Sanjib Narzary
commited on
Commit
·
a393828
1
Parent(s):
ee21de7
Pretrained Bodo LM using Roberta base configuration from scratch and line by line
Browse files- README.md +5 -6
- all_results.json +10 -10
- config.json +2 -2
- eval_results.json +6 -6
- merges.txt +0 -0
- pytorch_model.bin +2 -2
- special_tokens_map.json +43 -7
- tokenizer.json +0 -0
- tokenizer_config.json +58 -9
- train_results.json +4 -4
- trainer_state.json +342 -342
- training_args.bin +1 -1
- vocab.json +0 -0
README.md
CHANGED
|
@@ -1,23 +1,22 @@
|
|
| 1 |
---
|
| 2 |
-
license: mit
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
| 5 |
metrics:
|
| 6 |
- accuracy
|
| 7 |
model-index:
|
| 8 |
-
- name: brx-roberta-base-mlm
|
| 9 |
results: []
|
| 10 |
---
|
| 11 |
|
| 12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 13 |
should probably proofread and complete it, then remove this comment. -->
|
| 14 |
|
| 15 |
-
# brx-roberta-base-mlm
|
| 16 |
|
| 17 |
-
This model
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
-
- Loss:
|
| 20 |
-
- Accuracy: 0.
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 1 |
---
|
|
|
|
| 2 |
tags:
|
| 3 |
- generated_from_trainer
|
| 4 |
metrics:
|
| 5 |
- accuracy
|
| 6 |
model-index:
|
| 7 |
+
- name: brx-roberta-base-mlm-scratch-v1
|
| 8 |
results: []
|
| 9 |
---
|
| 10 |
|
| 11 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 12 |
should probably proofread and complete it, then remove this comment. -->
|
| 13 |
|
| 14 |
+
# brx-roberta-base-mlm-scratch-v1
|
| 15 |
|
| 16 |
+
This model was trained from scratch on an unknown dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: nan
|
| 19 |
+
- Accuracy: 0.1685
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
all_results.json
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss":
|
| 5 |
-
"eval_runtime":
|
| 6 |
"eval_samples": 8796,
|
| 7 |
-
"eval_samples_per_second":
|
| 8 |
-
"eval_steps_per_second":
|
| 9 |
-
"perplexity":
|
| 10 |
-
"train_loss":
|
| 11 |
-
"train_runtime":
|
| 12 |
"train_samples": 112168,
|
| 13 |
-
"train_samples_per_second":
|
| 14 |
-
"train_steps_per_second":
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.16853486319505737,
|
| 4 |
+
"eval_loss": NaN,
|
| 5 |
+
"eval_runtime": 84.0751,
|
| 6 |
"eval_samples": 8796,
|
| 7 |
+
"eval_samples_per_second": 104.621,
|
| 8 |
+
"eval_steps_per_second": 52.31,
|
| 9 |
+
"perplexity": NaN,
|
| 10 |
+
"train_loss": 4.168116396368988,
|
| 11 |
+
"train_runtime": 19036.7524,
|
| 12 |
"train_samples": 112168,
|
| 13 |
+
"train_samples_per_second": 17.677,
|
| 14 |
+
"train_steps_per_second": 8.838
|
| 15 |
}
|
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "roberta-base",
|
| 3 |
"architectures": [
|
| 4 |
"RobertaForMaskedLM"
|
| 5 |
],
|
|
@@ -23,5 +23,5 @@
|
|
| 23 |
"transformers_version": "4.26.0.dev0",
|
| 24 |
"type_vocab_size": 1,
|
| 25 |
"use_cache": true,
|
| 26 |
-
"vocab_size":
|
| 27 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "/home/sn/pretrained_models/huggingface_pretrained/brx-roberta-base-mlm-scratch",
|
| 3 |
"architectures": [
|
| 4 |
"RobertaForMaskedLM"
|
| 5 |
],
|
|
|
|
| 23 |
"transformers_version": "4.26.0.dev0",
|
| 24 |
"type_vocab_size": 1,
|
| 25 |
"use_cache": true,
|
| 26 |
+
"vocab_size": 20000
|
| 27 |
}
|
eval_results.json
CHANGED
|
@@ -1,10 +1,10 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss":
|
| 5 |
-
"eval_runtime":
|
| 6 |
"eval_samples": 8796,
|
| 7 |
-
"eval_samples_per_second":
|
| 8 |
-
"eval_steps_per_second":
|
| 9 |
-
"perplexity":
|
| 10 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.16853486319505737,
|
| 4 |
+
"eval_loss": NaN,
|
| 5 |
+
"eval_runtime": 84.0751,
|
| 6 |
"eval_samples": 8796,
|
| 7 |
+
"eval_samples_per_second": 104.621,
|
| 8 |
+
"eval_steps_per_second": 52.31,
|
| 9 |
+
"perplexity": NaN
|
| 10 |
}
|
merges.txt
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a797a26b0c9c83ac2216c436d3f706c401816906de2c6ccff43303317fc1f595
|
| 3 |
+
size 405766571
|
special_tokens_map.json
CHANGED
|
@@ -1,15 +1,51 @@
|
|
| 1 |
{
|
| 2 |
-
"bos_token":
|
| 3 |
-
|
| 4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
"mask_token": {
|
| 6 |
"content": "<mask>",
|
| 7 |
"lstrip": true,
|
| 8 |
-
"normalized":
|
| 9 |
"rstrip": false,
|
| 10 |
"single_word": false
|
| 11 |
},
|
| 12 |
-
"pad_token":
|
| 13 |
-
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": true,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
"mask_token": {
|
| 24 |
"content": "<mask>",
|
| 25 |
"lstrip": true,
|
| 26 |
+
"normalized": true,
|
| 27 |
"rstrip": false,
|
| 28 |
"single_word": false
|
| 29 |
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": true,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": true,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": true,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
}
|
tokenizer.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1,16 +1,65 @@
|
|
| 1 |
{
|
| 2 |
"add_prefix_space": false,
|
| 3 |
-
"bos_token":
|
| 4 |
-
|
| 5 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
"errors": "replace",
|
| 7 |
-
"mask_token":
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
"special_tokens_map_file": null,
|
| 13 |
"tokenizer_class": "RobertaTokenizer",
|
| 14 |
"trim_offsets": true,
|
| 15 |
-
"unk_token":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"add_prefix_space": false,
|
| 3 |
+
"bos_token": {
|
| 4 |
+
"__type": "AddedToken",
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false
|
| 10 |
+
},
|
| 11 |
+
"cls_token": {
|
| 12 |
+
"__type": "AddedToken",
|
| 13 |
+
"content": "<s>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false
|
| 18 |
+
},
|
| 19 |
+
"eos_token": {
|
| 20 |
+
"__type": "AddedToken",
|
| 21 |
+
"content": "</s>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": true,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false
|
| 26 |
+
},
|
| 27 |
"errors": "replace",
|
| 28 |
+
"mask_token": {
|
| 29 |
+
"__type": "AddedToken",
|
| 30 |
+
"content": "<mask>",
|
| 31 |
+
"lstrip": true,
|
| 32 |
+
"normalized": true,
|
| 33 |
+
"rstrip": false,
|
| 34 |
+
"single_word": false
|
| 35 |
+
},
|
| 36 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 37 |
+
"name_or_path": "/home/sn/pretrained_models/huggingface_pretrained/brx-roberta-base-mlm-scratch",
|
| 38 |
+
"pad_token": {
|
| 39 |
+
"__type": "AddedToken",
|
| 40 |
+
"content": "<pad>",
|
| 41 |
+
"lstrip": false,
|
| 42 |
+
"normalized": true,
|
| 43 |
+
"rstrip": false,
|
| 44 |
+
"single_word": false
|
| 45 |
+
},
|
| 46 |
+
"sep_token": {
|
| 47 |
+
"__type": "AddedToken",
|
| 48 |
+
"content": "</s>",
|
| 49 |
+
"lstrip": false,
|
| 50 |
+
"normalized": true,
|
| 51 |
+
"rstrip": false,
|
| 52 |
+
"single_word": false
|
| 53 |
+
},
|
| 54 |
"special_tokens_map_file": null,
|
| 55 |
"tokenizer_class": "RobertaTokenizer",
|
| 56 |
"trim_offsets": true,
|
| 57 |
+
"unk_token": {
|
| 58 |
+
"__type": "AddedToken",
|
| 59 |
+
"content": "<unk>",
|
| 60 |
+
"lstrip": false,
|
| 61 |
+
"normalized": true,
|
| 62 |
+
"rstrip": false,
|
| 63 |
+
"single_word": false
|
| 64 |
+
}
|
| 65 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 112168,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
+
"train_loss": 4.168116396368988,
|
| 4 |
+
"train_runtime": 19036.7524,
|
| 5 |
"train_samples": 112168,
|
| 6 |
+
"train_samples_per_second": 17.677,
|
| 7 |
+
"train_steps_per_second": 8.838
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -10,2032 +10,2032 @@
|
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
"learning_rate": 4.9851413356156244e-05,
|
| 13 |
-
"loss":
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.02,
|
| 18 |
"learning_rate": 4.9702826712312485e-05,
|
| 19 |
-
"loss":
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.03,
|
| 24 |
"learning_rate": 4.9554240068468726e-05,
|
| 25 |
-
"loss":
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.04,
|
| 30 |
"learning_rate": 4.940565342462497e-05,
|
| 31 |
-
"loss":
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.04,
|
| 36 |
"learning_rate": 4.9257066780781215e-05,
|
| 37 |
-
"loss":
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.05,
|
| 42 |
"learning_rate": 4.910848013693745e-05,
|
| 43 |
-
"loss":
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.06,
|
| 48 |
"learning_rate": 4.89598934930937e-05,
|
| 49 |
-
"loss":
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.07,
|
| 54 |
"learning_rate": 4.881130684924994e-05,
|
| 55 |
-
"loss":
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.08,
|
| 60 |
"learning_rate": 4.866272020540618e-05,
|
| 61 |
-
"loss":
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.09,
|
| 66 |
"learning_rate": 4.851413356156242e-05,
|
| 67 |
-
"loss":
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.1,
|
| 72 |
"learning_rate": 4.836554691771866e-05,
|
| 73 |
-
"loss":
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.11,
|
| 78 |
"learning_rate": 4.82169602738749e-05,
|
| 79 |
-
"loss":
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.12,
|
| 84 |
"learning_rate": 4.8068373630031144e-05,
|
| 85 |
-
"loss":
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.12,
|
| 90 |
"learning_rate": 4.7919786986187386e-05,
|
| 91 |
-
"loss":
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.13,
|
| 96 |
"learning_rate": 4.7771200342343634e-05,
|
| 97 |
-
"loss":
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.14,
|
| 102 |
"learning_rate": 4.7622613698499875e-05,
|
| 103 |
-
"loss":
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.15,
|
| 108 |
"learning_rate": 4.747402705465611e-05,
|
| 109 |
-
"loss":
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.16,
|
| 114 |
"learning_rate": 4.732544041081236e-05,
|
| 115 |
-
"loss":
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.17,
|
| 120 |
"learning_rate": 4.71768537669686e-05,
|
| 121 |
-
"loss":
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.18,
|
| 126 |
"learning_rate": 4.702826712312484e-05,
|
| 127 |
-
"loss":
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.19,
|
| 132 |
"learning_rate": 4.687968047928108e-05,
|
| 133 |
-
"loss":
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.2,
|
| 138 |
"learning_rate": 4.673109383543732e-05,
|
| 139 |
-
"loss":
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.21,
|
| 144 |
"learning_rate": 4.658250719159356e-05,
|
| 145 |
-
"loss":
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.21,
|
| 150 |
"learning_rate": 4.6433920547749804e-05,
|
| 151 |
-
"loss":
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.22,
|
| 156 |
"learning_rate": 4.6285333903906045e-05,
|
| 157 |
-
"loss":
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.23,
|
| 162 |
"learning_rate": 4.613674726006229e-05,
|
| 163 |
-
"loss":
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.24,
|
| 168 |
"learning_rate": 4.5988160616218534e-05,
|
| 169 |
-
"loss":
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.25,
|
| 174 |
"learning_rate": 4.583957397237477e-05,
|
| 175 |
-
"loss":
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.26,
|
| 180 |
"learning_rate": 4.569098732853102e-05,
|
| 181 |
-
"loss":
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.27,
|
| 186 |
"learning_rate": 4.554240068468726e-05,
|
| 187 |
-
"loss":
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.28,
|
| 192 |
"learning_rate": 4.53938140408435e-05,
|
| 193 |
-
"loss":
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.29,
|
| 198 |
"learning_rate": 4.524522739699974e-05,
|
| 199 |
-
"loss":
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.29,
|
| 204 |
"learning_rate": 4.509664075315598e-05,
|
| 205 |
-
"loss":
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.3,
|
| 210 |
"learning_rate": 4.494805410931223e-05,
|
| 211 |
-
"loss":
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.31,
|
| 216 |
"learning_rate": 4.4799467465468464e-05,
|
| 217 |
-
"loss":
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.32,
|
| 222 |
"learning_rate": 4.4650880821624705e-05,
|
| 223 |
-
"loss":
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.33,
|
| 228 |
"learning_rate": 4.450229417778095e-05,
|
| 229 |
-
"loss":
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.34,
|
| 234 |
"learning_rate": 4.435370753393719e-05,
|
| 235 |
-
"loss":
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.35,
|
| 240 |
"learning_rate": 4.4205120890093435e-05,
|
| 241 |
-
"loss":
|
| 242 |
"step": 19500
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.36,
|
| 246 |
"learning_rate": 4.4056534246249676e-05,
|
| 247 |
-
"loss":
|
| 248 |
"step": 20000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.37,
|
| 252 |
"learning_rate": 4.390794760240592e-05,
|
| 253 |
-
"loss":
|
| 254 |
"step": 20500
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.37,
|
| 258 |
"learning_rate": 4.375936095856216e-05,
|
| 259 |
-
"loss":
|
| 260 |
"step": 21000
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 0.38,
|
| 264 |
"learning_rate": 4.36107743147184e-05,
|
| 265 |
-
"loss":
|
| 266 |
"step": 21500
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 0.39,
|
| 270 |
"learning_rate": 4.346218767087464e-05,
|
| 271 |
-
"loss":
|
| 272 |
"step": 22000
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.4,
|
| 276 |
"learning_rate": 4.331360102703089e-05,
|
| 277 |
-
"loss":
|
| 278 |
"step": 22500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.41,
|
| 282 |
"learning_rate": 4.3165014383187123e-05,
|
| 283 |
-
"loss":
|
| 284 |
"step": 23000
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 0.42,
|
| 288 |
"learning_rate": 4.3016427739343365e-05,
|
| 289 |
-
"loss":
|
| 290 |
"step": 23500
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 0.43,
|
| 294 |
"learning_rate": 4.286784109549961e-05,
|
| 295 |
-
"loss":
|
| 296 |
"step": 24000
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.44,
|
| 300 |
"learning_rate": 4.271925445165585e-05,
|
| 301 |
-
"loss":
|
| 302 |
"step": 24500
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.45,
|
| 306 |
"learning_rate": 4.2570667807812095e-05,
|
| 307 |
