 |
-
| 0.
|
| 72 |
-
| 0.007 | 21.0 | 70014 | 0.0309 | 0 | 13 |
|
| 73 |
-
| 0.0066 | 22.0 | 73348 | 0.0306 | 0 | 11 |
|
| 74 |
-
| 0.0056 | 23.0 | 76682 | 0.0287 | 0 | 10 |
|
| 75 |
-
| 0.0053 | 24.0 | 80016 | 0.0312 | 0 | 12 |
|
| 76 |
-
| 0.0049 | 25.0 | 83350 | 0.0276 | 0 | 11 |
|
| 77 |
-
| 0.0053 | 26.0 | 86684 | 0.0308 | 0 | 10 |
|
| 78 |
-
| 0.0041 | 27.0 | 90018 | 0.0279 | 0 | 10 |
|
| 79 |
-
| 0.0041 | 28.0 | 93352 | 0.0292 | 0 | 11 |
|
| 80 |
-
| 0.0037 | 29.0 | 96686 | 0.0306 | 0 | 11 |
|
| 81 |
-
| 0.0035 | 30.0 | 100020 | 0.0272 | 0 | 12 |
|
| 82 |
-
| 0.0032 | 31.0 | 103354 | 0.0255 | 0 | 9 |
|
| 83 |
-
| 0.0031 | 32.0 | 106688 | 0.0293 | 0 | 10 |
|
| 84 |
-
| 0.0029 | 33.0 | 110022 | 0.0300 | 0 | 13 |
|
| 85 |
-
| 0.0026 | 34.0 | 113356 | 0.0305 | 0 | 11 |
|
| 86 |
-
| 0.0024 | 35.0 | 116690 | 0.0273 | 0 | 9 |
|
| 87 |
-
| 0.0023 | 36.0 | 120024 | 0.0284 | 0 | 10 |
|
| 88 |
-
| 0.0022 | 37.0 | 123358 | 0.0313 | 0 | 13 |
|
| 89 |
-
| 0.002 | 38.0 | 126692 | 0.0341 | 0 | 13 |
|
| 90 |
-
| 0.0017 | 39.0 | 130026 | 0.0301 | 0 | 13 |
|
| 91 |
-
| 0.0017 | 40.0 | 133360 | 0.0330 | 0 | 11 |
|
| 92 |
-
| 0.0016 | 41.0 | 136694 | 0.0344 | 0 | 11 |
|
| 93 |
-
| 0.0014 | 42.0 | 140028 | 0.0337 | 0 | 10 |
|
| 94 |
-
| 0.0013 | 43.0 | 143362 | 0.0292 | 0 | 12 |
|
| 95 |
-
| 0.0012 | 44.0 | 146696 | 0.0339 | 0 | 11 |
|
| 96 |
-
| 0.0012 | 45.0 | 150030 | 0.0330 | 0 | 11 |
|
| 97 |
-
| 0.001 | 46.0 | 153364 | 0.0307 | 0 | 11 |
|
| 98 |
-
| 0.001 | 47.0 | 156698 | 0.0330 | 0 | 10 |
|
| 99 |
-
| 0.0009 | 48.0 | 160032 | 0.0338 | 0 | 11 |
|
| 100 |
-
| 0.0009 | 49.0 | 163366 | 0.0288 | 0 | 10 |
|
| 101 |
-
| 0.0008 | 50.0 | 166700 | 0.0256 | 0 | 10 |
|
| 102 |
-
| 0.0007 | 51.0 | 170034 | 0.0284 | 0 | 11 |
|
| 103 |
-
| 0.0006 | 52.0 | 173368 | 0.0342 | 0 | 10 |
|
| 104 |
-
| 0.0006 | 53.0 | 176702 | 0.0312 | 0 | 10 |
|
| 105 |
-
| 0.0005 | 54.0 | 180036 | 0.0326 | 0 | 10 |
|
| 106 |
-
| 0.0005 | 55.0 | 183370 | 0.0304 | 0 | 11 |
|
| 107 |
-
| 0.0005 | 56.0 | 186704 | 0.0300 | 0 | 11 |
|
| 108 |
-
| 0.0004 | 57.0 | 190038 | 0.0313 | 0 | 11 |
|
| 109 |
-
| 0.0003 | 58.0 | 193372 | 0.0321 | 0 | 11 |
|
| 110 |
-
| 0.0003 | 59.0 | 196706 | 0.0316 | 0 | 10 |
|
| 111 |
-
| 0.0004 | 60.0 | 200040 | 0.0318 | 0 | 11 |
|
| 112 |
|
| 113 |
|
| 114 |
### Framework versions
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
+
base_model: alexue4/text-normalization-ru-new
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
model-index:
|
|
|
|
| 13 |
|
| 14 |
# text-normalization-ru-new
|
| 15 |
|
| 16 |
+
This model is a fine-tuned version of [alexue4/text-normalization-ru-new](https://huggingface.co/alexue4/text-normalization-ru-new) on an unknown dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 0.0007
|
| 19 |
- Mean Distance: 0
|
| 20 |
+
- Max Distance: 3
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 36 |
### Training hyperparameters
|
| 37 |
|
| 38 |
The following hyperparameters were used during training:
|
| 39 |
+
- learning_rate: 0.0001
|
| 40 |
- train_batch_size: 30
|
| 41 |
- eval_batch_size: 30
|
| 42 |
- seed: 42
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_ratio: 0.1
|
| 46 |
+
- num_epochs: 20
|
| 47 |
|
| 48 |
### Training results
|
| 49 |
|
| 50 |
+
| Training Loss | Epoch | Step | Validation Loss | Mean Distance | Max Distance |
|
| 51 |
+
|:-------------:|:-----:|:-----:|:---------------:|:-------------:|:------------:|
|
| 52 |
+
| 0.0028 | 1.0 | 3443 | 0.0015 | 0 | 3 |
|
| 53 |
+
| 0.0019 | 2.0 | 6886 | 0.0009 | 0 | 3 |
|
| 54 |
+
| 0.0016 | 3.0 | 10329 | 0.0013 | 0 | 3 |
|
| 55 |
+
| 0.0013 | 4.0 | 13772 | 0.0008 | 0 | 1 |
|
| 56 |
+
| 0.0012 | 5.0 | 17215 | 0.0011 | 0 | 3 |
|
| 57 |
+
| 0.0009 | 6.0 | 20658 | 0.0009 | 0 | 3 |
|
| 58 |
+
| 0.0008 | 7.0 | 24101 | 0.0011 | 0 | 3 |
|
| 59 |
+
| 0.0007 | 8.0 | 27544 | 0.0010 | 0 | 3 |
|
| 60 |
+
| 0.0006 | 9.0 | 30987 | 0.0012 | 0 | 3 |
|
| 61 |
+
| 0.0006 | 10.0 | 34430 | 0.0008 | 0 | 3 |
|
| 62 |
+
| 0.0006 | 11.0 | 37873 | 0.0005 | 0 | 0 |
|
| 63 |
+
| 0.0005 | 12.0 | 41316 | 0.0007 | 0 | 1 |
|
| 64 |
+
| 0.0004 | 13.0 | 44759 | 0.0007 | 0 | 0 |
|
| 65 |
+
| 0.0006 | 14.0 | 48202 | 0.0011 | 0 | 3 |
|
| 66 |
+
| 0.0005 | 15.0 | 51645 | 0.0008 | 0 | 3 |
|
| 67 |
+
| 0.0005 | 16.0 | 55088 | 0.0008 | 0 | 3 |
|
| 68 |
+
| 0.0005 | 17.0 | 58531 | 0.0008 | 0 | 3 |
|
| 69 |
+
| 0.0004 | 18.0 | 61974 | 0.0007 | 0 | 3 |
|
| 70 |
+
| 0.0004 | 19.0 | 65417 | 0.0007 | 0 | 3 |
|
| 71 |
+
| 0.0005 | 20.0 | 68860 | 0.0007 | 0 | 3 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
|
| 73 |
|
| 74 |
### Framework versions
|
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"architectures": [
|
| 4 |
"T5ForConditionalGeneration"
|
| 5 |
],
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "alexue4/text-normalization-ru-new",
|
| 3 |
"architectures": [
|
| 4 |
"T5ForConditionalGeneration"
|
| 5 |
],
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 258643461
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:426d8940304254b400c7865e3f92b2ed60ec87d2cec52df3644476d19c0451e2
|
| 3 |
size 258643461
|
special_tokens_map.json
CHANGED
|
@@ -1,4 +1,106 @@
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"eos_token": "</s>",
|
| 3 |
"pad_token": "<pad>",
|
| 4 |
"unk_token": "<unk>"
|
|
|
|
| 1 |
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<extra_id_0>",
|
| 4 |
+
"<extra_id_1>",
|
| 5 |
+
"<extra_id_2>",
|
| 6 |
+
"<extra_id_3>",
|
| 7 |
+
"<extra_id_4>",
|
| 8 |
+
"<extra_id_5>",
|
| 9 |
+
"<extra_id_6>",
|
| 10 |
+
"<extra_id_7>",
|
| 11 |
+
"<extra_id_8>",
|
| 12 |
+
"<extra_id_9>",
|
| 13 |
+
"<extra_id_10>",
|
| 14 |
+
"<extra_id_11>",
|
| 15 |
+
"<extra_id_12>",
|
| 16 |
+
"<extra_id_13>",
|
| 17 |
+
"<extra_id_14>",
|
| 18 |
+
"<extra_id_15>",
|
| 19 |
+
"<extra_id_16>",
|
| 20 |
+
"<extra_id_17>",
|
| 21 |
+
"<extra_id_18>",
|
| 22 |
+
"<extra_id_19>",
|
| 23 |
+
"<extra_id_20>",
|
| 24 |
+
"<extra_id_21>",
|
| 25 |
+
"<extra_id_22>",
|
| 26 |
+
"<extra_id_23>",
|
| 27 |
+
"<extra_id_24>",
|
| 28 |
+
"<extra_id_25>",
|
| 29 |
+
"<extra_id_26>",
|
| 30 |
+
"<extra_id_27>",
|
| 31 |
+
"<extra_id_28>",
|
| 32 |
+
"<extra_id_29>",
|
| 33 |
+
"<extra_id_30>",
|
| 34 |
+
"<extra_id_31>",
|
| 35 |
+
"<extra_id_32>",
|
| 36 |
+
"<extra_id_33>",
|
| 37 |
+
"<extra_id_34>",
|
| 38 |
+
"<extra_id_35>",
|
| 39 |
+
"<extra_id_36>",
|
| 40 |
+
"<extra_id_37>",
|
| 41 |
+
"<extra_id_38>",
|
| 42 |
+
"<extra_id_39>",
|
| 43 |
+
"<extra_id_40>",
|
| 44 |
+
"<extra_id_41>",
|
| 45 |
+
"<extra_id_42>",
|
| 46 |
+
"<extra_id_43>",
|
| 47 |
+
"<extra_id_44>",
|
| 48 |
+
"<extra_id_45>",
|
| 49 |
+
"<extra_id_46>",
|
| 50 |
+
"<extra_id_47>",
|
| 51 |
+
"<extra_id_48>",
|
| 52 |
+
"<extra_id_49>",
|
| 53 |
+
"<extra_id_50>",
|
| 54 |
+
"<extra_id_51>",
|
| 55 |
+
"<extra_id_52>",
|
| 56 |
+
"<extra_id_53>",
|
| 57 |
+
"<extra_id_54>",
|
| 58 |
+
"<extra_id_55>",
|
| 59 |
+
"<extra_id_56>",
|
| 60 |
+
"<extra_id_57>",
|
| 61 |
+
"<extra_id_58>",
|
| 62 |
+
"<extra_id_59>",
|
| 63 |
+
"<extra_id_60>",
|
| 64 |
+
"<extra_id_61>",
|
| 65 |
+
"<extra_id_62>",
|
| 66 |
+
"<extra_id_63>",
|
| 67 |
+
"<extra_id_64>",
|
| 68 |
+
"<extra_id_65>",
|
| 69 |
+
"<extra_id_66>",
|
| 70 |
+
"<extra_id_67>",
|
| 71 |
+
"<extra_id_68>",
|
| 72 |
+
"<extra_id_69>",
|
| 73 |
+
"<extra_id_70>",
|
| 74 |
+
"<extra_id_71>",
|
| 75 |
+
"<extra_id_72>",
|
| 76 |
+
"<extra_id_73>",
|
| 77 |
+
"<extra_id_74>",
|
| 78 |
+
"<extra_id_75>",
|
| 79 |
+
"<extra_id_76>",
|
| 80 |
+
"<extra_id_77>",
|
| 81 |
+
"<extra_id_78>",
|
| 82 |
+
"<extra_id_79>",
|
| 83 |
+
"<extra_id_80>",
|
| 84 |
+
"<extra_id_81>",
|
| 85 |
+
"<extra_id_82>",
|
| 86 |
+
"<extra_id_83>",
|
| 87 |
+
"<extra_id_84>",
|
| 88 |
+
"<extra_id_85>",
|
| 89 |
+
"<extra_id_86>",
|
| 90 |
+
"<extra_id_87>",
|
| 91 |
+
"<extra_id_88>",
|
| 92 |
+
"<extra_id_89>",
|
| 93 |
+
"<extra_id_90>",
|
| 94 |
+
"<extra_id_91>",
|
| 95 |
+
"<extra_id_92>",
|
| 96 |
+
"<extra_id_93>",
|
| 97 |
+
"<extra_id_94>",
|
| 98 |
+
"<extra_id_95>",
|
| 99 |
+
"<extra_id_96>",
|
| 100 |
+
"<extra_id_97>",
|
| 101 |
+
"<extra_id_98>",
|
| 102 |
+
"<extra_id_99>"
|
| 103 |
+
],
|
| 104 |
"eos_token": "</s>",
|
| 105 |
"pad_token": "<pad>",
|
| 106 |
"unk_token": "<unk>"
|
trainer_state.json
CHANGED
|
@@ -1,1828 +1,1428 @@
|
|
| 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": 0.0,
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"loss":
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
-
"learning_rate": 5.
|
| 20 |
-
"loss":
|
| 21 |
-
"step":
|
| 22 |
-
},
|
| 23 |
-
{
|
| 24 |
-
"epoch": 0.6,
|
| 25 |
-
"learning_rate": 0.00010007998400319936,
|
| 26 |
-
"loss": 0.3338,
|
| 27 |
-
"step": 2002
|
| 28 |
},
|
| 29 |
{
|
| 30 |
-
"epoch": 0.
|
| 31 |
-
"learning_rate":
|
| 32 |
-
"loss": 0.
|
| 33 |
-
"step":
|
| 34 |
-
},
|
| 35 |
-
{
|
| 36 |
-
"epoch": 1.0,
|
| 37 |
-
"eval_loss": 0.118980273604393,
|
| 38 |
-
"eval_max_distance": 29,
|
| 39 |
-
"eval_mean_distance": 3,
|
| 40 |
-
"eval_runtime": 0.3245,
|
| 41 |
-
"eval_samples_per_second": 154.076,
|
| 42 |
-
"eval_steps_per_second": 6.163,
|
| 43 |
-
"step": 3334
|
| 44 |
},
|
| 45 |
{
|
| 46 |
-
"epoch":
|
| 47 |
-
"learning_rate":
|
| 48 |
-
"loss": 0.
|
| 49 |
-
"step":
|
| 50 |
-
},
|
| 51 |
-
{
|
| 52 |
-
"epoch": 1.5,
|
| 53 |
-
"learning_rate": 0.0002501999600079984,
|
| 54 |
-
"loss": 0.1375,
|
| 55 |
-
"step": 5005
|
| 56 |
-
},
|
| 57 |
-
{
|
| 58 |
-
"epoch": 1.8,
|
| 59 |
-
"learning_rate": 0.0003002399520095981,
|
| 60 |
-
"loss": 0.1179,
|
| 61 |
-
"step": 6006
|
| 62 |
},
|
| 63 |
{
|
| 64 |
-
"epoch":
|
| 65 |
-
"
|
| 66 |
-
"
|
| 67 |
-
"
|
| 68 |
-
"eval_runtime": 0.2749,
|
| 69 |
-
"eval_samples_per_second": 181.873,
|
| 70 |
-
"eval_steps_per_second": 7.275,
|
| 71 |
-
"step": 6668
|
| 72 |
},
|
| 73 |
{
|
| 74 |
-
"epoch":
|
| 75 |
-
"learning_rate":
|
| 76 |
-
"loss": 0.
|
| 77 |
-
"step":
|
| 78 |
},
|
| 79 |
{
|
| 80 |
-
"epoch":
|
| 81 |
-
"learning_rate":
|
| 82 |
-
"loss": 0.
|
| 83 |
-
"step":
|
| 84 |
},
|
| 85 |
{
|
| 86 |
-
"epoch":
|
| 87 |
-
"learning_rate":
|
| 88 |
-
"loss": 0.
|
| 89 |
-
"step":
|
| 90 |
},
|
| 91 |
{
|
| 92 |
-
"epoch":
|
| 93 |
-
"
|
| 94 |
-
"
|
| 95 |
-
"
|
| 96 |
-
"eval_runtime": 0.2795,
|
| 97 |
-
"eval_samples_per_second": 178.881,
|
| 98 |
-
"eval_steps_per_second": 7.155,
|
| 99 |
-
"step": 10002
|
| 100 |
},
|
| 101 |
{
|
| 102 |
-
"epoch":
|
| 103 |
-
"learning_rate":
|
| 104 |
-
"loss": 0.
|
| 105 |
-
"step":
|
| 106 |
},
|
| 107 |
{
|
| 108 |
-
"epoch":
|
| 109 |
-
"
|
| 110 |
-
"
|
| 111 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 112 |
},
|
| 113 |
{
|
| 114 |
-
"epoch":
|
| 115 |
-
"learning_rate":
|
| 116 |
-
"loss": 0.
|
| 117 |
-
"step":
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch":
|
| 121 |
-
"learning_rate":
|
| 122 |
-
"loss": 0.
|
| 123 |
-
"step":
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch":
|
| 127 |
-
"
|
| 128 |
-
"
|
| 129 |
-
"
|
| 130 |
-
"eval_runtime": 0.2802,
|
| 131 |
-
"eval_samples_per_second": 178.422,
|
| 132 |
-
"eval_steps_per_second": 7.137,
|
| 133 |
-
"step": 13336
|
| 134 |
},
|
| 135 |
{
|
| 136 |
-
"epoch":
|
| 137 |
-
"learning_rate":
|
| 138 |
-
"loss": 0.
|
| 139 |
-
"step":
|
| 140 |
},
|
| 141 |
{
|
| 142 |
-
"epoch":
|
| 143 |
-
"learning_rate":
|
| 144 |
-
"loss": 0.
