Transformers
PyTorch
Graphcore
bert
Generated from Trainer
Inference Endpoints
sergiopperez commited on
Commit
387585b
1 Parent(s): b02bd0a

Update BERT large uncased checkpoint after running phase 1 (SL 128) and phase 2 (SL 512)

Browse files
README.md ADDED
@@ -0,0 +1,141 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ datasets:
5
+ - Graphcore/wikipedia-bert-128
6
+ - Graphcore/wikipedia-bert-512
7
+ model-index:
8
+ - name: Graphcore/bert-large-uncased
9
+ results: []
10
+ ---
11
+
12
+ # Graphcore/bert-large-uncased
13
+
14
+ This model is a pre-trained BERT-Large trained in two phases on the [Graphcore/wikipedia-bert-128](https://huggingface.co/datasets/Graphcore/wikipedia-bert-128) and [Graphcore/wikipedia-bert-512](https://huggingface.co/datasets/Graphcore/wikipedia-bert-512) datasets.
15
+
16
+ ## Model description
17
+
18
+ Pre-trained BERT Large model trained on Wikipedia data.
19
+
20
+
21
+ ## Training and evaluation data
22
+
23
+ Trained on wikipedia datasets:
24
+ - [Graphcore/wikipedia-bert-128](https://huggingface.co/datasets/Graphcore/wikipedia-bert-128)
25
+ - [Graphcore/wikipedia-bert-512](https://huggingface.co/datasets/Graphcore/wikipedia-bert-512)
26
+
27
+ ## Training procedure
28
+
29
+ Trained MLM and NSP pre-training scheme from [Large Batch Optimization for Deep Learning: Training BERT in 76 minutes](https://arxiv.org/abs/1904.00962).
30
+ Trained on 64 Graphcore Mk2 IPUs using [`optimum-graphcore`](https://github.com/huggingface/optimum-graphcore)
31
+
32
+ Command lines:
33
+
34
+ Phase 1:
35
+ ```
36
+ python examples/language-modeling/run_pretraining.py \
37
+ --config_name bert-large-uncased \
38
+ --tokenizer_name bert-large-uncased \
39
+ --ipu_config_name Graphcore/bert-large-ipu \
40
+ --dataset_name Graphcore/wikipedia-bert-128 \
41
+ --do_train \
42
+ --logging_steps 5 \
43
+ --max_seq_length 128 \
44
+ --max_steps 10550 \
45
+ --is_already_preprocessed \
46
+ --dataloader_num_workers 64 \
47
+ --dataloader_mode async_rebatched \
48
+ --lamb \
49
+ --lamb_no_bias_correction \
50
+ --per_device_train_batch_size 8 \
51
+ --gradient_accumulation_steps 512 \
52
+ --pod_type pod64 \
53
+ --learning_rate 0.006 \
54
+ --lr_scheduler_type linear \
55
+ --loss_scaling 32768 \
56
+ --weight_decay 0.01 \
57
+ --warmup_ratio 0.28 \
58
+ --config_overrides "layer_norm_eps=0.001" \
59
+ --ipu_config_overrides "matmul_proportion=[0.14 0.19 0.19 0.19]" \
60
+ --output_dir output-pretrain-bert-large-phase1
61
+ ```
62
+
63
+ Phase 2:
64
+ ```
65
+ python examples/language-modeling/run_pretraining.py \
66
+ --config_name bert-large-uncased \
67
+ --tokenizer_name bert-large-uncased \
68
+ --model_name_or_path ./output-pretrain-bert-large-phase1 \
69
+ --ipu_config_name Graphcore/bert-large-ipu \
70
+ --dataset_name Graphcore/wikipedia-bert-512 \
71
+ --do_train \
72
+ --logging_steps 5 \
73
+ --max_seq_length 512 \
74
+ --max_steps 2038 \
75
+ --is_already_preprocessed \
76
+ --dataloader_num_workers 96 \
77
+ --dataloader_mode async_rebatched \
78
+ --lamb \
79
+ --lamb_no_bias_correction \
80
+ --per_device_train_batch_size 2 \
81
+ --gradient_accumulation_steps 512 \
82
+ --pod_type pod64 \
83
+ --learning_rate 0.002828 \
84
+ --lr_scheduler_type linear \
85
+ --loss_scaling 16384 \
86
+ --weight_decay 0.01 \
87
+ --warmup_ratio 0.128 \
88
+ --config_overrides "layer_norm_eps=0.001" \
89
+ --ipu_config_overrides "matmul_proportion=[0.14 0.19 0.19 0.19]" \
90
+ --output_dir output-pretrain-bert-large-phase2
91
+ ```
92
+
93
+ ### Training hyperparameters
94
+
95
+ The following hyperparameters were used during phase 1 training:
96
+ - learning_rate: 0.006
97
+ - train_batch_size: 8
98
+ - eval_batch_size: 8
99
+ - seed: 42
100
+ - distributed_type: IPU
101
+ - gradient_accumulation_steps: 512
102
+ - total_train_batch_size: 65536
103
+ - total_eval_batch_size: 512
104
+ - optimizer: LAMB
105
+ - lr_scheduler_type: linear
106
+ - lr_scheduler_warmup_ratio: 0.28
107
+ - training_steps: 10550
108
+ - training precision: Mixed Precision
109
+
110
+ The following hyperparameters were used during phase 2 training:
111
+ - learning_rate: 0.002828
112
+ - train_batch_size: 2
113
+ - eval_batch_size: 8
114
+ - seed: 42
115
+ - distributed_type: IPU
116
+ - gradient_accumulation_steps: 512
117
+ - total_train_batch_size: 16384
118
+ - total_eval_batch_size: 512
119
+ - optimizer: LAMB
120
+ - lr_scheduler_type: linear
121
+ - lr_scheduler_warmup_ratio: 0.128
122
+ - training_steps: 2038
123
+ - training precision: Mixed Precision
124
+
125
+ ### Training results
126
+
127
+ ```
128
+ train/epoch: 2.04
129
+ train/global_step: 2038
130
+ train/loss: 1.2002
131
+ train/train_runtime: 12022.3897
132
+ train/train_steps_per_second: 0.17
133
+ train/train_samples_per_second: 2777.367
134
+ ```
135
+
136
+ ### Framework versions
137
+
138
+ - Transformers 4.17.0
139
+ - Pytorch 1.10.0+cpu
140
+ - Datasets 2.0.0
141
+ - Tokenizers 0.11.6
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.04,
3
+ "train_loss": 0.02294661615032911,
4
+ "train_runtime": 3034.1773,
5
+ "train_samples": 16407928,
6
+ "train_samples_per_second": 11004.826,
7
+ "train_steps_per_second": 0.672
8
+ }
config.json CHANGED
@@ -1,87 +1,26 @@
1
  {
2
- "_name_or_path": "/localdata/jamesbr/dev/pretrained_checkpoints/pytorch_bert_large_phase1/",
3
  "architectures": [
4
- "PoptorchPipelinedBertForPretraining"
5
  ],
6
- "async_dataloader": true,
7
- "attention_probs_dropout_prob": 0.0,
8
- "auto_loss_scaling": false,
9
- "batch_size": 2,
10
- "batches_per_step": 1,
11
- "checkpoint_output_dir": "/localdata/jamesbr/dev/pretrained_checkpoints/pytorch_bert_large_phase2",