-
"loss":
|
| 308 |
"step": 25000
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.45,
|
| 312 |
"learning_rate": 4.2422081163968336e-05,
|
| 313 |
-
"loss":
|
| 314 |
"step": 25500
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 0.46,
|
| 318 |
"learning_rate": 4.227349452012458e-05,
|
| 319 |
-
"loss":
|
| 320 |
"step": 26000
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 0.47,
|
| 324 |
"learning_rate": 4.212490787628082e-05,
|
| 325 |
-
"loss":
|
| 326 |
"step": 26500
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.48,
|
| 330 |
"learning_rate": 4.197632123243706e-05,
|
| 331 |
-
"loss":
|
| 332 |
"step": 27000
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 0.49,
|
| 336 |
"learning_rate": 4.18277345885933e-05,
|
| 337 |
-
"loss":
|
| 338 |
"step": 27500
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.5,
|
| 342 |
"learning_rate": 4.167914794474955e-05,
|
| 343 |
-
"loss":
|
| 344 |
"step": 28000
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 0.51,
|
| 348 |
"learning_rate": 4.153056130090578e-05,
|
| 349 |
-
"loss":
|
| 350 |
"step": 28500
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.52,
|
| 354 |
"learning_rate": 4.138197465706203e-05,
|
| 355 |
-
"loss":
|
| 356 |
"step": 29000
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 0.53,
|
| 360 |
"learning_rate": 4.123338801321827e-05,
|
| 361 |
-
"loss":
|
| 362 |
"step": 29500
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 0.53,
|
| 366 |
"learning_rate": 4.1084801369374507e-05,
|
| 367 |
-
"loss":
|
| 368 |
"step": 30000
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 0.54,
|
| 372 |
"learning_rate": 4.0936214725530755e-05,
|
| 373 |
-
"loss":
|
| 374 |
"step": 30500
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 0.55,
|
| 378 |
"learning_rate": 4.0787628081686996e-05,
|
| 379 |
-
"loss":
|
| 380 |
"step": 31000
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.56,
|
| 384 |
"learning_rate": 4.063904143784324e-05,
|
| 385 |
-
"loss":
|
| 386 |
"step": 31500
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 0.57,
|
| 390 |
"learning_rate": 4.049045479399948e-05,
|
| 391 |
-
"loss":
|
| 392 |
"step": 32000
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 0.58,
|
| 396 |
"learning_rate": 4.034186815015572e-05,
|
| 397 |
-
"loss":
|
| 398 |
"step": 32500
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 0.59,
|
| 402 |
"learning_rate": 4.019328150631197e-05,
|
| 403 |
-
"loss":
|
| 404 |
"step": 33000
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 0.6,
|
| 408 |
"learning_rate": 4.00446948624682e-05,
|
| 409 |
-
"loss":
|
| 410 |
"step": 33500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 0.61,
|
| 414 |
"learning_rate": 3.989610821862444e-05,
|
| 415 |
-
"loss":
|
| 416 |
"step": 34000
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 0.62,
|
| 420 |
"learning_rate": 3.974752157478069e-05,
|
| 421 |
-
"loss":
|
| 422 |
"step": 34500
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.62,
|
| 426 |
"learning_rate": 3.959893493093693e-05,
|
| 427 |
-
"loss":
|
| 428 |
"step": 35000
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 0.63,
|
| 432 |
"learning_rate": 3.9450348287093166e-05,
|
| 433 |
-
"loss":
|
| 434 |
"step": 35500
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 0.64,
|
| 438 |
"learning_rate": 3.9301761643249414e-05,
|
| 439 |
-
"loss":
|
| 440 |
"step": 36000
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 0.65,
|
| 444 |
"learning_rate": 3.9153174999405655e-05,
|
| 445 |
-
"loss":
|
| 446 |
"step": 36500
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 0.66,
|
| 450 |
"learning_rate": 3.9004588355561897e-05,
|
| 451 |
-
"loss":
|
| 452 |
"step": 37000
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 0.67,
|
| 456 |
"learning_rate": 3.885600171171814e-05,
|
| 457 |
-
"loss":
|
| 458 |
"step": 37500
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 0.68,
|
| 462 |
"learning_rate": 3.870741506787438e-05,
|
| 463 |
-
"loss":
|
| 464 |
"step": 38000
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.69,
|
| 468 |
"learning_rate": 3.855882842403063e-05,
|
| 469 |
-
"loss":
|
| 470 |
"step": 38500
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 0.7,
|
| 474 |
"learning_rate": 3.841024178018686e-05,
|
| 475 |
-
"loss":
|
| 476 |
"step": 39000
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 0.7,
|
| 480 |
"learning_rate": 3.82616551363431e-05,
|
| 481 |
-
"loss":
|
| 482 |
"step": 39500
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 0.71,
|
| 486 |
"learning_rate": 3.811306849249935e-05,
|
| 487 |
-
"loss":
|
| 488 |
"step": 40000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
"epoch": 0.72,
|
| 492 |
"learning_rate": 3.796448184865559e-05,
|
| 493 |
-
"loss":
|
| 494 |
"step": 40500
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 0.73,
|
| 498 |
"learning_rate": 3.781589520481183e-05,
|
| 499 |
-
"loss":
|
| 500 |
"step": 41000
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 0.74,
|
| 504 |
"learning_rate": 3.7667308560968074e-05,
|
| 505 |
-
"loss":
|
| 506 |
"step": 41500
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.75,
|
| 510 |
"learning_rate": 3.7518721917124315e-05,
|
| 511 |
-
"loss":
|
| 512 |
"step": 42000
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 0.76,
|
| 516 |
"learning_rate": 3.737013527328056e-05,
|
| 517 |
-
"loss":
|
| 518 |
"step": 42500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 0.77,
|
| 522 |
"learning_rate": 3.72215486294368e-05,
|
| 523 |
-
"loss":
|
| 524 |
"step": 43000
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 0.78,
|
| 528 |
"learning_rate": 3.707296198559304e-05,
|
| 529 |
-
"loss":
|
| 530 |
"step": 43500
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 0.78,
|
| 534 |
"learning_rate": 3.6924375341749286e-05,
|
| 535 |
-
"loss":
|
| 536 |
"step": 44000
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 0.79,
|
| 540 |
"learning_rate": 3.677578869790552e-05,
|
| 541 |
-
"loss":
|
| 542 |
"step": 44500
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 0.8,
|
| 546 |
"learning_rate": 3.662720205406177e-05,
|
| 547 |
-
"loss":
|
| 548 |
"step": 45000
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 0.81,
|
| 552 |
"learning_rate": 3.647861541021801e-05,
|
| 553 |
-
"loss":
|
| 554 |
"step": 45500
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 0.82,
|
| 558 |
"learning_rate": 3.633002876637425e-05,
|
| 559 |
-
"loss":
|
| 560 |
"step": 46000
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 0.83,
|
| 564 |
"learning_rate": 3.618144212253049e-05,
|
| 565 |
-
"loss":
|
| 566 |
"step": 46500
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 0.84,
|
| 570 |
"learning_rate": 3.6032855478686734e-05,
|
| 571 |
-
"loss":
|
| 572 |
"step": 47000
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 0.85,
|
| 576 |
"learning_rate": 3.5884268834842975e-05,
|
| 577 |
-
"loss":
|
| 578 |
"step": 47500
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 0.86,
|
| 582 |
"learning_rate": 3.5735682190999216e-05,
|
| 583 |
-
"loss":
|
| 584 |
"step": 48000
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 0.86,
|
| 588 |
"learning_rate": 3.558709554715546e-05,
|
| 589 |
-
"loss":
|
| 590 |
"step": 48500
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 0.87,
|
| 594 |
"learning_rate": 3.54385089033117e-05,
|
| 595 |
-
"loss":
|
| 596 |
"step": 49000
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 0.88,
|
| 600 |
"learning_rate": 3.5289922259467946e-05,
|
| 601 |
-
"loss":
|
| 602 |
"step": 49500
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 0.89,
|
| 606 |
"learning_rate": 3.514133561562418e-05,
|
| 607 |
-
"loss":
|
| 608 |
"step": 50000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 0.9,
|
| 612 |
"learning_rate": 3.499274897178043e-05,
|
| 613 |
-
"loss":
|
| 614 |
"step": 50500
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 0.91,
|
| 618 |
"learning_rate": 3.484416232793667e-05,
|
| 619 |
-
"loss":
|
| 620 |
"step": 51000
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 0.92,
|
| 624 |
"learning_rate": 3.469557568409291e-05,
|
| 625 |
-
"loss":
|
| 626 |
"step": 51500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 0.93,
|
| 630 |
"learning_rate": 3.454698904024915e-05,
|
| 631 |
-
"loss":
|
| 632 |
"step": 52000
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 0.94,
|
| 636 |
"learning_rate": 3.439840239640539e-05,
|
| 637 |
-
"loss":
|
| 638 |
"step": 52500
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 0.95,
|
| 642 |
"learning_rate": 3.4249815752561634e-05,
|
| 643 |
-
"loss":
|
| 644 |
"step": 53000
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 0.95,
|
| 648 |
"learning_rate": 3.4101229108717876e-05,
|
| 649 |
-
"loss":
|
| 650 |
"step": 53500
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 0.96,
|
| 654 |
"learning_rate": 3.395264246487412e-05,
|
| 655 |
-
"loss":
|
| 656 |
"step": 54000
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 0.97,
|
| 660 |
"learning_rate": 3.3804055821030365e-05,
|
| 661 |
-
"loss":
|
| 662 |
"step": 54500
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 0.98,
|
| 666 |
"learning_rate": 3.3655469177186606e-05,
|
| 667 |
-
"loss":
|
| 668 |
"step": 55000
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 0.99,
|
| 672 |
"learning_rate": 3.350688253334284e-05,
|
| 673 |
-
"loss":
|
| 674 |
"step": 55500
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 1.0,
|
| 678 |
"learning_rate": 3.335829588949909e-05,
|
| 679 |
-
"loss":
|
| 680 |
"step": 56000
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 1.01,
|
| 684 |
"learning_rate": 3.320970924565533e-05,
|
| 685 |
-
"loss":
|
| 686 |
"step": 56500
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 1.02,
|
| 690 |
"learning_rate": 3.306112260181157e-05,
|
| 691 |
-
"loss":
|
| 692 |
"step": 57000
|
| 693 |
},
|
| 694 |
{
|
| 695 |
"epoch": 1.03,
|
| 696 |
"learning_rate": 3.291253595796781e-05,
|
| 697 |
-
"loss":
|
| 698 |
"step": 57500
|
| 699 |
},
|
| 700 |
{
|
| 701 |
"epoch": 1.03,
|
| 702 |
"learning_rate": 3.276394931412405e-05,
|
| 703 |
-
"loss":
|
| 704 |
"step": 58000
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 1.04,
|
| 708 |
"learning_rate": 3.2615362670280294e-05,
|
| 709 |
-
"loss":
|
| 710 |
"step": 58500
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 1.05,
|
| 714 |
"learning_rate": 3.2466776026436535e-05,
|
| 715 |
-
"loss":
|
| 716 |
"step": 59000
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 1.06,
|
| 720 |
"learning_rate": 3.2318189382592776e-05,
|
| 721 |
-
"loss":
|
| 722 |
"step": 59500
|
| 723 |
},
|
| 724 |
{
|
| 725 |
"epoch": 1.07,
|
| 726 |
"learning_rate": 3.2169602738749024e-05,
|
| 727 |
-
"loss":
|
| 728 |
"step": 60000
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 1.08,
|
| 732 |
"learning_rate": 3.2021016094905265e-05,
|
| 733 |
-
"loss":
|
| 734 |
"step": 60500
|
| 735 |
},
|
| 736 |
{
|
| 737 |
"epoch": 1.09,
|
| 738 |
"learning_rate": 3.18724294510615e-05,
|
| 739 |
-
"loss":
|
| 740 |
"step": 61000
|
| 741 |
},
|
| 742 |
{
|
| 743 |
"epoch": 1.1,
|
| 744 |
"learning_rate": 3.172384280721775e-05,
|
| 745 |
-
"loss":
|
| 746 |
"step": 61500
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 1.11,
|
| 750 |
"learning_rate": 3.157525616337399e-05,
|
| 751 |
-
"loss":
|
| 752 |
"step": 62000
|
| 753 |
},
|
| 754 |
{
|
| 755 |
"epoch": 1.11,
|
| 756 |
"learning_rate": 3.142666951953023e-05,
|
| 757 |
-
"loss":
|
| 758 |
"step": 62500
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 1.12,
|
| 762 |
"learning_rate": 3.127808287568647e-05,
|
| 763 |
-
"loss":
|
| 764 |
"step": 63000
|
| 765 |
},
|
| 766 |
{
|
| 767 |
"epoch": 1.13,
|
| 768 |
"learning_rate": 3.112949623184271e-05,
|
| 769 |
-
"loss":
|
| 770 |
"step": 63500
|
| 771 |
},
|
| 772 |
{
|
| 773 |
"epoch": 1.14,
|
| 774 |
"learning_rate": 3.098090958799896e-05,
|
| 775 |
-
"loss":
|
| 776 |
"step": 64000
|
| 777 |
},
|
| 778 |
{
|
| 779 |
"epoch": 1.15,
|
| 780 |
"learning_rate": 3.0832322944155195e-05,
|
| 781 |
-
"loss":
|
| 782 |
"step": 64500
|
| 783 |
},
|
| 784 |
{
|
| 785 |
"epoch": 1.16,
|
| 786 |
"learning_rate": 3.0683736300311436e-05,
|
| 787 |
-
"loss":
|
| 788 |
"step": 65000
|
| 789 |
},
|
| 790 |
{
|
| 791 |
"epoch": 1.17,
|
| 792 |
"learning_rate": 3.0535149656467684e-05,
|
| 793 |
-
"loss":
|
| 794 |
"step": 65500
|
| 795 |
},
|
| 796 |
{
|
| 797 |
"epoch": 1.18,
|
| 798 |
"learning_rate": 3.0386563012623925e-05,
|
| 799 |
-
"loss":
|
| 800 |
"step": 66000
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 1.19,
|
| 804 |
"learning_rate": 3.0237976368780163e-05,
|
| 805 |
-
"loss":
|
| 806 |
"step": 66500
|
| 807 |
},
|
| 808 |
{
|
| 809 |
"epoch": 1.19,
|
| 810 |
"learning_rate": 3.0089389724936407e-05,
|
| 811 |
-
"loss":
|
| 812 |
"step": 67000
|
| 813 |
},
|
| 814 |
{
|
| 815 |
"epoch": 1.2,
|
| 816 |
"learning_rate": 2.994080308109265e-05,
|
| 817 |
-
"loss":
|
| 818 |
"step": 67500
|
| 819 |
},
|
| 820 |
{
|
| 821 |
"epoch": 1.21,
|
| 822 |
"learning_rate": 2.9792216437248886e-05,
|
| 823 |
-
"loss":
|
| 824 |
"step": 68000
|
| 825 |
},
|
| 826 |
{
|
| 827 |
"epoch": 1.22,
|
| 828 |
"learning_rate": 2.964362979340513e-05,
|
| 829 |
-
"loss":
|
| 830 |
"step": 68500
|
| 831 |
},
|
| 832 |
{
|
| 833 |
"epoch": 1.23,
|
| 834 |
"learning_rate": 2.9495043149561376e-05,
|
| 835 |
-
"loss":
|
| 836 |
"step": 69000
|
| 837 |
},
|
| 838 |
{
|
| 839 |
"epoch": 1.24,
|
| 840 |
"learning_rate": 2.9346456505717617e-05,
|
| 841 |
-
"loss":
|
| 842 |
"step": 69500
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 1.25,
|
| 846 |
"learning_rate": 2.9197869861873855e-05,
|
| 847 |
-
"loss":
|
| 848 |
"step": 70000
|
| 849 |
},
|
| 850 |
{
|
| 851 |
"epoch": 1.26,
|
| 852 |
"learning_rate": 2.90492832180301e-05,
|
| 853 |
-
"loss":
|
| 854 |
"step": 70500
|
| 855 |
},
|
| 856 |
{
|
| 857 |
"epoch": 1.27,
|
| 858 |
"learning_rate": 2.8900696574186344e-05,
|
| 859 |
-
"loss":
|
| 860 |
"step": 71000
|
| 861 |
},
|
| 862 |
{
|
| 863 |
"epoch": 1.27,
|
| 864 |
"learning_rate": 2.875210993034258e-05,
|
| 865 |
-
"loss":
|