|
| 145 |
-
"step":
|
| 146 |
},
|
| 147 |
{
|
| 148 |
-
"epoch":
|
| 149 |
-
"learning_rate":
|
| 150 |
-
"loss": 0.
|
| 151 |
-
"step":
|
| 152 |
},
|
| 153 |
{
|
| 154 |
-
"epoch":
|
| 155 |
-
"
|
| 156 |
-
"
|
| 157 |
-
"
|
| 158 |
-
"eval_runtime": 0.2717,
|
| 159 |
-
"eval_samples_per_second": 184.007,
|
| 160 |
-
"eval_steps_per_second": 7.36,
|
| 161 |
-
"step": 16670
|
| 162 |
},
|
| 163 |
{
|
| 164 |
-
"epoch":
|
| 165 |
-
"learning_rate":
|
| 166 |
-
"loss": 0.
|
| 167 |
-
"step":
|
| 168 |
},
|
| 169 |
{
|
| 170 |
-
"epoch":
|
| 171 |
-
"learning_rate":
|
| 172 |
-
"loss": 0.
|
| 173 |
-
"step":
|
| 174 |
},
|
| 175 |
{
|
| 176 |
-
"epoch":
|
| 177 |
-
"learning_rate":
|
| 178 |
-
"loss": 0.
|
| 179 |
-
"step":
|
| 180 |
},
|
| 181 |
{
|
| 182 |
-
"epoch":
|
| 183 |
-
"eval_loss": 0.
|
| 184 |
-
"eval_max_distance":
|
| 185 |
"eval_mean_distance": 0,
|
| 186 |
-
"eval_runtime":
|
| 187 |
-
"eval_samples_per_second":
|
| 188 |
-
"eval_steps_per_second":
|
| 189 |
-
"step":
|
| 190 |
-
},
|
| 191 |
-
{
|
| 192 |
-
"epoch": 6.0,
|
| 193 |
-
"learning_rate": 0.000999911128885334,
|
| 194 |
-
"loss": 0.0465,
|
| 195 |
-
"step": 20020
|
| 196 |
-
},
|
| 197 |
-
{
|
| 198 |
-
"epoch": 6.31,
|
| 199 |
-
"learning_rate": 0.0009943511297740451,
|
| 200 |
-
"loss": 0.0384,
|
| 201 |
-
"step": 21021
|
| 202 |
},
|
| 203 |
{
|
| 204 |
-
"epoch":
|
| 205 |
-
"learning_rate":
|
| 206 |
-
"loss": 0.
|
| 207 |
-
"step":
|
| 208 |
},
|
| 209 |
{
|
| 210 |
-
"epoch":
|
| 211 |
-
"learning_rate":
|
| 212 |
-
"loss": 0.
|
| 213 |
-
"step":
|
| 214 |
},
|
| 215 |
{
|
| 216 |
-
"epoch":
|
| 217 |
-
"
|
| 218 |
-
"
|
| 219 |
-
"
|
| 220 |
-
"eval_runtime": 0.2674,
|
| 221 |
-
"eval_samples_per_second": 187.01,
|
| 222 |
-
"eval_steps_per_second": 7.48,
|
| 223 |
-
"step": 23338
|
| 224 |
},
|
| 225 |
{
|
| 226 |
-
"epoch":
|
| 227 |
-
"learning_rate":
|
| 228 |
-
"loss": 0.
|
| 229 |
-
"step":
|
| 230 |
},
|
| 231 |
{
|
| 232 |
-
"epoch":
|
| 233 |
-
"learning_rate":
|
| 234 |
-
"loss": 0.
|
| 235 |
-
"step":
|
| 236 |
},
|
| 237 |
{
|
| 238 |
-
"epoch":
|
| 239 |
-
"learning_rate":
|
| 240 |
-
"loss": 0.
|
| 241 |
-
"step":
|
| 242 |
},
|
| 243 |
{
|
| 244 |
-
"epoch":
|
| 245 |
-
"
|
| 246 |
-
"
|
| 247 |
-
"
|
| 248 |
-
"eval_runtime": 0.2674,
|
| 249 |
-
"eval_samples_per_second": 186.98,
|
| 250 |
-
"eval_steps_per_second": 7.479,
|
| 251 |
-
"step": 26672
|
| 252 |
},
|
| 253 |
{
|
| 254 |
-
"epoch":
|
| 255 |
-
"learning_rate":
|
| 256 |
-
"loss": 0.
|
| 257 |
-
"step":
|
| 258 |
},
|
| 259 |
{
|
| 260 |
-
"epoch":
|
| 261 |
-
"learning_rate":
|
| 262 |
-
"loss": 0.
|
| 263 |
-
"step":
|
| 264 |
},
|
| 265 |
{
|
| 266 |
-
"epoch":
|
| 267 |
-
"learning_rate":
|
| 268 |
-
"loss": 0.
|
| 269 |
-
"step":
|
| 270 |
},
|
| 271 |
{
|
| 272 |
-
"epoch":
|
| 273 |
-
"eval_loss": 0.
|
| 274 |
-
"eval_max_distance":
|
| 275 |
"eval_mean_distance": 0,
|
| 276 |
-
"eval_runtime":
|
| 277 |
-
"eval_samples_per_second":
|
| 278 |
-
"eval_steps_per_second":
|
| 279 |
-
"step":
|
| 280 |
-
},
|
| 281 |
-
{
|
| 282 |
-
"epoch": 9.01,
|
| 283 |
-
"learning_rate": 0.0009443111377724454,
|
| 284 |
-
"loss": 0.0269,
|
| 285 |
-
"step": 30030
|
| 286 |
},
|
| 287 |
{
|
| 288 |
-
"epoch":
|
| 289 |
-
"learning_rate":
|
| 290 |
-
"loss": 0.
|
| 291 |
-
"step":
|
| 292 |
-
},
|
| 293 |
-
{
|
| 294 |
-
"epoch": 9.61,
|
| 295 |
-
"learning_rate": 0.0009331911395498677,
|
| 296 |
-
"loss": 0.0231,
|
| 297 |
-
"step": 32032
|
| 298 |
},
|
| 299 |
{
|
| 300 |
-
"epoch":
|
| 301 |
-
"learning_rate":
|
| 302 |
-
"loss": 0.
|
| 303 |
-
"step":
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch":
|
| 307 |
-
"
|
| 308 |
-
"
|
| 309 |
-
"
|
| 310 |
-
"eval_runtime": 0.2488,
|
| 311 |
-
"eval_samples_per_second": 200.959,
|
| 312 |
-
"eval_steps_per_second": 8.038,
|
| 313 |
-
"step": 33340
|
| 314 |
},
|
| 315 |
{
|
| 316 |
-
"epoch":
|
| 317 |
-
"learning_rate":
|
| 318 |
-
"loss": 0.
|
| 319 |
-
"step":
|
| 320 |
},
|
| 321 |
{
|
| 322 |
-
"epoch":
|
| 323 |
-
"learning_rate":
|
| 324 |
-
"loss": 0.
|
| 325 |
-
"step":
|
| 326 |
},
|
| 327 |
{
|
| 328 |
-
"epoch":
|
| 329 |
-
"learning_rate":
|
| 330 |
-
"loss": 0.
|
| 331 |
-
"step":
|
| 332 |
},
|
| 333 |
{
|
| 334 |
-
"epoch":
|
| 335 |
-
"
|
| 336 |
-
"
|
| 337 |
-
"
|
| 338 |
-
"eval_runtime": 0.2617,
|
| 339 |
-
"eval_samples_per_second": 191.039,
|
| 340 |
-
"eval_steps_per_second": 7.642,
|
| 341 |
-
"step": 36674
|
| 342 |
},
|
| 343 |
{
|
| 344 |
-
"epoch":
|
| 345 |
-
"learning_rate":
|
| 346 |
-
"loss": 0.
|
| 347 |
-
"step":
|
| 348 |
},
|
| 349 |
{
|
| 350 |
-
"epoch":
|
| 351 |
-
"learning_rate":
|
| 352 |
-
"loss": 0.
|
| 353 |
-
"step":
|
| 354 |
},
|
| 355 |
{
|
| 356 |
-
"epoch":
|
| 357 |
-
"learning_rate":
|
| 358 |
-
"loss": 0.
|
| 359 |
-
"step":
|
| 360 |
},
|
| 361 |
{
|
| 362 |
-
"epoch":
|
| 363 |
-
"eval_loss": 0.
|
| 364 |
-
"eval_max_distance":
|
| 365 |
"eval_mean_distance": 0,
|
| 366 |
-
"eval_runtime":
|
| 367 |
-
"eval_samples_per_second":
|
| 368 |
-
"eval_steps_per_second":
|
| 369 |
-
"step":
|
| 370 |
-
},
|
| 371 |
-
{
|
| 372 |
-
"epoch": 12.01,
|
| 373 |
-
"learning_rate": 0.000888711146659557,
|
| 374 |
-
"loss": 0.0178,
|
| 375 |
-
"step": 40040
|
| 376 |
},
|
| 377 |
{
|
| 378 |
-
"epoch":
|
| 379 |
-
"learning_rate":
|
| 380 |
-
"loss": 0.
|
| 381 |
-
"step":
|
| 382 |
-
},
|
| 383 |
-
{
|
| 384 |
-
"epoch": 12.61,
|
| 385 |
-
"learning_rate": 0.0008775911484369793,
|
| 386 |
-
"loss": 0.0149,
|
| 387 |
-
"step": 42042
|
| 388 |
},
|
| 389 |
{
|
| 390 |
-
"epoch":
|
| 391 |
-
"learning_rate":
|
| 392 |
-
"loss": 0.
|
| 393 |
-
"step":
|
| 394 |
},
|
| 395 |
{
|
| 396 |
-
"epoch":
|
| 397 |
-
"
|
| 398 |
-
"
|
| 399 |
-
"
|
| 400 |
-
"eval_runtime": 0.2582,
|
| 401 |
-
"eval_samples_per_second": 193.662,
|
| 402 |
-
"eval_steps_per_second": 7.746,
|
| 403 |
-
"step": 43342
|
| 404 |
},
|
| 405 |
{
|
| 406 |
-
"epoch":
|
| 407 |
-
"learning_rate":
|
| 408 |
-
"loss": 0.
|
| 409 |
-
"step":
|
| 410 |
},
|
| 411 |
{
|
| 412 |
-
"epoch":
|
| 413 |
-
"learning_rate":
|
| 414 |
-
"loss": 0.
|
| 415 |
-
"step":
|
| 416 |
},
|
| 417 |
{
|
| 418 |
-
"epoch":
|
| 419 |
-
"learning_rate":
|
| 420 |
-
"loss": 0.
|
| 421 |
-
"step":
|
| 422 |
},
|
| 423 |
{
|
| 424 |
-
"epoch":
|
| 425 |
-
"
|
| 426 |
-
"
|
| 427 |
-
"
|
| 428 |
-
"eval_runtime": 0.2465,
|
| 429 |
-
"eval_samples_per_second": 202.834,
|
| 430 |
-
"eval_steps_per_second": 8.113,
|
| 431 |
-
"step": 46676
|
| 432 |
},
|
| 433 |
{
|
| 434 |
-
"epoch":
|
| 435 |
-
"learning_rate":
|
| 436 |
-
"loss": 0.
|
| 437 |
-
"step":
|
| 438 |
},
|
| 439 |
{
|
| 440 |
-
"epoch":
|
| 441 |
-
"learning_rate":
|
| 442 |
-
"loss": 0.
|
| 443 |
-
"step":
|
| 444 |
},
|
| 445 |
{
|
| 446 |
-
"epoch":
|
| 447 |
-
"learning_rate":
|
| 448 |
-
"loss": 0.
|
| 449 |
-
"step":
|
| 450 |
},
|
| 451 |
{
|
| 452 |
-
"epoch":
|
| 453 |
-
"eval_loss": 0.
|
| 454 |
-
"eval_max_distance":
|
| 455 |
"eval_mean_distance": 0,
|
| 456 |
-
"eval_runtime":
|
| 457 |
-
"eval_samples_per_second":
|
| 458 |
-
"eval_steps_per_second":
|
| 459 |
-
"step":
|
| 460 |
},
|
| 461 |
{
|
| 462 |
-
"epoch":
|
| 463 |
-
"learning_rate":
|
| 464 |
-
"loss": 0.
|
| 465 |
-
"step":
|
| 466 |
-
},
|
| 467 |
-
{
|
| 468 |
-
"epoch": 15.31,
|
| 469 |
-
"learning_rate": 0.0008275511564353796,
|
| 470 |
-
"loss": 0.0101,
|
| 471 |
-
"step": 51051
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 15.61,
|
| 475 |
-
"learning_rate": 0.0008219911573240908,
|
| 476 |
-
"loss": 0.0108,
|
| 477 |
-
"step": 52052
|
| 478 |
},
|
| 479 |
{
|
| 480 |
-
"epoch":
|
| 481 |
-
"learning_rate":
|
| 482 |
-
"loss": 0.
|
| 483 |
-
"step":
|
| 484 |
},
|
| 485 |
{
|
| 486 |
-
"epoch":
|
| 487 |
-
"
|
| 488 |
-
"
|
| 489 |
-
"
|
| 490 |
-
"eval_runtime": 0.2455,
|
| 491 |
-
"eval_samples_per_second": 203.682,
|
| 492 |
-
"eval_steps_per_second": 8.147,
|
| 493 |
-
"step": 53344
|
| 494 |
},
|
| 495 |
{
|
| 496 |
-
"epoch":
|
| 497 |
-
"learning_rate":
|
| 498 |
-
"loss": 0.
|
| 499 |
-
"step":
|
| 500 |
},
|
| 501 |
{
|
| 502 |
-
"epoch":
|
| 503 |
-
"learning_rate":
|
| 504 |
-
"loss": 0.
|
| 505 |
-
"step":
|
| 506 |
},
|
| 507 |
{
|
| 508 |
-
"epoch":
|
| 509 |
-
"learning_rate":
|
| 510 |
-
"loss": 0.
|
| 511 |
-
"step":
|
| 512 |
},
|
| 513 |
{
|
| 514 |
-
"epoch":
|
| 515 |
-
"
|
| 516 |
-
"
|
| 517 |
-
"
|
| 518 |
-
"eval_runtime": 0.2506,
|
| 519 |
-
"eval_samples_per_second": 199.486,
|
| 520 |
-
"eval_steps_per_second": 7.979,
|
| 521 |
-
"step": 56678
|
| 522 |
},
|
| 523 |
{
|
| 524 |
-
"epoch":
|
| 525 |
-
"learning_rate":
|
| 526 |
-
"loss": 0.
|
| 527 |
-
"step":
|
| 528 |
},
|
| 529 |
{
|
| 530 |
-
"epoch":
|
| 531 |
-
"learning_rate":
|
| 532 |
-
"loss": 0.
|
| 533 |
-
"step":
|
| 534 |
},
|
| 535 |
{
|
| 536 |
-
"epoch":
|
| 537 |
-
"learning_rate":
|
| 538 |
-
"loss": 0.
|
| 539 |
-
"step":
|
| 540 |
},
|
| 541 |
{
|
| 542 |
-
"epoch":
|
| 543 |
-
"eval_loss": 0.
|
| 544 |
-
"eval_max_distance":
|
| 545 |
"eval_mean_distance": 0,
|
| 546 |
-
"eval_runtime":
|
| 547 |
-
"eval_samples_per_second":
|
| 548 |
-
"eval_steps_per_second":
|
| 549 |
-
"step":
|
| 550 |
},
|
| 551 |
{
|
| 552 |
-
"epoch":
|
| 553 |
-
"learning_rate":
|
| 554 |
-
"loss": 0.
|
| 555 |
-
"step":
|
| 556 |
-
},
|
| 557 |
-
{
|
| 558 |
-
"epoch": 18.31,
|
| 559 |
-
"learning_rate": 0.0007719511653224912,
|
| 560 |
-
"loss": 0.0075,
|
| 561 |
-
"step": 61061
|
| 562 |
-
},
|
| 563 |
-
{
|
| 564 |
-
"epoch": 18.61,
|
| 565 |
-
"learning_rate": 0.0007663911662112022,
|
| 566 |
-
"loss": 0.0079,
|
| 567 |
-
"step": 62062
|
| 568 |
},
|
| 569 |
{
|
| 570 |
-
"epoch":
|
| 571 |
-
"learning_rate":
|
| 572 |
-
"loss": 0.
|
| 573 |
-
"step":
|
| 574 |
},
|
| 575 |
{
|
| 576 |
-
"epoch":
|
| 577 |
-
"
|
| 578 |
-
"
|
| 579 |
-
"
|
| 580 |
-
"eval_runtime": 0.2416,
|
| 581 |
-
"eval_samples_per_second": 206.915,
|
| 582 |
-
"eval_steps_per_second": 8.277,
|
| 583 |
-
"step": 63346
|
| 584 |
},
|
| 585 |
{
|
| 586 |
-
"epoch":
|
| 587 |
-
"learning_rate":
|
| 588 |
-
"loss": 0.
|
| 589 |
-
"step":
|
| 590 |
},
|
| 591 |
{
|
| 592 |
-
"epoch":
|
| 593 |
-
"learning_rate":
|
| 594 |
-
"loss": 0.
|
| 595 |
-
"step":
|
| 596 |
},
|
| 597 |
{
|
| 598 |
-
"epoch":
|
| 599 |
-
"learning_rate":
|
| 600 |
-
"loss": 0.
|
| 601 |
-
"step":
|
| 602 |
},
|
| 603 |
{
|
| 604 |
-
"epoch":
|
| 605 |
-
"
|
| 606 |
-
"
|
| 607 |
-
"
|
| 608 |
-
"eval_runtime": 0.2481,
|
| 609 |
-
"eval_samples_per_second": 201.568,
|
| 610 |
-
"eval_steps_per_second": 8.063,
|
| 611 |
-
"step": 66680
|
| 612 |
},
|
| 613 |
{
|
| 614 |
-
"epoch":
|
| 615 |
-
"learning_rate":
|
| 616 |
-
"loss": 0.
|
| 617 |
-
"step":
|
| 618 |
},
|
| 619 |
{
|
| 620 |
-
"epoch":
|
| 621 |
-
"learning_rate":
|
| 622 |
-
"loss": 0.
|
| 623 |
-
"step":
|
| 624 |
},
|
| 625 |
{
|
| 626 |
-
"epoch":
|
| 627 |
-
"learning_rate":
|
| 628 |
-
"loss": 0.
|
| 629 |
-
"step":
|
| 630 |
},
|
| 631 |
{
|
| 632 |
-
"epoch":
|
| 633 |
-
"eval_loss": 0.