12
- "checkpoint_steps": null,
13
- "compile_only": false,
14
- "config": null,
15
- "custom_ops": true,
16
- "dataloader_workers": 64,
17
- "dataset": "pretraining",
18
- "disable_progress_bar": true,
19
- "embedding_serialization_factor": 2,
20
- "enable_half_first_order_momentum": false,
21
- "enable_half_partials": true,
22
- "executable_cache_dir": "",
23
- "file_buffer_size": 10,
24
- "global_batch_size": 16384,
25
- "gradient_accumulation": 2048,
26
  "gradient_checkpointing": false,
27
  "hidden_act": "gelu",
28
  "hidden_dropout_prob": 0.1,
29
  "hidden_size": 1024,
30
  "initializer_range": 0.02,
31
- "input_files": [
32
- "data/wikipedia/384/*.tfrecord"
33
- ],
34
  "intermediate_size": 4096,
35
- "ipus_per_replica": 4,
36
  "layer_norm_eps": 0.001,
37
- "layers_per_ipu": [
38
- 3,
39
- 7,
40
- 7,
41
- 7
42
- ],
43
- "learning_rate": 0.002828,
44
- "loss_scaling": 8192.0,
45
- "lr_schedule": "linear",
46
- "lr_warmup": 0.128,
47
- "mask_tokens": 56,
48
- "matmul_proportion": [
49
- 0.15,
50
- 0.25,
51
- 0.25,
52
- 0.25
53
- ],
54
  "max_position_embeddings": 512,
55
  "model_type": "bert",
56
  "num_attention_heads": 16,
57
- "num_epochs": null,
58
  "num_hidden_layers": 24,
59
- "optimizer": "LAMB",
60
- "optimizer_state_offchip": true,
61
  "pad_token_id": 0,
62
  "position_embedding_type": "absolute",
63
- "pred_head_transform": true,
64
- "pretrained_checkpoint": "/localdata/jamesbr/dev/pretrained_checkpoints/pytorch_bert_large_phase1/",
65
- "profile": false,
66
- "profile_dir": "profile",
67
- "random_seed": 42,
68
- "recompute_checkpoint_every_layer": true,
69
- "replicated_tensor_sharding": true,
70
- "replication_factor": 4,
71
- "restore_steps_and_optimizer": false,
72
- "samples_per_step": 16384,
73
- "sdk_version": "poplar_sdk-ubuntu_18_04-2.3.0-EA.1+716-757737e247",
74
- "sequence_length": 384,
75
- "squad_do_training": true,
76
- "squad_do_validation": true,
77
- "synthetic_data": false,
78
- "training_steps": 2137,
79
- "transformers_version": "4.7.0",
80
  "type_vocab_size": 2,
81
  "use_cache": true,
82
- "use_popdist": false,
83
- "vocab_size": 30522,
84
- "wandb": true,
85
- "wandb_param_steps": null,
86
- "weight_decay": 0.01
87
  }
 
1
  {
2
+ "_name_or_path": "./output-pretrain-bert-large-phase1",
3
  "architectures": [
4
+ "PoptorchPipelinedBertForPreTraining"
5
  ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8
  "gradient_checkpointing": false,
9
  "hidden_act": "gelu",
10
  "hidden_dropout_prob": 0.1,
11
  "hidden_size": 1024,
12
  "initializer_range": 0.02,
 
 
 
13
  "intermediate_size": 4096,
 
14
  "layer_norm_eps": 0.001,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
15
  "max_position_embeddings": 512,
16
  "model_type": "bert",
17
  "num_attention_heads": 16,
 
18
  "num_hidden_layers": 24,
 
 
19
  "pad_token_id": 0,
20
  "position_embedding_type": "absolute",
21
+ "torch_dtype": "float16",
22
+ "transformers_version": "4.17.0",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
  "type_vocab_size": 2,
24
  "use_cache": true,
25
+ "vocab_size": 30522
 
 
 
 
26
  }
ipu_config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "device_iterations": 1,
3
+ "embedding_serialization_factor": 2,
4
+ "enable_half_first_order_momentum": true,
5
+ "enable_half_partials": true,
6
+ "executable_cache_dir": "./exe_cache",
7
+ "gradient_accumulation_steps": 512,
8
+ "inference_device_iterations": 4,
9
+ "inference_replication_factor": 16,
10
+ "ipus_per_replica": 4,
11
+ "layers_per_ipu": [
12
+ 3,
13
+ 7,
14
+ 7,
15
+ 7
16
+ ],
17
+ "matmul_proportion": [
18
+ 0.1,
19
+ 0.15,
20
+ 0.15,
21
+ 0.15
22
+ ],
23
+ "optimizer_state_offchip": true,
24
+ "optimum_version": "1.0.0",
25
+ "output_mode": "final",
26
+ "profile_dir": "",
27
+ "recompute_checkpoint_every_layer": true,
28
+ "replicated_tensor_sharding": true,
29
+ "replication_factor": 16,
30
+ "seed": 42,
31
+ "use_popdist": false
32
+ }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0aa7af4e8c9c66b940751f91be95590f3dde4aba5160d65d5ce952dff2835a63
3
- size 672610403
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15c5db5802c5caced9e33d4bffcdec6a7616973bb7c477788ad7a595dd77f8c8
3
+ size 672706657
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "bert-large-uncased", "tokenizer_class": "BertTokenizer"}
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.04,
3
+ "train_loss": 0.02294661615032911,
4
+ "train_runtime": 3034.1773,
5
+ "train_samples": 16407928,
6
+ "train_samples_per_second": 11004.826,
7
+ "train_steps_per_second": 0.672
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2467 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.035964035964036,
5
+ "global_step": 2038,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 5.417624521072796e-05,
13
+ "loss": 3.9478,
14
+ "step": 5
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 0.00010835249042145592,
19
+ "loss": 3.4956,
20
+ "step": 10
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 0.00016252873563218388,
25
+ "loss": 3.5834,
26
+ "step": 15
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 0.00021670498084291185,
31
+ "loss": 3.6582,
32
+ "step": 20
33
+ },
34
+ {
35
+ "epoch": 0.02,
36
+ "learning_rate": 0.00027088122605363983,
37
+ "loss": 3.6931,
38
+ "step": 25
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 0.00032505747126436777,
43
+ "loss": 3.5358,
44
+ "step": 30
45
+ },
46
+ {
47
+ "epoch": 0.03,
48
+ "learning_rate": 0.00037923371647509576,
49
+ "loss": 3.4064,
50
+ "step": 35
51
+ },
52
+ {
53
+ "epoch": 0.04,
54
+ "learning_rate": 0.0004334099616858237,
55
+ "loss": 3.6459,