| 866 |
"step": 71500
|
| 867 |
},
|
| 868 |
{
|
| 869 |
"epoch": 1.28,
|
| 870 |
"learning_rate": 2.8603523286498823e-05,
|
| 871 |
-
"loss":
|
| 872 |
"step": 72000
|
| 873 |
},
|
| 874 |
{
|
| 875 |
"epoch": 1.29,
|
| 876 |
"learning_rate": 2.8454936642655067e-05,
|
| 877 |
-
"loss":
|
| 878 |
"step": 72500
|
| 879 |
},
|
| 880 |
{
|
| 881 |
"epoch": 1.3,
|
| 882 |
"learning_rate": 2.8306349998811312e-05,
|
| 883 |
-
"loss":
|
| 884 |
"step": 73000
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 1.31,
|
| 888 |
"learning_rate": 2.815776335496755e-05,
|
| 889 |
-
"loss":
|
| 890 |
"step": 73500
|
| 891 |
},
|
| 892 |
{
|
| 893 |
"epoch": 1.32,
|
| 894 |
"learning_rate": 2.800917671112379e-05,
|
| 895 |
-
"loss":
|
| 896 |
"step": 74000
|
| 897 |
},
|
| 898 |
{
|
| 899 |
"epoch": 1.33,
|
| 900 |
"learning_rate": 2.7860590067280035e-05,
|
| 901 |
-
"loss":
|
| 902 |
"step": 74500
|
| 903 |
},
|
| 904 |
{
|
| 905 |
"epoch": 1.34,
|
| 906 |
"learning_rate": 2.7712003423436276e-05,
|
| 907 |
-
"loss":
|
| 908 |
"step": 75000
|
| 909 |
},
|
| 910 |
{
|
| 911 |
"epoch": 1.35,
|
| 912 |
"learning_rate": 2.7563416779592514e-05,
|
| 913 |
-
"loss":
|
| 914 |
"step": 75500
|
| 915 |
},
|
| 916 |
{
|
| 917 |
"epoch": 1.36,
|
| 918 |
"learning_rate": 2.741483013574876e-05,
|
| 919 |
-
"loss":
|
| 920 |
"step": 76000
|
| 921 |
},
|
| 922 |
{
|
| 923 |
"epoch": 1.36,
|
| 924 |
"learning_rate": 2.7266243491905003e-05,
|
| 925 |
-
"loss":
|
| 926 |
"step": 76500
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 1.37,
|
| 930 |
"learning_rate": 2.711765684806124e-05,
|
| 931 |
-
"loss":
|
| 932 |
"step": 77000
|
| 933 |
},
|
| 934 |
{
|
| 935 |
"epoch": 1.38,
|
| 936 |
"learning_rate": 2.6969070204217482e-05,
|
| 937 |
-
"loss":
|
| 938 |
"step": 77500
|
| 939 |
},
|
| 940 |
{
|
| 941 |
"epoch": 1.39,
|
| 942 |
"learning_rate": 2.6820483560373727e-05,
|
| 943 |
-
"loss":
|
| 944 |
"step": 78000
|
| 945 |
},
|
| 946 |
{
|
| 947 |
"epoch": 1.4,
|
| 948 |
"learning_rate": 2.667189691652997e-05,
|
| 949 |
-
"loss":
|
| 950 |
"step": 78500
|
| 951 |
},
|
| 952 |
{
|
| 953 |
"epoch": 1.41,
|
| 954 |
"learning_rate": 2.652331027268621e-05,
|
| 955 |
-
"loss":
|
| 956 |
"step": 79000
|
| 957 |
},
|
| 958 |
{
|
| 959 |
"epoch": 1.42,
|
| 960 |
"learning_rate": 2.637472362884245e-05,
|
| 961 |
-
"loss":
|
| 962 |
"step": 79500
|
| 963 |
},
|
| 964 |
{
|
| 965 |
"epoch": 1.43,
|
| 966 |
"learning_rate": 2.6226136984998695e-05,
|
| 967 |
-
"loss":
|
| 968 |
"step": 80000
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 1.44,
|
| 972 |
"learning_rate": 2.607755034115494e-05,
|
| 973 |
-
"loss":
|
| 974 |
"step": 80500
|
| 975 |
},
|
| 976 |
{
|
| 977 |
"epoch": 1.44,
|
| 978 |
"learning_rate": 2.5928963697311177e-05,
|
| 979 |
-
"loss":
|
| 980 |
"step": 81000
|
| 981 |
},
|
| 982 |
{
|
| 983 |
"epoch": 1.45,
|
| 984 |
"learning_rate": 2.578037705346742e-05,
|
| 985 |
-
"loss":
|
| 986 |
"step": 81500
|
| 987 |
},
|
| 988 |
{
|
| 989 |
"epoch": 1.46,
|
| 990 |
"learning_rate": 2.5631790409623663e-05,
|
| 991 |
-
"loss":
|
| 992 |
"step": 82000
|
| 993 |
},
|
| 994 |
{
|
| 995 |
"epoch": 1.47,
|
| 996 |
"learning_rate": 2.54832037657799e-05,
|
| 997 |
-
"loss":
|
| 998 |
"step": 82500
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
"epoch": 1.48,
|
| 1002 |
"learning_rate": 2.5334617121936145e-05,
|
| 1003 |
-
"loss":
|
| 1004 |
"step": 83000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
"epoch": 1.49,
|
| 1008 |
"learning_rate": 2.5186030478092386e-05,
|
| 1009 |
-
"loss":
|
| 1010 |
"step": 83500
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 1.5,
|
| 1014 |
"learning_rate": 2.503744383424863e-05,
|
| 1015 |
-
"loss":
|
| 1016 |
"step": 84000
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
"epoch": 1.51,
|
| 1020 |
"learning_rate": 2.4888857190404872e-05,
|
| 1021 |
-
"loss":
|
| 1022 |
"step": 84500
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
"epoch": 1.52,
|
| 1026 |
"learning_rate": 2.4740270546561113e-05,
|
| 1027 |
-
"loss":
|
| 1028 |
"step": 85000
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
"epoch": 1.52,
|
| 1032 |
"learning_rate": 2.459168390271735e-05,
|
| 1033 |
-
"loss":
|
| 1034 |
"step": 85500
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
"epoch": 1.53,
|
| 1038 |
"learning_rate": 2.4443097258873596e-05,
|
| 1039 |
-
"loss":
|
| 1040 |
"step": 86000
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
"epoch": 1.54,
|
| 1044 |
"learning_rate": 2.4294510615029837e-05,
|
| 1045 |
-
"loss":
|
| 1046 |
"step": 86500
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
"epoch": 1.55,
|
| 1050 |
"learning_rate": 2.4145923971186078e-05,
|
| 1051 |
-
"loss":
|
| 1052 |
"step": 87000
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 1.56,
|
| 1056 |
"learning_rate": 2.399733732734232e-05,
|
| 1057 |
-
"loss":
|
| 1058 |
"step": 87500
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
"epoch": 1.57,
|
| 1062 |
"learning_rate": 2.3848750683498564e-05,
|
| 1063 |
-
"loss":
|
| 1064 |
"step": 88000
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
"epoch": 1.58,
|
| 1068 |
"learning_rate": 2.3700164039654805e-05,
|
| 1069 |
-
"loss":
|
| 1070 |
"step": 88500
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
"epoch": 1.59,
|
| 1074 |
"learning_rate": 2.3551577395811046e-05,
|
| 1075 |
-
"loss":
|
| 1076 |
"step": 89000
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
"epoch": 1.6,
|
| 1080 |
"learning_rate": 2.3402990751967287e-05,
|
| 1081 |
-
"loss":
|
| 1082 |
"step": 89500
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
"epoch": 1.6,
|
| 1086 |
"learning_rate": 2.3254404108123532e-05,
|
| 1087 |
-
"loss":
|
| 1088 |
"step": 90000
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
"epoch": 1.61,
|
| 1092 |
"learning_rate": 2.3105817464279773e-05,
|
| 1093 |
-
"loss":
|
| 1094 |
"step": 90500
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 1.62,
|
| 1098 |
"learning_rate": 2.2957230820436014e-05,
|
| 1099 |
-
"loss":
|
| 1100 |
"step": 91000
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
"epoch": 1.63,
|
| 1104 |
"learning_rate": 2.2808644176592255e-05,
|
| 1105 |
-
"loss":
|
| 1106 |
"step": 91500
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
"epoch": 1.64,
|
| 1110 |
"learning_rate": 2.2660057532748497e-05,
|
| 1111 |
-
"loss":
|
| 1112 |
"step": 92000
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
"epoch": 1.65,
|
| 1116 |
"learning_rate": 2.251147088890474e-05,
|
| 1117 |
-
"loss":
|
| 1118 |
"step": 92500
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
"epoch": 1.66,
|
| 1122 |
"learning_rate": 2.236288424506098e-05,
|
| 1123 |
-
"loss":
|
| 1124 |
"step": 93000
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
"epoch": 1.67,
|
| 1128 |
"learning_rate": 2.2214297601217223e-05,
|
| 1129 |
-
"loss":
|
| 1130 |
"step": 93500
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
"epoch": 1.68,
|
| 1134 |
"learning_rate": 2.2065710957373465e-05,
|
| 1135 |
-
"loss":
|
| 1136 |
"step": 94000
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 1.68,
|
| 1140 |
"learning_rate": 2.191712431352971e-05,
|
| 1141 |
-
"loss":
|
| 1142 |
"step": 94500
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
"epoch": 1.69,
|
| 1146 |
"learning_rate": 2.1768537669685947e-05,
|
| 1147 |
-
"loss":
|
| 1148 |
"step": 95000
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
"epoch": 1.7,
|
| 1152 |
"learning_rate": 2.1619951025842188e-05,
|
| 1153 |
-
"loss":
|
| 1154 |
"step": 95500
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
"epoch": 1.71,
|
| 1158 |
"learning_rate": 2.1471364381998433e-05,
|
| 1159 |
-
"loss":
|
| 1160 |
"step": 96000
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
"epoch": 1.72,
|
| 1164 |
"learning_rate": 2.1322777738154674e-05,
|
| 1165 |
-
"loss":
|
| 1166 |
"step": 96500
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
"epoch": 1.73,
|
| 1170 |
"learning_rate": 2.1174191094310915e-05,
|
| 1171 |
-
"loss":
|
| 1172 |
"step": 97000
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
"epoch": 1.74,
|
| 1176 |
"learning_rate": 2.1025604450467156e-05,
|
| 1177 |
-
"loss":
|
| 1178 |
"step": 97500
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
"epoch": 1.75,
|
| 1182 |
"learning_rate": 2.08770178066234e-05,
|
| 1183 |
-
"loss":
|
| 1184 |
"step": 98000
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
"epoch": 1.76,
|
| 1188 |
"learning_rate": 2.0728431162779642e-05,
|
| 1189 |
-
"loss":
|
| 1190 |
"step": 98500
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
"epoch": 1.77,
|
| 1194 |
"learning_rate": 2.0579844518935883e-05,
|
| 1195 |
-
"loss":
|
| 1196 |
"step": 99000
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
"epoch": 1.77,
|
| 1200 |
"learning_rate": 2.0431257875092124e-05,
|
| 1201 |
-
"loss":
|
| 1202 |
"step": 99500
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
"epoch": 1.78,
|
| 1206 |
"learning_rate": 2.0282671231248365e-05,
|
| 1207 |
-
"loss":
|
| 1208 |
"step": 100000
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
"epoch": 1.79,
|
| 1212 |
"learning_rate": 2.013408458740461e-05,
|
| 1213 |
-
"loss":
|
| 1214 |
"step": 100500
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
"epoch": 1.8,
|
| 1218 |
"learning_rate": 1.9985497943560848e-05,
|
| 1219 |
-
"loss":
|
| 1220 |
"step": 101000
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
"epoch": 1.81,
|
| 1224 |
"learning_rate": 1.9836911299717092e-05,
|
| 1225 |
-
"loss":
|
| 1226 |
"step": 101500
|
| 1227 |
},
|
| 1228 |
{
|
| 1229 |
"epoch": 1.82,
|
| 1230 |
"learning_rate": 1.9688324655873334e-05,
|
| 1231 |
-
"loss":
|
| 1232 |
"step": 102000
|
| 1233 |
},
|
| 1234 |
{
|
| 1235 |
"epoch": 1.83,
|
| 1236 |
"learning_rate": 1.9539738012029578e-05,
|
| 1237 |
-
"loss":
|
| 1238 |
"step": 102500
|
| 1239 |
},
|
| 1240 |
{
|
| 1241 |
"epoch": 1.84,
|
| 1242 |
"learning_rate": 1.9391151368185816e-05,
|
| 1243 |
-
"loss":
|
| 1244 |
"step": 103000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
"epoch": 1.85,
|
| 1248 |
"learning_rate": 1.924256472434206e-05,
|
| 1249 |
-
"loss":
|
| 1250 |
"step": 103500
|
| 1251 |
},
|
| 1252 |
{
|
| 1253 |
"epoch": 1.85,
|
| 1254 |
"learning_rate": 1.90939780804983e-05,
|
| 1255 |
-
"loss":
|
| 1256 |
"step": 104000
|
| 1257 |
},
|
| 1258 |
{
|
| 1259 |
"epoch": 1.86,
|
| 1260 |
"learning_rate": 1.8945391436654543e-05,
|
| 1261 |
-
"loss":
|
| 1262 |
"step": 104500
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
"epoch": 1.87,
|
| 1266 |
"learning_rate": 1.8796804792810784e-05,
|
| 1267 |
-
"loss":
|
| 1268 |
"step": 105000
|
| 1269 |
},
|
| 1270 |
{
|
| 1271 |
"epoch": 1.88,
|
| 1272 |
"learning_rate": 1.8648218148967025e-05,
|
| 1273 |
-
"loss":
|
| 1274 |
"step": 105500
|
| 1275 |
},
|
| 1276 |
{
|
| 1277 |
"epoch": 1.89,
|
| 1278 |
"learning_rate": 1.849963150512327e-05,
|
| 1279 |
-
"loss":
|
| 1280 |
"step": 106000
|
| 1281 |
},
|
| 1282 |
{
|
| 1283 |
"epoch": 1.9,
|
| 1284 |
"learning_rate": 1.835104486127951e-05,
|
| 1285 |
-
"loss":
|
| 1286 |
"step": 106500
|
| 1287 |
},
|
| 1288 |
{
|
| 1289 |
"epoch": 1.91,
|
| 1290 |
"learning_rate": 1.8202458217435752e-05,
|
| 1291 |
-
"loss":
|
| 1292 |
"step": 107000
|
| 1293 |
},
|
| 1294 |
{
|
| 1295 |
"epoch": 1.92,
|
| 1296 |
"learning_rate": 1.8053871573591993e-05,
|
| 1297 |
-
"loss":
|
| 1298 |
"step": 107500
|
| 1299 |
},
|
| 1300 |
{
|
| 1301 |
"epoch": 1.93,
|
| 1302 |
"learning_rate": 1.7905284929748238e-05,
|
| 1303 |
-
"loss":
|
| 1304 |
"step": 108000
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
"epoch": 1.93,
|
| 1308 |
"learning_rate": 1.775669828590448e-05,
|
| 1309 |
-
"loss":
|
| 1310 |
"step": 108500
|
| 1311 |
},
|
| 1312 |
{
|
| 1313 |
"epoch": 1.94,
|
| 1314 |
"learning_rate": 1.760811164206072e-05,
|
| 1315 |
-
"loss":
|
| 1316 |
"step": 109000
|
| 1317 |
},
|
| 1318 |
{
|
| 1319 |
"epoch": 1.95,
|
| 1320 |
"learning_rate": 1.745952499821696e-05,
|
| 1321 |
-
"loss":
|
| 1322 |
"step": 109500
|
| 1323 |
},
|
| 1324 |
{
|
| 1325 |
"epoch": 1.96,
|
| 1326 |
"learning_rate": 1.7310938354373202e-05,
|
| 1327 |
-
"loss":
|
| 1328 |
"step": 110000
|
| 1329 |
},
|
| 1330 |
{
|
| 1331 |
"epoch": 1.97,
|
| 1332 |
"learning_rate": 1.7162351710529444e-05,
|
| 1333 |
-
"loss":
|
| 1334 |
"step": 110500
|
| 1335 |
},
|
| 1336 |
{
|
| 1337 |
"epoch": 1.98,
|
| 1338 |
"learning_rate": 1.7013765066685685e-05,
|
| 1339 |
-
"loss":
|
| 1340 |
"step": 111000
|
| 1341 |
},
|
| 1342 |
{
|
| 1343 |
"epoch": 1.99,
|
| 1344 |
"learning_rate": 1.686517842284193e-05,
|
| 1345 |
-
"loss":
|
| 1346 |
"step": 111500
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
"epoch": 2.0,
|
| 1350 |
"learning_rate": 1.671659177899817e-05,
|
| 1351 |
-
"loss":
|
| 1352 |
"step": 112000
|
| 1353 |
},
|
| 1354 |
{
|
| 1355 |
"epoch": 2.01,
|
| 1356 |
"learning_rate": 1.656800513515441e-05,
|
| 1357 |
-
"loss":
|
| 1358 |
"step": 112500
|
| 1359 |
},
|
| 1360 |
{
|
| 1361 |
"epoch": 2.01,
|
| 1362 |
"learning_rate": 1.6419418491310653e-05,
|
| 1363 |
-
"loss":
|
| 1364 |
"step": 113000
|
| 1365 |
},
|
| 1366 |
{
|
| 1367 |
"epoch": 2.02,
|
| 1368 |
"learning_rate": 1.6270831847466897e-05,
|
| 1369 |
-
"loss":
|
| 1370 |
"step": 113500
|
| 1371 |
},
|
| 1372 |
{
|
| 1373 |
"epoch": 2.03,
|
| 1374 |
"learning_rate": 1.612224520362314e-05,
|
| 1375 |
-
"loss":
|
| 1376 |
"step": 114000
|
| 1377 |
},
|
| 1378 |
{
|
| 1379 |
"epoch": 2.04,
|
| 1380 |
"learning_rate": 1.597365855977938e-05,
|
| 1381 |
-
"loss":
|
| 1382 |
"step": 114500
|
| 1383 |
},
|
| 1384 |
{
|
| 1385 |
"epoch": 2.05,
|
| 1386 |
"learning_rate": 1.582507191593562e-05,
|
| 1387 |
-
"loss":
|
| 1388 |
"step": 115000
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
"epoch": 2.06,
|
| 1392 |
"learning_rate": 1.5676485272091862e-05,
|
| 1393 |
-
"loss":
|
| 1394 |
"step": 115500
|
| 1395 |
},
|
| 1396 |
{
|
| 1397 |
"epoch": 2.07,
|
| 1398 |
"learning_rate": 1.5527898628248107e-05,
|
| 1399 |
-
"loss":
|
| 1400 |
"step": 116000
|
| 1401 |
},
|
| 1402 |
{
|
| 1403 |
"epoch": 2.08,
|
| 1404 |
"learning_rate": 1.5379311984404344e-05,
|
| 1405 |
-
"loss":
|
| 1406 |