|
| 634 |
-
"eval_max_distance":
|
| 635 |
"eval_mean_distance": 0,
|
| 636 |
-
"eval_runtime":
|
| 637 |
-
"eval_samples_per_second":
|
| 638 |
-
"eval_steps_per_second":
|
| 639 |
-
"step":
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 21.02,
|
| 643 |
-
"learning_rate": 0.0007219111733208914,
|
| 644 |
-
"loss": 0.0069,
|
| 645 |
-
"step": 70070
|
| 646 |
},
|
| 647 |
{
|
| 648 |
-
"epoch":
|
| 649 |
-
"learning_rate":
|
| 650 |
-
"loss": 0.
|
| 651 |
-
"step":
|
| 652 |
-
},
|
| 653 |
-
{
|
| 654 |
-
"epoch": 21.62,
|
| 655 |
-
"learning_rate": 0.0007107911750983137,
|
| 656 |
-
"loss": 0.0061,
|
| 657 |
-
"step": 72072
|
| 658 |
},
|
| 659 |
{
|
| 660 |
-
"epoch":
|
| 661 |
-
"learning_rate":
|
| 662 |
-
"loss": 0.
|
| 663 |
-
"step":
|
| 664 |
},
|
| 665 |
{
|
| 666 |
-
"epoch":
|
| 667 |
-
"
|
| 668 |
-
"
|
| 669 |
-
"
|
| 670 |
-
"eval_runtime": 0.2419,
|
| 671 |
-
"eval_samples_per_second": 206.734,
|
| 672 |
-
"eval_steps_per_second": 8.269,
|
| 673 |
-
"step": 73348
|
| 674 |
},
|
| 675 |
{
|
| 676 |
-
"epoch":
|
| 677 |
-
"learning_rate":
|
| 678 |
-
"loss": 0.
|
| 679 |
-
"step":
|
| 680 |
},
|
| 681 |
{
|
| 682 |
-
"epoch":
|
| 683 |
-
"learning_rate":
|
| 684 |
-
"loss": 0.
|
| 685 |
-
"step":
|
| 686 |
},
|
| 687 |
{
|
| 688 |
-
"epoch":
|
| 689 |
-
"learning_rate":
|
| 690 |
-
"loss": 0.
|
| 691 |
-
"step":
|
| 692 |
},
|
| 693 |
{
|
| 694 |
-
"epoch":
|
| 695 |
-
"
|
| 696 |
-
"
|
| 697 |
-
"
|
| 698 |
-
"eval_runtime": 0.2431,
|
| 699 |
-
"eval_samples_per_second": 205.684,
|
| 700 |
-
"eval_steps_per_second": 8.227,
|
| 701 |
-
"step": 76682
|
| 702 |
},
|
| 703 |
{
|
| 704 |
-
"epoch":
|
| 705 |
-
"learning_rate":
|
| 706 |
-
"loss": 0.
|
| 707 |
-
"step":
|
| 708 |
},
|
| 709 |
{
|
| 710 |
-
"epoch":
|
| 711 |
-
"learning_rate":
|
| 712 |
-
"loss": 0.
|
| 713 |
-
"step":
|
| 714 |
},
|
| 715 |
{
|
| 716 |
-
"epoch":
|
| 717 |
-
"learning_rate":
|
| 718 |
-
"loss": 0.
|
| 719 |
-
"step":
|
| 720 |
},
|
| 721 |
{
|
| 722 |
-
"epoch":
|
| 723 |
-
"eval_loss": 0.
|
| 724 |
-
"eval_max_distance":
|
| 725 |
"eval_mean_distance": 0,
|
| 726 |
-
"eval_runtime":
|
| 727 |
-
"eval_samples_per_second":
|
| 728 |
-
"eval_steps_per_second":
|
| 729 |
-
"step":
|
| 730 |
-
},
|
| 731 |
-
{
|
| 732 |
-
"epoch": 24.02,
|
| 733 |
-
"learning_rate": 0.0006663111822080029,
|
| 734 |
-
"loss": 0.0054,
|
| 735 |
-
"step": 80080
|
| 736 |
},
|
| 737 |
{
|
| 738 |
-
"epoch":
|
| 739 |
-
"learning_rate":
|
| 740 |
-
"loss": 0.
|
| 741 |
-
"step":
|
| 742 |
-
},
|
| 743 |
-
{
|
| 744 |
-
"epoch": 24.62,
|
| 745 |
-
"learning_rate": 0.000655191183985425,
|
| 746 |
-
"loss": 0.0048,
|
| 747 |
-
"step": 82082
|
| 748 |
},
|
| 749 |
{
|
| 750 |
-
"epoch":
|
| 751 |
-
"learning_rate":
|
| 752 |
-
"loss": 0.
|
| 753 |
-
"step":
|
| 754 |
},
|
| 755 |
{
|
| 756 |
-
"epoch":
|
| 757 |
-
"
|
| 758 |
-
"
|
| 759 |
-
"
|
| 760 |
-
"eval_runtime": 0.2475,
|
| 761 |
-
"eval_samples_per_second": 202.046,
|
| 762 |
-
"eval_steps_per_second": 8.082,
|
| 763 |
-
"step": 83350
|
| 764 |
},
|
| 765 |
{
|
| 766 |
-
"epoch":
|
| 767 |
-
"learning_rate":
|
| 768 |
-
"loss": 0.
|
| 769 |
-
"step":
|
| 770 |
},
|
| 771 |
{
|
| 772 |
-
"epoch":
|
| 773 |
-
"learning_rate":
|
| 774 |
-
"loss": 0.
|
| 775 |
-
"step":
|
| 776 |
},
|
| 777 |
{
|
| 778 |
-
"epoch":
|
| 779 |
-
"learning_rate":
|
| 780 |
-
"loss": 0.
|
| 781 |
-
"step":
|
| 782 |
},
|
| 783 |
{
|
| 784 |
-
"epoch":
|
| 785 |
-
"
|
| 786 |
-
"
|
| 787 |
-
"
|
| 788 |
-
"eval_runtime": 0.2424,
|
| 789 |
-
"eval_samples_per_second": 206.301,
|
| 790 |
-
"eval_steps_per_second": 8.252,
|
| 791 |
-
"step": 86684
|
| 792 |
},
|
| 793 |
{
|
| 794 |
-
"epoch":
|
| 795 |
-
"learning_rate":
|
| 796 |
-
"loss": 0.
|
| 797 |
-
"step":
|
| 798 |
},
|
| 799 |
{
|
| 800 |
-
"epoch":
|
| 801 |
-
"learning_rate":
|
| 802 |
-
"loss": 0.
|
| 803 |
-
"step":
|
| 804 |
},
|
| 805 |
{
|
| 806 |
-
"epoch":
|
| 807 |
-
"learning_rate":
|
| 808 |
-
"loss": 0.
|
| 809 |
-
"step":
|
| 810 |
},
|
| 811 |
{
|
| 812 |
-
"epoch":
|
| 813 |
-
"eval_loss": 0.
|
| 814 |
-
"eval_max_distance":
|
| 815 |
"eval_mean_distance": 0,
|
| 816 |
-
"eval_runtime":
|
| 817 |
-
"eval_samples_per_second":
|
| 818 |
-
"eval_steps_per_second":
|
| 819 |
-
"step":
|
| 820 |
-
},
|
| 821 |
-
{
|
| 822 |
-
"epoch": 27.02,
|
| 823 |
-
"learning_rate": 0.0006107111910951143,
|
| 824 |
-
"loss": 0.0043,
|
| 825 |
-
"step": 90090
|
| 826 |
},
|
| 827 |
{
|
| 828 |
-
"epoch":
|
| 829 |
-
"learning_rate":
|
| 830 |
-
"loss": 0.
|
| 831 |
-
"step":
|
| 832 |
-
},
|
| 833 |
-
{
|
| 834 |
-
"epoch": 27.62,
|
| 835 |
-
"learning_rate": 0.0005995911928725366,
|
| 836 |
-
"loss": 0.0038,
|
| 837 |
-
"step": 92092
|
| 838 |
},
|
| 839 |
{
|
| 840 |
-
"epoch":
|
| 841 |
-
"learning_rate":
|
| 842 |
-
"loss": 0.
|
| 843 |
-
"step":
|
| 844 |
},
|
| 845 |
{
|
| 846 |
-
"epoch":
|
| 847 |
-
"
|
| 848 |
-
"
|
| 849 |
-
"
|
| 850 |
-
"eval_runtime": 0.2482,
|
| 851 |
-
"eval_samples_per_second": 201.481,
|
| 852 |
-
"eval_steps_per_second": 8.059,
|
| 853 |
-
"step": 93352
|
| 854 |
},
|
| 855 |
{
|
| 856 |
-
"epoch":
|
| 857 |
-
"learning_rate":
|
| 858 |
-
"loss": 0.
|
| 859 |
-
"step":
|
| 860 |
},
|
| 861 |
{
|
| 862 |
-
"epoch":
|
| 863 |
-
"learning_rate":
|
| 864 |
-
"loss": 0.
|
| 865 |
-
"step":
|
| 866 |
},
|
| 867 |
{
|
| 868 |
-
"epoch":
|
| 869 |
-
"learning_rate":
|
| 870 |
-
"loss": 0.
|
| 871 |
-
"step":
|
| 872 |
},
|
| 873 |
{
|
| 874 |
-
"epoch":
|
| 875 |
-
"
|
| 876 |
-
"
|
| 877 |
-
"
|
| 878 |
-
"eval_runtime": 0.2429,
|
| 879 |
-
"eval_samples_per_second": 205.838,
|
| 880 |
-
"eval_steps_per_second": 8.234,
|
| 881 |
-
"step": 96686
|
| 882 |
},
|
| 883 |
{
|
| 884 |
-
"epoch":
|
| 885 |
-
"learning_rate":
|
| 886 |
-
"loss": 0.
|
| 887 |
-
"step":
|
| 888 |
},
|
| 889 |
{
|
| 890 |
-
"epoch":
|
| 891 |
-
"learning_rate":
|
| 892 |
-
"loss": 0.
|
| 893 |
-
"step":
|
| 894 |
},
|
| 895 |
{
|
| 896 |
-
"epoch":
|
| 897 |
-
"learning_rate":
|
| 898 |
-
"loss": 0.
|
| 899 |
-
"step":
|
| 900 |
},
|
| 901 |
{
|
| 902 |
-
"epoch":
|
| 903 |
-
"eval_loss": 0.
|
| 904 |
-
"eval_max_distance":
|
| 905 |
"eval_mean_distance": 0,
|
| 906 |
-
"eval_runtime":
|
| 907 |
-
"eval_samples_per_second":
|
| 908 |
-
"eval_steps_per_second":
|
| 909 |
-
"step":
|
| 910 |
-
},
|
| 911 |
-
{
|
| 912 |
-
"epoch": 30.02,
|
| 913 |
-
"learning_rate": 0.0005551111999822258,
|
| 914 |
-
"loss": 0.0033,
|
| 915 |
-
"step": 100100
|
| 916 |
},
|
| 917 |
{
|
| 918 |
-
"epoch":
|
| 919 |
-
"learning_rate":
|
| 920 |
-
"loss": 0.
|
| 921 |
-
"step":
|
| 922 |
-
},
|
| 923 |
-
{
|
| 924 |
-
"epoch": 30.62,
|
| 925 |
-
"learning_rate": 0.0005439912017596481,
|
| 926 |
-
"loss": 0.0031,
|
| 927 |
-
"step": 102102
|
| 928 |
},
|
| 929 |
{
|
| 930 |
-
"epoch":
|
| 931 |
-
"learning_rate":
|
| 932 |
-
"loss": 0.
|
| 933 |
-
"step":
|
| 934 |
},
|
| 935 |
{
|
| 936 |
-
"epoch":
|
| 937 |
-
"
|
| 938 |
-
"
|
| 939 |
-
"
|
| 940 |
-
"eval_runtime": 0.2446,
|
| 941 |
-
"eval_samples_per_second": 204.388,
|
| 942 |
-
"eval_steps_per_second": 8.176,
|
| 943 |
-
"step": 103354
|
| 944 |
},
|
| 945 |
{
|
| 946 |
-
"epoch":
|
| 947 |
-
"learning_rate":
|
| 948 |
-
"loss": 0.
|
| 949 |
-
"step":
|
| 950 |
},
|
| 951 |
{
|
| 952 |
-
"epoch":
|
| 953 |
-
"learning_rate":
|
| 954 |
-
"loss": 0.
|
| 955 |
-
"step":
|
| 956 |
},
|
| 957 |
{
|
| 958 |
-
"epoch":
|
| 959 |
-
"learning_rate":
|
| 960 |
-
"loss": 0.
|
| 961 |
-
"step":
|
| 962 |
},
|
| 963 |
{
|
| 964 |
-
"epoch":
|
| 965 |
-
"
|
| 966 |
-
"
|
| 967 |
-
"
|
| 968 |
-
"eval_runtime": 0.2518,
|
| 969 |
-
"eval_samples_per_second": 198.594,
|
| 970 |
-
"eval_steps_per_second": 7.944,
|
| 971 |
-
"step": 106688
|
| 972 |
},
|
| 973 |
{
|
| 974 |
-
"epoch":
|
| 975 |
-
"learning_rate":
|
| 976 |
-
"loss": 0.
|
| 977 |
-
"step":
|
| 978 |
},
|
| 979 |
{
|
| 980 |
-
"epoch":
|
| 981 |
-
"learning_rate":
|
| 982 |
-
"loss": 0.
|
| 983 |
-
"step":
|
| 984 |
},
|
| 985 |
{
|
| 986 |
-
"epoch":
|
| 987 |
-
"learning_rate":
|
| 988 |
-
"loss": 0.
|
| 989 |
-
"step":
|
| 990 |
},
|
| 991 |
{
|
| 992 |
-
"epoch":
|
| 993 |
-
"eval_loss": 0.
|
| 994 |
-
"eval_max_distance":
|
| 995 |
"eval_mean_distance": 0,
|
| 996 |
-
"eval_runtime":
|
| 997 |
-
"eval_samples_per_second":
|
| 998 |
-
"eval_steps_per_second":
|
| 999 |
-
"step":
|
| 1000 |
-
},
|
| 1001 |
-
{
|
| 1002 |
-
"epoch": 33.03,
|
| 1003 |
-
"learning_rate": 0.0004995112088693373,
|
| 1004 |
-
"loss": 0.0027,
|
| 1005 |
-
"step": 110110
|
| 1006 |
},
|
| 1007 |
{
|
| 1008 |
-
"epoch":
|
| 1009 |
-
"learning_rate":
|
| 1010 |
-
"loss": 0.
|
| 1011 |
-
"step":
|
| 1012 |
-
},
|
| 1013 |
-
{
|
| 1014 |
-
"epoch": 33.63,
|
| 1015 |
-
"learning_rate": 0.0004883912106467596,
|
| 1016 |
-
"loss": 0.0026,
|
| 1017 |
-
"step": 112112
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
-
"epoch":
|
| 1021 |
-
"learning_rate":
|
| 1022 |
-
"loss": 0.
|
| 1023 |
-
"step":
|
| 1024 |
},
|
| 1025 |
{
|
| 1026 |
-
"epoch":
|
| 1027 |
-
"
|
| 1028 |
-
"
|
| 1029 |
-
"
|
| 1030 |
-
"eval_runtime": 0.2507,
|
| 1031 |
-
"eval_samples_per_second": 199.458,
|
| 1032 |
-
"eval_steps_per_second": 7.978,
|
| 1033 |
-
"step": 113356
|
| 1034 |
},
|
| 1035 |
{
|
| 1036 |
-
"epoch":
|
| 1037 |
-
"learning_rate":
|
| 1038 |
-
"loss": 0.
|
| 1039 |
-
"step":
|
| 1040 |
},
|
| 1041 |
{
|
| 1042 |
-
"epoch":
|
| 1043 |
-
"learning_rate":
|
| 1044 |
-
"loss": 0.
|
| 1045 |
-
"step":
|
| 1046 |
},
|
| 1047 |
{
|
| 1048 |
-
"epoch":
|
| 1049 |
-
"learning_rate":
|
| 1050 |
-
"loss": 0.
|
| 1051 |
-
"step":
|
| 1052 |
},
|
| 1053 |
{
|
| 1054 |
-
"epoch":
|
| 1055 |
-
"
|
| 1056 |
-
"
|
| 1057 |
-
"
|
| 1058 |
-
"eval_runtime": 0.2447,
|
| 1059 |
-
"eval_samples_per_second": 204.372,
|
| 1060 |
-
"eval_steps_per_second": 8.175,
|
| 1061 |
-
"step": 116690
|
| 1062 |
},
|
| 1063 |
{
|
| 1064 |
-
"epoch":
|
| 1065 |
-
"learning_rate":
|
| 1066 |
-
"loss": 0.
|
| 1067 |
-
"step":
|
| 1068 |
},
|
| 1069 |
{
|
| 1070 |
-
"epoch":
|
| 1071 |
-
"learning_rate":
|
| 1072 |
-
"loss": 0.
|
| 1073 |
-
"step":
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
-
"epoch":
|
| 1077 |
-
"learning_rate":
|
| 1078 |
-
"loss": 0.
|
| 1079 |
-
"step":
|
| 1080 |
},
|
| 1081 |
{
|
| 1082 |
-
"epoch":
|
| 1083 |
-
"eval_loss": 0.
|
| 1084 |
-
"eval_max_distance":
|
| 1085 |
"eval_mean_distance": 0,
|
| 1086 |
-
"eval_runtime":
|
| 1087 |
-
"eval_samples_per_second":
|
| 1088 |
-
"eval_steps_per_second":
|
| 1089 |
-
"step":
|
| 1090 |
-
},
|
| 1091 |
-
{
|
| 1092 |
-
"epoch": 36.03,
|
| 1093 |
-
"learning_rate": 0.00044391121775644874,
|
| 1094 |
-
"loss": 0.0022,
|
| 1095 |
-
"step": 120120
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
-
"epoch":
|
| 1099 |
-
"learning_rate":
|
| 1100 |
-
"loss": 0.
|
| 1101 |
-
"step":
|
| 1102 |
-
},
|
| 1103 |
-
{
|
| 1104 |
-
"epoch": 36.63,
|
| 1105 |
-
"learning_rate": 0.00043279121953387103,
|
| 1106 |
-
"loss": 0.0022,
|
| 1107 |
-
"step": 122122
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
-
"epoch":
|
| 1111 |
-
"learning_rate":
|
| 1112 |
-
"loss": 0.
|
| 1113 |
-
"step":
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
-
"epoch":
|
| 1117 |
-
"
|
| 1118 |
-
"
|
| 1119 |
-
"
|
| 1120 |
-
"eval_runtime": 0.2436,
|
| 1121 |
-
"eval_samples_per_second": 205.289,
|
| 1122 |
-
"eval_steps_per_second": 8.212,
|
| 1123 |
-
"step": 123358
|
| 1124 |
},
|
| 1125 |
{
|
| 1126 |
-
"epoch":
|
| 1127 |
-
"learning_rate":
|
| 1128 |
-
"loss": 0.