56
+ "step": 40
57
+ },
58
+ {
59
+ "epoch": 0.04,
60
+ "learning_rate": 0.00048758620689655173,
61
+ "loss": 3.537,
62
+ "step": 45
63
+ },
64
+ {
65
+ "epoch": 0.05,
66
+ "learning_rate": 0.0005417624521072797,
67
+ "loss": 3.5901,
68
+ "step": 50
69
+ },
70
+ {
71
+ "epoch": 0.05,
72
+ "learning_rate": 0.0005959386973180077,
73
+ "loss": 3.354,
74
+ "step": 55
75
+ },
76
+ {
77
+ "epoch": 0.06,
78
+ "learning_rate": 0.0006501149425287355,
79
+ "loss": 3.5718,
80
+ "step": 60
81
+ },
82
+ {
83
+ "epoch": 0.06,
84
+ "learning_rate": 0.0007042911877394635,
85
+ "loss": 3.4404,
86
+ "step": 65
87
+ },
88
+ {
89
+ "epoch": 0.07,
90
+ "learning_rate": 0.0007584674329501915,
91
+ "loss": 3.6794,
92
+ "step": 70
93
+ },
94
+ {
95
+ "epoch": 0.07,
96
+ "learning_rate": 0.0008126436781609194,
97
+ "loss": 3.4177,
98
+ "step": 75
99
+ },
100
+ {
101
+ "epoch": 0.08,
102
+ "learning_rate": 0.0008668199233716474,
103
+ "loss": 3.318,
104
+ "step": 80
105
+ },
106
+ {
107
+ "epoch": 0.08,
108
+ "learning_rate": 0.0009209961685823754,
109
+ "loss": 3.2787,
110
+ "step": 85
111
+ },
112
+ {
113
+ "epoch": 0.09,
114
+ "learning_rate": 0.0009751724137931035,
115
+ "loss": 3.4573,
116
+ "step": 90
117
+ },
118
+ {
119
+ "epoch": 0.09,
120
+ "learning_rate": 0.0010293486590038313,
121
+ "loss": 3.5262,
122
+ "step": 95
123
+ },
124
+ {
125
+ "epoch": 0.1,
126
+ "learning_rate": 0.0010835249042145593,
127
+ "loss": 3.5019,
128
+ "step": 100
129
+ },
130
+ {
131
+ "epoch": 0.1,
132
+ "learning_rate": 0.0011377011494252873,
133
+ "loss": 3.4242,
134
+ "step": 105
135
+ },
136
+ {
137
+ "epoch": 0.11,
138
+ "learning_rate": 0.0011918773946360153,
139
+ "loss": 3.321,
140
+ "step": 110
141
+ },
142
+ {
143
+ "epoch": 0.11,
144
+ "learning_rate": 0.0012460536398467433,
145
+ "loss": 3.2642,
146
+ "step": 115
147
+ },
148
+ {
149
+ "epoch": 0.12,
150
+ "learning_rate": 0.001300229885057471,
151
+ "loss": 3.4654,
152
+ "step": 120
153
+ },
154
+ {
155
+ "epoch": 0.12,
156
+ "learning_rate": 0.001354406130268199,
157
+ "loss": 3.5731,
158
+ "step": 125
159
+ },
160
+ {
161
+ "epoch": 0.13,
162
+ "learning_rate": 0.001408582375478927,
163
+ "loss": 3.5807,
164
+ "step": 130
165
+ },
166
+ {
167
+ "epoch": 0.13,
168
+ "learning_rate": 0.0014627586206896553,
169
+ "loss": 3.3095,
170
+ "step": 135
171
+ },
172
+ {
173
+ "epoch": 0.14,
174
+ "learning_rate": 0.001516934865900383,
175
+ "loss": 3.2068,
176
+ "step": 140
177
+ },
178
+ {
179
+ "epoch": 0.14,
180
+ "learning_rate": 0.001571111111111111,
181
+ "loss": 3.3576,
182
+ "step": 145
183
+ },
184
+ {
185
+ "epoch": 0.15,
186
+ "learning_rate": 0.0016252873563218388,
187
+ "loss": 3.1022,
188
+ "step": 150
189
+ },
190
+ {
191
+ "epoch": 0.15,
192
+ "learning_rate": 0.001679463601532567,
193
+ "loss": 3.3052,
194
+ "step": 155
195
+ },
196
+ {
197
+ "epoch": 0.16,
198
+ "learning_rate": 0.0017336398467432948,
199
+ "loss": 3.1333,
200
+ "step": 160
201
+ },
202
+ {
203
+ "epoch": 0.16,
204
+ "learning_rate": 0.001787816091954023,
205
+ "loss": 3.1107,
206
+ "step": 165
207
+ },
208
+ {
209
+ "epoch": 0.17,
210
+ "learning_rate": 0.0018419923371647507,
211
+ "loss": 3.2723,
212
+ "step": 170
213
+ },
214
+ {
215
+ "epoch": 0.17,
216
+ "learning_rate": 0.0018961685823754787,
217
+ "loss": 3.0396,
218
+ "step": 175
219
+ },
220
+ {
221
+ "epoch": 0.18,
222
+ "learning_rate": 0.001950344827586207,
223
+ "loss": 3.1269,
224
+ "step": 180
225
+ },
226
+ {
227
+ "epoch": 0.18,
228
+ "learning_rate": 0.0020045210727969347,
229
+ "loss": 3.4973,
230
+ "step": 185
231
+ },
232
+ {
233
+ "epoch": 0.19,
234
+ "learning_rate": 0.0020586973180076627,
235
+ "loss": 3.2676,
236
+ "step": 190
237
+ },
238
+ {
239
+ "epoch": 0.19,
240
+ "learning_rate": 0.0021128735632183907,
241
+ "loss": 2.7068,
242
+ "step": 195
243
+ },
244
+ {
245
+ "epoch": 0.2,
246
+ "learning_rate": 0.0021670498084291187,
247
+ "loss": 3.0751,
248
+ "step": 200
249
+ },
250
+ {
251
+ "epoch": 0.2,
252
+ "learning_rate": 0.0022212260536398467,
253
+ "loss": 3.0416,
254
+ "step": 205
255
+ },
256
+ {
257
+ "epoch": 0.21,
258
+ "learning_rate": 0.0022754022988505746,
259
+ "loss": 3.1006,
260
+ "step": 210
261
+ },
262
+ {
263
+ "epoch": 0.21,
264
+ "learning_rate": 0.002329578544061302,
265
+ "loss": 2.8475,
266
+ "step": 215
267
+ },
268
+ {
269
+ "epoch": 0.22,
270
+ "learning_rate": 0.0023837547892720306,
271
+ "loss": 3.1305,
272
+ "step": 220
273
+ },
274
+ {
275
+ "epoch": 0.22,
276
+ "learning_rate": 0.002437931034482758,
277
+ "loss": 3.0844,
278
+ "step": 225
279
+ },
280
+ {
281
+ "epoch": 0.23,
282
+ "learning_rate": 0.0024921072796934866,
283
+ "loss": 2.8136,
284
+ "step": 230
285
+ },
286
+ {
287
+ "epoch": 0.23,
288
+ "learning_rate": 0.0025462835249042146,
289
+ "loss": 2.9123,
290
+ "step": 235
291
+ },
292
+ {
293
+ "epoch": 0.24,
294
+ "learning_rate": 0.002600459770114942,
295
+ "loss": 3.0378,
296
+ "step": 240
297
+ },
298
+ {
299
+ "epoch": 0.24,
300
+ "learning_rate": 0.0026546360153256706,
301
+ "loss": 2.6106,
302
+ "step": 245
303
+ },
304
+ {
305
+ "epoch": 0.25,
306
+ "learning_rate": 0.002708812260536398,
307
+ "loss": 2.7494,
308
+ "step": 250
309
+ },
310
+ {
311
+ "epoch": 0.25,
312
+ "learning_rate": 0.002762988505747126,
313
+ "loss": 2.8047,
314
+ "step": 255
315
+ },
316
+ {
317
+ "epoch": 0.26,
318
+ "learning_rate": 0.002817164750957854,
319
+ "loss": 2.4465,
320
+ "step": 260
321
+ },
322
+ {
323
+ "epoch": 0.26,
324
+ "learning_rate": 0.002821634214969049,
325
+ "loss": 2.6294,
326
+ "step": 265
327
+ },
328
+ {
329
+ "epoch": 0.27,
330
+ "learning_rate": 0.00281367698368036,
331
+ "loss": 2.7954,