"step": 116500
|
| 1407 |
},
|
| 1408 |
{
|
| 1409 |
"epoch": 2.09,
|
| 1410 |
"learning_rate": 1.5230725340560589e-05,
|
| 1411 |
-
"loss":
|
| 1412 |
"step": 117000
|
| 1413 |
},
|
| 1414 |
{
|
| 1415 |
"epoch": 2.1,
|
| 1416 |
"learning_rate": 1.508213869671683e-05,
|
| 1417 |
-
"loss":
|
| 1418 |
"step": 117500
|
| 1419 |
},
|
| 1420 |
{
|
| 1421 |
"epoch": 2.1,
|
| 1422 |
"learning_rate": 1.4933552052873073e-05,
|
| 1423 |
-
"loss":
|
| 1424 |
"step": 118000
|
| 1425 |
},
|
| 1426 |
{
|
| 1427 |
"epoch": 2.11,
|
| 1428 |
"learning_rate": 1.4784965409029314e-05,
|
| 1429 |
-
"loss":
|
| 1430 |
"step": 118500
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
"epoch": 2.12,
|
| 1434 |
"learning_rate": 1.4636378765185554e-05,
|
| 1435 |
-
"loss":
|
| 1436 |
"step": 119000
|
| 1437 |
},
|
| 1438 |
{
|
| 1439 |
"epoch": 2.13,
|
| 1440 |
"learning_rate": 1.4487792121341798e-05,
|
| 1441 |
-
"loss":
|
| 1442 |
"step": 119500
|
| 1443 |
},
|
| 1444 |
{
|
| 1445 |
"epoch": 2.14,
|
| 1446 |
"learning_rate": 1.4339205477498038e-05,
|
| 1447 |
-
"loss":
|
| 1448 |
"step": 120000
|
| 1449 |
},
|
| 1450 |
{
|
| 1451 |
"epoch": 2.15,
|
| 1452 |
"learning_rate": 1.4190618833654282e-05,
|
| 1453 |
-
"loss":
|
| 1454 |
"step": 120500
|
| 1455 |
},
|
| 1456 |
{
|
| 1457 |
"epoch": 2.16,
|
| 1458 |
"learning_rate": 1.4042032189810522e-05,
|
| 1459 |
-
"loss":
|
| 1460 |
"step": 121000
|
| 1461 |
},
|
| 1462 |
{
|
| 1463 |
"epoch": 2.17,
|
| 1464 |
"learning_rate": 1.3893445545966766e-05,
|
| 1465 |
-
"loss":
|
| 1466 |
"step": 121500
|
| 1467 |
},
|
| 1468 |
{
|
| 1469 |
"epoch": 2.18,
|
| 1470 |
"learning_rate": 1.3744858902123006e-05,
|
| 1471 |
-
"loss":
|
| 1472 |
"step": 122000
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
"epoch": 2.18,
|
| 1476 |
"learning_rate": 1.359627225827925e-05,
|
| 1477 |
-
"loss":
|
| 1478 |
"step": 122500
|
| 1479 |
},
|
| 1480 |
{
|
| 1481 |
"epoch": 2.19,
|
| 1482 |
"learning_rate": 1.344768561443549e-05,
|
| 1483 |
-
"loss":
|
| 1484 |
"step": 123000
|
| 1485 |
},
|
| 1486 |
{
|
| 1487 |
"epoch": 2.2,
|
| 1488 |
"learning_rate": 1.3299098970591731e-05,
|
| 1489 |
-
"loss":
|
| 1490 |
"step": 123500
|
| 1491 |
},
|
| 1492 |
{
|
| 1493 |
"epoch": 2.21,
|
| 1494 |
"learning_rate": 1.3150512326747974e-05,
|
| 1495 |
-
"loss":
|
| 1496 |
"step": 124000
|
| 1497 |
},
|
| 1498 |
{
|
| 1499 |
"epoch": 2.22,
|
| 1500 |
"learning_rate": 1.3001925682904215e-05,
|
| 1501 |
-
"loss":
|
| 1502 |
"step": 124500
|
| 1503 |
},
|
| 1504 |
{
|
| 1505 |
"epoch": 2.23,
|
| 1506 |
"learning_rate": 1.2853339039060458e-05,
|
| 1507 |
-
"loss":
|
| 1508 |
"step": 125000
|
| 1509 |
},
|
| 1510 |
{
|
| 1511 |
"epoch": 2.24,
|
| 1512 |
"learning_rate": 1.2704752395216699e-05,
|
| 1513 |
-
"loss":
|
| 1514 |
"step": 125500
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
"epoch": 2.25,
|
| 1518 |
"learning_rate": 1.2556165751372942e-05,
|
| 1519 |
-
"loss":
|
| 1520 |
"step": 126000
|
| 1521 |
},
|
| 1522 |
{
|
| 1523 |
"epoch": 2.26,
|
| 1524 |
"learning_rate": 1.2407579107529183e-05,
|
| 1525 |
-
"loss":
|
| 1526 |
"step": 126500
|
| 1527 |
},
|
| 1528 |
{
|
| 1529 |
"epoch": 2.26,
|
| 1530 |
"learning_rate": 1.2258992463685424e-05,
|
| 1531 |
-
"loss":
|
| 1532 |
"step": 127000
|
| 1533 |
},
|
| 1534 |
{
|
| 1535 |
"epoch": 2.27,
|
| 1536 |
"learning_rate": 1.2110405819841667e-05,
|
| 1537 |
-
"loss":
|
| 1538 |
"step": 127500
|
| 1539 |
},
|
| 1540 |
{
|
| 1541 |
"epoch": 2.28,
|
| 1542 |
"learning_rate": 1.1961819175997908e-05,
|
| 1543 |
-
"loss":
|
| 1544 |
"step": 128000
|
| 1545 |
},
|
| 1546 |
{
|
| 1547 |
"epoch": 2.29,
|
| 1548 |
"learning_rate": 1.1813232532154151e-05,
|
| 1549 |
-
"loss":
|
| 1550 |
"step": 128500
|
| 1551 |
},
|
| 1552 |
{
|
| 1553 |
"epoch": 2.3,
|
| 1554 |
"learning_rate": 1.1664645888310392e-05,
|
| 1555 |
-
"loss":
|
| 1556 |
"step": 129000
|
| 1557 |
},
|
| 1558 |
{
|
| 1559 |
"epoch": 2.31,
|
| 1560 |
"learning_rate": 1.1516059244466634e-05,
|
| 1561 |
-
"loss":
|
| 1562 |
"step": 129500
|
| 1563 |
},
|
| 1564 |
{
|
| 1565 |
"epoch": 2.32,
|
| 1566 |
"learning_rate": 1.1367472600622876e-05,
|
| 1567 |
-
"loss":
|
| 1568 |
"step": 130000
|
| 1569 |
},
|
| 1570 |
{
|
| 1571 |
"epoch": 2.33,
|
| 1572 |
"learning_rate": 1.1218885956779118e-05,
|
| 1573 |
-
"loss":
|
| 1574 |
"step": 130500
|
| 1575 |
},
|
| 1576 |
{
|
| 1577 |
"epoch": 2.34,
|
| 1578 |
"learning_rate": 1.1070299312935359e-05,
|
| 1579 |
-
"loss":
|
| 1580 |
"step": 131000
|
| 1581 |
},
|
| 1582 |
{
|
| 1583 |
"epoch": 2.34,
|
| 1584 |
"learning_rate": 1.0921712669091602e-05,
|
| 1585 |
-
"loss":
|
| 1586 |
"step": 131500
|
| 1587 |
},
|
| 1588 |
{
|
| 1589 |
"epoch": 2.35,
|
| 1590 |
"learning_rate": 1.0773126025247843e-05,
|
| 1591 |
-
"loss":
|
| 1592 |
"step": 132000
|
| 1593 |
},
|
| 1594 |
{
|
| 1595 |
"epoch": 2.36,
|
| 1596 |
"learning_rate": 1.0624539381404084e-05,
|
| 1597 |
-
"loss":
|
| 1598 |
"step": 132500
|
| 1599 |
},
|
| 1600 |
{
|
| 1601 |
"epoch": 2.37,
|
| 1602 |
"learning_rate": 1.0475952737560327e-05,
|
| 1603 |
-
"loss":
|
| 1604 |
"step": 133000
|
| 1605 |
},
|
| 1606 |
{
|
| 1607 |
"epoch": 2.38,
|
| 1608 |
"learning_rate": 1.0327366093716568e-05,
|
| 1609 |
-
"loss":
|
| 1610 |
"step": 133500
|
| 1611 |
},
|
| 1612 |
{
|
| 1613 |
"epoch": 2.39,
|
| 1614 |
"learning_rate": 1.0178779449872811e-05,
|
| 1615 |
-
"loss":
|
| 1616 |
"step": 134000
|
| 1617 |
},
|
| 1618 |
{
|
| 1619 |
"epoch": 2.4,
|
| 1620 |
"learning_rate": 1.0030192806029052e-05,
|
| 1621 |
-
"loss":
|
| 1622 |
"step": 134500
|
| 1623 |
},
|
| 1624 |
{
|
| 1625 |
"epoch": 2.41,
|
| 1626 |
"learning_rate": 9.881606162185295e-06,
|
| 1627 |
-
"loss":
|
| 1628 |
"step": 135000
|
| 1629 |
},
|
| 1630 |
{
|
| 1631 |
"epoch": 2.42,
|
| 1632 |
"learning_rate": 9.733019518341534e-06,
|
| 1633 |
-
"loss":
|
| 1634 |
"step": 135500
|
| 1635 |
},
|
| 1636 |
{
|
| 1637 |
"epoch": 2.42,
|
| 1638 |
"learning_rate": 9.584432874497777e-06,
|
| 1639 |
-
"loss":
|
| 1640 |
"step": 136000
|
| 1641 |
},
|
| 1642 |
{
|
| 1643 |
"epoch": 2.43,
|
| 1644 |
"learning_rate": 9.435846230654018e-06,
|
| 1645 |
-
"loss":
|
| 1646 |
"step": 136500
|
| 1647 |
},
|
| 1648 |
{
|
| 1649 |
"epoch": 2.44,
|
| 1650 |
"learning_rate": 9.287259586810261e-06,
|
| 1651 |
-
"loss":
|
| 1652 |
"step": 137000
|
| 1653 |
},
|
| 1654 |
{
|
| 1655 |
"epoch": 2.45,
|
| 1656 |
"learning_rate": 9.138672942966502e-06,
|
| 1657 |
-
"loss":
|
| 1658 |
"step": 137500
|
| 1659 |
},
|
| 1660 |
{
|
| 1661 |
"epoch": 2.46,
|
| 1662 |
"learning_rate": 8.990086299122745e-06,
|
| 1663 |
-
"loss":
|
| 1664 |
"step": 138000
|
| 1665 |
},
|
| 1666 |
{
|
| 1667 |
"epoch": 2.47,
|
| 1668 |
"learning_rate": 8.841499655278986e-06,
|
| 1669 |
-
"loss":
|
| 1670 |
"step": 138500
|
| 1671 |
},
|
| 1672 |
{
|
| 1673 |
"epoch": 2.48,
|
| 1674 |
"learning_rate": 8.69291301143523e-06,
|
| 1675 |
-
"loss":
|
| 1676 |
"step": 139000
|
| 1677 |
},
|
| 1678 |
{
|
| 1679 |
"epoch": 2.49,
|
| 1680 |
"learning_rate": 8.54432636759147e-06,
|
| 1681 |
-
"loss":
|
| 1682 |
"step": 139500
|
| 1683 |
},
|
| 1684 |
{
|
| 1685 |
"epoch": 2.5,
|
| 1686 |
"learning_rate": 8.395739723747713e-06,
|
| 1687 |
-
"loss":
|
| 1688 |
"step": 140000
|
| 1689 |
},
|
| 1690 |
{
|
| 1691 |
"epoch": 2.51,
|
| 1692 |
"learning_rate": 8.247153079903953e-06,
|
| 1693 |
-
"loss":
|
| 1694 |
"step": 140500
|
| 1695 |
},
|
| 1696 |
{
|
| 1697 |
"epoch": 2.51,
|
| 1698 |
"learning_rate": 8.098566436060196e-06,
|
| 1699 |
-
"loss":
|
| 1700 |
"step": 141000
|
| 1701 |
},
|
| 1702 |
{
|
| 1703 |
"epoch": 2.52,
|
| 1704 |
"learning_rate": 7.949979792216437e-06,
|
| 1705 |
-
"loss":
|
| 1706 |
"step": 141500
|
| 1707 |
},
|
| 1708 |
{
|
| 1709 |
"epoch": 2.53,
|
| 1710 |
"learning_rate": 7.80139314837268e-06,
|
| 1711 |
-
"loss":
|
| 1712 |
"step": 142000
|
| 1713 |
},
|
| 1714 |
{
|
| 1715 |
"epoch": 2.54,
|
| 1716 |
"learning_rate": 7.652806504528921e-06,
|
| 1717 |
-
"loss":
|
| 1718 |
"step": 142500
|
| 1719 |
},
|
| 1720 |
{
|
| 1721 |
"epoch": 2.55,
|
| 1722 |
"learning_rate": 7.504219860685163e-06,
|
| 1723 |
-
"loss":
|
| 1724 |
"step": 143000
|
| 1725 |
},
|
| 1726 |
{
|
| 1727 |
"epoch": 2.56,
|
| 1728 |
"learning_rate": 7.355633216841405e-06,
|
| 1729 |
-
"loss":
|
| 1730 |
"step": 143500
|
| 1731 |
},
|
| 1732 |
{
|
| 1733 |
"epoch": 2.57,
|
| 1734 |
"learning_rate": 7.207046572997647e-06,
|
| 1735 |
-
"loss":
|
| 1736 |
"step": 144000
|
| 1737 |
},
|
| 1738 |
{
|
| 1739 |
"epoch": 2.58,
|
| 1740 |
"learning_rate": 7.058459929153889e-06,
|
| 1741 |
-
"loss":
|
| 1742 |
"step": 144500
|
| 1743 |
},
|
| 1744 |
{
|
| 1745 |
"epoch": 2.59,
|
| 1746 |
"learning_rate": 6.909873285310129e-06,
|
| 1747 |
-
"loss":
|
| 1748 |
"step": 145000
|
| 1749 |
},
|
| 1750 |
{
|
| 1751 |
"epoch": 2.59,
|
| 1752 |
"learning_rate": 6.761286641466371e-06,
|
| 1753 |
-
"loss":
|
| 1754 |
"step": 145500
|
| 1755 |
},
|
| 1756 |
{
|
| 1757 |
"epoch": 2.6,
|
| 1758 |
"learning_rate": 6.612699997622613e-06,
|
| 1759 |
-
"loss":
|
| 1760 |
"step": 146000
|
| 1761 |
},
|
| 1762 |
{
|
| 1763 |
"epoch": 2.61,
|
| 1764 |
"learning_rate": 6.464113353778855e-06,
|
| 1765 |
-
"loss":
|
| 1766 |
"step": 146500
|
| 1767 |
},
|
| 1768 |
{
|
| 1769 |
"epoch": 2.62,
|
| 1770 |
"learning_rate": 6.315526709935097e-06,
|
| 1771 |
-
"loss":
|
| 1772 |
"step": 147000
|
| 1773 |
},
|
| 1774 |
{
|
| 1775 |
"epoch": 2.63,
|
| 1776 |
"learning_rate": 6.1669400660913394e-06,
|
| 1777 |
-
"loss":
|
| 1778 |
"step": 147500
|
| 1779 |
},
|
| 1780 |
{
|
| 1781 |
"epoch": 2.64,
|
| 1782 |
"learning_rate": 6.0183534222475815e-06,
|
| 1783 |
-
"loss":
|
| 1784 |
"step": 148000
|
| 1785 |
},
|
| 1786 |
{
|
| 1787 |
"epoch": 2.65,
|
| 1788 |
"learning_rate": 5.869766778403823e-06,
|
| 1789 |
-
"loss":
|
| 1790 |
"step": 148500
|
| 1791 |
},
|
| 1792 |
{
|
| 1793 |
"epoch": 2.66,
|
| 1794 |
"learning_rate": 5.721180134560065e-06,
|
| 1795 |
-
"loss":
|
| 1796 |
"step": 149000
|
| 1797 |
},
|
| 1798 |
{
|
| 1799 |
"epoch": 2.67,
|
| 1800 |
"learning_rate": 5.572593490716307e-06,
|
| 1801 |
-
"loss":
|
| 1802 |
"step": 149500
|
| 1803 |
},
|
| 1804 |
{
|
| 1805 |
"epoch": 2.67,
|
| 1806 |
"learning_rate": 5.424006846872549e-06,
|
| 1807 |
-
"loss":
|
| 1808 |
"step": 150000
|
| 1809 |
},
|
| 1810 |
{
|
| 1811 |
"epoch": 2.68,
|
| 1812 |
"learning_rate": 5.275420203028791e-06,
|
| 1813 |
-
"loss":
|
| 1814 |
"step": 150500
|
| 1815 |
},
|
| 1816 |
{
|
| 1817 |
"epoch": 2.69,
|
| 1818 |
"learning_rate": 5.126833559185032e-06,
|
| 1819 |
-
"loss":
|
| 1820 |
"step": 151000
|
| 1821 |
},
|
| 1822 |
{
|
| 1823 |
"epoch": 2.7,
|
| 1824 |
"learning_rate": 4.978246915341274e-06,
|
| 1825 |
-
"loss":
|
| 1826 |
"step": 151500
|
| 1827 |
},
|
| 1828 |
{
|
| 1829 |
"epoch": 2.71,
|
| 1830 |
"learning_rate": 4.829660271497516e-06,
|
| 1831 |
-
"loss":
|
| 1832 |
"step": 152000
|
| 1833 |
},
|
| 1834 |
{
|
| 1835 |
"epoch": 2.72,
|
| 1836 |
"learning_rate": 4.681073627653758e-06,
|
| 1837 |
-
"loss":
|
| 1838 |
"step": 152500
|
| 1839 |
},
|
| 1840 |
{
|
| 1841 |
"epoch": 2.73,
|
| 1842 |
"learning_rate": 4.53248698381e-06,
|
| 1843 |
-
"loss":
|
| 1844 |
"step": 153000
|
| 1845 |
},
|
| 1846 |
{
|
| 1847 |
"epoch": 2.74,
|
| 1848 |
"learning_rate": 4.383900339966241e-06,
|
| 1849 |
-
"loss":
|
| 1850 |
"step": 153500
|
| 1851 |
},
|
| 1852 |
{
|
| 1853 |
"epoch": 2.75,
|
| 1854 |
"learning_rate": 4.235313696122483e-06,
|
| 1855 |
-
"loss":
|
| 1856 |
"step": 154000
|
| 1857 |
},
|
| 1858 |
{
|
| 1859 |
"epoch": 2.75,
|
| 1860 |
"learning_rate": 4.086727052278725e-06,
|
| 1861 |
-
"loss":
|
| 1862 |
"step": 154500
|
| 1863 |
},
|
| 1864 |
{
|
| 1865 |
"epoch": 2.76,
|
| 1866 |
"learning_rate": 3.938140408434967e-06,
|
| 1867 |
-
"loss":
|
| 1868 |
"step": 155000
|
| 1869 |
},
|
| 1870 |
{
|
| 1871 |
"epoch": 2.77,
|
| 1872 |
"learning_rate": 3.7895537645912088e-06,
|
| 1873 |
-
"loss":
|
| 1874 |
"step": 155500
|
| 1875 |
},
|
| 1876 |
{
|
| 1877 |
"epoch": 2.78,
|
| 1878 |
"learning_rate": 3.6409671207474504e-06,
|
| 1879 |
-
"loss":
|
| 1880 |
"step": 156000
|
| 1881 |
},
|
| 1882 |
{
|
| 1883 |
"epoch": 2.79,
|
| 1884 |
"learning_rate": 3.4923804769036924e-06,
|
| 1885 |
-
"loss":
|
| 1886 |
"step": 156500
|
| 1887 |
},
|
| 1888 |
{
|
| 1889 |
"epoch": 2.8,
|
| 1890 |
"learning_rate": 3.343793833059934e-06,
|
| 1891 |
-
"loss":
|
| 1892 |
"step": 157000
|
| 1893 |
},
|
| 1894 |
{
|
| 1895 |
"epoch": 2.81,
|
| 1896 |
"learning_rate": 3.195207189216176e-06,
|
| 1897 |
-
"loss":
|
| 1898 |
"step": 157500
|
| 1899 |
},
|
| 1900 |
{
|
| 1901 |
"epoch": 2.82,
|
| 1902 |
"learning_rate": 3.0466205453724176e-06,
|
| 1903 |
-
"loss":
|
| 1904 |
"step": 158000
|
| 1905 |
},
|
| 1906 |
{
|
| 1907 |
"epoch": 2.83,
|
| 1908 |
"learning_rate": 2.898033901528659e-06,
|
| 1909 |
-
"loss":
|
| 1910 |
"step": 158500
|
| 1911 |
},
|
| 1912 |
{
|
| 1913 |
"epoch": 2.84,
|
| 1914 |
"learning_rate": 2.749447257684901e-06,
|
| 1915 |
-
"loss":
|
| 1916 |
"step": 159000
|
| 1917 |
},
|
| 1918 |
{
|
| 1919 |
"epoch": 2.84,
|
| 1920 |
"learning_rate": 2.6008606138411432e-06,
|
| 1921 |
-
"loss":
|
| 1922 |
"step": 159500
|
| 1923 |
},
|
| 1924 |
{
|
| 1925 |
"epoch": 2.85,
|
| 1926 |
"learning_rate": 2.452273969997385e-06,
|
| 1927 |
-
"loss":
|
| 1928 |
"step": 160000
|
| 1929 |
},
|
| 1930 |
{
|
| 1931 |
"epoch": 2.86,
|
| 1932 |
"learning_rate": 2.303687326153627e-06,
|
| 1933 |
-
"loss":
|
| 1934 |
"step": 160500
|
| 1935 |
},
|
| 1936 |
{
|
| 1937 |
"epoch": 2.87,
|
| 1938 |
"learning_rate": 2.1551006823098684e-06,
|
| 1939 |
-
"loss":
|
| 1940 |
"step": 161000
|
| 1941 |
},
|
| 1942 |
{
|
| 1943 |
"epoch": 2.88,
|
| 1944 |
"learning_rate": 2.0065140384661104e-06,
|
| 1945 |
-
"loss":
|
| 1946 |
"step": 161500
|
| 1947 |
},
|
| 1948 |
{
|
| 1949 |
"epoch": 2.89,
|
| 1950 |
"learning_rate": 1.8579273946223525e-06,
|
| 1951 |
-
"loss":
|
| 1952 |
"step": 162000
|
| 1953 |
},
|
| 1954 |
{
|
| 1955 |
"epoch": 2.9,
|
| 1956 |
"learning_rate": 1.709340750778594e-06,
|
| 1957 |
-
"loss":
|
| 1958 |
"step": 162500
|
| 1959 |
},
|
| 1960 |
{
|
| 1961 |
"epoch": 2.91,
|
| 1962 |
"learning_rate": 1.5607541069348359e-06,
|
| 1963 |
-
"loss":
|
| 1964 |
"step": 163000
|