|
| 1129 |
-
"step":
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
-
"epoch":
|
| 1133 |
-
"learning_rate":
|
| 1134 |
-
"loss": 0.
|
| 1135 |
-
"step":
|
| 1136 |
},
|
| 1137 |
{
|
| 1138 |
-
"epoch":
|
| 1139 |
-
"learning_rate":
|
| 1140 |
-
"loss": 0.
|
| 1141 |
-
"step":
|
| 1142 |
},
|
| 1143 |
{
|
| 1144 |
-
"epoch":
|
| 1145 |
-
"
|
| 1146 |
-
"
|
| 1147 |
-
"
|
| 1148 |
-
"eval_runtime": 0.242,
|
| 1149 |
-
"eval_samples_per_second": 206.579,
|
| 1150 |
-
"eval_steps_per_second": 8.263,
|
| 1151 |
-
"step": 126692
|
| 1152 |
},
|
| 1153 |
{
|
| 1154 |
-
"epoch":
|
| 1155 |
-
"learning_rate":
|
| 1156 |
-
"loss": 0.
|
| 1157 |
-
"step":
|
| 1158 |
},
|
| 1159 |
{
|
| 1160 |
-
"epoch":
|
| 1161 |
-
"learning_rate":
|
| 1162 |
-
"loss": 0.
|
| 1163 |
-
"step":
|
| 1164 |
},
|
| 1165 |
{
|
| 1166 |
-
"epoch":
|
| 1167 |
-
"learning_rate":
|
| 1168 |
-
"loss": 0.
|
| 1169 |
-
"step":
|
| 1170 |
},
|
| 1171 |
{
|
| 1172 |
-
"epoch":
|
| 1173 |
-
"eval_loss": 0.
|
| 1174 |
-
"eval_max_distance":
|
| 1175 |
"eval_mean_distance": 0,
|
| 1176 |
-
"eval_runtime":
|
| 1177 |
-
"eval_samples_per_second":
|
| 1178 |
-
"eval_steps_per_second":
|
| 1179 |
-
"step":
|
| 1180 |
-
},
|
| 1181 |
-
{
|
| 1182 |
-
"epoch": 39.03,
|
| 1183 |
-
"learning_rate": 0.00038831122664356016,
|
| 1184 |
-
"loss": 0.0018,
|
| 1185 |
-
"step": 130130
|
| 1186 |
-
},
|
| 1187 |
-
{
|
| 1188 |
-
"epoch": 39.33,
|
| 1189 |
-
"learning_rate": 0.0003827512275322713,
|
| 1190 |
-
"loss": 0.0016,
|
| 1191 |
-
"step": 131131
|
| 1192 |
},
|
| 1193 |
{
|
| 1194 |
-
"epoch":
|
| 1195 |
-
"learning_rate":
|
| 1196 |
-
"loss": 0.
|
| 1197 |
-
"step":
|
| 1198 |
},
|
| 1199 |
{
|
| 1200 |
-
"epoch":
|
| 1201 |
-
"learning_rate":
|
| 1202 |
-
"loss": 0.
|
| 1203 |
-
"step":
|
| 1204 |
},
|
| 1205 |
{
|
| 1206 |
-
"epoch":
|
| 1207 |
-
"
|
| 1208 |
-
"
|
| 1209 |
-
"
|
| 1210 |
-
"eval_runtime": 0.2478,
|
| 1211 |
-
"eval_samples_per_second": 201.796,
|
| 1212 |
-
"eval_steps_per_second": 8.072,
|
| 1213 |
-
"step": 133360
|
| 1214 |
},
|
| 1215 |
{
|
| 1216 |
-
"epoch":
|
| 1217 |
-
"learning_rate":
|
| 1218 |
-
"loss": 0.
|
| 1219 |
-
"step":
|
| 1220 |
},
|
| 1221 |
{
|
| 1222 |
-
"epoch":
|
| 1223 |
-
"learning_rate":
|
| 1224 |
-
"loss": 0.
|
| 1225 |
-
"step":
|
| 1226 |
},
|
| 1227 |
{
|
| 1228 |
-
"epoch":
|
| 1229 |
-
"learning_rate":
|
| 1230 |
-
"loss": 0.
|
| 1231 |
-
"step":
|
| 1232 |
},
|
| 1233 |
{
|
| 1234 |
-
"epoch":
|
| 1235 |
-
"
|
| 1236 |
-
"
|
| 1237 |
-
"
|
| 1238 |
-
"eval_runtime": 0.2543,
|
| 1239 |
-
"eval_samples_per_second": 196.583,
|
| 1240 |
-
"eval_steps_per_second": 7.863,
|
| 1241 |
-
"step": 136694
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
-
"epoch":
|
| 1245 |
-
"learning_rate":
|
| 1246 |
-
"loss": 0.
|
| 1247 |
-
"step":
|
| 1248 |
},
|
| 1249 |
{
|
| 1250 |
-
"epoch":
|
| 1251 |
-
"learning_rate":
|
| 1252 |
-
"loss": 0.
|
| 1253 |
-
"step":
|
| 1254 |
},
|
| 1255 |
{
|
| 1256 |
-
"epoch":
|
| 1257 |
-
"learning_rate":
|
| 1258 |
-
"loss": 0.
|
| 1259 |
-
"step":
|
| 1260 |
},
|
| 1261 |
{
|
| 1262 |
-
"epoch":
|
| 1263 |
-
"eval_loss": 0.
|
| 1264 |
-
"eval_max_distance":
|
| 1265 |
"eval_mean_distance": 0,
|
| 1266 |
-
"eval_runtime":
|
| 1267 |
-
"eval_samples_per_second":
|
| 1268 |
-
"eval_steps_per_second":
|
| 1269 |
-
"step":
|
| 1270 |
-
},
|
| 1271 |
-
{
|
| 1272 |
-
"epoch": 42.03,
|
| 1273 |
-
"learning_rate": 0.0003327112355306717,
|
| 1274 |
-
"loss": 0.0015,
|
| 1275 |
-
"step": 140140
|
| 1276 |
-
},
|
| 1277 |
-
{
|
| 1278 |
-
"epoch": 42.33,
|
| 1279 |
-
"learning_rate": 0.0003271512364193828,
|
| 1280 |
-
"loss": 0.0014,
|
| 1281 |
-
"step": 141141
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
-
"epoch":
|
| 1285 |
-
"learning_rate":
|
| 1286 |
-
"loss": 0.
|
| 1287 |
-
"step":
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
-
"epoch":
|
| 1291 |
-
"learning_rate":
|
| 1292 |
-
"loss": 0.
|
| 1293 |
-
"step":
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
-
"epoch":
|
| 1297 |
-
"
|
| 1298 |
-
"
|
| 1299 |
-
"
|
| 1300 |
-
"eval_runtime": 0.2458,
|
| 1301 |
-
"eval_samples_per_second": 203.394,
|
| 1302 |
-
"eval_steps_per_second": 8.136,
|
| 1303 |
-
"step": 143362
|
| 1304 |
},
|
| 1305 |
{
|
| 1306 |
-
"epoch":
|
| 1307 |
-
"learning_rate":
|
| 1308 |
-
"loss": 0.
|
| 1309 |
-
"step":
|
| 1310 |
},
|
| 1311 |
{
|
| 1312 |
-
"epoch":
|
| 1313 |
-
"learning_rate":
|
| 1314 |
-
"loss": 0.
|
| 1315 |
-
"step":
|
| 1316 |
},
|
| 1317 |
{
|
| 1318 |
-
"epoch":
|
| 1319 |
-
"learning_rate":
|
| 1320 |
-
"loss": 0.
|
| 1321 |
-
"step":
|
| 1322 |
},
|
| 1323 |
{
|
| 1324 |
-
"epoch":
|
| 1325 |
-
"
|
| 1326 |
-
"
|
| 1327 |
-
"
|
| 1328 |
-
"eval_runtime": 0.248,
|
| 1329 |
-
"eval_samples_per_second": 201.602,
|
| 1330 |
-
"eval_steps_per_second": 8.064,
|
| 1331 |
-
"step": 146696
|
| 1332 |
},
|
| 1333 |
{
|
| 1334 |
-
"epoch":
|
| 1335 |
-
"learning_rate":
|
| 1336 |
-
"loss": 0.
|
| 1337 |
-
"step":
|
| 1338 |
},
|
| 1339 |
{
|
| 1340 |
-
"epoch":
|
| 1341 |
-
"learning_rate":
|
| 1342 |
-
"loss": 0.
|
| 1343 |
-
"step":
|
| 1344 |
},
|
| 1345 |
{
|
| 1346 |
-
"epoch":
|
| 1347 |
-
"learning_rate":
|
| 1348 |
-
"loss": 0.
|
| 1349 |
-
"step":
|
| 1350 |
},
|
| 1351 |
{
|
| 1352 |
-
"epoch":
|
| 1353 |
-
"eval_loss": 0.
|
| 1354 |
-
"eval_max_distance":
|
| 1355 |
"eval_mean_distance": 0,
|
| 1356 |
-
"eval_runtime":
|
| 1357 |
-
"eval_samples_per_second":
|
| 1358 |
-
"eval_steps_per_second":
|
| 1359 |
-
"step":
|
| 1360 |
-
},
|
| 1361 |
-
{
|
| 1362 |
-
"epoch": 45.04,
|
| 1363 |
-
"learning_rate": 0.00027711124441778313,
|
| 1364 |
-
"loss": 0.0012,
|
| 1365 |
-
"step": 150150
|
| 1366 |
-
},
|
| 1367 |
-
{
|
| 1368 |
-
"epoch": 45.34,
|
| 1369 |
-
"learning_rate": 0.0002715512453064943,
|
| 1370 |
-
"loss": 0.001,
|
| 1371 |
-
"step": 151151
|
| 1372 |
},
|
| 1373 |
{
|
| 1374 |
-
"epoch":
|
| 1375 |
-
"learning_rate":
|
| 1376 |
-
"loss": 0.
|
| 1377 |
-
"step":
|
| 1378 |
},
|
| 1379 |
{
|
| 1380 |
-
"epoch":
|
| 1381 |
-
"learning_rate":
|
| 1382 |
-
"loss": 0.
|
| 1383 |
-
"step":
|
| 1384 |
},
|
| 1385 |
{
|
| 1386 |
-
"epoch":
|
| 1387 |
-
"
|
| 1388 |
-
"
|
| 1389 |
-
"
|
| 1390 |
-
"eval_runtime": 0.2486,
|
| 1391 |
-
"eval_samples_per_second": 201.091,
|
| 1392 |
-
"eval_steps_per_second": 8.044,
|
| 1393 |
-
"step": 153364
|
| 1394 |
},
|
| 1395 |
{
|
| 1396 |
-
"epoch":
|
| 1397 |
-
"learning_rate":
|
| 1398 |
-
"loss": 0.
|
| 1399 |
-
"step":
|
| 1400 |
},
|
| 1401 |
{
|
| 1402 |
-
"epoch":
|
| 1403 |
-
"learning_rate":
|
| 1404 |
-
"loss": 0.
|
| 1405 |
-
"step":
|
| 1406 |
},
|
| 1407 |
{
|
| 1408 |
-
"epoch":
|
| 1409 |
-
"learning_rate":
|
| 1410 |
-
"loss": 0.
|
| 1411 |
-
"step":
|
| 1412 |
},
|
| 1413 |
{
|
| 1414 |
-
"epoch":
|
| 1415 |
-
"
|
| 1416 |
-
"
|
| 1417 |
-
"
|
| 1418 |
-
"eval_runtime": 0.2471,
|
| 1419 |
-
"eval_samples_per_second": 202.373,
|
| 1420 |
-
"eval_steps_per_second": 8.095,
|
| 1421 |
-
"step": 156698
|
| 1422 |
},
|
| 1423 |
{
|
| 1424 |
-
"epoch":
|
| 1425 |
-
"learning_rate":
|
| 1426 |
-
"loss": 0.
|
| 1427 |
-
"step":
|
| 1428 |
},
|
| 1429 |
{
|
| 1430 |
-
"epoch":
|
| 1431 |
-
"learning_rate":
|
| 1432 |
-
"loss": 0.
|
| 1433 |
-
"step":
|
| 1434 |
},
|
| 1435 |
{
|
| 1436 |
-
"epoch":
|
| 1437 |
-
"learning_rate":
|
| 1438 |
-
"loss": 0.
|
| 1439 |
-
"step":
|
| 1440 |
},
|
| 1441 |
{
|
| 1442 |
-
"epoch":
|
| 1443 |
-
"eval_loss": 0.
|
| 1444 |
-
"eval_max_distance":
|
| 1445 |
"eval_mean_distance": 0,
|
| 1446 |
-
"eval_runtime":
|
| 1447 |
-
"eval_samples_per_second":
|
| 1448 |
-
"eval_steps_per_second":
|
| 1449 |
-
"step":
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
-
"epoch":
|
| 1453 |
-
"learning_rate":
|
| 1454 |
-
"loss": 0.
|
| 1455 |
-
"step":
|
| 1456 |
-
},
|
| 1457 |
-
{
|
| 1458 |
-
"epoch": 48.34,
|
| 1459 |
-
"learning_rate": 0.00021595125419360573,
|
| 1460 |
-
"loss": 0.0008,
|
| 1461 |
-
"step": 161161
|
| 1462 |
-
},
|
| 1463 |
-
{
|
| 1464 |
-
"epoch": 48.64,
|
| 1465 |
-
"learning_rate": 0.00021039125508231688,
|
| 1466 |
-
"loss": 0.0009,
|
| 1467 |
-
"step": 162162
|
| 1468 |
},
|
| 1469 |
{
|
| 1470 |
-
"epoch":
|
| 1471 |
-
"learning_rate":
|
| 1472 |
-
"loss": 0.
|
| 1473 |
-
"step":
|
| 1474 |
},
|
| 1475 |
{
|
| 1476 |
-
"epoch":
|
| 1477 |
-
"
|
| 1478 |
-
"
|
| 1479 |
-
"
|
| 1480 |
-
"eval_runtime": 0.2518,
|
| 1481 |
-
"eval_samples_per_second": 198.574,
|
| 1482 |
-
"eval_steps_per_second": 7.943,
|
| 1483 |
-
"step": 163366
|
| 1484 |
},
|
| 1485 |
{
|
| 1486 |
-
"epoch":
|
| 1487 |
-
"learning_rate":
|
| 1488 |
-
"loss": 0.
|
| 1489 |
-
"step":
|
| 1490 |
},
|
| 1491 |
{
|
| 1492 |
-
"epoch":
|
| 1493 |
-
"learning_rate":
|
| 1494 |
-
"loss": 0.
|
| 1495 |
-
"step":
|
| 1496 |
},
|
| 1497 |
{
|
| 1498 |
-
"epoch":
|
| 1499 |
-
"learning_rate":
|
| 1500 |
-
"loss": 0.
|
| 1501 |
-
"step":
|
| 1502 |
},
|
| 1503 |
{
|
| 1504 |
-
"epoch":
|
| 1505 |
-
"
|
| 1506 |
-
"
|
| 1507 |
-
"
|
| 1508 |
-
"eval_runtime": 0.2461,
|
| 1509 |
-
"eval_samples_per_second": 203.155,
|
| 1510 |
-
"eval_steps_per_second": 8.126,
|
| 1511 |
-
"step": 166700
|
| 1512 |
},
|
| 1513 |
{
|
| 1514 |
-
"epoch":
|
| 1515 |
-
"learning_rate":
|
| 1516 |
-
"loss": 0.
|
| 1517 |
-
"step":
|
| 1518 |
},
|
| 1519 |
{
|
| 1520 |
-
"epoch":
|
| 1521 |
-
"learning_rate":
|
| 1522 |
-
"loss": 0.
|
| 1523 |
-
"step":
|
| 1524 |
},
|
| 1525 |
{
|
| 1526 |
-
"epoch":
|
| 1527 |
-
"learning_rate":
|
| 1528 |
-
"loss": 0.
|
| 1529 |
-
"step":
|
| 1530 |
},
|
| 1531 |
{
|
| 1532 |
-
"epoch":
|
| 1533 |
-
"eval_loss": 0.
|
| 1534 |
-
"eval_max_distance":
|
| 1535 |
"eval_mean_distance": 0,
|
| 1536 |
-
"eval_runtime":
|
| 1537 |
-
"eval_samples_per_second":
|
| 1538 |
-
"eval_steps_per_second":
|
| 1539 |
-
"step":
|
| 1540 |
-
},
|
| 1541 |
-
{
|
| 1542 |
-
"epoch": 51.04,
|
| 1543 |
-
"learning_rate": 0.00016591126219200607,
|
| 1544 |
-
"loss": 0.0007,
|
| 1545 |
-
"step": 170170
|
| 1546 |
},
|
| 1547 |
{
|
| 1548 |
-
"epoch":
|
| 1549 |
-
"learning_rate":
|
| 1550 |
-
"loss": 0.
|
| 1551 |
-
"step":
|
| 1552 |
-
},
|
| 1553 |
-
{
|
| 1554 |
-
"epoch": 51.64,
|
| 1555 |
-
"learning_rate": 0.00015479126396942834,
|
| 1556 |
-
"loss": 0.0006,
|
| 1557 |
-
"step": 172172
|
| 1558 |
},
|
| 1559 |
{
|
| 1560 |
-
"epoch":
|
| 1561 |
-
"learning_rate":
|
| 1562 |
-
"loss": 0.
|
| 1563 |
-
"step":
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
-
"epoch":
|
| 1567 |
-
"
|
| 1568 |
-
"
|
| 1569 |
-
"
|
| 1570 |
-
"eval_runtime": 0.2536,
|
| 1571 |
-
"eval_samples_per_second": 197.147,
|
| 1572 |
-
"eval_steps_per_second": 7.886,
|
| 1573 |
-
"step": 173368
|
| 1574 |
},
|
| 1575 |
{
|
| 1576 |
-
"epoch":
|
| 1577 |
-
"learning_rate":
|
| 1578 |
-
"loss": 0.
|
| 1579 |
-
"step":
|
| 1580 |
},
|
| 1581 |
{
|
| 1582 |
-
"epoch":
|
| 1583 |
-
"learning_rate":
|
| 1584 |
-
"loss": 0.
|
| 1585 |
-
"step":
|
| 1586 |
},
|
| 1587 |
{
|
| 1588 |
-
"epoch":
|
| 1589 |
-
"learning_rate":
|
| 1590 |
-
"loss": 0.