332
+ "step": 270
333
+ },
334
+ {
335
+ "epoch": 0.27,
336
+ "learning_rate": 0.002805719752391671,
337
+ "loss": 2.3837,
338
+ "step": 275
339
+ },
340
+ {
341
+ "epoch": 0.28,
342
+ "learning_rate": 0.0027977625211029825,
343
+ "loss": 2.0739,
344
+ "step": 280
345
+ },
346
+ {
347
+ "epoch": 0.28,
348
+ "learning_rate": 0.0027898052898142935,
349
+ "loss": 2.08,
350
+ "step": 285
351
+ },
352
+ {
353
+ "epoch": 0.29,
354
+ "learning_rate": 0.002781848058525605,
355
+ "loss": 2.2492,
356
+ "step": 290
357
+ },
358
+ {
359
+ "epoch": 0.29,
360
+ "learning_rate": 0.002773890827236916,
361
+ "loss": 1.9265,
362
+ "step": 295
363
+ },
364
+ {
365
+ "epoch": 0.3,
366
+ "learning_rate": 0.002765933595948227,
367
+ "loss": 1.9779,
368
+ "step": 300
369
+ },
370
+ {
371
+ "epoch": 0.3,
372
+ "learning_rate": 0.0027579763646595383,
373
+ "loss": 1.8238,
374
+ "step": 305
375
+ },
376
+ {
377
+ "epoch": 0.31,
378
+ "learning_rate": 0.0027500191333708493,
379
+ "loss": 1.6509,
380
+ "step": 310
381
+ },
382
+ {
383
+ "epoch": 0.31,
384
+ "learning_rate": 0.0027420619020821608,
385
+ "loss": 1.5633,
386
+ "step": 315
387
+ },
388
+ {
389
+ "epoch": 0.32,
390
+ "learning_rate": 0.0027341046707934718,
391
+ "loss": 1.5235,
392
+ "step": 320
393
+ },
394
+ {
395
+ "epoch": 0.32,
396
+ "learning_rate": 0.002726147439504783,
397
+ "loss": 1.4549,
398
+ "step": 325
399
+ },
400
+ {
401
+ "epoch": 0.33,
402
+ "learning_rate": 0.002718190208216094,
403
+ "loss": 1.4057,
404
+ "step": 330
405
+ },
406
+ {
407
+ "epoch": 0.33,
408
+ "learning_rate": 0.0027102329769274056,
409
+ "loss": 1.4699,
410
+ "step": 335
411
+ },
412
+ {
413
+ "epoch": 0.34,
414
+ "learning_rate": 0.002702275745638717,
415
+ "loss": 1.3995,
416
+ "step": 340
417
+ },
418
+ {
419
+ "epoch": 0.34,
420
+ "learning_rate": 0.002694318514350028,
421
+ "loss": 1.2831,
422
+ "step": 345
423
+ },
424
+ {
425
+ "epoch": 0.35,
426
+ "learning_rate": 0.002686361283061339,
427
+ "loss": 1.4057,
428
+ "step": 350
429
+ },
430
+ {
431
+ "epoch": 0.35,
432
+ "learning_rate": 0.0026784040517726505,
433
+ "loss": 1.3694,
434
+ "step": 355
435
+ },
436
+ {
437
+ "epoch": 0.36,
438
+ "learning_rate": 0.0026704468204839615,
439
+ "loss": 1.3262,
440
+ "step": 360
441
+ },
442
+ {
443
+ "epoch": 0.36,
444
+ "learning_rate": 0.002662489589195273,
445
+ "loss": 1.3559,
446
+ "step": 365
447
+ },
448
+ {
449
+ "epoch": 0.37,
450
+ "learning_rate": 0.002654532357906584,
451
+ "loss": 1.3316,
452
+ "step": 370
453
+ },
454
+ {
455
+ "epoch": 0.37,
456
+ "learning_rate": 0.002646575126617895,
457
+ "loss": 1.4259,
458
+ "step": 375
459
+ },
460
+ {
461
+ "epoch": 0.38,
462
+ "learning_rate": 0.0026386178953292064,
463
+ "loss": 1.3915,
464
+ "step": 380
465
+ },
466
+ {
467
+ "epoch": 0.38,
468
+ "learning_rate": 0.0026306606640405174,
469
+ "loss": 1.3454,
470
+ "step": 385
471
+ },
472
+ {
473
+ "epoch": 0.39,
474
+ "learning_rate": 0.002622703432751829,
475
+ "loss": 1.4387,
476
+ "step": 390
477
+ },
478
+ {
479
+ "epoch": 0.39,
480
+ "learning_rate": 0.00261474620146314,
481
+ "loss": 1.3051,
482
+ "step": 395
483
+ },
484
+ {
485
+ "epoch": 0.4,
486
+ "learning_rate": 0.002606788970174451,
487
+ "loss": 1.4805,
488
+ "step": 400
489
+ },
490
+ {
491
+ "epoch": 0.4,
492
+ "learning_rate": 0.0025988317388857623,
493
+ "loss": 1.4211,
494
+ "step": 405
495
+ },
496
+ {
497
+ "epoch": 0.41,
498
+ "learning_rate": 0.0025908745075970733,
499
+ "loss": 1.3394,
500
+ "step": 410
501
+ },
502
+ {
503
+ "epoch": 0.41,
504
+ "learning_rate": 0.0025829172763083847,
505
+ "loss": 1.4346,
506
+ "step": 415
507
+ },
508
+ {
509
+ "epoch": 0.42,
510
+ "learning_rate": 0.002574960045019696,
511
+ "loss": 1.4137,
512
+ "step": 420
513
+ },
514
+ {
515
+ "epoch": 0.42,
516
+ "learning_rate": 0.002567002813731007,
517
+ "loss": 1.407,
518
+ "step": 425
519
+ },
520
+ {
521
+ "epoch": 0.43,
522
+ "learning_rate": 0.0025590455824423186,
523
+ "loss": 1.4121,
524
+ "step": 430
525
+ },
526
+ {
527
+ "epoch": 0.43,
528
+ "learning_rate": 0.0025510883511536296,
529
+ "loss": 1.3267,
530
+ "step": 435
531
+ },
532
+ {
533
+ "epoch": 0.44,
534
+ "learning_rate": 0.002543131119864941,
535
+ "loss": 1.4152,
536
+ "step": 440
537
+ },
538
+ {
539
+ "epoch": 0.44,
540
+ "learning_rate": 0.002535173888576252,
541
+ "loss": 1.3324,
542
+ "step": 445
543
+ },
544
+ {
545
+ "epoch": 0.45,
546
+ "learning_rate": 0.002527216657287563,
547
+ "loss": 1.3688,
548
+ "step": 450
549
+ },
550
+ {
551
+ "epoch": 0.45,
552
+ "learning_rate": 0.0025192594259988745,
553
+ "loss": 1.4605,
554
+ "step": 455
555
+ },
556
+ {
557
+ "epoch": 0.46,
558
+ "learning_rate": 0.0025113021947101855,
559
+ "loss": 1.3467,
560
+ "step": 460
561
+ },
562
+ {
563
+ "epoch": 0.46,
564
+ "learning_rate": 0.002503344963421497,
565
+ "loss": 1.3955,
566
+ "step": 465
567
+ },
568
+ {
569
+ "epoch": 0.47,
570
+ "learning_rate": 0.002495387732132808,
571
+ "loss": 1.3075,
572
+ "step": 470
573
+ },
574
+ {
575
+ "epoch": 0.47,
576
+ "learning_rate": 0.002487430500844119,
577
+ "loss": 1.4311,
578
+ "step": 475
579
+ },
580
+ {
581
+ "epoch": 0.48,
582
+ "learning_rate": 0.0024794732695554303,
583
+ "loss": 1.454,
584
+ "step": 480
585
+ },
586
+ {
587
+ "epoch": 0.48,
588
+ "learning_rate": 0.0024715160382667413,
589
+ "loss": 1.422,
590
+ "step": 485
591
+ },
592
+ {
593
+ "epoch": 0.49,
594
+ "learning_rate": 0.0024635588069780528,
595
+ "loss": 1.2557,
596
+ "step": 490
597
+ },
598
+ {
599
+ "epoch": 0.49,
600
+ "learning_rate": 0.0024556015756893638,
601
+ "loss": 1.3965,
602
+ "step": 495
603
+ },
604
+ {
605
+ "epoch": 0.5,
606
+ "learning_rate": 0.002447644344400675,