| 1965 |
},
|
| 1966 |
{
|
| 1967 |
"epoch": 2.92,
|
| 1968 |
"learning_rate": 1.4121674630910779e-06,
|
| 1969 |
-
"loss":
|
| 1970 |
"step": 163500
|
| 1971 |
},
|
| 1972 |
{
|
| 1973 |
"epoch": 2.92,
|
| 1974 |
"learning_rate": 1.2635808192473197e-06,
|
| 1975 |
-
"loss":
|
| 1976 |
"step": 164000
|
| 1977 |
},
|
| 1978 |
{
|
| 1979 |
"epoch": 2.93,
|
| 1980 |
"learning_rate": 1.1149941754035615e-06,
|
| 1981 |
-
"loss":
|
| 1982 |
"step": 164500
|
| 1983 |
},
|
| 1984 |
{
|
| 1985 |
"epoch": 2.94,
|
| 1986 |
"learning_rate": 9.66407531559803e-07,
|
| 1987 |
-
"loss":
|
| 1988 |
"step": 165000
|
| 1989 |
},
|
| 1990 |
{
|
| 1991 |
"epoch": 2.95,
|
| 1992 |
"learning_rate": 8.17820887716045e-07,
|
| 1993 |
-
"loss":
|
| 1994 |
"step": 165500
|
| 1995 |
},
|
| 1996 |
{
|
| 1997 |
"epoch": 2.96,
|
| 1998 |
"learning_rate": 6.692342438722869e-07,
|
| 1999 |
-
"loss":
|
| 2000 |
"step": 166000
|
| 2001 |
},
|
| 2002 |
{
|
| 2003 |
"epoch": 2.97,
|
| 2004 |
"learning_rate": 5.206476000285286e-07,
|
| 2005 |
-
"loss":
|
| 2006 |
"step": 166500
|
| 2007 |
},
|
| 2008 |
{
|
| 2009 |
"epoch": 2.98,
|
| 2010 |
"learning_rate": 3.720609561847705e-07,
|
| 2011 |
-
"loss":
|
| 2012 |
"step": 167000
|
| 2013 |
},
|
| 2014 |
{
|
| 2015 |
"epoch": 2.99,
|
| 2016 |
"learning_rate": 2.234743123410123e-07,
|
| 2017 |
-
"loss":
|
| 2018 |
"step": 167500
|
| 2019 |
},
|
| 2020 |
{
|
| 2021 |
"epoch": 3.0,
|
| 2022 |
"learning_rate": 7.488766849725412e-08,
|
| 2023 |
-
"loss":
|
| 2024 |
"step": 168000
|
| 2025 |
},
|
| 2026 |
{
|
| 2027 |
"epoch": 3.0,
|
| 2028 |
"step": 168252,
|
| 2029 |
-
"total_flos":
|
| 2030 |
-
"train_loss":
|
| 2031 |
-
"train_runtime":
|
| 2032 |
-
"train_samples_per_second":
|
| 2033 |
-
"train_steps_per_second":
|
| 2034 |
}
|
| 2035 |
],
|
| 2036 |
"max_steps": 168252,
|
| 2037 |
"num_train_epochs": 3,
|
| 2038 |
-
"total_flos":
|
| 2039 |
"trial_name": null,
|
| 2040 |
"trial_params": null
|
| 2041 |
}
|
|
|
|
| 10 |
{
|
| 11 |
"epoch": 0.01,
|
| 12 |
"learning_rate": 4.9851413356156244e-05,
|
| 13 |
+
"loss": 4.6864,
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.02,
|
| 18 |
"learning_rate": 4.9702826712312485e-05,
|
| 19 |
+
"loss": 4.4314,
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.03,
|
| 24 |
"learning_rate": 4.9554240068468726e-05,
|
| 25 |
+
"loss": 4.1833,
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.04,
|
| 30 |
"learning_rate": 4.940565342462497e-05,
|
| 31 |
+
"loss": 4.1583,
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.04,
|
| 36 |
"learning_rate": 4.9257066780781215e-05,
|
| 37 |
+
"loss": 4.2004,
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.05,
|
| 42 |
"learning_rate": 4.910848013693745e-05,
|
| 43 |
+
"loss": 4.2117,
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.06,
|
| 48 |
"learning_rate": 4.89598934930937e-05,
|
| 49 |
+
"loss": 4.2172,
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.07,
|
| 54 |
"learning_rate": 4.881130684924994e-05,
|
| 55 |
+
"loss": 4.2154,
|
| 56 |
"step": 4000
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.08,
|
| 60 |
"learning_rate": 4.866272020540618e-05,
|
| 61 |
+
"loss": 4.1957,
|
| 62 |
"step": 4500
|
| 63 |
},
|
| 64 |
{
|
| 65 |
"epoch": 0.09,
|
| 66 |
"learning_rate": 4.851413356156242e-05,
|
| 67 |
+
"loss": 4.199,
|
| 68 |
"step": 5000
|
| 69 |
},
|
| 70 |
{
|
| 71 |
"epoch": 0.1,
|
| 72 |
"learning_rate": 4.836554691771866e-05,
|
| 73 |
+
"loss": 4.2366,
|
| 74 |
"step": 5500
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.11,
|
| 78 |
"learning_rate": 4.82169602738749e-05,
|
| 79 |
+
"loss": 4.1593,
|
| 80 |
"step": 6000
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.12,
|
| 84 |
"learning_rate": 4.8068373630031144e-05,
|
| 85 |
+
"loss": 4.1763,
|
| 86 |
"step": 6500
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.12,
|
| 90 |
"learning_rate": 4.7919786986187386e-05,
|
| 91 |
+
"loss": 4.1563,
|
| 92 |
"step": 7000
|
| 93 |
},
|
| 94 |
{
|
| 95 |
"epoch": 0.13,
|
| 96 |
"learning_rate": 4.7771200342343634e-05,
|
| 97 |
+
"loss": 4.18,
|
| 98 |
"step": 7500
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.14,
|
| 102 |
"learning_rate": 4.7622613698499875e-05,
|
| 103 |
+
"loss": 4.1764,
|
| 104 |
"step": 8000
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.15,
|
| 108 |
"learning_rate": 4.747402705465611e-05,
|
| 109 |
+
"loss": 4.1724,
|
| 110 |
"step": 8500
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.16,
|
| 114 |
"learning_rate": 4.732544041081236e-05,
|
| 115 |
+
"loss": 4.1814,
|
| 116 |
"step": 9000
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.17,
|
| 120 |
"learning_rate": 4.71768537669686e-05,
|
| 121 |
+
"loss": 4.1784,
|
| 122 |
"step": 9500
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.18,
|
| 126 |
"learning_rate": 4.702826712312484e-05,
|
| 127 |
+
"loss": 4.19,
|
| 128 |
"step": 10000
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.19,
|
| 132 |
"learning_rate": 4.687968047928108e-05,
|
| 133 |
+
"loss": 4.165,
|
| 134 |
"step": 10500
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.2,
|
| 138 |
"learning_rate": 4.673109383543732e-05,
|
| 139 |
+
"loss": 4.1738,
|
| 140 |
"step": 11000
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.21,
|
| 144 |
"learning_rate": 4.658250719159356e-05,
|
| 145 |
+
"loss": 4.1795,
|
| 146 |
"step": 11500
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.21,
|
| 150 |
"learning_rate": 4.6433920547749804e-05,
|
| 151 |
+
"loss": 4.1975,
|
| 152 |
"step": 12000
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.22,
|
| 156 |
"learning_rate": 4.6285333903906045e-05,
|
| 157 |
+
"loss": 4.1582,
|
| 158 |
"step": 12500
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.23,
|
| 162 |
"learning_rate": 4.613674726006229e-05,
|
| 163 |
+
"loss": 4.1763,
|
| 164 |
"step": 13000
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.24,
|
| 168 |
"learning_rate": 4.5988160616218534e-05,
|
| 169 |
+
"loss": 4.1703,
|
| 170 |
"step": 13500
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.25,
|
| 174 |
"learning_rate": 4.583957397237477e-05,
|
| 175 |
+
"loss": 4.1839,
|
| 176 |
"step": 14000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.26,
|
| 180 |
"learning_rate": 4.569098732853102e-05,
|
| 181 |
+
"loss": 4.2002,
|
| 182 |
"step": 14500
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.27,
|
| 186 |
"learning_rate": 4.554240068468726e-05,
|
| 187 |
+
"loss": 4.1428,
|
| 188 |
"step": 15000
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.28,
|
| 192 |
"learning_rate": 4.53938140408435e-05,
|
| 193 |
+
"loss": 4.187,
|
| 194 |
"step": 15500
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.29,
|
| 198 |
"learning_rate": 4.524522739699974e-05,
|
| 199 |
+
"loss": 4.1869,
|
| 200 |
"step": 16000
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.29,
|
| 204 |
"learning_rate": 4.509664075315598e-05,
|
| 205 |
+
"loss": 4.1829,
|
| 206 |
"step": 16500
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.3,
|
| 210 |
"learning_rate": 4.494805410931223e-05,
|
| 211 |
+
"loss": 4.1572,
|
| 212 |
"step": 17000
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.31,
|
| 216 |
"learning_rate": 4.4799467465468464e-05,
|
| 217 |
+
"loss": 4.1367,
|
| 218 |
"step": 17500
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.32,
|
| 222 |
"learning_rate": 4.4650880821624705e-05,
|
| 223 |
+
"loss": 4.1692,
|
| 224 |
"step": 18000
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.33,
|
| 228 |
"learning_rate": 4.450229417778095e-05,
|
| 229 |
+
"loss": 4.182,
|
| 230 |
"step": 18500
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.34,
|
| 234 |
"learning_rate": 4.435370753393719e-05,
|
| 235 |
+
"loss": 4.2536,
|
| 236 |
"step": 19000
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.35,
|
| 240 |
"learning_rate": 4.4205120890093435e-05,
|
| 241 |
+
"loss": 4.2086,
|
| 242 |
"step": 19500
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 0.36,
|
| 246 |
"learning_rate": 4.4056534246249676e-05,
|
| 247 |
+
"loss": 4.1256,
|
| 248 |
"step": 20000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 0.37,
|
| 252 |
"learning_rate": 4.390794760240592e-05,
|
| 253 |
+
"loss": 4.1798,
|
| 254 |
"step": 20500
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.37,
|
| 258 |
"learning_rate": 4.375936095856216e-05,
|
| 259 |
+
"loss": 4.1822,
|
| 260 |
"step": 21000
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 0.38,
|
| 264 |
"learning_rate": 4.36107743147184e-05,
|
| 265 |
+
"loss": 4.1925,
|
| 266 |
"step": 21500
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 0.39,
|
| 270 |
"learning_rate": 4.346218767087464e-05,
|
| 271 |
+
"loss": 4.1702,
|
| 272 |
"step": 22000
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 0.4,
|
| 276 |
"learning_rate": 4.331360102703089e-05,
|
| 277 |
+
"loss": 4.1983,
|
| 278 |
"step": 22500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 0.41,
|
| 282 |
"learning_rate": 4.3165014383187123e-05,
|
| 283 |
+
"loss": 4.2146,
|
| 284 |
"step": 23000
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 0.42,
|
| 288 |
"learning_rate": 4.3016427739343365e-05,
|
| 289 |
+
"loss": 4.2053,
|
| 290 |
"step": 23500
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 0.43,
|
| 294 |
"learning_rate": 4.286784109549961e-05,
|
| 295 |
+
"loss": 4.2043,
|
| 296 |
"step": 24000
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.44,
|
| 300 |
"learning_rate": 4.271925445165585e-05,
|
| 301 |
+
"loss": 4.1945,
|
| 302 |
"step": 24500
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 0.45,
|
| 306 |
"learning_rate": 4.2570667807812095e-05,
|
| 307 |
+
"loss": 4.1583,
|
| 308 |
"step": 25000
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 0.45,
|
| 312 |
"learning_rate": 4.2422081163968336e-05,
|
| 313 |
+
"loss": 4.2209,
|
| 314 |
"step": 25500
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 0.46,
|
| 318 |
"learning_rate": 4.227349452012458e-05,
|
| 319 |
+
"loss": 4.1892,
|
| 320 |
"step": 26000
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 0.47,
|
| 324 |
"learning_rate": 4.212490787628082e-05,
|
| 325 |
+
"loss": 4.1542,
|
| 326 |
"step": 26500
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 0.48,
|
| 330 |
"learning_rate": 4.197632123243706e-05,
|
| 331 |
+
"loss": 4.1594,
|
| 332 |
"step": 27000
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 0.49,
|
| 336 |
"learning_rate": 4.18277345885933e-05,
|
| 337 |
+
"loss": 4.1926,
|
| 338 |
"step": 27500
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.5,
|
| 342 |
"learning_rate": 4.167914794474955e-05,
|
| 343 |
+
"loss": 4.1596,
|
| 344 |
"step": 28000
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 0.51,
|
| 348 |
"learning_rate": 4.153056130090578e-05,
|
| 349 |
+
"loss": 4.1525,
|
| 350 |
"step": 28500
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 0.52,
|
| 354 |
"learning_rate": 4.138197465706203e-05,
|
| 355 |
+
"loss": 4.1282,
|
| 356 |
"step": 29000
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 0.53,
|
| 360 |
"learning_rate": 4.123338801321827e-05,
|
| 361 |
+
"loss": 4.1569,
|
| 362 |
"step": 29500
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 0.53,
|
| 366 |
"learning_rate": 4.1084801369374507e-05,
|
| 367 |
+
"loss": 4.1544,
|
| 368 |
"step": 30000
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 0.54,
|
| 372 |
"learning_rate": 4.0936214725530755e-05,
|
| 373 |
+
"loss": 4.1984,
|
| 374 |
"step": 30500
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 0.55,
|
| 378 |
"learning_rate": 4.0787628081686996e-05,
|
| 379 |
+
"loss": 4.1907,
|
| 380 |
"step": 31000
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.56,
|
| 384 |
"learning_rate": 4.063904143784324e-05,
|
| 385 |
+
"loss": 4.2402,
|
| 386 |
"step": 31500
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 0.57,
|
| 390 |
"learning_rate": 4.049045479399948e-05,
|
| 391 |
+
"loss": 4.1914,
|
| 392 |
"step": 32000
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 0.58,
|
| 396 |
"learning_rate": 4.034186815015572e-05,
|
| 397 |
+
"loss": 4.1622,
|
| 398 |
"step": 32500
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 0.59,
|
| 402 |
"learning_rate": 4.019328150631197e-05,
|
| 403 |
+
"loss": 4.1444,
|
| 404 |
"step": 33000
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 0.6,
|
| 408 |
"learning_rate": 4.00446948624682e-05,
|
| 409 |
+
"loss": 4.1581,
|
| 410 |
"step": 33500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 0.61,
|
| 414 |
"learning_rate": 3.989610821862444e-05,
|
| 415 |
+
"loss": 4.2014,
|
| 416 |
"step": 34000
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 0.62,
|
| 420 |
"learning_rate": 3.974752157478069e-05,
|
| 421 |
+
"loss": 4.1797,
|
| 422 |
"step": 34500
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.62,
|
| 426 |
"learning_rate": 3.959893493093693e-05,
|
| 427 |
+
"loss": 4.1669,
|
| 428 |
"step": 35000
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 0.63,
|
| 432 |
"learning_rate": 3.9450348287093166e-05,
|
| 433 |
+
"loss": 4.1538,
|
| 434 |
"step": 35500
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 0.64,
|
| 438 |
"learning_rate": 3.9301761643249414e-05,
|
| 439 |
+
"loss": 4.1359,
|
| 440 |
"step": 36000
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 0.65,
|
| 444 |
"learning_rate": 3.9153174999405655e-05,
|
| 445 |
+
"loss": 4.1551,
|
| 446 |
"step": 36500
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 0.66,
|
| 450 |
"learning_rate": 3.9004588355561897e-05,
|
| 451 |
+
"loss": 4.2037,
|
| 452 |
"step": 37000
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 0.67,
|
| 456 |
"learning_rate": 3.885600171171814e-05,
|
| 457 |
+
"loss": 4.1428,
|
| 458 |
"step": 37500
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 0.68,
|
| 462 |
"learning_rate": 3.870741506787438e-05,
|
| 463 |
+
"loss": 4.1649,
|
| 464 |
"step": 38000
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.69,
|
| 468 |
"learning_rate": 3.855882842403063e-05,
|
| 469 |
+
"loss": 4.1858,
|
| 470 |
"step": 38500
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 0.7,
|
| 474 |