|
| 1591 |
-
"step":
|
| 1592 |
},
|
| 1593 |
{
|
| 1594 |
-
"epoch":
|
| 1595 |
-
"
|
| 1596 |
-
"
|
| 1597 |
-
"
|
| 1598 |
-
"eval_runtime": 0.2541,
|
| 1599 |
-
"eval_samples_per_second": 196.804,
|
| 1600 |
-
"eval_steps_per_second": 7.872,
|
| 1601 |
-
"step": 176702
|
| 1602 |
},
|
| 1603 |
{
|
| 1604 |
-
"epoch":
|
| 1605 |
-
"learning_rate":
|
| 1606 |
-
"loss": 0.
|
| 1607 |
-
"step":
|
| 1608 |
},
|
| 1609 |
{
|
| 1610 |
-
"epoch":
|
| 1611 |
-
"learning_rate":
|
| 1612 |
-
"loss": 0.
|
| 1613 |
-
"step":
|
| 1614 |
},
|
| 1615 |
{
|
| 1616 |
-
"epoch":
|
| 1617 |
-
"learning_rate":
|
| 1618 |
-
"loss": 0.
|
| 1619 |
-
"step":
|
| 1620 |
},
|
| 1621 |
{
|
| 1622 |
-
"epoch":
|
| 1623 |
-
"eval_loss": 0.
|
| 1624 |
-
"eval_max_distance":
|
| 1625 |
"eval_mean_distance": 0,
|
| 1626 |
-
"eval_runtime":
|
| 1627 |
-
"eval_samples_per_second":
|
| 1628 |
-
"eval_steps_per_second":
|
| 1629 |
-
"step":
|
| 1630 |
},
|
| 1631 |
{
|
| 1632 |
-
"epoch":
|
| 1633 |
-
"learning_rate":
|
| 1634 |
-
"loss": 0.
|
| 1635 |
-
"step":
|
| 1636 |
},
|
| 1637 |
{
|
| 1638 |
-
"epoch":
|
| 1639 |
-
"learning_rate":
|
| 1640 |
-
"loss": 0.
|
| 1641 |
-
"step":
|
| 1642 |
},
|
| 1643 |
{
|
| 1644 |
-
"epoch":
|
| 1645 |
-
"learning_rate": 9.
|
| 1646 |
-
"loss": 0.
|
| 1647 |
-
"step":
|
| 1648 |
},
|
| 1649 |
{
|
| 1650 |
-
"epoch":
|
| 1651 |
-
"learning_rate": 9.
|
| 1652 |
-
"loss": 0.
|
| 1653 |
-
"step":
|
| 1654 |
},
|
| 1655 |
{
|
| 1656 |
-
"epoch":
|
| 1657 |
-
"
|
| 1658 |
-
"
|
| 1659 |
-
"
|
| 1660 |
-
"eval_runtime": 0.2417,
|
| 1661 |
-
"eval_samples_per_second": 206.906,
|
| 1662 |
-
"eval_steps_per_second": 8.276,
|
| 1663 |
-
"step": 183370
|
| 1664 |
},
|
| 1665 |
{
|
| 1666 |
-
"epoch":
|
| 1667 |
-
"learning_rate": 8.
|
| 1668 |
-
"loss": 0.
|
| 1669 |
-
"step":
|
| 1670 |
},
|
| 1671 |
{
|
| 1672 |
-
"epoch":
|
| 1673 |
-
"learning_rate":
|
| 1674 |
"loss": 0.0004,
|
| 1675 |
-
"step":
|
| 1676 |
},
|
| 1677 |
{
|
| 1678 |
-
"epoch":
|
| 1679 |
-
"learning_rate": 7.
|
| 1680 |
"loss": 0.0005,
|
| 1681 |
-
"step":
|
| 1682 |
-
},
|
| 1683 |
-
{
|
| 1684 |
-
"epoch": 56.0,
|
| 1685 |
-
"eval_loss": 0.02997196838259697,
|
| 1686 |
-
"eval_max_distance": 11,
|
| 1687 |
-
"eval_mean_distance": 0,
|
| 1688 |
-
"eval_runtime": 0.2484,
|
| 1689 |
-
"eval_samples_per_second": 201.291,
|
| 1690 |
-
"eval_steps_per_second": 8.052,
|
| 1691 |
-
"step": 186704
|
| 1692 |
-
},
|
| 1693 |
-
{
|
| 1694 |
-
"epoch": 56.14,
|
| 1695 |
-
"learning_rate": 7.139127730009553e-05,
|
| 1696 |
-
"loss": 0.0004,
|
| 1697 |
-
"step": 187187
|
| 1698 |
},
|
| 1699 |
{
|
| 1700 |
-
"epoch":
|
| 1701 |
-
"learning_rate": 6.
|
| 1702 |
"loss": 0.0004,
|
| 1703 |
-
"step":
|
| 1704 |
},
|
| 1705 |
{
|
| 1706 |
-
"epoch":
|
| 1707 |
-
"learning_rate":
|
| 1708 |
"loss": 0.0004,
|
| 1709 |
-
"step":
|
| 1710 |
},
|
| 1711 |
{
|
| 1712 |
-
"epoch":
|
| 1713 |
-
"eval_loss": 0.
|
| 1714 |
-
"eval_max_distance":
|
| 1715 |
"eval_mean_distance": 0,
|
| 1716 |
-
"eval_runtime":
|
| 1717 |
-
"eval_samples_per_second":
|
| 1718 |
-
"eval_steps_per_second":
|
| 1719 |
-
"step":
|
| 1720 |
},
|
| 1721 |
{
|
| 1722 |
-
"epoch":
|
| 1723 |
-
"learning_rate": 5.
|
| 1724 |
"loss": 0.0004,
|
| 1725 |
-
"step":
|
| 1726 |
},
|
| 1727 |
{
|
| 1728 |
-
"epoch":
|
| 1729 |
-
"learning_rate": 4.
|
| 1730 |
"loss": 0.0004,
|
| 1731 |
-
"step":
|
| 1732 |
},
|
| 1733 |
{
|
| 1734 |
-
"epoch":
|
| 1735 |
-
"learning_rate": 4.
|
| 1736 |
"loss": 0.0004,
|
| 1737 |
-
"step":
|
| 1738 |
},
|
| 1739 |
{
|
| 1740 |
-
"epoch":
|
| 1741 |
-
"learning_rate": 3.
|
| 1742 |
"loss": 0.0003,
|
| 1743 |
-
"step":
|
| 1744 |
},
|
| 1745 |
{
|
| 1746 |
-
"epoch":
|
| 1747 |
-
"
|
| 1748 |
-
"
|
| 1749 |
-
"
|
| 1750 |
-
"eval_runtime": 0.236,
|
| 1751 |
-
"eval_samples_per_second": 211.858,
|
| 1752 |
-
"eval_steps_per_second": 8.474,
|
| 1753 |
-
"step": 193372
|
| 1754 |
-
},
|
| 1755 |
-
{
|
| 1756 |
-
"epoch": 58.25,
|
| 1757 |
-
"learning_rate": 3.247128352107356e-05,
|
| 1758 |
-
"loss": 0.0003,
|
| 1759 |
-
"step": 194194
|
| 1760 |
},
|
| 1761 |
{
|
| 1762 |
-
"epoch":
|
| 1763 |
-
"learning_rate": 2.
|
| 1764 |
"loss": 0.0004,
|
| 1765 |
-
"step":
|
| 1766 |
},
|
| 1767 |
{
|
| 1768 |
-
"epoch":
|
| 1769 |
-
"learning_rate": 2.
|
| 1770 |
-
"loss": 0.
|
| 1771 |
-
"step":
|
| 1772 |
},
|
| 1773 |
{
|
| 1774 |
-
"epoch":
|
| 1775 |
-
"
|
| 1776 |
-
"
|
| 1777 |
-
"
|
| 1778 |
-
"eval_runtime": 0.2475,
|
| 1779 |
-
"eval_samples_per_second": 201.99,
|
| 1780 |
-
"eval_steps_per_second": 8.08,
|
| 1781 |
-
"step": 196706
|
| 1782 |
},
|
| 1783 |
{
|
| 1784 |
-
"epoch":
|
| 1785 |
-
"learning_rate":
|
| 1786 |
-
"loss": 0.
|
| 1787 |
-
"step":
|
| 1788 |
},
|
| 1789 |
{
|
| 1790 |
-
"epoch":
|
| 1791 |
-
"learning_rate":
|
| 1792 |
-
"loss": 0.
|
| 1793 |
-
"step":
|
| 1794 |
},
|
| 1795 |
{
|
| 1796 |
-
"epoch":
|
| 1797 |
-
"
|
| 1798 |
-
"
|
| 1799 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1800 |
},
|
| 1801 |
{
|
| 1802 |
-
"epoch":
|
| 1803 |
-
"
|
| 1804 |
-
"
|
| 1805 |
-
"
|
| 1806 |
-
"
|
| 1807 |
-
"
|
| 1808 |
-
"
|
| 1809 |
-
"step": 200040
|
| 1810 |
-
},
|
| 1811 |
-
{
|
| 1812 |
-
"epoch": 60.0,
|
| 1813 |
-
"step": 200040,
|
| 1814 |
-
"total_flos": 1.1617191885791232e+17,
|
| 1815 |
-
"train_loss": 0.03170474885008116,
|
| 1816 |
-
"train_runtime": 15592.8332,
|
| 1817 |
-
"train_samples_per_second": 384.846,
|
| 1818 |
-
"train_steps_per_second": 12.829
|
| 1819 |
}
|
| 1820 |
],
|
| 1821 |
-
"logging_steps":
|
| 1822 |
-
"max_steps":
|
| 1823 |
-
"num_train_epochs":
|
| 1824 |
-
"save_steps":
|
| 1825 |
-
"total_flos":
|
| 1826 |
"trial_name": null,
|
| 1827 |
"trial_params": null
|
| 1828 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 68860,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 1.4522218995062446e-08,
|
| 14 |
+
"loss": 0.0,
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.1,
|
| 19 |
+
"learning_rate": 5.010165553296544e-06,
|
| 20 |
+
"loss": 0.0114,
|
| 21 |
+
"step": 345
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
},
|
| 23 |
{
|
| 24 |
+
"epoch": 0.2,
|
| 25 |
+
"learning_rate": 1.0020331106593089e-05,
|
| 26 |
+
"loss": 0.0115,
|
| 27 |
+
"step": 690
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
},
|
| 29 |
{
|
| 30 |
+
"epoch": 0.3,
|
| 31 |
+
"learning_rate": 1.5030496659889632e-05,
|
| 32 |
+
"loss": 0.0099,
|
| 33 |
+
"step": 1035
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
},
|
| 35 |
{
|
| 36 |
+
"epoch": 0.4,
|
| 37 |
+
"learning_rate": 2.0040662213186177e-05,
|
| 38 |
+
"loss": 0.0076,
|
| 39 |
+
"step": 1380
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 0.5,
|
| 43 |
+
"learning_rate": 2.505082776648272e-05,
|
| 44 |
+
"loss": 0.0047,
|
| 45 |
+
"step": 1725
|
| 46 |
},
|
| 47 |
{
|
| 48 |
+
"epoch": 0.6,
|
| 49 |
+
"learning_rate": 3.0060993319779264e-05,
|
| 50 |
+
"loss": 0.0039,
|
| 51 |
+
"step": 2070
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.7,
|
| 55 |
+
"learning_rate": 3.507115887307581e-05,
|
| 56 |
+
"loss": 0.0036,
|
| 57 |
+
"step": 2415
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.8,
|
| 61 |
+
"learning_rate": 4.0081324426372355e-05,
|
| 62 |
+
"loss": 0.0031,
|
| 63 |
+
"step": 2760
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.9,
|
| 67 |
+
"learning_rate": 4.509148997966889e-05,
|
| 68 |
+
"loss": 0.0028,
|
| 69 |
+
"step": 3105
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 1.0,
|
| 73 |
+
"eval_loss": 0.001530332607217133,
|
| 74 |
+
"eval_max_distance": 3,
|
| 75 |
+
"eval_mean_distance": 0,
|
| 76 |
+
"eval_runtime": 2.4239,
|
| 77 |
+
"eval_samples_per_second": 20.628,
|
| 78 |
+
"eval_steps_per_second": 0.825,
|
| 79 |
+
"step": 3443
|
| 80 |
},
|
| 81 |
{
|
| 82 |
+
"epoch": 1.0,
|
| 83 |
+
"learning_rate": 5.010165553296544e-05,
|
| 84 |
+
"loss": 0.0025,
|
| 85 |
+
"step": 3450
|
| 86 |
},
|
| 87 |
{
|
| 88 |
+
"epoch": 1.1,
|
| 89 |
+
"learning_rate": 5.5111821086261985e-05,
|
| 90 |
+
"loss": 0.0022,
|
| 91 |
+
"step": 3795
|
| 92 |
},
|
| 93 |
{
|
| 94 |
+
"epoch": 1.2,
|
| 95 |
+
"learning_rate": 6.012198663955853e-05,
|
| 96 |
+
"loss": 0.002,
|
| 97 |
+
"step": 4140
|
|
|
|
|
|
|
|
|
|
|
|
|
| 98 |
},
|
| 99 |
{
|
| 100 |
+
"epoch": 1.3,
|
| 101 |
+
"learning_rate": 6.513215219285507e-05,
|
| 102 |
+
"loss": 0.0022,
|
| 103 |
+
"step": 4485
|
| 104 |
},
|
| 105 |
{
|
| 106 |
+
"epoch": 1.4,
|
| 107 |
+
"learning_rate": 7.014231774615162e-05,
|
| 108 |
+
"loss": 0.0022,
|
| 109 |
+
"step": 4830
|
| 110 |
},
|
| 111 |
{
|
| 112 |
+
"epoch": 1.5,
|
| 113 |
+
"learning_rate": 7.515248329944817e-05,
|
| 114 |
+
"loss": 0.002,
|
| 115 |
+
"step": 5175
|
| 116 |
},
|
| 117 |
{
|
| 118 |
+
"epoch": 1.6,
|
| 119 |
+
"learning_rate": 8.016264885274471e-05,
|
| 120 |
+
"loss": 0.0021,
|
| 121 |
+
"step": 5520
|
|
|
|
|
|
|
|
|
|
|
|
|
| 122 |
},
|
| 123 |
{
|
| 124 |
+
"epoch": 1.7,
|
| 125 |
+
"learning_rate": 8.517281440604125e-05,
|
| 126 |
+
"loss": 0.0021,
|
| 127 |
+
"step": 5865
|
| 128 |
},
|
| 129 |
{
|
| 130 |
+
"epoch": 1.8,
|
| 131 |
+
"learning_rate": 9.018297995933778e-05,
|
| 132 |
+
"loss": 0.0019,
|
| 133 |
+
"step": 6210
|
| 134 |
},
|
| 135 |
{
|
| 136 |
+
"epoch": 1.9,
|
| 137 |
+
"learning_rate": 9.519314551263433e-05,
|
| 138 |
+
"loss": 0.0019,
|
| 139 |
+
"step": 6555
|
| 140 |
},
|
| 141 |
{
|
| 142 |
+
"epoch": 2.0,
|
| 143 |
+
"eval_loss": 0.0008699939935468137,
|
| 144 |
+
"eval_max_distance": 3,
|
| 145 |
"eval_mean_distance": 0,
|
| 146 |
+
"eval_runtime": 2.4743,
|
| 147 |
+
"eval_samples_per_second": 20.208,
|
| 148 |
+
"eval_steps_per_second": 0.808,
|
| 149 |
+
"step": 6886
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 150 |
},
|
| 151 |
{
|
| 152 |
+
"epoch": 2.0,
|
| 153 |
+
"learning_rate": 9.997740988156324e-05,
|
| 154 |
+
"loss": 0.0021,
|
| 155 |
+
"step": 6900
|
| 156 |
},
|
| 157 |
{
|
| 158 |
+
"epoch": 2.1,
|
| 159 |
+
"learning_rate": 9.942072482008585e-05,
|
| 160 |
+
"loss": 0.0014,
|
| 161 |
+
"step": 7245
|
| 162 |
},
|
| 163 |
{
|
| 164 |
+
"epoch": 2.2,
|
| 165 |
+
"learning_rate": 9.886403975860845e-05,
|
| 166 |
+
"loss": 0.002,
|
| 167 |
+
"step": 7590
|
|
|
|
|
|
|
|
|
|
|
|
|
| 168 |
},
|
| 169 |
{
|
| 170 |
+
"epoch": 2.3,
|
| 171 |
+
"learning_rate": 9.830735469713106e-05,
|
| 172 |
+
"loss": 0.0015,
|
| 173 |
+
"step": 7935
|
| 174 |
},
|
| 175 |
{
|
| 176 |
+
"epoch": 2.4,
|
| 177 |
+
"learning_rate": 9.775066963565367e-05,
|
| 178 |
+
"loss": 0.0012,
|
| 179 |
+
"step": 8280
|
| 180 |
},
|
| 181 |
{
|
| 182 |
+
"epoch": 2.51,
|
| 183 |
+
"learning_rate": 9.719398457417627e-05,
|
| 184 |
+
"loss": 0.0017,
|
| 185 |
+
"step": 8625
|
| 186 |
},
|
| 187 |
{
|
| 188 |
+
"epoch": 2.61,
|
| 189 |
+
"learning_rate": 9.663729951269888e-05,
|
| 190 |
+
"loss": 0.0013,
|
| 191 |
+
"step": 8970
|
|
|
|
|
|
|
|
|
|
|
|
|
| 192 |
},
|
| 193 |
{
|
| 194 |
+
"epoch": 2.71,
|
| 195 |
+
"learning_rate": 9.608061445122149e-05,
|
| 196 |
+
"loss": 0.0015,
|
| 197 |
+
"step": 9315
|
| 198 |
},
|
| 199 |
{
|
| 200 |
+
"epoch": 2.81,
|
| 201 |
+
"learning_rate": 9.552392938974409e-05,
|
| 202 |
+
"loss": 0.0018,
|
| 203 |
+
"step": 9660
|
| 204 |
},
|
| 205 |
{
|
| 206 |
+
"epoch": 2.91,
|
| 207 |
+
"learning_rate": 9.49672443282667e-05,
|
| 208 |
+
"loss": 0.0016,
|
| 209 |
+
"step": 10005
|
| 210 |
},
|
| 211 |
{
|
| 212 |
+
"epoch": 3.0,
|
| 213 |
+
"eval_loss": 0.0013113931054249406,
|
| 214 |
+
"eval_max_distance": 3,
|
| 215 |
"eval_mean_distance": 0,
|
| 216 |
+
"eval_runtime": 2.4418,
|
| 217 |
+
"eval_samples_per_second": 20.477,
|
| 218 |
+
"eval_steps_per_second": 0.819,
|
| 219 |
+
"step": 10329
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 3.01,
|
| 223 |
+
"learning_rate": 9.44105592667893e-05,
|
| 224 |
+
"loss": 0.0013,
|
| 225 |
+
"step": 10350
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 226 |
},
|
| 227 |
{
|
| 228 |
+
"epoch": 3.11,
|
| 229 |
+
"learning_rate": 9.385387420531191e-05,