607
+ "loss": 1.3865,
608
+ "step": 500
609
+ },
610
+ {
611
+ "epoch": 0.5,
612
+ "learning_rate": 0.002439687113111986,
613
+ "loss": 1.3908,
614
+ "step": 505
615
+ },
616
+ {
617
+ "epoch": 0.51,
618
+ "learning_rate": 0.0024317298818232976,
619
+ "loss": 1.349,
620
+ "step": 510
621
+ },
622
+ {
623
+ "epoch": 0.51,
624
+ "learning_rate": 0.0024237726505346086,
625
+ "loss": 1.2927,
626
+ "step": 515
627
+ },
628
+ {
629
+ "epoch": 0.52,
630
+ "learning_rate": 0.00241581541924592,
631
+ "loss": 1.3646,
632
+ "step": 520
633
+ },
634
+ {
635
+ "epoch": 0.52,
636
+ "learning_rate": 0.002407858187957231,
637
+ "loss": 1.3899,
638
+ "step": 525
639
+ },
640
+ {
641
+ "epoch": 0.53,
642
+ "learning_rate": 0.0023999009566685425,
643
+ "loss": 1.3826,
644
+ "step": 530
645
+ },
646
+ {
647
+ "epoch": 0.53,
648
+ "learning_rate": 0.0023919437253798535,
649
+ "loss": 1.3454,
650
+ "step": 535
651
+ },
652
+ {
653
+ "epoch": 0.54,
654
+ "learning_rate": 0.002383986494091165,
655
+ "loss": 1.2858,
656
+ "step": 540
657
+ },
658
+ {
659
+ "epoch": 0.54,
660
+ "learning_rate": 0.002376029262802476,
661
+ "loss": 1.3471,
662
+ "step": 545
663
+ },
664
+ {
665
+ "epoch": 0.55,
666
+ "learning_rate": 0.002368072031513787,
667
+ "loss": 1.3412,
668
+ "step": 550
669
+ },
670
+ {
671
+ "epoch": 0.55,
672
+ "learning_rate": 0.0023601148002250984,
673
+ "loss": 1.4073,
674
+ "step": 555
675
+ },
676
+ {
677
+ "epoch": 0.56,
678
+ "learning_rate": 0.0023521575689364094,
679
+ "loss": 1.4214,
680
+ "step": 560
681
+ },
682
+ {
683
+ "epoch": 0.56,
684
+ "learning_rate": 0.002344200337647721,
685
+ "loss": 1.3379,
686
+ "step": 565
687
+ },
688
+ {
689
+ "epoch": 0.57,
690
+ "learning_rate": 0.002336243106359032,
691
+ "loss": 1.4228,
692
+ "step": 570
693
+ },
694
+ {
695
+ "epoch": 0.57,
696
+ "learning_rate": 0.002328285875070343,
697
+ "loss": 1.433,
698
+ "step": 575
699
+ },
700
+ {
701
+ "epoch": 0.58,
702
+ "learning_rate": 0.0023203286437816543,
703
+ "loss": 1.3152,
704
+ "step": 580
705
+ },
706
+ {
707
+ "epoch": 0.58,
708
+ "learning_rate": 0.0023123714124929653,
709
+ "loss": 1.3381,
710
+ "step": 585
711
+ },
712
+ {
713
+ "epoch": 0.59,
714
+ "learning_rate": 0.0023044141812042767,
715
+ "loss": 1.3229,
716
+ "step": 590
717
+ },
718
+ {
719
+ "epoch": 0.59,
720
+ "learning_rate": 0.0022964569499155877,
721
+ "loss": 1.331,
722
+ "step": 595
723
+ },
724
+ {
725
+ "epoch": 0.6,
726
+ "learning_rate": 0.002288499718626899,
727
+ "loss": 1.3322,
728
+ "step": 600
729
+ },
730
+ {
731
+ "epoch": 0.6,
732
+ "learning_rate": 0.0022805424873382106,
733
+ "loss": 1.284,
734
+ "step": 605
735
+ },
736
+ {
737
+ "epoch": 0.61,
738
+ "learning_rate": 0.0022725852560495216,
739
+ "loss": 1.4003,
740
+ "step": 610
741
+ },
742
+ {
743
+ "epoch": 0.61,
744
+ "learning_rate": 0.002264628024760833,
745
+ "loss": 1.2798,
746
+ "step": 615
747
+ },
748
+ {
749
+ "epoch": 0.62,
750
+ "learning_rate": 0.002256670793472144,
751
+ "loss": 1.3699,
752
+ "step": 620
753
+ },
754
+ {
755
+ "epoch": 0.62,
756
+ "learning_rate": 0.002248713562183455,
757
+ "loss": 1.2648,
758
+ "step": 625
759
+ },
760
+ {
761
+ "epoch": 0.63,
762
+ "learning_rate": 0.0022407563308947664,
763
+ "loss": 1.2905,
764
+ "step": 630
765
+ },
766
+ {
767
+ "epoch": 0.63,
768
+ "learning_rate": 0.0022327990996060774,
769
+ "loss": 1.3432,
770
+ "step": 635
771
+ },
772
+ {
773
+ "epoch": 0.64,
774
+ "learning_rate": 0.002224841868317389,
775
+ "loss": 1.3604,
776
+ "step": 640
777
+ },
778
+ {
779
+ "epoch": 0.64,
780
+ "learning_rate": 0.0022168846370287,
781
+ "loss": 1.3513,
782
+ "step": 645
783
+ },
784
+ {
785
+ "epoch": 0.65,
786
+ "learning_rate": 0.002208927405740011,
787
+ "loss": 1.3222,
788
+ "step": 650
789
+ },
790
+ {
791
+ "epoch": 0.65,
792
+ "learning_rate": 0.0022009701744513223,
793
+ "loss": 1.4611,
794
+ "step": 655
795
+ },
796
+ {
797
+ "epoch": 0.66,
798
+ "learning_rate": 0.0021930129431626333,
799
+ "loss": 1.3717,
800
+ "step": 660
801
+ },
802
+ {
803
+ "epoch": 0.66,
804
+ "learning_rate": 0.0021850557118739448,
805
+ "loss": 1.3321,
806
+ "step": 665
807
+ },
808
+ {
809
+ "epoch": 0.67,
810
+ "learning_rate": 0.0021770984805852558,
811
+ "loss": 1.242,
812
+ "step": 670
813
+ },
814
+ {
815
+ "epoch": 0.67,
816
+ "learning_rate": 0.0021691412492965668,
817
+ "loss": 1.3239,
818
+ "step": 675
819
+ },
820
+ {
821
+ "epoch": 0.68,
822
+ "learning_rate": 0.002161184018007878,
823
+ "loss": 1.3978,
824
+ "step": 680
825
+ },
826
+ {
827
+ "epoch": 0.68,
828
+ "learning_rate": 0.0021532267867191896,
829
+ "loss": 1.3098,
830
+ "step": 685
831
+ },
832
+ {
833
+ "epoch": 0.69,
834
+ "learning_rate": 0.0021452695554305006,
835
+ "loss": 1.3272,
836
+ "step": 690
837
+ },
838
+ {
839
+ "epoch": 0.69,
840
+ "learning_rate": 0.002137312324141812,
841
+ "loss": 1.3465,
842
+ "step": 695
843
+ },
844
+ {
845
+ "epoch": 0.7,
846
+ "learning_rate": 0.002129355092853123,
847
+ "loss": 1.3863,
848
+ "step": 700
849
+ },
850
+ {
851
+ "epoch": 0.7,
852
+ "learning_rate": 0.0021213978615644345,
853
+ "loss": 1.2466,
854
+ "step": 705
855
+ },
856
+ {
857
+ "epoch": 0.71,
858
+ "learning_rate": 0.0021134406302757455,
859
+ "loss": 1.2819,
860
+ "step": 710
861
+ },
862
+ {
863
+ "epoch": 0.71,
864
+ "learning_rate": 0.002105483398987057,
865
+ "loss": 1.2984,
866
+ "step": 715
867
+ },
868
+ {
869
+ "epoch": 0.72,
870
+ "learning_rate": 0.002097526167698368,
871
+ "loss": 1.3937,
872
+ "step": 720
873
+ },
874
+ {
875
+ "epoch": 0.72,
876
+ "learning_rate": 0.002089568936409679,
877
+ "loss": 1.3559,
878
+ "step": 725
879
+ },
880
+ {
881