"learning_rate": 3.841024178018686e-05,
|
| 475 |
+
"loss": 4.1586,
|
| 476 |
"step": 39000
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 0.7,
|
| 480 |
"learning_rate": 3.82616551363431e-05,
|
| 481 |
+
"loss": 4.1849,
|
| 482 |
"step": 39500
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 0.71,
|
| 486 |
"learning_rate": 3.811306849249935e-05,
|
| 487 |
+
"loss": 4.1875,
|
| 488 |
"step": 40000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
"epoch": 0.72,
|
| 492 |
"learning_rate": 3.796448184865559e-05,
|
| 493 |
+
"loss": 4.1367,
|
| 494 |
"step": 40500
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 0.73,
|
| 498 |
"learning_rate": 3.781589520481183e-05,
|
| 499 |
+
"loss": 4.1681,
|
| 500 |
"step": 41000
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 0.74,
|
| 504 |
"learning_rate": 3.7667308560968074e-05,
|
| 505 |
+
"loss": 4.1701,
|
| 506 |
"step": 41500
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.75,
|
| 510 |
"learning_rate": 3.7518721917124315e-05,
|
| 511 |
+
"loss": 4.1509,
|
| 512 |
"step": 42000
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 0.76,
|
| 516 |
"learning_rate": 3.737013527328056e-05,
|
| 517 |
+
"loss": 4.2155,
|
| 518 |
"step": 42500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 0.77,
|
| 522 |
"learning_rate": 3.72215486294368e-05,
|
| 523 |
+
"loss": 4.2056,
|
| 524 |
"step": 43000
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 0.78,
|
| 528 |
"learning_rate": 3.707296198559304e-05,
|
| 529 |
+
"loss": 4.1598,
|
| 530 |
"step": 43500
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 0.78,
|
| 534 |
"learning_rate": 3.6924375341749286e-05,
|
| 535 |
+
"loss": 4.1701,
|
| 536 |
"step": 44000
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 0.79,
|
| 540 |
"learning_rate": 3.677578869790552e-05,
|
| 541 |
+
"loss": 4.1735,
|
| 542 |
"step": 44500
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 0.8,
|
| 546 |
"learning_rate": 3.662720205406177e-05,
|
| 547 |
+
"loss": 4.1457,
|
| 548 |
"step": 45000
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 0.81,
|
| 552 |
"learning_rate": 3.647861541021801e-05,
|
| 553 |
+
"loss": 4.1469,
|
| 554 |
"step": 45500
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 0.82,
|
| 558 |
"learning_rate": 3.633002876637425e-05,
|
| 559 |
+
"loss": 4.1014,
|
| 560 |
"step": 46000
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 0.83,
|
| 564 |
"learning_rate": 3.618144212253049e-05,
|
| 565 |
+
"loss": 4.1308,
|
| 566 |
"step": 46500
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 0.84,
|
| 570 |
"learning_rate": 3.6032855478686734e-05,
|
| 571 |
+
"loss": 4.1282,
|
| 572 |
"step": 47000
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 0.85,
|
| 576 |
"learning_rate": 3.5884268834842975e-05,
|
| 577 |
+
"loss": 4.1384,
|
| 578 |
"step": 47500
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 0.86,
|
| 582 |
"learning_rate": 3.5735682190999216e-05,
|
| 583 |
+
"loss": 4.1709,
|
| 584 |
"step": 48000
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 0.86,
|
| 588 |
"learning_rate": 3.558709554715546e-05,
|
| 589 |
+
"loss": 4.1861,
|
| 590 |
"step": 48500
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 0.87,
|
| 594 |
"learning_rate": 3.54385089033117e-05,
|
| 595 |
+
"loss": 4.1939,
|
| 596 |
"step": 49000
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 0.88,
|
| 600 |
"learning_rate": 3.5289922259467946e-05,
|
| 601 |
+
"loss": 4.186,
|
| 602 |
"step": 49500
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 0.89,
|
| 606 |
"learning_rate": 3.514133561562418e-05,
|
| 607 |
+
"loss": 4.1661,
|
| 608 |
"step": 50000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 0.9,
|
| 612 |
"learning_rate": 3.499274897178043e-05,
|
| 613 |
+
"loss": 4.1291,
|
| 614 |
"step": 50500
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 0.91,
|
| 618 |
"learning_rate": 3.484416232793667e-05,
|
| 619 |
+
"loss": 4.1377,
|
| 620 |
"step": 51000
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 0.92,
|
| 624 |
"learning_rate": 3.469557568409291e-05,
|
| 625 |
+
"loss": 4.2083,
|
| 626 |
"step": 51500
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 0.93,
|
| 630 |
"learning_rate": 3.454698904024915e-05,
|
| 631 |
+
"loss": 4.1944,
|
| 632 |
"step": 52000
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 0.94,
|
| 636 |
"learning_rate": 3.439840239640539e-05,
|
| 637 |
+
"loss": 4.1597,
|
| 638 |
"step": 52500
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 0.95,
|
| 642 |
"learning_rate": 3.4249815752561634e-05,
|
| 643 |
+
"loss": 4.1432,
|
| 644 |
"step": 53000
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 0.95,
|
| 648 |
"learning_rate": 3.4101229108717876e-05,
|
| 649 |
+
"loss": 4.1659,
|
| 650 |
"step": 53500
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 0.96,
|
| 654 |
"learning_rate": 3.395264246487412e-05,
|
| 655 |
+
"loss": 4.1568,
|
| 656 |
"step": 54000
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 0.97,
|
| 660 |
"learning_rate": 3.3804055821030365e-05,
|
| 661 |
+
"loss": 4.1673,
|
| 662 |
"step": 54500
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 0.98,
|
| 666 |
"learning_rate": 3.3655469177186606e-05,
|
| 667 |
+
"loss": 4.1522,
|
| 668 |
"step": 55000
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 0.99,
|
| 672 |
"learning_rate": 3.350688253334284e-05,
|
| 673 |
+
"loss": 4.1354,
|
| 674 |
"step": 55500
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 1.0,
|
| 678 |
"learning_rate": 3.335829588949909e-05,
|
| 679 |
+
"loss": 4.172,
|
| 680 |
"step": 56000
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 1.01,
|
| 684 |
"learning_rate": 3.320970924565533e-05,
|
| 685 |
+
"loss": 4.1534,
|
| 686 |
"step": 56500
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 1.02,
|
| 690 |
"learning_rate": 3.306112260181157e-05,
|
| 691 |
+
"loss": 4.1092,
|
| 692 |
"step": 57000
|
| 693 |
},
|
| 694 |
{
|
| 695 |
"epoch": 1.03,
|
| 696 |
"learning_rate": 3.291253595796781e-05,
|
| 697 |
+
"loss": 4.1695,
|
| 698 |
"step": 57500
|
| 699 |
},
|
| 700 |
{
|
| 701 |
"epoch": 1.03,
|
| 702 |
"learning_rate": 3.276394931412405e-05,
|
| 703 |
+
"loss": 4.1512,
|
| 704 |
"step": 58000
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 1.04,
|
| 708 |
"learning_rate": 3.2615362670280294e-05,
|
| 709 |
+
"loss": 4.1747,
|
| 710 |
"step": 58500
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 1.05,
|
| 714 |
"learning_rate": 3.2466776026436535e-05,
|
| 715 |
+
"loss": 4.1872,
|
| 716 |
"step": 59000
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 1.06,
|
| 720 |
"learning_rate": 3.2318189382592776e-05,
|
| 721 |
+
"loss": 4.1634,
|
| 722 |
"step": 59500
|
| 723 |
},
|
| 724 |
{
|
| 725 |
"epoch": 1.07,
|
| 726 |
"learning_rate": 3.2169602738749024e-05,
|
| 727 |
+
"loss": 4.1033,
|
| 728 |
"step": 60000
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 1.08,
|
| 732 |
"learning_rate": 3.2021016094905265e-05,
|
| 733 |
+
"loss": 4.1247,
|
| 734 |
"step": 60500
|
| 735 |
},
|
| 736 |
{
|
| 737 |
"epoch": 1.09,
|
| 738 |
"learning_rate": 3.18724294510615e-05,
|
| 739 |
+
"loss": 4.1823,
|
| 740 |
"step": 61000
|
| 741 |
},
|
| 742 |
{
|
| 743 |
"epoch": 1.1,
|
| 744 |
"learning_rate": 3.172384280721775e-05,
|
| 745 |
+
"loss": 4.1718,
|
| 746 |
"step": 61500
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 1.11,
|
| 750 |
"learning_rate": 3.157525616337399e-05,
|
| 751 |
+
"loss": 4.1982,
|
| 752 |
"step": 62000
|
| 753 |
},
|
| 754 |
{
|
| 755 |
"epoch": 1.11,
|
| 756 |
"learning_rate": 3.142666951953023e-05,
|
| 757 |
+
"loss": 4.1956,
|
| 758 |
"step": 62500
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 1.12,
|
| 762 |
"learning_rate": 3.127808287568647e-05,
|
| 763 |
+
"loss": 4.162,
|
| 764 |
"step": 63000
|
| 765 |
},
|
| 766 |
{
|
| 767 |
"epoch": 1.13,
|
| 768 |
"learning_rate": 3.112949623184271e-05,
|
| 769 |
+
"loss": 4.1438,
|
| 770 |
"step": 63500
|
| 771 |
},
|
| 772 |
{
|
| 773 |
"epoch": 1.14,
|
| 774 |
"learning_rate": 3.098090958799896e-05,
|
| 775 |
+
"loss": 4.2248,
|
| 776 |
"step": 64000
|
| 777 |
},
|
| 778 |
{
|
| 779 |
"epoch": 1.15,
|
| 780 |
"learning_rate": 3.0832322944155195e-05,
|
| 781 |
+
"loss": 4.1655,
|
| 782 |
"step": 64500
|
| 783 |
},
|
| 784 |
{
|
| 785 |
"epoch": 1.16,
|
| 786 |
"learning_rate": 3.0683736300311436e-05,
|
| 787 |
+
"loss": 4.152,
|
| 788 |
"step": 65000
|
| 789 |
},
|
| 790 |
{
|
| 791 |
"epoch": 1.17,
|
| 792 |
"learning_rate": 3.0535149656467684e-05,
|
| 793 |
+
"loss": 4.1784,
|
| 794 |
"step": 65500
|
| 795 |
},
|
| 796 |
{
|
| 797 |
"epoch": 1.18,
|
| 798 |
"learning_rate": 3.0386563012623925e-05,
|
| 799 |
+
"loss": 4.1958,
|
| 800 |
"step": 66000
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 1.19,
|
| 804 |
"learning_rate": 3.0237976368780163e-05,
|
| 805 |
+
"loss": 4.156,
|
| 806 |
"step": 66500
|
| 807 |
},
|
| 808 |
{
|
| 809 |
"epoch": 1.19,
|
| 810 |
"learning_rate": 3.0089389724936407e-05,
|
| 811 |
+
"loss": 4.178,
|
| 812 |
"step": 67000
|
| 813 |
},
|
| 814 |
{
|
| 815 |
"epoch": 1.2,
|
| 816 |
"learning_rate": 2.994080308109265e-05,
|
| 817 |
+
"loss": 4.1562,
|
| 818 |
"step": 67500
|
| 819 |
},
|
| 820 |
{
|
| 821 |
"epoch": 1.21,
|
| 822 |
"learning_rate": 2.9792216437248886e-05,
|
| 823 |
+
"loss": 4.1535,
|
| 824 |
"step": 68000
|
| 825 |
},
|
| 826 |
{
|
| 827 |
"epoch": 1.22,
|
| 828 |
"learning_rate": 2.964362979340513e-05,
|
| 829 |
+
"loss": 4.1224,
|
| 830 |
"step": 68500
|
| 831 |
},
|
| 832 |
{
|
| 833 |
"epoch": 1.23,
|
| 834 |
"learning_rate": 2.9495043149561376e-05,
|
| 835 |
+
"loss": 4.1547,
|
| 836 |
"step": 69000
|
| 837 |
},
|
| 838 |
{
|
| 839 |
"epoch": 1.24,
|
| 840 |
"learning_rate": 2.9346456505717617e-05,
|
| 841 |
+
"loss": 4.1378,
|
| 842 |
"step": 69500
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 1.25,
|
| 846 |
"learning_rate": 2.9197869861873855e-05,
|
| 847 |
+
"loss": 4.1573,
|
| 848 |
"step": 70000
|
| 849 |
},
|
| 850 |
{
|
| 851 |
"epoch": 1.26,
|
| 852 |
"learning_rate": 2.90492832180301e-05,
|
| 853 |
+
"loss": 4.1575,
|
| 854 |
"step": 70500
|
| 855 |
},
|
| 856 |
{
|
| 857 |
"epoch": 1.27,
|
| 858 |
"learning_rate": 2.8900696574186344e-05,
|
| 859 |
+
"loss": 4.1726,
|
| 860 |
"step": 71000
|
| 861 |
},
|
| 862 |
{
|
| 863 |
"epoch": 1.27,
|
| 864 |
"learning_rate": 2.875210993034258e-05,
|
| 865 |
+
"loss": 4.1971,
|
| 866 |
"step": 71500
|
| 867 |
},
|
| 868 |
{
|
| 869 |
"epoch": 1.28,
|
| 870 |
"learning_rate": 2.8603523286498823e-05,
|
| 871 |
+
"loss": 4.1289,
|
| 872 |
"step": 72000
|
| 873 |
},
|
| 874 |
{
|
| 875 |
"epoch": 1.29,
|
| 876 |
"learning_rate": 2.8454936642655067e-05,
|
| 877 |
+
"loss": 4.1571,
|
| 878 |
"step": 72500
|
| 879 |
},
|
| 880 |
{
|
| 881 |
"epoch": 1.3,
|
| 882 |
"learning_rate": 2.8306349998811312e-05,
|
| 883 |
+
"loss": 4.124,
|
| 884 |
"step": 73000
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 1.31,
|
| 888 |
"learning_rate": 2.815776335496755e-05,
|
| 889 |
+
"loss": 4.1769,
|
| 890 |
"step": 73500
|
| 891 |
},
|
| 892 |
{
|
| 893 |
"epoch": 1.32,
|
| 894 |
"learning_rate": 2.800917671112379e-05,
|
| 895 |
+
"loss": 4.1581,
|
| 896 |
"step": 74000
|
| 897 |
},
|
| 898 |
{
|
| 899 |
"epoch": 1.33,
|
| 900 |
"learning_rate": 2.7860590067280035e-05,
|
| 901 |
+
"loss": 4.1828,
|
| 902 |
"step": 74500
|
| 903 |
},
|
| 904 |
{
|
| 905 |
"epoch": 1.34,
|
| 906 |
"learning_rate": 2.7712003423436276e-05,
|
| 907 |
+
"loss": 4.1448,
|
| 908 |
"step": 75000
|
| 909 |
},
|
| 910 |
{
|
| 911 |
"epoch": 1.35,
|
| 912 |
"learning_rate": 2.7563416779592514e-05,
|
| 913 |
+
"loss": 4.2025,
|
| 914 |
"step": 75500
|
| 915 |
},
|
| 916 |
{
|
| 917 |
"epoch": 1.36,
|
| 918 |
"learning_rate": 2.741483013574876e-05,
|
| 919 |
+
"loss": 4.1474,
|
| 920 |
"step": 76000
|
| 921 |
},
|
| 922 |
{
|
| 923 |
"epoch": 1.36,
|
| 924 |
"learning_rate": 2.7266243491905003e-05,
|
| 925 |
+
"loss": 4.2199,
|
| 926 |
"step": 76500
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 1.37,
|
| 930 |
"learning_rate": 2.711765684806124e-05,
|
| 931 |
+
"loss": 4.1517,
|
| 932 |
"step": 77000
|
| 933 |
},
|
| 934 |
{
|
| 935 |
"epoch": 1.38,
|
| 936 |
"learning_rate": 2.6969070204217482e-05,
|
| 937 |
+
"loss": 4.1469,
|
| 938 |
"step": 77500
|
| 939 |
},
|
| 940 |
{
|
| 941 |
"epoch": 1.39,
|
| 942 |
"learning_rate": 2.6820483560373727e-05,
|
| 943 |
+
"loss": 4.1104,
|
| 944 |
"step": 78000
|
| 945 |
},
|
| 946 |
{
|
| 947 |
"epoch": 1.4,
|
| 948 |
"learning_rate": 2.667189691652997e-05,
|
| 949 |
+
"loss": 4.1482,
|
| 950 |
"step": 78500
|
| 951 |
},
|
| 952 |
{
|
| 953 |
"epoch": 1.41,
|
| 954 |
"learning_rate": 2.652331027268621e-05,
|
| 955 |
+
"loss": 4.1674,
|
| 956 |
"step": 79000
|
| 957 |
},
|
| 958 |
{
|
| 959 |
"epoch": 1.42,
|
| 960 |
"learning_rate": 2.637472362884245e-05,
|
| 961 |
+
"loss": 4.1601,
|
| 962 |
"step": 79500
|
| 963 |
},
|
| 964 |
{
|
| 965 |
"epoch": 1.43,
|
| 966 |
"learning_rate": 2.6226136984998695e-05,
|
| 967 |
+
"loss": 4.2056,
|
| 968 |
"step": 80000
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 1.44,
|
| 972 |
"learning_rate": 2.607755034115494e-05,
|
| 973 |
+
"loss": 4.1223,
|
| 974 |
"step": 80500
|
| 975 |
},
|
| 976 |
{
|
| 977 |
"epoch": 1.44,
|
| 978 |
"learning_rate": 2.5928963697311177e-05,
|
| 979 |
+
"loss": 4.1843,
|
| 980 |
"step": 81000
|
| 981 |
},
|
| 982 |
{
|
| 983 |
"epoch": 1.45,
|
| 984 |
"learning_rate": 2.578037705346742e-05,
|
| 985 |
+
"loss": 4.1564,
|
| 986 |
"step": 81500
|
| 987 |
},
|
| 988 |
{
|
| 989 |
"epoch": 1.46,
|
| 990 |
"learning_rate": 2.5631790409623663e-05,
|
| 991 |
+
"loss": 4.1476,
|
| 992 |
"step": 82000
|
| 993 |
},
|
| 994 |
{
|
| 995 |
"epoch": 1.47,
|
| 996 |
"learning_rate": 2.54832037657799e-05,
|
| 997 |
+
"loss": 4.139,
|
| 998 |
"step": 82500
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
"epoch": 1.48,
|
| 1002 |
"learning_rate": 2.5334617121936145e-05,
|
| 1003 |
+
"loss": 4.1541,
|
| 1004 |
"step": 83000
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
"epoch": 1.49,
|
| 1008 |
"learning_rate": 2.5186030478092386e-05,
|
| 1009 |
+
"loss": 4.1593,
|
| 1010 |