|
| 230 |
+
"loss": 0.0011,
|
| 231 |
+
"step": 10695
|
| 232 |
},
|
| 233 |
{
|
| 234 |
+
"epoch": 3.21,
|
| 235 |
+
"learning_rate": 9.329718914383452e-05,
|
| 236 |
+
"loss": 0.0011,
|
| 237 |
+
"step": 11040
|
|
|
|
|
|
|
|
|
|
|
|
|
| 238 |
},
|
| 239 |
{
|
| 240 |
+
"epoch": 3.31,
|
| 241 |
+
"learning_rate": 9.274050408235712e-05,
|
| 242 |
+
"loss": 0.0012,
|
| 243 |
+
"step": 11385
|
| 244 |
},
|
| 245 |
{
|
| 246 |
+
"epoch": 3.41,
|
| 247 |
+
"learning_rate": 9.218381902087973e-05,
|
| 248 |
+
"loss": 0.0013,
|
| 249 |
+
"step": 11730
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 3.51,
|
| 253 |
+
"learning_rate": 9.162713395940234e-05,
|
| 254 |
+
"loss": 0.0012,
|
| 255 |
+
"step": 12075
|
| 256 |
},
|
| 257 |
{
|
| 258 |
+
"epoch": 3.61,
|
| 259 |
+
"learning_rate": 9.107044889792494e-05,
|
| 260 |
+
"loss": 0.0012,
|
| 261 |
+
"step": 12420
|
|
|
|
|
|
|
|
|
|
|
|
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 3.71,
|
| 265 |
+
"learning_rate": 9.051376383644755e-05,
|
| 266 |
+
"loss": 0.0015,
|
| 267 |
+
"step": 12765
|
| 268 |
},
|
| 269 |
{
|
| 270 |
+
"epoch": 3.81,
|
| 271 |
+
"learning_rate": 8.995707877497016e-05,
|
| 272 |
+
"loss": 0.0011,
|
| 273 |
+
"step": 13110
|
| 274 |
},
|
| 275 |
{
|
| 276 |
+
"epoch": 3.91,
|
| 277 |
+
"learning_rate": 8.940039371349276e-05,
|
| 278 |
+
"loss": 0.0013,
|
| 279 |
+
"step": 13455
|
| 280 |
},
|
| 281 |
{
|
| 282 |
+
"epoch": 4.0,
|
| 283 |
+
"eval_loss": 0.0007705892785452306,
|
| 284 |
+
"eval_max_distance": 1,
|
| 285 |
"eval_mean_distance": 0,
|
| 286 |
+
"eval_runtime": 2.3823,
|
| 287 |
+
"eval_samples_per_second": 20.989,
|
| 288 |
+
"eval_steps_per_second": 0.84,
|
| 289 |
+
"step": 13772
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 290 |
},
|
| 291 |
{
|
| 292 |
+
"epoch": 4.01,
|
| 293 |
+
"learning_rate": 8.884370865201537e-05,
|
| 294 |
+
"loss": 0.0012,
|
| 295 |
+
"step": 13800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 296 |
},
|
| 297 |
{
|
| 298 |
+
"epoch": 4.11,
|
| 299 |
+
"learning_rate": 8.828702359053797e-05,
|
| 300 |
+
"loss": 0.0011,
|
| 301 |
+
"step": 14145
|
| 302 |
},
|
| 303 |
{
|
| 304 |
+
"epoch": 4.21,
|
| 305 |
+
"learning_rate": 8.773033852906058e-05,
|
| 306 |
+
"loss": 0.001,
|
| 307 |
+
"step": 14490
|
|
|
|
|
|
|
|
|
|
|
|
|
| 308 |
},
|
| 309 |
{
|
| 310 |
+
"epoch": 4.31,
|
| 311 |
+
"learning_rate": 8.717365346758319e-05,
|
| 312 |
+
"loss": 0.0009,
|
| 313 |
+
"step": 14835
|
| 314 |
},
|
| 315 |
{
|
| 316 |
+
"epoch": 4.41,
|
| 317 |
+
"learning_rate": 8.661696840610579e-05,
|
| 318 |
+
"loss": 0.001,
|
| 319 |
+
"step": 15180
|
| 320 |
},
|
| 321 |
{
|
| 322 |
+
"epoch": 4.51,
|
| 323 |
+
"learning_rate": 8.60602833446284e-05,
|
| 324 |
+
"loss": 0.001,
|
| 325 |
+
"step": 15525
|
| 326 |
},
|
| 327 |
{
|
| 328 |
+
"epoch": 4.61,
|
| 329 |
+
"learning_rate": 8.550359828315099e-05,
|
| 330 |
+
"loss": 0.001,
|
| 331 |
+
"step": 15870
|
|
|
|
|
|
|
|
|
|
|
|
|
| 332 |
},
|
| 333 |
{
|
| 334 |
+
"epoch": 4.71,
|
| 335 |
+
"learning_rate": 8.494691322167361e-05,
|
| 336 |
+
"loss": 0.0011,
|
| 337 |
+
"step": 16215
|
| 338 |
},
|
| 339 |
{
|
| 340 |
+
"epoch": 4.81,
|
| 341 |
+
"learning_rate": 8.439022816019622e-05,
|
| 342 |
+
"loss": 0.001,
|
| 343 |
+
"step": 16560
|
| 344 |
},
|
| 345 |
{
|
| 346 |
+
"epoch": 4.91,
|
| 347 |
+
"learning_rate": 8.383354309871882e-05,
|
| 348 |
+
"loss": 0.0012,
|
| 349 |
+
"step": 16905
|
| 350 |
},
|
| 351 |
{
|
| 352 |
+
"epoch": 5.0,
|
| 353 |
+
"eval_loss": 0.0010777737479656935,
|
| 354 |
+
"eval_max_distance": 3,
|
| 355 |
"eval_mean_distance": 0,
|
| 356 |
+
"eval_runtime": 2.413,
|
| 357 |
+
"eval_samples_per_second": 20.721,
|
| 358 |
+
"eval_steps_per_second": 0.829,
|
| 359 |
+
"step": 17215
|
| 360 |
},
|
| 361 |
{
|
| 362 |
+
"epoch": 5.01,
|
| 363 |
+
"learning_rate": 8.327685803724142e-05,
|
| 364 |
+
"loss": 0.001,
|
| 365 |
+
"step": 17250
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 366 |
},
|
| 367 |
{
|
| 368 |
+
"epoch": 5.11,
|
| 369 |
+
"learning_rate": 8.272017297576404e-05,
|
| 370 |
+
"loss": 0.0008,
|
| 371 |
+
"step": 17595
|
| 372 |
},
|
| 373 |
{
|
| 374 |
+
"epoch": 5.21,
|
| 375 |
+
"learning_rate": 8.216348791428664e-05,
|
| 376 |
+
"loss": 0.0009,
|
| 377 |
+
"step": 17940
|
|
|
|
|
|
|
|
|
|
|
|
|
| 378 |
},
|
| 379 |
{
|
| 380 |
+
"epoch": 5.31,
|
| 381 |
+
"learning_rate": 8.160680285280925e-05,
|
| 382 |
+
"loss": 0.0009,
|
| 383 |
+
"step": 18285
|
| 384 |
},
|
| 385 |
{
|
| 386 |
+
"epoch": 5.41,
|
| 387 |
+
"learning_rate": 8.105011779133184e-05,
|
| 388 |
+
"loss": 0.0009,
|
| 389 |
+
"step": 18630
|
| 390 |
},
|
| 391 |
{
|
| 392 |
+
"epoch": 5.51,
|
| 393 |
+
"learning_rate": 8.049343272985446e-05,
|
| 394 |
+
"loss": 0.0009,
|
| 395 |
+
"step": 18975
|
| 396 |
},
|
| 397 |
{
|
| 398 |
+
"epoch": 5.61,
|
| 399 |
+
"learning_rate": 7.993674766837707e-05,
|
| 400 |
+
"loss": 0.0008,
|
| 401 |
+
"step": 19320
|
|
|
|
|
|
|
|
|
|
|
|
|
| 402 |
},
|
| 403 |
{
|
| 404 |
+
"epoch": 5.71,
|
| 405 |
+
"learning_rate": 7.938006260689967e-05,
|
| 406 |
+
"loss": 0.0009,
|
| 407 |
+
"step": 19665
|
| 408 |
},
|
| 409 |
{
|
| 410 |
+
"epoch": 5.81,
|
| 411 |
+
"learning_rate": 7.882337754542227e-05,
|
| 412 |
+
"loss": 0.0011,
|
| 413 |
+
"step": 20010
|
| 414 |
},
|
| 415 |
{
|
| 416 |
+
"epoch": 5.91,
|
| 417 |
+
"learning_rate": 7.826669248394489e-05,
|
| 418 |
+
"loss": 0.0009,
|
| 419 |
+
"step": 20355
|
| 420 |
},
|
| 421 |
{
|
| 422 |
+
"epoch": 6.0,
|
| 423 |
+
"eval_loss": 0.0008971834322437644,
|
| 424 |
+
"eval_max_distance": 3,
|
| 425 |
"eval_mean_distance": 0,
|
| 426 |
+
"eval_runtime": 2.4223,
|
| 427 |
+
"eval_samples_per_second": 20.641,
|
| 428 |
+
"eval_steps_per_second": 0.826,
|
| 429 |
+
"step": 20658
|
| 430 |
},
|
| 431 |
{
|
| 432 |
+
"epoch": 6.01,
|
| 433 |
+
"learning_rate": 7.77100074224675e-05,
|
| 434 |
+
"loss": 0.001,
|
| 435 |
+
"step": 20700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 436 |
},
|
| 437 |
{
|
| 438 |
+
"epoch": 6.11,
|
| 439 |
+
"learning_rate": 7.71533223609901e-05,
|
| 440 |
+
"loss": 0.0007,
|
| 441 |
+
"step": 21045
|
| 442 |
},
|
| 443 |
{
|
| 444 |
+
"epoch": 6.21,
|
| 445 |
+
"learning_rate": 7.65966372995127e-05,
|
| 446 |
+
"loss": 0.0008,
|
| 447 |
+
"step": 21390
|
|
|
|
|
|
|
|
|
|
|
|
|
| 448 |
},
|
| 449 |
{
|
| 450 |
+
"epoch": 6.31,
|
| 451 |
+
"learning_rate": 7.603995223803531e-05,
|
| 452 |
+
"loss": 0.0008,
|
| 453 |
+
"step": 21735
|
| 454 |
},
|
| 455 |
{
|
| 456 |
+
"epoch": 6.41,
|
| 457 |
+
"learning_rate": 7.548326717655792e-05,
|
| 458 |
+
"loss": 0.0008,
|
| 459 |
+
"step": 22080
|
| 460 |
},
|
| 461 |
{
|
| 462 |
+
"epoch": 6.51,
|
| 463 |
+
"learning_rate": 7.492658211508052e-05,
|
| 464 |
+
"loss": 0.0008,
|
| 465 |
+
"step": 22425
|
| 466 |
},
|
| 467 |
{
|
| 468 |
+
"epoch": 6.61,
|
| 469 |
+
"learning_rate": 7.436989705360313e-05,
|
| 470 |
+
"loss": 0.0008,
|
| 471 |
+
"step": 22770
|
|
|
|
|
|
|
|
|
|
|
|
|
| 472 |
},
|
| 473 |
{
|
| 474 |
+
"epoch": 6.71,
|
| 475 |
+
"learning_rate": 7.381321199212572e-05,
|
| 476 |
+
"loss": 0.0008,
|
| 477 |
+
"step": 23115
|
| 478 |
},
|
| 479 |
{
|
| 480 |
+
"epoch": 6.81,
|
| 481 |
+
"learning_rate": 7.325652693064834e-05,
|
| 482 |
+
"loss": 0.0008,
|
| 483 |
+
"step": 23460
|
| 484 |
},
|
| 485 |
{
|
| 486 |
+
"epoch": 6.91,
|
| 487 |
+
"learning_rate": 7.269984186917095e-05,
|
| 488 |
+
"loss": 0.0008,
|
| 489 |
+
"step": 23805
|
| 490 |
},
|
| 491 |
{
|
| 492 |
+
"epoch": 7.0,
|
| 493 |
+
"eval_loss": 0.0011104686418548226,
|
| 494 |
+
"eval_max_distance": 3,
|
| 495 |
"eval_mean_distance": 0,
|
| 496 |
+
"eval_runtime": 2.4364,
|
| 497 |
+
"eval_samples_per_second": 20.522,
|
| 498 |
+
"eval_steps_per_second": 0.821,
|
| 499 |
+
"step": 24101
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 500 |
},
|
| 501 |
{
|
| 502 |
+
"epoch": 7.01,
|
| 503 |
+
"learning_rate": 7.214315680769356e-05,
|
| 504 |
+
"loss": 0.0007,
|
| 505 |
+
"step": 24150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 506 |
},
|
| 507 |
{
|
| 508 |
+
"epoch": 7.11,
|
| 509 |
+
"learning_rate": 7.158647174621615e-05,
|
| 510 |
+
"loss": 0.0007,
|
| 511 |
+
"step": 24495
|
| 512 |
},
|
| 513 |
{
|
| 514 |
+
"epoch": 7.21,
|
| 515 |
+
"learning_rate": 7.102978668473877e-05,
|
| 516 |
+
"loss": 0.0007,
|
| 517 |
+
"step": 24840
|
|
|
|
|
|
|
|
|
|
|
|
|
| 518 |
},
|
| 519 |
{
|
| 520 |
+
"epoch": 7.31,
|
| 521 |
+
"learning_rate": 7.047310162326137e-05,
|
| 522 |
+
"loss": 0.0007,
|
| 523 |
+
"step": 25185
|
| 524 |
},
|
| 525 |
{
|
| 526 |
+
"epoch": 7.42,
|
| 527 |
+
"learning_rate": 6.991641656178398e-05,
|
| 528 |
+
"loss": 0.0007,
|
| 529 |
+
"step": 25530
|
| 530 |
},
|
| 531 |
{
|
| 532 |
+
"epoch": 7.52,
|
| 533 |
+
"learning_rate": 6.935973150030657e-05,
|
| 534 |
+
"loss": 0.0008,
|
| 535 |
+
"step": 25875
|
| 536 |
},
|
| 537 |
{
|
| 538 |
+
"epoch": 7.62,
|
| 539 |
+
"learning_rate": 6.88030464388292e-05,
|
| 540 |
+
"loss": 0.0006,
|
| 541 |
+
"step": 26220
|
|
|
|
|
|
|
|
|
|
|
|
|
| 542 |
},
|
| 543 |
{
|
| 544 |
+
"epoch": 7.72,
|
| 545 |
+
"learning_rate": 6.82463613773518e-05,
|
| 546 |
+
"loss": 0.0008,
|
| 547 |
+
"step": 26565
|
| 548 |
},
|
| 549 |
{
|
| 550 |
+
"epoch": 7.82,
|
| 551 |
+
"learning_rate": 6.76896763158744e-05,
|
| 552 |
+
"loss": 0.0008,
|
| 553 |
+
"step": 26910
|
| 554 |
},
|
| 555 |
{
|
| 556 |
+
"epoch": 7.92,
|
| 557 |
+
"learning_rate": 6.7132991254397e-05,
|
| 558 |
+
"loss": 0.0007,
|
| 559 |
+
"step": 27255
|
| 560 |
},
|
| 561 |
{
|
| 562 |
+
"epoch": 8.0,
|
| 563 |
+
"eval_loss": 0.0009568997193127871,
|
| 564 |
+
"eval_max_distance": 3,
|
| 565 |
"eval_mean_distance": 0,
|
| 566 |
+
"eval_runtime": 2.353,
|
| 567 |
+
"eval_samples_per_second": 21.249,
|
| 568 |
+
"eval_steps_per_second": 0.85,
|
| 569 |
+
"step": 27544
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 570 |
},
|
| 571 |
{
|
| 572 |
+
"epoch": 8.02,
|
| 573 |
+
"learning_rate": 6.657630619291962e-05,
|
| 574 |
+
"loss": 0.0008,
|
| 575 |
+
"step": 27600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 576 |
},
|
| 577 |
{
|
| 578 |
+
"epoch": 8.12,
|
| 579 |
+
"learning_rate": 6.601962113144223e-05,
|
| 580 |
+
"loss": 0.0006,
|
| 581 |
+
"step": 27945
|
| 582 |
},
|
| 583 |
{
|
| 584 |
+
"epoch": 8.22,
|
| 585 |
+
"learning_rate": 6.546293606996483e-05,
|
| 586 |
+
"loss": 0.0006,
|
| 587 |
+
"step": 28290
|
|
|
|
|
|
|
|
|
|
|
|
|
| 588 |
},
|
| 589 |
{
|
| 590 |
+
"epoch": 8.32,
|
| 591 |
+
"learning_rate": 6.490625100848742e-05,
|
| 592 |
+
"loss": 0.0007,
|
| 593 |
+
"step": 28635
|
| 594 |
},
|
| 595 |
{
|
| 596 |
+
"epoch": 8.42,
|
| 597 |
+
"learning_rate": 6.434956594701004e-05,
|
| 598 |
+
"loss": 0.0007,
|
| 599 |
+
"step": 28980
|
| 600 |
},
|
| 601 |
{
|
| 602 |
+
"epoch": 8.52,
|
| 603 |
+
"learning_rate": 6.379288088553265e-05,
|
| 604 |
+
"loss": 0.0008,
|
| 605 |
+
"step": 29325
|
| 606 |
},
|
| 607 |
{
|
| 608 |
+
"epoch": 8.62,
|
| 609 |
+
"learning_rate": 6.323619582405526e-05,
|
| 610 |
+
"loss": 0.0006,
|
| 611 |
+
"step": 29670
|
|
|
|
|
|
|
|
|
|
|
|
|
| 612 |
},
|
| 613 |
{
|
| 614 |
+
"epoch": 8.72,
|
| 615 |
+
"learning_rate": 6.267951076257785e-05,
|
| 616 |
+
"loss": 0.0006,
|
| 617 |
+
"step": 30015
|
| 618 |
},
|
| 619 |
{
|
| 620 |
+
"epoch": 8.82,
|
| 621 |
+
"learning_rate": 6.212282570110046e-05,
|
| 622 |
+
"loss": 0.0007,
|
| 623 |
+
"step": 30360
|
| 624 |
},
|
| 625 |
{
|
| 626 |
+
"epoch": 8.92,
|
| 627 |
+
"learning_rate": 6.156614063962308e-05,
|
| 628 |
+
"loss": 0.0006,
|
| 629 |
+
"step": 30705
|
| 630 |
},
|
| 631 |
{
|
| 632 |
+
"epoch": 9.0,
|
| 633 |
+
"eval_loss": 0.001235798466950655,
|
| 634 |
+
"eval_max_distance": 3,
|
| 635 |
"eval_mean_distance": 0,
|
| 636 |
+
"eval_runtime": 2.3847,
|
| 637 |
+
"eval_samples_per_second": 20.967,
|
| 638 |
+
"eval_steps_per_second": 0.839,
|
| 639 |
+
"step": 30987
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 640 |