+ "epoch": 0.73,
882
+ "learning_rate": 0.0020816117051209904,
883
+ "loss": 1.3167,
884
+ "step": 730
885
+ },
886
+ {
887
+ "epoch": 0.73,
888
+ "learning_rate": 0.0020736544738323014,
889
+ "loss": 1.312,
890
+ "step": 735
891
+ },
892
+ {
893
+ "epoch": 0.74,
894
+ "learning_rate": 0.002065697242543613,
895
+ "loss": 1.4823,
896
+ "step": 740
897
+ },
898
+ {
899
+ "epoch": 0.74,
900
+ "learning_rate": 0.002057740011254924,
901
+ "loss": 1.3505,
902
+ "step": 745
903
+ },
904
+ {
905
+ "epoch": 0.75,
906
+ "learning_rate": 0.002049782779966235,
907
+ "loss": 1.3069,
908
+ "step": 750
909
+ },
910
+ {
911
+ "epoch": 0.75,
912
+ "learning_rate": 0.0020418255486775463,
913
+ "loss": 1.2876,
914
+ "step": 755
915
+ },
916
+ {
917
+ "epoch": 0.76,
918
+ "learning_rate": 0.0020338683173888573,
919
+ "loss": 1.3353,
920
+ "step": 760
921
+ },
922
+ {
923
+ "epoch": 0.76,
924
+ "learning_rate": 0.0020259110861001687,
925
+ "loss": 1.3717,
926
+ "step": 765
927
+ },
928
+ {
929
+ "epoch": 0.77,
930
+ "learning_rate": 0.0020179538548114797,
931
+ "loss": 1.3586,
932
+ "step": 770
933
+ },
934
+ {
935
+ "epoch": 0.77,
936
+ "learning_rate": 0.002009996623522791,
937
+ "loss": 1.3533,
938
+ "step": 775
939
+ },
940
+ {
941
+ "epoch": 0.78,
942
+ "learning_rate": 0.002002039392234102,
943
+ "loss": 1.237,
944
+ "step": 780
945
+ },
946
+ {
947
+ "epoch": 0.78,
948
+ "learning_rate": 0.0019940821609454136,
949
+ "loss": 1.2833,
950
+ "step": 785
951
+ },
952
+ {
953
+ "epoch": 0.79,
954
+ "learning_rate": 0.001986124929656725,
955
+ "loss": 1.3615,
956
+ "step": 790
957
+ },
958
+ {
959
+ "epoch": 0.79,
960
+ "learning_rate": 0.001978167698368036,
961
+ "loss": 1.461,
962
+ "step": 795
963
+ },
964
+ {
965
+ "epoch": 0.8,
966
+ "learning_rate": 0.001970210467079347,
967
+ "loss": 1.3725,
968
+ "step": 800
969
+ },
970
+ {
971
+ "epoch": 0.8,
972
+ "learning_rate": 0.0019622532357906584,
973
+ "loss": 1.2839,
974
+ "step": 805
975
+ },
976
+ {
977
+ "epoch": 0.81,
978
+ "learning_rate": 0.0019542960045019694,
979
+ "loss": 1.2933,
980
+ "step": 810
981
+ },
982
+ {
983
+ "epoch": 0.81,
984
+ "learning_rate": 0.0019463387732132809,
985
+ "loss": 1.2416,
986
+ "step": 815
987
+ },
988
+ {
989
+ "epoch": 0.82,
990
+ "learning_rate": 0.0019383815419245919,
991
+ "loss": 1.3515,
992
+ "step": 820
993
+ },
994
+ {
995
+ "epoch": 0.82,
996
+ "learning_rate": 0.0019304243106359029,
997
+ "loss": 1.2728,
998
+ "step": 825
999
+ },
1000
+ {
1001
+ "epoch": 0.83,
1002
+ "learning_rate": 0.0019224670793472143,
1003
+ "loss": 1.2913,
1004
+ "step": 830
1005
+ },
1006
+ {
1007
+ "epoch": 0.83,
1008
+ "learning_rate": 0.0019145098480585253,
1009
+ "loss": 1.2836,
1010
+ "step": 835
1011
+ },
1012
+ {
1013
+ "epoch": 0.84,
1014
+ "learning_rate": 0.0019065526167698368,
1015
+ "loss": 1.3717,
1016
+ "step": 840
1017
+ },
1018
+ {
1019
+ "epoch": 0.84,
1020
+ "learning_rate": 0.0018985953854811478,
1021
+ "loss": 1.2558,
1022
+ "step": 845
1023
+ },
1024
+ {
1025
+ "epoch": 0.85,
1026
+ "learning_rate": 0.001890638154192459,
1027
+ "loss": 1.3507,
1028
+ "step": 850
1029
+ },
1030
+ {
1031
+ "epoch": 0.85,
1032
+ "learning_rate": 0.0018826809229037704,
1033
+ "loss": 1.3441,
1034
+ "step": 855
1035
+ },
1036
+ {
1037
+ "epoch": 0.86,
1038
+ "learning_rate": 0.0018747236916150814,
1039
+ "loss": 1.307,
1040
+ "step": 860
1041
+ },
1042
+ {
1043
+ "epoch": 0.86,
1044
+ "learning_rate": 0.0018667664603263928,
1045
+ "loss": 1.3997,
1046
+ "step": 865
1047
+ },
1048
+ {
1049
+ "epoch": 0.87,
1050
+ "learning_rate": 0.0018588092290377038,
1051
+ "loss": 1.3367,
1052
+ "step": 870
1053
+ },
1054
+ {
1055
+ "epoch": 0.87,
1056
+ "learning_rate": 0.0018508519977490148,
1057
+ "loss": 1.3763,
1058
+ "step": 875
1059
+ },
1060
+ {
1061
+ "epoch": 0.88,
1062
+ "learning_rate": 0.0018428947664603263,
1063
+ "loss": 1.4019,
1064
+ "step": 880
1065
+ },
1066
+ {
1067
+ "epoch": 0.88,
1068
+ "learning_rate": 0.0018349375351716373,
1069
+ "loss": 1.1833,
1070
+ "step": 885
1071
+ },
1072
+ {
1073
+ "epoch": 0.89,
1074
+ "learning_rate": 0.0018269803038829487,
1075
+ "loss": 1.3296,
1076
+ "step": 890
1077
+ },
1078
+ {
1079
+ "epoch": 0.89,
1080
+ "learning_rate": 0.00181902307259426,
1081
+ "loss": 1.3258,
1082
+ "step": 895
1083
+ },
1084
+ {
1085
+ "epoch": 0.9,
1086
+ "learning_rate": 0.001811065841305571,
1087
+ "loss": 1.3643,
1088
+ "step": 900
1089
+ },
1090
+ {
1091
+ "epoch": 0.9,
1092
+ "learning_rate": 0.0018031086100168824,
1093
+ "loss": 1.3896,
1094
+ "step": 905
1095
+ },
1096
+ {
1097
+ "epoch": 0.91,
1098
+ "learning_rate": 0.0017951513787281934,
1099
+ "loss": 1.3775,
1100
+ "step": 910
1101
+ },
1102
+ {
1103
+ "epoch": 0.91,
1104
+ "learning_rate": 0.0017871941474395048,
1105
+ "loss": 1.3878,
1106
+ "step": 915
1107
+ },
1108
+ {
1109
+ "epoch": 0.92,
1110
+ "learning_rate": 0.0017792369161508158,
1111
+ "loss": 1.339,
1112
+ "step": 920
1113
+ },
1114
+ {
1115
+ "epoch": 0.92,
1116
+ "learning_rate": 0.0017712796848621268,
1117
+ "loss": 1.3225,
1118
+ "step": 925
1119
+ },
1120
+ {
1121
+ "epoch": 0.93,
1122
+ "learning_rate": 0.0017633224535734382,
1123
+ "loss": 1.2899,
1124
+ "step": 930
1125
+ },
1126
+ {
1127
+ "epoch": 0.93,
1128
+ "learning_rate": 0.0017553652222847495,
1129
+ "loss": 1.3449,
1130
+ "step": 935
1131
+ },
1132
+ {
1133
+ "epoch": 0.94,
1134
+ "learning_rate": 0.0017474079909960607,
1135
+ "loss": 1.3988,
1136
+ "step": 940
1137
+ },
1138
+ {
1139
+ "epoch": 0.94,
1140
+ "learning_rate": 0.001739450759707372,
1141
+ "loss": 1.2508,
1142
+ "step": 945
1143
+ },
1144
+ {
1145
+ "epoch": 0.95,
1146
+ "learning_rate": 0.001731493528418683,