"step": 83500
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 1.5,
|
| 1014 |
"learning_rate": 2.503744383424863e-05,
|
| 1015 |
+
"loss": 4.1922,
|
| 1016 |
"step": 84000
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
"epoch": 1.51,
|
| 1020 |
"learning_rate": 2.4888857190404872e-05,
|
| 1021 |
+
"loss": 4.1539,
|
| 1022 |
"step": 84500
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
"epoch": 1.52,
|
| 1026 |
"learning_rate": 2.4740270546561113e-05,
|
| 1027 |
+
"loss": 4.1665,
|
| 1028 |
"step": 85000
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
"epoch": 1.52,
|
| 1032 |
"learning_rate": 2.459168390271735e-05,
|
| 1033 |
+
"loss": 4.1476,
|
| 1034 |
"step": 85500
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
"epoch": 1.53,
|
| 1038 |
"learning_rate": 2.4443097258873596e-05,
|
| 1039 |
+
"loss": 4.1933,
|
| 1040 |
"step": 86000
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
"epoch": 1.54,
|
| 1044 |
"learning_rate": 2.4294510615029837e-05,
|
| 1045 |
+
"loss": 4.2239,
|
| 1046 |
"step": 86500
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
"epoch": 1.55,
|
| 1050 |
"learning_rate": 2.4145923971186078e-05,
|
| 1051 |
+
"loss": 4.1581,
|
| 1052 |
"step": 87000
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 1.56,
|
| 1056 |
"learning_rate": 2.399733732734232e-05,
|
| 1057 |
+
"loss": 4.217,
|
| 1058 |
"step": 87500
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
"epoch": 1.57,
|
| 1062 |
"learning_rate": 2.3848750683498564e-05,
|
| 1063 |
+
"loss": 4.1504,
|
| 1064 |
"step": 88000
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
"epoch": 1.58,
|
| 1068 |
"learning_rate": 2.3700164039654805e-05,
|
| 1069 |
+
"loss": 4.1499,
|
| 1070 |
"step": 88500
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
"epoch": 1.59,
|
| 1074 |
"learning_rate": 2.3551577395811046e-05,
|
| 1075 |
+
"loss": 4.1879,
|
| 1076 |
"step": 89000
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
"epoch": 1.6,
|
| 1080 |
"learning_rate": 2.3402990751967287e-05,
|
| 1081 |
+
"loss": 4.1592,
|
| 1082 |
"step": 89500
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
"epoch": 1.6,
|
| 1086 |
"learning_rate": 2.3254404108123532e-05,
|
| 1087 |
+
"loss": 4.1356,
|
| 1088 |
"step": 90000
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
"epoch": 1.61,
|
| 1092 |
"learning_rate": 2.3105817464279773e-05,
|
| 1093 |
+
"loss": 4.1924,
|
| 1094 |
"step": 90500
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 1.62,
|
| 1098 |
"learning_rate": 2.2957230820436014e-05,
|
| 1099 |
+
"loss": 4.1122,
|
| 1100 |
"step": 91000
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
"epoch": 1.63,
|
| 1104 |
"learning_rate": 2.2808644176592255e-05,
|
| 1105 |
+
"loss": 4.1681,
|
| 1106 |
"step": 91500
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
"epoch": 1.64,
|
| 1110 |
"learning_rate": 2.2660057532748497e-05,
|
| 1111 |
+
"loss": 4.1771,
|
| 1112 |
"step": 92000
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
"epoch": 1.65,
|
| 1116 |
"learning_rate": 2.251147088890474e-05,
|
| 1117 |
+
"loss": 4.1596,
|
| 1118 |
"step": 92500
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
"epoch": 1.66,
|
| 1122 |
"learning_rate": 2.236288424506098e-05,
|
| 1123 |
+
"loss": 4.1591,
|
| 1124 |
"step": 93000
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
"epoch": 1.67,
|
| 1128 |
"learning_rate": 2.2214297601217223e-05,
|
| 1129 |
+
"loss": 4.1259,
|
| 1130 |
"step": 93500
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
"epoch": 1.68,
|
| 1134 |
"learning_rate": 2.2065710957373465e-05,
|
| 1135 |
+
"loss": 4.1673,
|
| 1136 |
"step": 94000
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 1.68,
|
| 1140 |
"learning_rate": 2.191712431352971e-05,
|
| 1141 |
+
"loss": 4.1816,
|
| 1142 |
"step": 94500
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
"epoch": 1.69,
|
| 1146 |
"learning_rate": 2.1768537669685947e-05,
|
| 1147 |
+
"loss": 4.1738,
|
| 1148 |
"step": 95000
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
"epoch": 1.7,
|
| 1152 |
"learning_rate": 2.1619951025842188e-05,
|
| 1153 |
+
"loss": 4.1417,
|
| 1154 |
"step": 95500
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
"epoch": 1.71,
|
| 1158 |
"learning_rate": 2.1471364381998433e-05,
|
| 1159 |
+
"loss": 4.141,
|
| 1160 |
"step": 96000
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
"epoch": 1.72,
|
| 1164 |
"learning_rate": 2.1322777738154674e-05,
|
| 1165 |
+
"loss": 4.1754,
|
| 1166 |
"step": 96500
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
"epoch": 1.73,
|
| 1170 |
"learning_rate": 2.1174191094310915e-05,
|
| 1171 |
+
"loss": 4.1311,
|
| 1172 |
"step": 97000
|
| 1173 |
},
|
| 1174 |
{
|
| 1175 |
"epoch": 1.74,
|
| 1176 |
"learning_rate": 2.1025604450467156e-05,
|
| 1177 |
+
"loss": 4.2245,
|
| 1178 |
"step": 97500
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
"epoch": 1.75,
|
| 1182 |
"learning_rate": 2.08770178066234e-05,
|
| 1183 |
+
"loss": 4.2167,
|
| 1184 |
"step": 98000
|
| 1185 |
},
|
| 1186 |
{
|
| 1187 |
"epoch": 1.76,
|
| 1188 |
"learning_rate": 2.0728431162779642e-05,
|
| 1189 |
+
"loss": 4.1694,
|
| 1190 |
"step": 98500
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
"epoch": 1.77,
|
| 1194 |
"learning_rate": 2.0579844518935883e-05,
|
| 1195 |
+
"loss": 4.1671,
|
| 1196 |
"step": 99000
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
"epoch": 1.77,
|
| 1200 |
"learning_rate": 2.0431257875092124e-05,
|
| 1201 |
+
"loss": 4.1839,
|
| 1202 |
"step": 99500
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
"epoch": 1.78,
|
| 1206 |
"learning_rate": 2.0282671231248365e-05,
|
| 1207 |
+
"loss": 4.1632,
|
| 1208 |
"step": 100000
|
| 1209 |
},
|
| 1210 |
{
|
| 1211 |
"epoch": 1.79,
|
| 1212 |
"learning_rate": 2.013408458740461e-05,
|
| 1213 |
+
"loss": 4.2308,
|
| 1214 |
"step": 100500
|
| 1215 |
},
|
| 1216 |
{
|
| 1217 |
"epoch": 1.8,
|
| 1218 |
"learning_rate": 1.9985497943560848e-05,
|
| 1219 |
+
"loss": 4.1291,
|
| 1220 |
"step": 101000
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
"epoch": 1.81,
|
| 1224 |
"learning_rate": 1.9836911299717092e-05,
|
| 1225 |
+
"loss": 4.1498,
|
| 1226 |
"step": 101500
|
| 1227 |
},
|
| 1228 |
{
|
| 1229 |
"epoch": 1.82,
|
| 1230 |
"learning_rate": 1.9688324655873334e-05,
|
| 1231 |
+
"loss": 4.1277,
|
| 1232 |
"step": 102000
|
| 1233 |
},
|
| 1234 |
{
|
| 1235 |
"epoch": 1.83,
|
| 1236 |
"learning_rate": 1.9539738012029578e-05,
|
| 1237 |
+
"loss": 4.1425,
|
| 1238 |
"step": 102500
|
| 1239 |
},
|
| 1240 |
{
|
| 1241 |
"epoch": 1.84,
|
| 1242 |
"learning_rate": 1.9391151368185816e-05,
|
| 1243 |
+
"loss": 4.1518,
|
| 1244 |
"step": 103000
|
| 1245 |
},
|
| 1246 |
{
|
| 1247 |
"epoch": 1.85,
|
| 1248 |
"learning_rate": 1.924256472434206e-05,
|
| 1249 |
+
"loss": 4.1907,
|
| 1250 |
"step": 103500
|
| 1251 |
},
|
| 1252 |
{
|
| 1253 |
"epoch": 1.85,
|
| 1254 |
"learning_rate": 1.90939780804983e-05,
|
| 1255 |
+
"loss": 4.1234,
|
| 1256 |
"step": 104000
|
| 1257 |
},
|
| 1258 |
{
|
| 1259 |
"epoch": 1.86,
|
| 1260 |
"learning_rate": 1.8945391436654543e-05,
|
| 1261 |
+
"loss": 4.1702,
|
| 1262 |
"step": 104500
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
"epoch": 1.87,
|
| 1266 |
"learning_rate": 1.8796804792810784e-05,
|
| 1267 |
+
"loss": 4.1586,
|
| 1268 |
"step": 105000
|
| 1269 |
},
|
| 1270 |
{
|
| 1271 |
"epoch": 1.88,
|
| 1272 |
"learning_rate": 1.8648218148967025e-05,
|
| 1273 |
+
"loss": 4.1564,
|
| 1274 |
"step": 105500
|
| 1275 |
},
|
| 1276 |
{
|
| 1277 |
"epoch": 1.89,
|
| 1278 |
"learning_rate": 1.849963150512327e-05,
|
| 1279 |
+
"loss": 4.1591,
|
| 1280 |
"step": 106000
|
| 1281 |
},
|
| 1282 |
{
|
| 1283 |
"epoch": 1.9,
|
| 1284 |
"learning_rate": 1.835104486127951e-05,
|
| 1285 |
+
"loss": 4.1811,
|
| 1286 |
"step": 106500
|
| 1287 |
},
|
| 1288 |
{
|
| 1289 |
"epoch": 1.91,
|
| 1290 |
"learning_rate": 1.8202458217435752e-05,
|
| 1291 |
+
"loss": 4.1795,
|
| 1292 |
"step": 107000
|
| 1293 |
},
|
| 1294 |
{
|
| 1295 |
"epoch": 1.92,
|
| 1296 |
"learning_rate": 1.8053871573591993e-05,
|
| 1297 |
+
"loss": 4.1605,
|
| 1298 |
"step": 107500
|
| 1299 |
},
|
| 1300 |
{
|
| 1301 |
"epoch": 1.93,
|
| 1302 |
"learning_rate": 1.7905284929748238e-05,
|
| 1303 |
+
"loss": 4.1761,
|
| 1304 |
"step": 108000
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
"epoch": 1.93,
|
| 1308 |
"learning_rate": 1.775669828590448e-05,
|
| 1309 |
+
"loss": 4.186,
|
| 1310 |
"step": 108500
|
| 1311 |
},
|
| 1312 |
{
|
| 1313 |
"epoch": 1.94,
|
| 1314 |
"learning_rate": 1.760811164206072e-05,
|
| 1315 |
+
"loss": 4.2381,
|
| 1316 |
"step": 109000
|
| 1317 |
},
|
| 1318 |
{
|
| 1319 |
"epoch": 1.95,
|
| 1320 |
"learning_rate": 1.745952499821696e-05,
|
| 1321 |
+
"loss": 4.1839,
|
| 1322 |
"step": 109500
|
| 1323 |
},
|
| 1324 |
{
|
| 1325 |
"epoch": 1.96,
|
| 1326 |
"learning_rate": 1.7310938354373202e-05,
|
| 1327 |
+
"loss": 4.2121,
|
| 1328 |
"step": 110000
|
| 1329 |
},
|
| 1330 |
{
|
| 1331 |
"epoch": 1.97,
|
| 1332 |
"learning_rate": 1.7162351710529444e-05,
|
| 1333 |
+
"loss": 4.1832,
|
| 1334 |
"step": 110500
|
| 1335 |
},
|
| 1336 |
{
|
| 1337 |
"epoch": 1.98,
|
| 1338 |
"learning_rate": 1.7013765066685685e-05,
|
| 1339 |
+
"loss": 4.1467,
|
| 1340 |
"step": 111000
|
| 1341 |
},
|
| 1342 |
{
|
| 1343 |
"epoch": 1.99,
|
| 1344 |
"learning_rate": 1.686517842284193e-05,
|
| 1345 |
+
"loss": 4.1932,
|
| 1346 |
"step": 111500
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
"epoch": 2.0,
|
| 1350 |
"learning_rate": 1.671659177899817e-05,
|
| 1351 |
+
"loss": 4.1277,
|
| 1352 |
"step": 112000
|
| 1353 |
},
|
| 1354 |
{
|
| 1355 |
"epoch": 2.01,
|
| 1356 |
"learning_rate": 1.656800513515441e-05,
|
| 1357 |
+
"loss": 4.0758,
|
| 1358 |
"step": 112500
|
| 1359 |
},
|
| 1360 |
{
|
| 1361 |
"epoch": 2.01,
|
| 1362 |
"learning_rate": 1.6419418491310653e-05,
|
| 1363 |
+
"loss": 4.1676,
|
| 1364 |
"step": 113000
|
| 1365 |
},
|
| 1366 |
{
|
| 1367 |
"epoch": 2.02,
|
| 1368 |
"learning_rate": 1.6270831847466897e-05,
|
| 1369 |
+
"loss": 4.2034,
|
| 1370 |
"step": 113500
|
| 1371 |
},
|
| 1372 |
{
|
| 1373 |
"epoch": 2.03,
|
| 1374 |
"learning_rate": 1.612224520362314e-05,
|
| 1375 |
+
"loss": 4.1162,
|
| 1376 |
"step": 114000
|
| 1377 |
},
|
| 1378 |
{
|
| 1379 |
"epoch": 2.04,
|
| 1380 |
"learning_rate": 1.597365855977938e-05,
|
| 1381 |
+
"loss": 4.1963,
|
| 1382 |
"step": 114500
|
| 1383 |
},
|
| 1384 |
{
|
| 1385 |
"epoch": 2.05,
|
| 1386 |
"learning_rate": 1.582507191593562e-05,
|
| 1387 |
+
"loss": 4.1657,
|
| 1388 |
"step": 115000
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
"epoch": 2.06,
|
| 1392 |
"learning_rate": 1.5676485272091862e-05,
|
| 1393 |
+
"loss": 4.1668,
|
| 1394 |
"step": 115500
|
| 1395 |
},
|
| 1396 |
{
|
| 1397 |
"epoch": 2.07,
|
| 1398 |
"learning_rate": 1.5527898628248107e-05,
|
| 1399 |
+
"loss": 4.153,
|
| 1400 |
"step": 116000
|
| 1401 |
},
|
| 1402 |
{
|
| 1403 |
"epoch": 2.08,
|
| 1404 |
"learning_rate": 1.5379311984404344e-05,
|
| 1405 |
+
"loss": 4.1883,
|
| 1406 |
"step": 116500
|
| 1407 |
},
|
| 1408 |
{
|
| 1409 |
"epoch": 2.09,
|
| 1410 |
"learning_rate": 1.5230725340560589e-05,
|
| 1411 |
+
"loss": 4.1552,
|
| 1412 |
"step": 117000
|
| 1413 |
},
|
| 1414 |
{
|
| 1415 |
"epoch": 2.1,
|
| 1416 |
"learning_rate": 1.508213869671683e-05,
|
| 1417 |
+
"loss": 4.1509,
|
| 1418 |
"step": 117500
|
| 1419 |
},
|
| 1420 |
{
|
| 1421 |
"epoch": 2.1,
|
| 1422 |
"learning_rate": 1.4933552052873073e-05,
|
| 1423 |
+
"loss": 4.1463,
|
| 1424 |
"step": 118000
|
| 1425 |
},
|
| 1426 |
{
|
| 1427 |
"epoch": 2.11,
|
| 1428 |
"learning_rate": 1.4784965409029314e-05,
|
| 1429 |
+
"loss": 4.1101,
|
| 1430 |
"step": 118500
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
"epoch": 2.12,
|
| 1434 |
"learning_rate": 1.4636378765185554e-05,
|
| 1435 |
+
"loss": 4.085,
|
| 1436 |
"step": 119000
|
| 1437 |
},
|
| 1438 |
{
|
| 1439 |
"epoch": 2.13,
|
| 1440 |
"learning_rate": 1.4487792121341798e-05,
|
| 1441 |
+
"loss": 4.1117,
|
| 1442 |
"step": 119500
|
| 1443 |
},
|
| 1444 |
{
|
| 1445 |
"epoch": 2.14,
|
| 1446 |
"learning_rate": 1.4339205477498038e-05,
|
| 1447 |
+
"loss": 4.1806,
|
| 1448 |
"step": 120000
|
| 1449 |
},
|
| 1450 |
{
|
| 1451 |
"epoch": 2.15,
|
| 1452 |
"learning_rate": 1.4190618833654282e-05,
|
| 1453 |
+
"loss": 4.1498,
|
| 1454 |
"step": 120500
|
| 1455 |
},
|
| 1456 |
{
|
| 1457 |
"epoch": 2.16,
|
| 1458 |
"learning_rate": 1.4042032189810522e-05,
|
| 1459 |
+
"loss": 4.1366,
|
| 1460 |
"step": 121000
|
| 1461 |
},
|
| 1462 |
{
|
| 1463 |
"epoch": 2.17,
|
| 1464 |
"learning_rate": 1.3893445545966766e-05,
|
| 1465 |
+
"loss": 4.1916,
|
| 1466 |
"step": 121500
|
| 1467 |
},
|
| 1468 |
{
|
| 1469 |
"epoch": 2.18,
|
| 1470 |
"learning_rate": 1.3744858902123006e-05,
|
| 1471 |
+
"loss": 4.1232,
|
| 1472 |
"step": 122000
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
"epoch": 2.18,
|
| 1476 |
"learning_rate": 1.359627225827925e-05,
|
| 1477 |
+
"loss": 4.1785,
|
| 1478 |
"step": 122500
|
| 1479 |
},
|
| 1480 |
{
|
| 1481 |
"epoch": 2.19,
|
| 1482 |
"learning_rate": 1.344768561443549e-05,
|
| 1483 |
+
"loss": 4.1444,
|
| 1484 |
"step": 123000
|
| 1485 |
},
|
| 1486 |
{
|
| 1487 |
"epoch": 2.2,
|
| 1488 |
"learning_rate": 1.3299098970591731e-05,
|
| 1489 |
+
"loss": 4.1848,
|
| 1490 |
"step": 123500
|
| 1491 |
},
|
| 1492 |
{
|
| 1493 |
"epoch": 2.21,
|
| 1494 |
"learning_rate": 1.3150512326747974e-05,
|
| 1495 |
+
"loss": 4.1975,
|
| 1496 |
"step": 124000
|
| 1497 |
},
|
| 1498 |
{
|
| 1499 |
"epoch": 2.22,
|
| 1500 |
"learning_rate": 1.3001925682904215e-05,
|
| 1501 |
+
"loss": 4.1828,
|
| 1502 |
"step": 124500
|
| 1503 |
},
|
| 1504 |
{
|
| 1505 |
"epoch": 2.23,
|
| 1506 |
"learning_rate": 1.2853339039060458e-05,
|
| 1507 |
+
"loss": 4.1857,
|
| 1508 |
"step": 125000
|
| 1509 |
},
|
| 1510 |
{
|
| 1511 |
"epoch": 2.24,
|
| 1512 |
"learning_rate": 1.2704752395216699e-05,
|
| 1513 |
+
"loss": 4.1146,
|
| 1514 |
"step": 125500
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
"epoch": 2.25,
|
| 1518 |
"learning_rate": 1.2556165751372942e-05,
|
| 1519 |
+
"loss": 4.1367,
|
| 1520 |
"step": 126000
|