},
|
| 641 |
{
|
| 642 |
+
"epoch": 9.02,
|
| 643 |
+
"learning_rate": 6.100945557814568e-05,
|
| 644 |
+
"loss": 0.0007,
|
| 645 |
+
"step": 31050
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 646 |
},
|
| 647 |
{
|
| 648 |
+
"epoch": 9.12,
|
| 649 |
+
"learning_rate": 6.045277051666828e-05,
|
| 650 |
+
"loss": 0.0006,
|
| 651 |
+
"step": 31395
|
| 652 |
},
|
| 653 |
{
|
| 654 |
+
"epoch": 9.22,
|
| 655 |
+
"learning_rate": 5.989608545519089e-05,
|
| 656 |
+
"loss": 0.0005,
|
| 657 |
+
"step": 31740
|
|
|
|
|
|
|
|
|
|
|
|
|
| 658 |
},
|
| 659 |
{
|
| 660 |
+
"epoch": 9.32,
|
| 661 |
+
"learning_rate": 5.9339400393713494e-05,
|
| 662 |
+
"loss": 0.0006,
|
| 663 |
+
"step": 32085
|
| 664 |
},
|
| 665 |
{
|
| 666 |
+
"epoch": 9.42,
|
| 667 |
+
"learning_rate": 5.878271533223611e-05,
|
| 668 |
+
"loss": 0.0007,
|
| 669 |
+
"step": 32430
|
| 670 |
},
|
| 671 |
{
|
| 672 |
+
"epoch": 9.52,
|
| 673 |
+
"learning_rate": 5.822603027075871e-05,
|
| 674 |
+
"loss": 0.0006,
|
| 675 |
+
"step": 32775
|
| 676 |
},
|
| 677 |
{
|
| 678 |
+
"epoch": 9.62,
|
| 679 |
+
"learning_rate": 5.766934520928131e-05,
|
| 680 |
+
"loss": 0.0006,
|
| 681 |
+
"step": 33120
|
|
|
|
|
|
|
|
|
|
|
|
|
| 682 |
},
|
| 683 |
{
|
| 684 |
+
"epoch": 9.72,
|
| 685 |
+
"learning_rate": 5.711266014780392e-05,
|
| 686 |
+
"loss": 0.0006,
|
| 687 |
+
"step": 33465
|
| 688 |
},
|
| 689 |
{
|
| 690 |
+
"epoch": 9.82,
|
| 691 |
+
"learning_rate": 5.655597508632653e-05,
|
| 692 |
+
"loss": 0.0006,
|
| 693 |
+
"step": 33810
|
| 694 |
},
|
| 695 |
{
|
| 696 |
+
"epoch": 9.92,
|
| 697 |
+
"learning_rate": 5.599929002484914e-05,
|
| 698 |
+
"loss": 0.0006,
|
| 699 |
+
"step": 34155
|
| 700 |
},
|
| 701 |
{
|
| 702 |
+
"epoch": 10.0,
|
| 703 |
+
"eval_loss": 0.0008162627927958965,
|
| 704 |
+
"eval_max_distance": 3,
|
| 705 |
"eval_mean_distance": 0,
|
| 706 |
+
"eval_runtime": 2.4197,
|
| 707 |
+
"eval_samples_per_second": 20.663,
|
| 708 |
+
"eval_steps_per_second": 0.827,
|
| 709 |
+
"step": 34430
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 710 |
},
|
| 711 |
{
|
| 712 |
+
"epoch": 10.02,
|
| 713 |
+
"learning_rate": 5.544260496337174e-05,
|
| 714 |
+
"loss": 0.0006,
|
| 715 |
+
"step": 34500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 716 |
},
|
| 717 |
{
|
| 718 |
+
"epoch": 10.12,
|
| 719 |
+
"learning_rate": 5.4885919901894344e-05,
|
| 720 |
+
"loss": 0.0005,
|
| 721 |
+
"step": 34845
|
| 722 |
},
|
| 723 |
{
|
| 724 |
+
"epoch": 10.22,
|
| 725 |
+
"learning_rate": 5.432923484041696e-05,
|
| 726 |
+
"loss": 0.0006,
|
| 727 |
+
"step": 35190
|
|
|
|
|
|
|
|
|
|
|
|
|
| 728 |
},
|
| 729 |
{
|
| 730 |
+
"epoch": 10.32,
|
| 731 |
+
"learning_rate": 5.377254977893956e-05,
|
| 732 |
+
"loss": 0.0005,
|
| 733 |
+
"step": 35535
|
| 734 |
},
|
| 735 |
{
|
| 736 |
+
"epoch": 10.42,
|
| 737 |
+
"learning_rate": 5.321586471746216e-05,
|
| 738 |
+
"loss": 0.0005,
|
| 739 |
+
"step": 35880
|
| 740 |
},
|
| 741 |
{
|
| 742 |
+
"epoch": 10.52,
|
| 743 |
+
"learning_rate": 5.265917965598477e-05,
|
| 744 |
+
"loss": 0.0006,
|
| 745 |
+
"step": 36225
|
| 746 |
},
|
| 747 |
{
|
| 748 |
+
"epoch": 10.62,
|
| 749 |
+
"learning_rate": 5.2102494594507375e-05,
|
| 750 |
+
"loss": 0.0006,
|
| 751 |
+
"step": 36570
|
|
|
|
|
|
|
|
|
|
|
|
|
| 752 |
},
|
| 753 |
{
|
| 754 |
+
"epoch": 10.72,
|
| 755 |
+
"learning_rate": 5.154580953302999e-05,
|
| 756 |
+
"loss": 0.0006,
|
| 757 |
+
"step": 36915
|
| 758 |
},
|
| 759 |
{
|
| 760 |
+
"epoch": 10.82,
|
| 761 |
+
"learning_rate": 5.098912447155259e-05,
|
| 762 |
+
"loss": 0.0005,
|
| 763 |
+
"step": 37260
|
| 764 |
},
|
| 765 |
{
|
| 766 |
+
"epoch": 10.92,
|
| 767 |
+
"learning_rate": 5.0432439410075194e-05,
|
| 768 |
+
"loss": 0.0006,
|
| 769 |
+
"step": 37605
|
| 770 |
},
|
| 771 |
{
|
| 772 |
+
"epoch": 11.0,
|
| 773 |
+
"eval_loss": 0.000511307327542454,
|
| 774 |
+
"eval_max_distance": 0,
|
| 775 |
"eval_mean_distance": 0,
|
| 776 |
+
"eval_runtime": 2.378,
|
| 777 |
+
"eval_samples_per_second": 21.026,
|
| 778 |
+
"eval_steps_per_second": 0.841,
|
| 779 |
+
"step": 37873
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 780 |
},
|
| 781 |
{
|
| 782 |
+
"epoch": 11.02,
|
| 783 |
+
"learning_rate": 4.98757543485978e-05,
|
| 784 |
+
"loss": 0.0005,
|
| 785 |
+
"step": 37950
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 786 |
},
|
| 787 |
{
|
| 788 |
+
"epoch": 11.12,
|
| 789 |
+
"learning_rate": 4.931906928712041e-05,
|
| 790 |
+
"loss": 0.0006,
|
| 791 |
+
"step": 38295
|
| 792 |
},
|
| 793 |
{
|
| 794 |
+
"epoch": 11.22,
|
| 795 |
+
"learning_rate": 4.876238422564301e-05,
|
| 796 |
+
"loss": 0.0005,
|
| 797 |
+
"step": 38640
|
|
|
|
|
|
|
|
|
|
|
|
|
| 798 |
},
|
| 799 |
{
|
| 800 |
+
"epoch": 11.32,
|
| 801 |
+
"learning_rate": 4.820569916416562e-05,
|
| 802 |
+
"loss": 0.0005,
|
| 803 |
+
"step": 38985
|
| 804 |
},
|
| 805 |
{
|
| 806 |
+
"epoch": 11.42,
|
| 807 |
+
"learning_rate": 4.7649014102688226e-05,
|
| 808 |
+
"loss": 0.0005,
|
| 809 |
+
"step": 39330
|
| 810 |
},
|
| 811 |
{
|
| 812 |
+
"epoch": 11.52,
|
| 813 |
+
"learning_rate": 4.709232904121083e-05,
|
| 814 |
+
"loss": 0.0005,
|
| 815 |
+
"step": 39675
|
| 816 |
},
|
| 817 |
{
|
| 818 |
+
"epoch": 11.62,
|
| 819 |
+
"learning_rate": 4.653564397973344e-05,
|
| 820 |
+
"loss": 0.0005,
|
| 821 |
+
"step": 40020
|
|
|
|
|
|
|
|
|
|
|
|
|
| 822 |
},
|
| 823 |
{
|
| 824 |
+
"epoch": 11.72,
|
| 825 |
+
"learning_rate": 4.5978958918256045e-05,
|
| 826 |
+
"loss": 0.0006,
|
| 827 |
+
"step": 40365
|
| 828 |
},
|
| 829 |
{
|
| 830 |
+
"epoch": 11.82,
|
| 831 |
+
"learning_rate": 4.542227385677865e-05,
|
| 832 |
+
"loss": 0.0007,
|
| 833 |
+
"step": 40710
|
| 834 |
},
|
| 835 |
{
|
| 836 |
+
"epoch": 11.92,
|
| 837 |
+
"learning_rate": 4.486558879530126e-05,
|
| 838 |
+
"loss": 0.0005,
|
| 839 |
+
"step": 41055
|
| 840 |
},
|
| 841 |
{
|
| 842 |
+
"epoch": 12.0,
|
| 843 |
+
"eval_loss": 0.0007161315297707915,
|
| 844 |
+
"eval_max_distance": 1,
|
| 845 |
"eval_mean_distance": 0,
|
| 846 |
+
"eval_runtime": 2.3647,
|
| 847 |
+
"eval_samples_per_second": 21.145,
|
| 848 |
+
"eval_steps_per_second": 0.846,
|
| 849 |
+
"step": 41316
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 850 |
},
|
| 851 |
{
|
| 852 |
+
"epoch": 12.02,
|
| 853 |
+
"learning_rate": 4.430890373382386e-05,
|
| 854 |
+
"loss": 0.0005,
|
| 855 |
+
"step": 41400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 856 |
},
|
| 857 |
{
|
| 858 |
+
"epoch": 12.12,
|
| 859 |
+
"learning_rate": 4.375221867234647e-05,
|
| 860 |
+
"loss": 0.0005,
|
| 861 |
+
"step": 41745
|
| 862 |
},
|
| 863 |
{
|
| 864 |
+
"epoch": 12.22,
|
| 865 |
+
"learning_rate": 4.3195533610869076e-05,
|
| 866 |
+
"loss": 0.0005,
|
| 867 |
+
"step": 42090
|
|
|
|
|
|
|
|
|
|
|
|
|
| 868 |
},
|
| 869 |
{
|
| 870 |
+
"epoch": 12.33,
|
| 871 |
+
"learning_rate": 4.263884854939168e-05,
|
| 872 |
+
"loss": 0.0006,
|
| 873 |
+
"step": 42435
|
| 874 |
},
|
| 875 |
{
|
| 876 |
+
"epoch": 12.43,
|
| 877 |
+
"learning_rate": 4.208216348791429e-05,
|
| 878 |
+
"loss": 0.0006,
|
| 879 |
+
"step": 42780
|
| 880 |
},
|
| 881 |
{
|
| 882 |
+
"epoch": 12.53,
|
| 883 |
+
"learning_rate": 4.1525478426436895e-05,
|
| 884 |
+
"loss": 0.0005,
|
| 885 |
+
"step": 43125
|
| 886 |
},
|
| 887 |
{
|
| 888 |
+
"epoch": 12.63,
|
| 889 |
+
"learning_rate": 4.09687933649595e-05,
|
| 890 |
+
"loss": 0.0005,
|
| 891 |
+
"step": 43470
|
|
|
|
|
|
|
|
|
|
|
|
|
| 892 |
},
|
| 893 |
{
|
| 894 |
+
"epoch": 12.73,
|
| 895 |
+
"learning_rate": 4.041210830348211e-05,
|
| 896 |
+
"loss": 0.0005,
|
| 897 |
+
"step": 43815
|
| 898 |
},
|
| 899 |
{
|
| 900 |
+
"epoch": 12.83,
|
| 901 |
+
"learning_rate": 3.9855423242004714e-05,
|
| 902 |
+
"loss": 0.0007,
|
| 903 |
+
"step": 44160
|
| 904 |
},
|
| 905 |
{
|
| 906 |
+
"epoch": 12.93,
|
| 907 |
+
"learning_rate": 3.929873818052732e-05,
|
| 908 |
+
"loss": 0.0004,
|
| 909 |
+
"step": 44505
|
| 910 |
},
|
| 911 |
{
|
| 912 |
+
"epoch": 13.0,
|
| 913 |
+
"eval_loss": 0.0006984297069720924,
|
| 914 |
+
"eval_max_distance": 0,
|
| 915 |
"eval_mean_distance": 0,
|
| 916 |
+
"eval_runtime": 2.3638,
|
| 917 |
+
"eval_samples_per_second": 21.153,
|
| 918 |
+
"eval_steps_per_second": 0.846,
|
| 919 |
+
"step": 44759
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 920 |
},
|
| 921 |
{
|
| 922 |
+
"epoch": 13.03,
|
| 923 |
+
"learning_rate": 3.8742053119049926e-05,
|
| 924 |
+
"loss": 0.0005,
|
| 925 |
+
"step": 44850
|
| 926 |
},
|
| 927 |
{
|
| 928 |
+
"epoch": 13.13,
|
| 929 |
+
"learning_rate": 3.818536805757253e-05,
|
| 930 |
+
"loss": 0.0006,
|
| 931 |
+
"step": 45195
|
| 932 |
},
|
| 933 |
{
|
| 934 |
+
"epoch": 13.23,
|
| 935 |
+
"learning_rate": 3.762868299609514e-05,
|
| 936 |
+
"loss": 0.0005,
|
| 937 |
+
"step": 45540
|
|
|
|
|
|
|
|
|
|
|
|
|
| 938 |
},
|
| 939 |
{
|
| 940 |
+
"epoch": 13.33,
|
| 941 |
+
"learning_rate": 3.7071997934617745e-05,
|
| 942 |
+
"loss": 0.0005,
|
| 943 |
+
"step": 45885
|
| 944 |
},
|
| 945 |
{
|
| 946 |
+
"epoch": 13.43,
|
| 947 |
+
"learning_rate": 3.651531287314035e-05,
|
| 948 |
+
"loss": 0.0005,
|
| 949 |
+
"step": 46230
|
| 950 |
},
|
| 951 |
{
|
| 952 |
+
"epoch": 13.53,
|
| 953 |
+
"learning_rate": 3.595862781166296e-05,
|
| 954 |
+
"loss": 0.0006,
|
| 955 |
+
"step": 46575
|
| 956 |
},
|
| 957 |
{
|
| 958 |
+
"epoch": 13.63,
|
| 959 |
+
"learning_rate": 3.5401942750185564e-05,
|
| 960 |
+
"loss": 0.0005,
|
| 961 |
+
"step": 46920
|
|
|
|
|
|
|
|
|
|
|
|
|
| 962 |
},
|
| 963 |
{
|
| 964 |
+
"epoch": 13.73,
|
| 965 |
+
"learning_rate": 3.484525768870817e-05,
|
| 966 |
+
"loss": 0.0005,
|
| 967 |
+
"step": 47265
|
| 968 |
},
|
| 969 |
{
|
| 970 |
+
"epoch": 13.83,
|
| 971 |
+
"learning_rate": 3.4288572627230776e-05,
|
| 972 |
+
"loss": 0.0004,
|
| 973 |
+
"step": 47610
|
| 974 |
},
|
| 975 |
{
|
| 976 |
+
"epoch": 13.93,
|
| 977 |
+
"learning_rate": 3.373188756575338e-05,
|
| 978 |
+
"loss": 0.0006,
|
| 979 |
+
"step": 47955
|
| 980 |
},
|
| 981 |
{
|
| 982 |
+
"epoch": 14.0,
|
| 983 |
+
"eval_loss": 0.001082880888134241,
|
| 984 |
+
"eval_max_distance": 3,
|
| 985 |
"eval_mean_distance": 0,
|
| 986 |
+
"eval_runtime": 2.3639,
|
| 987 |
+
"eval_samples_per_second": 21.152,
|
| 988 |
+
"eval_steps_per_second": 0.846,
|
| 989 |
+
"step": 48202
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 990 |
},
|
| 991 |
{
|
| 992 |
+
"epoch": 14.03,
|
| 993 |
+
"learning_rate": 3.317520250427599e-05,
|
| 994 |
+
"loss": 0.0005,
|
| 995 |
+
"step": 48300
|
| 996 |
},
|
| 997 |
{
|
| 998 |
+
"epoch": 14.13,
|
| 999 |
+
"learning_rate": 3.2618517442798595e-05,
|
| 1000 |
+
"loss": 0.0005,
|
| 1001 |
+
"step": 48645
|
| 1002 |
},
|
| 1003 |
{
|
| 1004 |
+
"epoch": 14.23,
|
| 1005 |
+
"learning_rate": 3.20618323813212e-05,
|
| 1006 |
+
"loss": 0.0005,
|
| 1007 |
+
"step": 48990
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1008 |
},
|
| 1009 |
{
|
| 1010 |
+
"epoch": 14.33,
|
| 1011 |
+
"learning_rate": 3.150514731984381e-05,
|
| 1012 |
+
"loss": 0.0004,
|
| 1013 |
+
"step": 49335
|
| 1014 |
},
|
| 1015 |
{
|
| 1016 |
+
"epoch": 14.43,
|
| 1017 |
+
"learning_rate": 3.0948462258366414e-05,
|
| 1018 |
+
"loss": 0.0005,
|
| 1019 |
+
"step": 49680
|
| 1020 |
},
|
| 1021 |
{
|
| 1022 |
+
"epoch": 14.53,
|
| 1023 |
+
"learning_rate": 3.039177719688902e-05,
|
| 1024 |
+
"loss": 0.0004,
|
| 1025 |
+
"step": 50025
|
| 1026 |
},
|
| 1027 |
{
|
| 1028 |
+
"epoch": 14.63,
|
| 1029 |
+
"learning_rate": 2.9835092135411623e-05,
|
| 1030 |
+
"loss": 0.0005,
|
| 1031 |
+
"step": 50370
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
+
"epoch": 14.73,
|
| 1035 |
+
"learning_rate": 2.9278407073934233e-05,
|
| 1036 |
+
"loss": 0.0005,
|
| 1037 |
+
"step": 50715
|
| 1038 |
},
|
| 1039 |
{
|
| 1040 |
+
"epoch": 14.83,
|
| 1041 |
+
"learning_rate": 2.8721722012456836e-05,
|
| 1042 |
+
"loss": 0.0005,
|
| 1043 |
+
"step": 51060
|
| 1044 |
},
|
| 1045 |
{
|
| 1046 |
+
"epoch": 14.93,
|
| 1047 |
+