1147
+ "loss": 1.2946,
1148
+ "step": 950
1149
+ },
1150
+ {
1151
+ "epoch": 0.95,
1152
+ "learning_rate": 0.0017235362971299943,
1153
+ "loss": 1.2393,
1154
+ "step": 955
1155
+ },
1156
+ {
1157
+ "epoch": 0.96,
1158
+ "learning_rate": 0.0017155790658413053,
1159
+ "loss": 1.3437,
1160
+ "step": 960
1161
+ },
1162
+ {
1163
+ "epoch": 0.96,
1164
+ "learning_rate": 0.0017076218345526168,
1165
+ "loss": 1.2589,
1166
+ "step": 965
1167
+ },
1168
+ {
1169
+ "epoch": 0.97,
1170
+ "learning_rate": 0.0016996646032639278,
1171
+ "loss": 1.3129,
1172
+ "step": 970
1173
+ },
1174
+ {
1175
+ "epoch": 0.97,
1176
+ "learning_rate": 0.001691707371975239,
1177
+ "loss": 1.3064,
1178
+ "step": 975
1179
+ },
1180
+ {
1181
+ "epoch": 0.98,
1182
+ "learning_rate": 0.0016837501406865502,
1183
+ "loss": 1.3121,
1184
+ "step": 980
1185
+ },
1186
+ {
1187
+ "epoch": 0.98,
1188
+ "learning_rate": 0.0016757929093978614,
1189
+ "loss": 1.316,
1190
+ "step": 985
1191
+ },
1192
+ {
1193
+ "epoch": 0.99,
1194
+ "learning_rate": 0.0016678356781091727,
1195
+ "loss": 1.3004,
1196
+ "step": 990
1197
+ },
1198
+ {
1199
+ "epoch": 0.99,
1200
+ "learning_rate": 0.0016598784468204839,
1201
+ "loss": 1.3423,
1202
+ "step": 995
1203
+ },
1204
+ {
1205
+ "epoch": 1.0,
1206
+ "learning_rate": 0.0016519212155317949,
1207
+ "loss": 1.2852,
1208
+ "step": 1000
1209
+ },
1210
+ {
1211
+ "epoch": 1.0,
1212
+ "learning_rate": 0.0016439639842431063,
1213
+ "loss": 1.3119,
1214
+ "step": 1005
1215
+ },
1216
+ {
1217
+ "epoch": 1.01,
1218
+ "learning_rate": 0.0016360067529544173,
1219
+ "loss": 1.2672,
1220
+ "step": 1010
1221
+ },
1222
+ {
1223
+ "epoch": 1.01,
1224
+ "learning_rate": 0.0016280495216657287,
1225
+ "loss": 1.2331,
1226
+ "step": 1015
1227
+ },
1228
+ {
1229
+ "epoch": 1.02,
1230
+ "learning_rate": 0.0016200922903770397,
1231
+ "loss": 1.2987,
1232
+ "step": 1020
1233
+ },
1234
+ {
1235
+ "epoch": 1.02,
1236
+ "learning_rate": 0.001612135059088351,
1237
+ "loss": 1.3455,
1238
+ "step": 1025
1239
+ },
1240
+ {
1241
+ "epoch": 1.03,
1242
+ "learning_rate": 0.0016041778277996622,
1243
+ "loss": 1.2899,
1244
+ "step": 1030
1245
+ },
1246
+ {
1247
+ "epoch": 1.03,
1248
+ "learning_rate": 0.0015962205965109734,
1249
+ "loss": 1.2729,
1250
+ "step": 1035
1251
+ },
1252
+ {
1253
+ "epoch": 1.04,
1254
+ "learning_rate": 0.0015882633652222848,
1255
+ "loss": 1.2688,
1256
+ "step": 1040
1257
+ },
1258
+ {
1259
+ "epoch": 1.04,
1260
+ "learning_rate": 0.0015803061339335958,
1261
+ "loss": 1.2367,
1262
+ "step": 1045
1263
+ },
1264
+ {
1265
+ "epoch": 1.05,
1266
+ "learning_rate": 0.0015723489026449068,
1267
+ "loss": 1.302,
1268
+ "step": 1050
1269
+ },
1270
+ {
1271
+ "epoch": 1.05,
1272
+ "learning_rate": 0.0015643916713562183,
1273
+ "loss": 1.3786,
1274
+ "step": 1055
1275
+ },
1276
+ {
1277
+ "epoch": 1.06,
1278
+ "learning_rate": 0.0015564344400675293,
1279
+ "loss": 1.2653,
1280
+ "step": 1060
1281
+ },
1282
+ {
1283
+ "epoch": 1.06,
1284
+ "learning_rate": 0.0015484772087788407,
1285
+ "loss": 1.2689,
1286
+ "step": 1065
1287
+ },
1288
+ {
1289
+ "epoch": 1.07,
1290
+ "learning_rate": 0.0015405199774901517,
1291
+ "loss": 1.2673,
1292
+ "step": 1070
1293
+ },
1294
+ {
1295
+ "epoch": 1.07,
1296
+ "learning_rate": 0.001532562746201463,
1297
+ "loss": 1.261,
1298
+ "step": 1075
1299
+ },
1300
+ {
1301
+ "epoch": 1.08,
1302
+ "learning_rate": 0.0015246055149127744,
1303
+ "loss": 1.1957,
1304
+ "step": 1080
1305
+ },
1306
+ {
1307
+ "epoch": 1.08,
1308
+ "learning_rate": 0.0015166482836240854,
1309
+ "loss": 1.4158,
1310
+ "step": 1085
1311
+ },
1312
+ {
1313
+ "epoch": 1.09,
1314
+ "learning_rate": 0.0015086910523353968,
1315
+ "loss": 1.1961,
1316
+ "step": 1090
1317
+ },
1318
+ {
1319
+ "epoch": 1.09,
1320
+ "learning_rate": 0.0015007338210467078,
1321
+ "loss": 1.3019,
1322
+ "step": 1095
1323
+ },
1324
+ {
1325
+ "epoch": 1.1,
1326
+ "learning_rate": 0.0014927765897580188,
1327
+ "loss": 1.2551,
1328
+ "step": 1100
1329
+ },
1330
+ {
1331
+ "epoch": 1.1,
1332
+ "learning_rate": 0.0014848193584693302,
1333
+ "loss": 1.2763,
1334
+ "step": 1105
1335
+ },
1336
+ {
1337
+ "epoch": 1.11,
1338
+ "learning_rate": 0.0014768621271806412,
1339
+ "loss": 1.2999,
1340
+ "step": 1110
1341
+ },
1342
+ {
1343
+ "epoch": 1.11,
1344
+ "learning_rate": 0.0014689048958919527,
1345
+ "loss": 1.2929,
1346
+ "step": 1115
1347
+ },
1348
+ {
1349
+ "epoch": 1.12,
1350
+ "learning_rate": 0.001460947664603264,
1351
+ "loss": 1.2321,
1352
+ "step": 1120
1353
+ },
1354
+ {
1355
+ "epoch": 1.12,
1356
+ "learning_rate": 0.001452990433314575,
1357
+ "loss": 1.3635,
1358
+ "step": 1125
1359
+ },
1360
+ {
1361
+ "epoch": 1.13,
1362
+ "learning_rate": 0.0014450332020258863,
1363
+ "loss": 1.2883,
1364
+ "step": 1130
1365
+ },
1366
+ {
1367
+ "epoch": 1.13,
1368
+ "learning_rate": 0.0014370759707371973,
1369
+ "loss": 1.3663,
1370
+ "step": 1135
1371
+ },
1372
+ {
1373
+ "epoch": 1.14,
1374
+ "learning_rate": 0.0014291187394485088,
1375
+ "loss": 1.268,
1376
+ "step": 1140
1377
+ },
1378
+ {
1379
+ "epoch": 1.14,
1380
+ "learning_rate": 0.0014211615081598198,
1381
+ "loss": 1.2976,
1382
+ "step": 1145
1383
+ },
1384
+ {
1385
+ "epoch": 1.15,
1386
+ "learning_rate": 0.001413204276871131,
1387
+ "loss": 1.3624,
1388
+ "step": 1150
1389
+ },
1390
+ {
1391
+ "epoch": 1.15,
1392
+ "learning_rate": 0.0014052470455824422,
1393
+ "loss": 1.3659,
1394
+ "step": 1155
1395
+ },
1396
+ {
1397
+ "epoch": 1.16,
1398
+ "learning_rate": 0.0013972898142937534,
1399
+ "loss": 1.2711,
1400
+ "step": 1160
1401
+ },
1402
+ {
1403
+ "epoch": 1.16,
1404
+ "learning_rate": 0.0013893325830050646,