| 1521 |
},
|
| 1522 |
{
|
| 1523 |
"epoch": 2.26,
|
| 1524 |
"learning_rate": 1.2407579107529183e-05,
|
| 1525 |
+
"loss": 4.1608,
|
| 1526 |
"step": 126500
|
| 1527 |
},
|
| 1528 |
{
|
| 1529 |
"epoch": 2.26,
|
| 1530 |
"learning_rate": 1.2258992463685424e-05,
|
| 1531 |
+
"loss": 4.1655,
|
| 1532 |
"step": 127000
|
| 1533 |
},
|
| 1534 |
{
|
| 1535 |
"epoch": 2.27,
|
| 1536 |
"learning_rate": 1.2110405819841667e-05,
|
| 1537 |
+
"loss": 4.158,
|
| 1538 |
"step": 127500
|
| 1539 |
},
|
| 1540 |
{
|
| 1541 |
"epoch": 2.28,
|
| 1542 |
"learning_rate": 1.1961819175997908e-05,
|
| 1543 |
+
"loss": 4.1302,
|
| 1544 |
"step": 128000
|
| 1545 |
},
|
| 1546 |
{
|
| 1547 |
"epoch": 2.29,
|
| 1548 |
"learning_rate": 1.1813232532154151e-05,
|
| 1549 |
+
"loss": 4.1538,
|
| 1550 |
"step": 128500
|
| 1551 |
},
|
| 1552 |
{
|
| 1553 |
"epoch": 2.3,
|
| 1554 |
"learning_rate": 1.1664645888310392e-05,
|
| 1555 |
+
"loss": 4.1838,
|
| 1556 |
"step": 129000
|
| 1557 |
},
|
| 1558 |
{
|
| 1559 |
"epoch": 2.31,
|
| 1560 |
"learning_rate": 1.1516059244466634e-05,
|
| 1561 |
+
"loss": 4.178,
|
| 1562 |
"step": 129500
|
| 1563 |
},
|
| 1564 |
{
|
| 1565 |
"epoch": 2.32,
|
| 1566 |
"learning_rate": 1.1367472600622876e-05,
|
| 1567 |
+
"loss": 4.1766,
|
| 1568 |
"step": 130000
|
| 1569 |
},
|
| 1570 |
{
|
| 1571 |
"epoch": 2.33,
|
| 1572 |
"learning_rate": 1.1218885956779118e-05,
|
| 1573 |
+
"loss": 4.1533,
|
| 1574 |
"step": 130500
|
| 1575 |
},
|
| 1576 |
{
|
| 1577 |
"epoch": 2.34,
|
| 1578 |
"learning_rate": 1.1070299312935359e-05,
|
| 1579 |
+
"loss": 4.1715,
|
| 1580 |
"step": 131000
|
| 1581 |
},
|
| 1582 |
{
|
| 1583 |
"epoch": 2.34,
|
| 1584 |
"learning_rate": 1.0921712669091602e-05,
|
| 1585 |
+
"loss": 4.1212,
|
| 1586 |
"step": 131500
|
| 1587 |
},
|
| 1588 |
{
|
| 1589 |
"epoch": 2.35,
|
| 1590 |
"learning_rate": 1.0773126025247843e-05,
|
| 1591 |
+
"loss": 4.1308,
|
| 1592 |
"step": 132000
|
| 1593 |
},
|
| 1594 |
{
|
| 1595 |
"epoch": 2.36,
|
| 1596 |
"learning_rate": 1.0624539381404084e-05,
|
| 1597 |
+
"loss": 4.1403,
|
| 1598 |
"step": 132500
|
| 1599 |
},
|
| 1600 |
{
|
| 1601 |
"epoch": 2.37,
|
| 1602 |
"learning_rate": 1.0475952737560327e-05,
|
| 1603 |
+
"loss": 4.1551,
|
| 1604 |
"step": 133000
|
| 1605 |
},
|
| 1606 |
{
|
| 1607 |
"epoch": 2.38,
|
| 1608 |
"learning_rate": 1.0327366093716568e-05,
|
| 1609 |
+
"loss": 4.1915,
|
| 1610 |
"step": 133500
|
| 1611 |
},
|
| 1612 |
{
|
| 1613 |
"epoch": 2.39,
|
| 1614 |
"learning_rate": 1.0178779449872811e-05,
|
| 1615 |
+
"loss": 4.1381,
|
| 1616 |
"step": 134000
|
| 1617 |
},
|
| 1618 |
{
|
| 1619 |
"epoch": 2.4,
|
| 1620 |
"learning_rate": 1.0030192806029052e-05,
|
| 1621 |
+
"loss": 4.1527,
|
| 1622 |
"step": 134500
|
| 1623 |
},
|
| 1624 |
{
|
| 1625 |
"epoch": 2.41,
|
| 1626 |
"learning_rate": 9.881606162185295e-06,
|
| 1627 |
+
"loss": 4.1663,
|
| 1628 |
"step": 135000
|
| 1629 |
},
|
| 1630 |
{
|
| 1631 |
"epoch": 2.42,
|
| 1632 |
"learning_rate": 9.733019518341534e-06,
|
| 1633 |
+
"loss": 4.2214,
|
| 1634 |
"step": 135500
|
| 1635 |
},
|
| 1636 |
{
|
| 1637 |
"epoch": 2.42,
|
| 1638 |
"learning_rate": 9.584432874497777e-06,
|
| 1639 |
+
"loss": 4.1587,
|
| 1640 |
"step": 136000
|
| 1641 |
},
|
| 1642 |
{
|
| 1643 |
"epoch": 2.43,
|
| 1644 |
"learning_rate": 9.435846230654018e-06,
|
| 1645 |
+
"loss": 4.1345,
|
| 1646 |
"step": 136500
|
| 1647 |
},
|
| 1648 |
{
|
| 1649 |
"epoch": 2.44,
|
| 1650 |
"learning_rate": 9.287259586810261e-06,
|
| 1651 |
+
"loss": 4.189,
|
| 1652 |
"step": 137000
|
| 1653 |
},
|
| 1654 |
{
|
| 1655 |
"epoch": 2.45,
|
| 1656 |
"learning_rate": 9.138672942966502e-06,
|
| 1657 |
+
"loss": 4.1978,
|
| 1658 |
"step": 137500
|
| 1659 |
},
|
| 1660 |
{
|
| 1661 |
"epoch": 2.46,
|
| 1662 |
"learning_rate": 8.990086299122745e-06,
|
| 1663 |
+
"loss": 4.1403,
|
| 1664 |
"step": 138000
|
| 1665 |
},
|
| 1666 |
{
|
| 1667 |
"epoch": 2.47,
|
| 1668 |
"learning_rate": 8.841499655278986e-06,
|
| 1669 |
+
"loss": 4.1012,
|
| 1670 |
"step": 138500
|
| 1671 |
},
|
| 1672 |
{
|
| 1673 |
"epoch": 2.48,
|
| 1674 |
"learning_rate": 8.69291301143523e-06,
|
| 1675 |
+
"loss": 4.1605,
|
| 1676 |
"step": 139000
|
| 1677 |
},
|
| 1678 |
{
|
| 1679 |
"epoch": 2.49,
|
| 1680 |
"learning_rate": 8.54432636759147e-06,
|
| 1681 |
+
"loss": 4.1597,
|
| 1682 |
"step": 139500
|
| 1683 |
},
|
| 1684 |
{
|
| 1685 |
"epoch": 2.5,
|
| 1686 |
"learning_rate": 8.395739723747713e-06,
|
| 1687 |
+
"loss": 4.1998,
|
| 1688 |
"step": 140000
|
| 1689 |
},
|
| 1690 |
{
|
| 1691 |
"epoch": 2.51,
|
| 1692 |
"learning_rate": 8.247153079903953e-06,
|
| 1693 |
+
"loss": 4.1804,
|
| 1694 |
"step": 140500
|
| 1695 |
},
|
| 1696 |
{
|
| 1697 |
"epoch": 2.51,
|
| 1698 |
"learning_rate": 8.098566436060196e-06,
|
| 1699 |
+
"loss": 4.1377,
|
| 1700 |
"step": 141000
|
| 1701 |
},
|
| 1702 |
{
|
| 1703 |
"epoch": 2.52,
|
| 1704 |
"learning_rate": 7.949979792216437e-06,
|
| 1705 |
+
"loss": 4.1452,
|
| 1706 |
"step": 141500
|
| 1707 |
},
|
| 1708 |
{
|
| 1709 |
"epoch": 2.53,
|
| 1710 |
"learning_rate": 7.80139314837268e-06,
|
| 1711 |
+
"loss": 4.1304,
|
| 1712 |
"step": 142000
|
| 1713 |
},
|
| 1714 |
{
|
| 1715 |
"epoch": 2.54,
|
| 1716 |
"learning_rate": 7.652806504528921e-06,
|
| 1717 |
+
"loss": 4.15,
|
| 1718 |
"step": 142500
|
| 1719 |
},
|
| 1720 |
{
|
| 1721 |
"epoch": 2.55,
|
| 1722 |
"learning_rate": 7.504219860685163e-06,
|
| 1723 |
+
"loss": 4.1462,
|
| 1724 |
"step": 143000
|
| 1725 |
},
|
| 1726 |
{
|
| 1727 |
"epoch": 2.56,
|
| 1728 |
"learning_rate": 7.355633216841405e-06,
|
| 1729 |
+
"loss": 4.1741,
|
| 1730 |
"step": 143500
|
| 1731 |
},
|
| 1732 |
{
|
| 1733 |
"epoch": 2.57,
|
| 1734 |
"learning_rate": 7.207046572997647e-06,
|
| 1735 |
+
"loss": 4.187,
|
| 1736 |
"step": 144000
|
| 1737 |
},
|
| 1738 |
{
|
| 1739 |
"epoch": 2.58,
|
| 1740 |
"learning_rate": 7.058459929153889e-06,
|
| 1741 |
+
"loss": 4.183,
|
| 1742 |
"step": 144500
|
| 1743 |
},
|
| 1744 |
{
|
| 1745 |
"epoch": 2.59,
|
| 1746 |
"learning_rate": 6.909873285310129e-06,
|
| 1747 |
+
"loss": 4.0993,
|
| 1748 |
"step": 145000
|
| 1749 |
},
|
| 1750 |
{
|
| 1751 |
"epoch": 2.59,
|
| 1752 |
"learning_rate": 6.761286641466371e-06,
|
| 1753 |
+
"loss": 4.2255,
|
| 1754 |
"step": 145500
|
| 1755 |
},
|
| 1756 |
{
|
| 1757 |
"epoch": 2.6,
|
| 1758 |
"learning_rate": 6.612699997622613e-06,
|
| 1759 |
+
"loss": 4.1642,
|
| 1760 |
"step": 146000
|
| 1761 |
},
|
| 1762 |
{
|
| 1763 |
"epoch": 2.61,
|
| 1764 |
"learning_rate": 6.464113353778855e-06,
|
| 1765 |
+
"loss": 4.1559,
|
| 1766 |
"step": 146500
|
| 1767 |
},
|
| 1768 |
{
|
| 1769 |
"epoch": 2.62,
|
| 1770 |
"learning_rate": 6.315526709935097e-06,
|
| 1771 |
+
"loss": 4.1859,
|
| 1772 |
"step": 147000
|
| 1773 |
},
|
| 1774 |
{
|
| 1775 |
"epoch": 2.63,
|
| 1776 |
"learning_rate": 6.1669400660913394e-06,
|
| 1777 |
+
"loss": 4.1573,
|
| 1778 |
"step": 147500
|
| 1779 |
},
|
| 1780 |
{
|
| 1781 |
"epoch": 2.64,
|
| 1782 |
"learning_rate": 6.0183534222475815e-06,
|
| 1783 |
+
"loss": 4.1746,
|
| 1784 |
"step": 148000
|
| 1785 |
},
|
| 1786 |
{
|
| 1787 |
"epoch": 2.65,
|
| 1788 |
"learning_rate": 5.869766778403823e-06,
|
| 1789 |
+
"loss": 4.2273,
|
| 1790 |
"step": 148500
|
| 1791 |
},
|
| 1792 |
{
|
| 1793 |
"epoch": 2.66,
|
| 1794 |
"learning_rate": 5.721180134560065e-06,
|
| 1795 |
+
"loss": 4.1844,
|
| 1796 |
"step": 149000
|
| 1797 |
},
|
| 1798 |
{
|
| 1799 |
"epoch": 2.67,
|
| 1800 |
"learning_rate": 5.572593490716307e-06,
|
| 1801 |
+
"loss": 4.1636,
|
| 1802 |
"step": 149500
|
| 1803 |
},
|
| 1804 |
{
|
| 1805 |
"epoch": 2.67,
|
| 1806 |
"learning_rate": 5.424006846872549e-06,
|
| 1807 |
+
"loss": 4.1785,
|
| 1808 |
"step": 150000
|
| 1809 |
},
|
| 1810 |
{
|
| 1811 |
"epoch": 2.68,
|
| 1812 |
"learning_rate": 5.275420203028791e-06,
|
| 1813 |
+
"loss": 4.2064,
|
| 1814 |
"step": 150500
|
| 1815 |
},
|
| 1816 |
{
|
| 1817 |
"epoch": 2.69,
|
| 1818 |
"learning_rate": 5.126833559185032e-06,
|
| 1819 |
+
"loss": 4.1103,
|
| 1820 |
"step": 151000
|
| 1821 |
},
|
| 1822 |
{
|
| 1823 |
"epoch": 2.7,
|
| 1824 |
"learning_rate": 4.978246915341274e-06,
|
| 1825 |
+
"loss": 4.1814,
|
| 1826 |
"step": 151500
|
| 1827 |
},
|
| 1828 |
{
|
| 1829 |
"epoch": 2.71,
|
| 1830 |
"learning_rate": 4.829660271497516e-06,
|
| 1831 |
+
"loss": 4.1903,
|
| 1832 |
"step": 152000
|
| 1833 |
},
|
| 1834 |
{
|
| 1835 |
"epoch": 2.72,
|
| 1836 |
"learning_rate": 4.681073627653758e-06,
|
| 1837 |
+
"loss": 4.1319,
|
| 1838 |
"step": 152500
|
| 1839 |
},
|
| 1840 |
{
|
| 1841 |
"epoch": 2.73,
|
| 1842 |
"learning_rate": 4.53248698381e-06,
|
| 1843 |
+
"loss": 4.127,
|
| 1844 |
"step": 153000
|
| 1845 |
},
|
| 1846 |
{
|
| 1847 |
"epoch": 2.74,
|
| 1848 |
"learning_rate": 4.383900339966241e-06,
|
| 1849 |
+
"loss": 4.1278,
|
| 1850 |
"step": 153500
|
| 1851 |
},
|
| 1852 |
{
|
| 1853 |
"epoch": 2.75,
|
| 1854 |
"learning_rate": 4.235313696122483e-06,
|
| 1855 |
+
"loss": 4.1209,
|
| 1856 |
"step": 154000
|
| 1857 |
},
|
| 1858 |
{
|
| 1859 |
"epoch": 2.75,
|
| 1860 |
"learning_rate": 4.086727052278725e-06,
|
| 1861 |
+
"loss": 4.145,
|
| 1862 |
"step": 154500
|
| 1863 |
},
|
| 1864 |
{
|
| 1865 |
"epoch": 2.76,
|
| 1866 |
"learning_rate": 3.938140408434967e-06,
|
| 1867 |
+
"loss": 4.1624,
|
| 1868 |
"step": 155000
|
| 1869 |
},
|
| 1870 |
{
|
| 1871 |
"epoch": 2.77,
|
| 1872 |
"learning_rate": 3.7895537645912088e-06,
|
| 1873 |
+
"loss": 4.1436,
|
| 1874 |
"step": 155500
|
| 1875 |
},
|
| 1876 |
{
|
| 1877 |
"epoch": 2.78,
|
| 1878 |
"learning_rate": 3.6409671207474504e-06,
|
| 1879 |
+
"loss": 4.1485,
|
| 1880 |
"step": 156000
|
| 1881 |
},
|
| 1882 |
{
|
| 1883 |
"epoch": 2.79,
|
| 1884 |
"learning_rate": 3.4923804769036924e-06,
|
| 1885 |
+
"loss": 4.1695,
|
| 1886 |
"step": 156500
|
| 1887 |
},
|
| 1888 |
{
|
| 1889 |
"epoch": 2.8,
|
| 1890 |
"learning_rate": 3.343793833059934e-06,
|
| 1891 |
+
"loss": 4.0884,
|
| 1892 |
"step": 157000
|
| 1893 |
},
|
| 1894 |
{
|
| 1895 |
"epoch": 2.81,
|
| 1896 |
"learning_rate": 3.195207189216176e-06,
|
| 1897 |
+
"loss": 4.1719,
|
| 1898 |
"step": 157500
|
| 1899 |
},
|
| 1900 |
{
|
| 1901 |
"epoch": 2.82,
|
| 1902 |
"learning_rate": 3.0466205453724176e-06,
|
| 1903 |
+
"loss": 4.1232,
|
| 1904 |
"step": 158000
|
| 1905 |
},
|
| 1906 |
{
|
| 1907 |
"epoch": 2.83,
|
| 1908 |
"learning_rate": 2.898033901528659e-06,
|
| 1909 |
+
"loss": 4.1327,
|
| 1910 |
"step": 158500
|
| 1911 |
},
|
| 1912 |
{
|
| 1913 |
"epoch": 2.84,
|
| 1914 |
"learning_rate": 2.749447257684901e-06,
|
| 1915 |
+
"loss": 4.1568,
|
| 1916 |
"step": 159000
|
| 1917 |
},
|
| 1918 |
{
|
| 1919 |
"epoch": 2.84,
|
| 1920 |
"learning_rate": 2.6008606138411432e-06,
|
| 1921 |
+
"loss": 4.0888,
|
| 1922 |
"step": 159500
|
| 1923 |
},
|
| 1924 |
{
|
| 1925 |
"epoch": 2.85,
|
| 1926 |
"learning_rate": 2.452273969997385e-06,
|
| 1927 |
+
"loss": 4.1711,
|
| 1928 |
"step": 160000
|
| 1929 |
},
|
| 1930 |
{
|
| 1931 |
"epoch": 2.86,
|
| 1932 |
"learning_rate": 2.303687326153627e-06,
|
| 1933 |
+
"loss": 4.2205,
|
| 1934 |
"step": 160500
|
| 1935 |
},
|
| 1936 |
{
|
| 1937 |
"epoch": 2.87,
|
| 1938 |
"learning_rate": 2.1551006823098684e-06,
|
| 1939 |
+
"loss": 4.1221,
|
| 1940 |
"step": 161000
|
| 1941 |
},
|
| 1942 |
{
|
| 1943 |
"epoch": 2.88,
|
| 1944 |
"learning_rate": 2.0065140384661104e-06,
|
| 1945 |
+
"loss": 4.1887,
|
| 1946 |
"step": 161500
|
| 1947 |
},
|
| 1948 |
{
|
| 1949 |
"epoch": 2.89,
|
| 1950 |
"learning_rate": 1.8579273946223525e-06,
|
| 1951 |
+
"loss": 4.138,
|
| 1952 |
"step": 162000
|
| 1953 |
},
|
| 1954 |
{
|
| 1955 |
"epoch": 2.9,
|
| 1956 |
"learning_rate": 1.709340750778594e-06,
|
| 1957 |
+
"loss": 4.166,
|
| 1958 |
"step": 162500
|
| 1959 |
},
|
| 1960 |
{
|
| 1961 |
"epoch": 2.91,
|
| 1962 |
"learning_rate": 1.5607541069348359e-06,
|
| 1963 |
+
"loss": 4.1599,
|
| 1964 |
"step": 163000
|
| 1965 |
},
|
| 1966 |
{
|
| 1967 |
"epoch": 2.92,
|
| 1968 |
"learning_rate": 1.4121674630910779e-06,
|
| 1969 |
+
"loss": 4.1308,
|
| 1970 |
"step": 163500
|
| 1971 |
},
|
| 1972 |
{
|
| 1973 |
"epoch": 2.92,
|
| 1974 |
"learning_rate": 1.2635808192473197e-06,
|
| 1975 |
+
"loss": 4.2161,
|
| 1976 |
"step": 164000
|
| 1977 |
},
|
| 1978 |
{
|
| 1979 |
"epoch": 2.93,
|
| 1980 |
"learning_rate": 1.1149941754035615e-06,
|
| 1981 |
+
"loss": 4.1883,
|
| 1982 |
"step": 164500
|
| 1983 |
},
|
| 1984 |
{
|
| 1985 |
"epoch": 2.94,
|
| 1986 |
"learning_rate": 9.66407531559803e-07,
|
| 1987 |
+
"loss": 4.17,
|
| 1988 |
"step": 165000
|
| 1989 |
},
|
| 1990 |
{
|
| 1991 |
"epoch": 2.95,
|
| 1992 |
"learning_rate": 8.17820887716045e-07,
|
| 1993 |
+
"loss": 4.1741,
|
| 1994 |
"step": 165500
|
| 1995 |
},
|
| 1996 |
{
|
| 1997 |
"epoch": 2.96,
|
| 1998 |
"learning_rate": 6.692342438722869e-07,
|
| 1999 |
+
"loss": 4.1797,
|
| 2000 |
"step": 166000
|
| 2001 |
},
|
| 2002 |
{
|
| 2003 |
"epoch": 2.97,
|
| 2004 |
"learning_rate": 5.206476000285286e-07,
|
| 2005 |
+
"loss": 4.1768,
|
| 2006 |
"step": 166500
|
| 2007 |
},
|
| 2008 |
{
|
| 2009 |
"epoch": 2.98,
|
| 2010 |
"learning_rate": 3.720609561847705e-07,
|
| 2011 |
+
"loss": 4.1909,
|
| 2012 |
"step": 167000
|
| 2013 |
},
|
| 2014 |
{
|
| 2015 |
"epoch": 2.99,
|
| 2016 |
"learning_rate": 2.234743123410123e-07,
|
| 2017 |
+
"loss": 4.1689,
|
| 2018 |
"step": 167500
|
| 2019 |
},
|
| 2020 |
{
|
| 2021 |
"epoch": 3.0,
|
| 2022 |
"learning_rate": 7.488766849725412e-08,
|
| 2023 |
+
"loss": 4.1115,
|
| 2024 |
"step": 168000
|
| 2025 |
},
|
| 2026 |
{
|
| 2027 |
"epoch": 3.0,
|
| 2028 |
"step": 168252,
|
| 2029 |
+
"total_flos": 2.098762564848768e+16,
|
| 2030 |
+
"train_loss": 4.168116396368988,
|
| 2031 |
+
"train_runtime": 19036.7524,
|
| 2032 |
+
"train_samples_per_second": 17.677,
|
| 2033 |
+
"train_steps_per_second": 8.838
|
| 2034 |
}
|
| 2035 |
],
|
| 2036 |
"max_steps": 168252,
|
| 2037 |
"num_train_epochs": 3,
|
| 2038 |
+
"total_flos": 2.098762564848768e+16,
|
| 2039 |
"trial_name": null,
|
| 2040 |
"trial_params": null
|
| 2041 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3567
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:20a50fe822855aeb5b17d149f2c34cd69ab6301adcc00442b5e2675797edd1d6
|
| 3 |
size 3567
|
vocab.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|