"learning_rate": 2.8165036950979445e-05,
|
| 1048 |
+
"loss": 0.0005,
|
| 1049 |
+
"step": 51405
|
| 1050 |
},
|
| 1051 |
{
|
| 1052 |
+
"epoch": 15.0,
|
| 1053 |
+
"eval_loss": 0.0008490388281643391,
|
| 1054 |
+
"eval_max_distance": 3,
|
| 1055 |
"eval_mean_distance": 0,
|
| 1056 |
+
"eval_runtime": 2.3846,
|
| 1057 |
+
"eval_samples_per_second": 20.967,
|
| 1058 |
+
"eval_steps_per_second": 0.839,
|
| 1059 |
+
"step": 51645
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
+
"epoch": 15.03,
|
| 1063 |
+
"learning_rate": 2.760835188950205e-05,
|
| 1064 |
+
"loss": 0.0005,
|
| 1065 |
+
"step": 51750
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
+
"epoch": 15.13,
|
| 1069 |
+
"learning_rate": 2.7051666828024658e-05,
|
| 1070 |
+
"loss": 0.0005,
|
| 1071 |
+
"step": 52095
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
+
"epoch": 15.23,
|
| 1075 |
+
"learning_rate": 2.649498176654726e-05,
|
| 1076 |
+
"loss": 0.0004,
|
| 1077 |
+
"step": 52440
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1078 |
},
|
| 1079 |
{
|
| 1080 |
+
"epoch": 15.33,
|
| 1081 |
+
"learning_rate": 2.593829670506987e-05,
|
| 1082 |
+
"loss": 0.0005,
|
| 1083 |
+
"step": 52785
|
| 1084 |
},
|
| 1085 |
{
|
| 1086 |
+
"epoch": 15.43,
|
| 1087 |
+
"learning_rate": 2.5381611643592473e-05,
|
| 1088 |
+
"loss": 0.0004,
|
| 1089 |
+
"step": 53130
|
| 1090 |
},
|
| 1091 |
{
|
| 1092 |
+
"epoch": 15.53,
|
| 1093 |
+
"learning_rate": 2.4824926582115083e-05,
|
| 1094 |
+
"loss": 0.0004,
|
| 1095 |
+
"step": 53475
|
| 1096 |
},
|
| 1097 |
{
|
| 1098 |
+
"epoch": 15.63,
|
| 1099 |
+
"learning_rate": 2.426824152063769e-05,
|
| 1100 |
+
"loss": 0.0005,
|
| 1101 |
+
"step": 53820
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
+
"epoch": 15.73,
|
| 1105 |
+
"learning_rate": 2.3711556459160296e-05,
|
| 1106 |
+
"loss": 0.0005,
|
| 1107 |
+
"step": 54165
|
| 1108 |
},
|
| 1109 |
{
|
| 1110 |
+
"epoch": 15.83,
|
| 1111 |
+
"learning_rate": 2.3154871397682902e-05,
|
| 1112 |
+
"loss": 0.0004,
|
| 1113 |
+
"step": 54510
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
+
"epoch": 15.93,
|
| 1117 |
+
"learning_rate": 2.2598186336205508e-05,
|
| 1118 |
+
"loss": 0.0005,
|
| 1119 |
+
"step": 54855
|
| 1120 |
},
|
| 1121 |
{
|
| 1122 |
+
"epoch": 16.0,
|
| 1123 |
+
"eval_loss": 0.0008035104838199914,
|
| 1124 |
+
"eval_max_distance": 3,
|
| 1125 |
"eval_mean_distance": 0,
|
| 1126 |
+
"eval_runtime": 2.3835,
|
| 1127 |
+
"eval_samples_per_second": 20.978,
|
| 1128 |
+
"eval_steps_per_second": 0.839,
|
| 1129 |
+
"step": 55088
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
+
"epoch": 16.03,
|
| 1133 |
+
"learning_rate": 2.2041501274728115e-05,
|
| 1134 |
+
"loss": 0.0005,
|
| 1135 |
+
"step": 55200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1136 |
},
|
| 1137 |
{
|
| 1138 |
+
"epoch": 16.13,
|
| 1139 |
+
"learning_rate": 2.1484816213250717e-05,
|
| 1140 |
+
"loss": 0.0004,
|
| 1141 |
+
"step": 55545
|
| 1142 |
},
|
| 1143 |
{
|
| 1144 |
+
"epoch": 16.23,
|
| 1145 |
+
"learning_rate": 2.0928131151773324e-05,
|
| 1146 |
+
"loss": 0.0004,
|
| 1147 |
+
"step": 55890
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1148 |
},
|
| 1149 |
{
|
| 1150 |
+
"epoch": 16.33,
|
| 1151 |
+
"learning_rate": 2.037144609029593e-05,
|
| 1152 |
+
"loss": 0.0005,
|
| 1153 |
+
"step": 56235
|
| 1154 |
},
|
| 1155 |
{
|
| 1156 |
+
"epoch": 16.43,
|
| 1157 |
+
"learning_rate": 1.9814761028818536e-05,
|
| 1158 |
+
"loss": 0.0005,
|
| 1159 |
+
"step": 56580
|
| 1160 |
},
|
| 1161 |
{
|
| 1162 |
+
"epoch": 16.53,
|
| 1163 |
+
"learning_rate": 1.9258075967341143e-05,
|
| 1164 |
+
"loss": 0.0004,
|
| 1165 |
+
"step": 56925
|
| 1166 |
},
|
| 1167 |
{
|
| 1168 |
+
"epoch": 16.63,
|
| 1169 |
+
"learning_rate": 1.870139090586375e-05,
|
| 1170 |
+
"loss": 0.0005,
|
| 1171 |
+
"step": 57270
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1172 |
},
|
| 1173 |
{
|
| 1174 |
+
"epoch": 16.73,
|
| 1175 |
+
"learning_rate": 1.8144705844386355e-05,
|
| 1176 |
+
"loss": 0.0004,
|
| 1177 |
+
"step": 57615
|
| 1178 |
},
|
| 1179 |
{
|
| 1180 |
+
"epoch": 16.83,
|
| 1181 |
+
"learning_rate": 1.758802078290896e-05,
|
| 1182 |
+
"loss": 0.0004,
|
| 1183 |
+
"step": 57960
|
| 1184 |
},
|
| 1185 |
{
|
| 1186 |
+
"epoch": 16.93,
|
| 1187 |
+
"learning_rate": 1.7031335721431568e-05,
|
| 1188 |
+
"loss": 0.0005,
|
| 1189 |
+
"step": 58305
|
| 1190 |
},
|
| 1191 |
{
|
| 1192 |
+
"epoch": 17.0,
|
| 1193 |
+
"eval_loss": 0.000769978913012892,
|
| 1194 |
+
"eval_max_distance": 3,
|
| 1195 |
"eval_mean_distance": 0,
|
| 1196 |
+
"eval_runtime": 2.3886,
|
| 1197 |
+
"eval_samples_per_second": 20.933,
|
| 1198 |
+
"eval_steps_per_second": 0.837,
|
| 1199 |
+
"step": 58531
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1200 |
},
|
| 1201 |
{
|
| 1202 |
+
"epoch": 17.03,
|
| 1203 |
+
"learning_rate": 1.6474650659954174e-05,
|
| 1204 |
+
"loss": 0.0005,
|
| 1205 |
+
"step": 58650
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1206 |
},
|
| 1207 |
{
|
| 1208 |
+
"epoch": 17.13,
|
| 1209 |
+
"learning_rate": 1.591796559847678e-05,
|
| 1210 |
+
"loss": 0.0004,
|
| 1211 |
+
"step": 58995
|
| 1212 |
},
|
| 1213 |
{
|
| 1214 |
+
"epoch": 17.23,
|
| 1215 |
+
"learning_rate": 1.5361280536999387e-05,
|
| 1216 |
+
"loss": 0.0005,
|
| 1217 |
+
"step": 59340
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1218 |
},
|
| 1219 |
{
|
| 1220 |
+
"epoch": 17.34,
|
| 1221 |
+
"learning_rate": 1.4804595475521993e-05,
|
| 1222 |
+
"loss": 0.0005,
|
| 1223 |
+
"step": 59685
|
| 1224 |
},
|
| 1225 |
{
|
| 1226 |
+
"epoch": 17.44,
|
| 1227 |
+
"learning_rate": 1.4247910414044599e-05,
|
| 1228 |
+
"loss": 0.0004,
|
| 1229 |
+
"step": 60030
|
| 1230 |
},
|
| 1231 |
{
|
| 1232 |
+
"epoch": 17.54,
|
| 1233 |
+
"learning_rate": 1.3691225352567205e-05,
|
| 1234 |
+
"loss": 0.0004,
|
| 1235 |
+
"step": 60375
|
| 1236 |
},
|
| 1237 |
{
|
| 1238 |
+
"epoch": 17.64,
|
| 1239 |
+
"learning_rate": 1.3134540291089812e-05,
|
| 1240 |
+
"loss": 0.0004,
|
| 1241 |
+
"step": 60720
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
+
"epoch": 17.74,
|
| 1245 |
+
"learning_rate": 1.2577855229612418e-05,
|
| 1246 |
+
"loss": 0.0004,
|
| 1247 |
+
"step": 61065
|
| 1248 |
},
|
| 1249 |
{
|
| 1250 |
+
"epoch": 17.84,
|
| 1251 |
+
"learning_rate": 1.2021170168135024e-05,
|
| 1252 |
+
"loss": 0.0004,
|
| 1253 |
+
"step": 61410
|
| 1254 |
},
|
| 1255 |
{
|
| 1256 |
+
"epoch": 17.94,
|
| 1257 |
+
"learning_rate": 1.146448510665763e-05,
|
| 1258 |
+
"loss": 0.0004,
|
| 1259 |
+
"step": 61755
|
| 1260 |
},
|
| 1261 |
{
|
| 1262 |
+
"epoch": 18.0,
|
| 1263 |
+
"eval_loss": 0.0007153275073505938,
|
| 1264 |
+
"eval_max_distance": 3,
|
| 1265 |
"eval_mean_distance": 0,
|
| 1266 |
+
"eval_runtime": 2.3591,
|
| 1267 |
+
"eval_samples_per_second": 21.194,
|
| 1268 |
+
"eval_steps_per_second": 0.848,
|
| 1269 |
+
"step": 61974
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
+
"epoch": 18.04,
|
| 1273 |
+
"learning_rate": 1.0907800045180237e-05,
|
| 1274 |
+
"loss": 0.0003,
|
| 1275 |
+
"step": 62100
|
| 1276 |
},
|
| 1277 |
{
|
| 1278 |
+
"epoch": 18.14,
|
| 1279 |
+
"learning_rate": 1.0351114983702843e-05,
|
| 1280 |
+
"loss": 0.0004,
|
| 1281 |
+
"step": 62445
|
| 1282 |
},
|
| 1283 |
{
|
| 1284 |
+
"epoch": 18.24,
|
| 1285 |
+
"learning_rate": 9.79442992222545e-06,
|
| 1286 |
+
"loss": 0.0004,
|
| 1287 |
+
"step": 62790
|
| 1288 |
},
|
| 1289 |
{
|
| 1290 |
+
"epoch": 18.34,
|
| 1291 |
+
"learning_rate": 9.237744860748056e-06,
|
| 1292 |
+
"loss": 0.0004,
|
| 1293 |
+
"step": 63135
|
| 1294 |
},
|
| 1295 |
{
|
| 1296 |
+
"epoch": 18.44,
|
| 1297 |
+
"learning_rate": 8.681059799270662e-06,
|
| 1298 |
+
"loss": 0.0004,
|
| 1299 |
+
"step": 63480
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1300 |
},
|
| 1301 |
{
|
| 1302 |
+
"epoch": 18.54,
|
| 1303 |
+
"learning_rate": 8.124374737793268e-06,
|
| 1304 |
+
"loss": 0.0004,
|
| 1305 |
+
"step": 63825
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
+
"epoch": 18.64,
|
| 1309 |
+
"learning_rate": 7.5676896763158745e-06,
|
| 1310 |
"loss": 0.0004,
|
| 1311 |
+
"step": 64170
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
+
"epoch": 18.74,
|
| 1315 |
+
"learning_rate": 7.01100461483848e-06,
|
| 1316 |
"loss": 0.0005,
|
| 1317 |
+
"step": 64515
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1318 |
},
|
| 1319 |
{
|
| 1320 |
+
"epoch": 18.84,
|
| 1321 |
+
"learning_rate": 6.454319553361088e-06,
|
| 1322 |
"loss": 0.0004,
|
| 1323 |
+
"step": 64860
|
| 1324 |
},
|
| 1325 |
{
|
| 1326 |
+
"epoch": 18.94,
|
| 1327 |
+
"learning_rate": 5.897634491883693e-06,
|
| 1328 |
"loss": 0.0004,
|
| 1329 |
+
"step": 65205
|
| 1330 |
},
|
| 1331 |
{
|
| 1332 |
+
"epoch": 19.0,
|
| 1333 |
+
"eval_loss": 0.0007207673625089228,
|
| 1334 |
+
"eval_max_distance": 3,
|
| 1335 |
"eval_mean_distance": 0,
|
| 1336 |
+
"eval_runtime": 2.3645,
|
| 1337 |
+
"eval_samples_per_second": 21.147,
|
| 1338 |
+
"eval_steps_per_second": 0.846,
|
| 1339 |
+
"step": 65417
|
| 1340 |
},
|
| 1341 |
{
|
| 1342 |
+
"epoch": 19.04,
|
| 1343 |
+
"learning_rate": 5.3409494304063e-06,
|
| 1344 |
"loss": 0.0004,
|
| 1345 |
+
"step": 65550
|
| 1346 |
},
|
| 1347 |
{
|
| 1348 |
+
"epoch": 19.14,
|
| 1349 |
+
"learning_rate": 4.784264368928906e-06,
|
| 1350 |
"loss": 0.0004,
|
| 1351 |
+
"step": 65895
|
| 1352 |
},
|
| 1353 |
{
|
| 1354 |
+
"epoch": 19.24,
|
| 1355 |
+
"learning_rate": 4.227579307451512e-06,
|
| 1356 |
"loss": 0.0004,
|
| 1357 |
+
"step": 66240
|
| 1358 |
},
|
| 1359 |
{
|
| 1360 |
+
"epoch": 19.34,
|
| 1361 |
+
"learning_rate": 3.6708942459741184e-06,
|
| 1362 |
"loss": 0.0003,
|
| 1363 |
+
"step": 66585
|
| 1364 |
},
|
| 1365 |
{
|
| 1366 |
+
"epoch": 19.44,
|
| 1367 |
+
"learning_rate": 3.1142091844967247e-06,
|
| 1368 |
+
"loss": 0.0004,
|
| 1369 |
+
"step": 66930
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1370 |
},
|
| 1371 |
{
|
| 1372 |
+
"epoch": 19.54,
|
| 1373 |
+
"learning_rate": 2.557524123019331e-06,
|
| 1374 |
"loss": 0.0004,
|
| 1375 |
+
"step": 67275
|
| 1376 |
},
|
| 1377 |
{
|
| 1378 |
+
"epoch": 19.64,
|
| 1379 |
+
"learning_rate": 2.000839061541937e-06,
|
| 1380 |
+
"loss": 0.0005,
|
| 1381 |
+
"step": 67620
|
| 1382 |
},
|
| 1383 |
{
|
| 1384 |
+
"epoch": 19.74,
|
| 1385 |
+
"learning_rate": 1.4441540000645434e-06,
|
| 1386 |
+
"loss": 0.0005,
|
| 1387 |
+
"step": 67965
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1388 |
},
|
| 1389 |
{
|
| 1390 |
+
"epoch": 19.84,
|
| 1391 |
+
"learning_rate": 8.874689385871494e-07,
|
| 1392 |
+
"loss": 0.0004,
|
| 1393 |
+
"step": 68310
|
| 1394 |
},
|
| 1395 |
{
|
| 1396 |
+
"epoch": 19.94,
|
| 1397 |
+
"learning_rate": 3.307838771097557e-07,
|
| 1398 |
+
"loss": 0.0005,
|
| 1399 |
+
"step": 68655
|
| 1400 |
},
|
| 1401 |
{
|
| 1402 |
+
"epoch": 20.0,
|
| 1403 |
+
"eval_loss": 0.0007007673266343772,
|
| 1404 |
+
"eval_max_distance": 3,
|
| 1405 |
+
"eval_mean_distance": 0,
|
| 1406 |
+
"eval_runtime": 2.376,
|
| 1407 |
+
"eval_samples_per_second": 21.044,
|
| 1408 |
+
"eval_steps_per_second": 0.842,
|
| 1409 |
+
"step": 68860
|
| 1410 |
},
|
| 1411 |
{
|
| 1412 |
+
"epoch": 20.0,
|
| 1413 |
+
"step": 68860,
|
| 1414 |
+
"total_flos": 4.025961075861504e+16,
|
| 1415 |
+
"train_loss": 0.0010230464439000071,
|
| 1416 |
+
"train_runtime": 6513.3096,
|
| 1417 |
+
"train_samples_per_second": 317.166,
|
| 1418 |
+
"train_steps_per_second": 10.572
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1419 |
}
|
| 1420 |
],
|
| 1421 |
+
"logging_steps": 345,
|
| 1422 |
+
"max_steps": 68860,
|
| 1423 |
+
"num_train_epochs": 20,
|
| 1424 |
+
"save_steps": 689,
|
| 1425 |
+
"total_flos": 4.025961075861504e+16,
|
| 1426 |
"trial_name": null,
|
| 1427 |
"trial_params": null
|
| 1428 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4091
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:586ed9c845a3561a92f038b449b118f6c3411340e86cc6c2b3ab167a6c2d8141
|
| 3 |
size 4091
|