1405
+ "loss": 1.2882,
1406
+ "step": 1165
1407
+ },
1408
+ {
1409
+ "epoch": 1.17,
1410
+ "learning_rate": 0.0013813753517163759,
1411
+ "loss": 1.321,
1412
+ "step": 1170
1413
+ },
1414
+ {
1415
+ "epoch": 1.17,
1416
+ "learning_rate": 0.001373418120427687,
1417
+ "loss": 1.3505,
1418
+ "step": 1175
1419
+ },
1420
+ {
1421
+ "epoch": 1.18,
1422
+ "learning_rate": 0.0013654608891389983,
1423
+ "loss": 1.2967,
1424
+ "step": 1180
1425
+ },
1426
+ {
1427
+ "epoch": 1.18,
1428
+ "learning_rate": 0.0013575036578503093,
1429
+ "loss": 1.2868,
1430
+ "step": 1185
1431
+ },
1432
+ {
1433
+ "epoch": 1.19,
1434
+ "learning_rate": 0.0013495464265616205,
1435
+ "loss": 1.2786,
1436
+ "step": 1190
1437
+ },
1438
+ {
1439
+ "epoch": 1.19,
1440
+ "learning_rate": 0.0013415891952729317,
1441
+ "loss": 1.2839,
1442
+ "step": 1195
1443
+ },
1444
+ {
1445
+ "epoch": 1.2,
1446
+ "learning_rate": 0.001333631963984243,
1447
+ "loss": 1.309,
1448
+ "step": 1200
1449
+ },
1450
+ {
1451
+ "epoch": 1.2,
1452
+ "learning_rate": 0.0013256747326955542,
1453
+ "loss": 1.3182,
1454
+ "step": 1205
1455
+ },
1456
+ {
1457
+ "epoch": 1.21,
1458
+ "learning_rate": 0.0013177175014068654,
1459
+ "loss": 1.3086,
1460
+ "step": 1210
1461
+ },
1462
+ {
1463
+ "epoch": 1.21,
1464
+ "learning_rate": 0.0013097602701181766,
1465
+ "loss": 1.189,
1466
+ "step": 1215
1467
+ },
1468
+ {
1469
+ "epoch": 1.22,
1470
+ "learning_rate": 0.0013018030388294878,
1471
+ "loss": 1.3253,
1472
+ "step": 1220
1473
+ },
1474
+ {
1475
+ "epoch": 1.22,
1476
+ "learning_rate": 0.001293845807540799,
1477
+ "loss": 1.2408,
1478
+ "step": 1225
1479
+ },
1480
+ {
1481
+ "epoch": 1.23,
1482
+ "learning_rate": 0.0012858885762521103,
1483
+ "loss": 1.3168,
1484
+ "step": 1230
1485
+ },
1486
+ {
1487
+ "epoch": 1.23,
1488
+ "learning_rate": 0.0012779313449634215,
1489
+ "loss": 1.2883,
1490
+ "step": 1235
1491
+ },
1492
+ {
1493
+ "epoch": 1.24,
1494
+ "learning_rate": 0.0012699741136747325,
1495
+ "loss": 1.3386,
1496
+ "step": 1240
1497
+ },
1498
+ {
1499
+ "epoch": 1.24,
1500
+ "learning_rate": 0.0012620168823860437,
1501
+ "loss": 1.4053,
1502
+ "step": 1245
1503
+ },
1504
+ {
1505
+ "epoch": 1.25,
1506
+ "learning_rate": 0.001254059651097355,
1507
+ "loss": 1.2768,
1508
+ "step": 1250
1509
+ },
1510
+ {
1511
+ "epoch": 1.25,
1512
+ "learning_rate": 0.0012461024198086661,
1513
+ "loss": 1.2946,
1514
+ "step": 1255
1515
+ },
1516
+ {
1517
+ "epoch": 1.26,
1518
+ "learning_rate": 0.0012381451885199776,
1519
+ "loss": 1.1944,
1520
+ "step": 1260
1521
+ },
1522
+ {
1523
+ "epoch": 1.26,
1524
+ "learning_rate": 0.0012301879572312886,
1525
+ "loss": 1.1929,
1526
+ "step": 1265
1527
+ },
1528
+ {
1529
+ "epoch": 1.27,
1530
+ "learning_rate": 0.0012222307259425998,
1531
+ "loss": 1.3793,
1532
+ "step": 1270
1533
+ },
1534
+ {
1535
+ "epoch": 1.27,
1536
+ "learning_rate": 0.001214273494653911,
1537
+ "loss": 1.2986,
1538
+ "step": 1275
1539
+ },
1540
+ {
1541
+ "epoch": 1.28,
1542
+ "learning_rate": 0.0012063162633652222,
1543
+ "loss": 1.3376,
1544
+ "step": 1280
1545
+ },
1546
+ {
1547
+ "epoch": 1.28,
1548
+ "learning_rate": 0.0011983590320765335,
1549
+ "loss": 1.2406,
1550
+ "step": 1285
1551
+ },
1552
+ {
1553
+ "epoch": 1.29,
1554
+ "learning_rate": 0.0011904018007878445,
1555
+ "loss": 1.3542,
1556
+ "step": 1290
1557
+ },
1558
+ {
1559
+ "epoch": 1.29,
1560
+ "learning_rate": 0.0011824445694991557,
1561
+ "loss": 1.3042,
1562
+ "step": 1295
1563
+ },
1564
+ {
1565
+ "epoch": 1.3,
1566
+ "learning_rate": 0.001174487338210467,
1567
+ "loss": 1.3189,
1568
+ "step": 1300
1569
+ },
1570
+ {
1571
+ "epoch": 1.3,
1572
+ "learning_rate": 0.0011665301069217783,
1573
+ "loss": 1.2538,
1574
+ "step": 1305
1575
+ },
1576
+ {
1577
+ "epoch": 1.31,
1578
+ "learning_rate": 0.0011585728756330895,
1579
+ "loss": 1.3152,
1580
+ "step": 1310
1581
+ },
1582
+ {
1583
+ "epoch": 1.31,
1584
+ "learning_rate": 0.0011506156443444005,
1585
+ "loss": 1.2528,
1586
+ "step": 1315
1587
+ },
1588
+ {
1589
+ "epoch": 1.32,
1590
+ "learning_rate": 0.0011426584130557118,
1591
+ "loss": 1.2752,
1592
+ "step": 1320
1593
+ },
1594
+ {
1595
+ "epoch": 1.32,
1596
+ "learning_rate": 0.001134701181767023,
1597
+ "loss": 1.3363,
1598
+ "step": 1325
1599
+ },
1600
+ {
1601
+ "epoch": 1.33,
1602
+ "learning_rate": 0.0011267439504783342,
1603
+ "loss": 1.2964,
1604
+ "step": 1330
1605
+ },
1606
+ {
1607
+ "epoch": 1.33,
1608
+ "learning_rate": 0.0011187867191896454,
1609
+ "loss": 1.2031,
1610
+ "step": 1335
1611
+ },
1612
+ {
1613
+ "epoch": 1.34,
1614
+ "learning_rate": 0.0011108294879009564,
1615
+ "loss": 1.3499,
1616
+ "step": 1340
1617
+ },
1618
+ {
1619
+ "epoch": 1.34,
1620
+ "learning_rate": 0.0011028722566122679,
1621
+ "loss": 1.2842,
1622
+ "step": 1345
1623
+ },
1624
+ {
1625
+ "epoch": 1.35,
1626
+ "learning_rate": 0.001094915025323579,
1627
+ "loss": 1.2108,
1628
+ "step": 1350
1629
+ },
1630
+ {
1631
+ "epoch": 1.35,
1632
+ "learning_rate": 0.0010869577940348903,
1633
+ "loss": 1.3822,
1634
+ "step": 1355
1635
+ },
1636
+ {
1637
+ "epoch": 1.36,
1638
+ "learning_rate": 0.0010790005627462015,
1639
+ "loss": 1.303,
1640
+ "step": 1360
1641
+ },
1642
+ {
1643
+ "epoch": 1.36,
1644
+ "learning_rate": 0.0010710433314575125,
1645
+ "loss": 1.25,
1646
+ "step": 1365
1647
+ },
1648
+ {
1649
+ "epoch": 1.37,
1650
+ "learning_rate": 0.0010630861001688237,
1651
+ "loss": 1.2506,
1652
+ "step": 1370
1653
+ },
1654
+ {
1655
+ "epoch": 1.37,
1656
+ "learning_rate": 0.001055128868880135,
1657
+ "loss": 1.3144,
1658
+ "step": 1375
1659
+ },
1660
+ {
1661
+ "epoch": 1.38,
1662
+ "learning_rate": 0.0010471716375914462,
1663
+ "loss": 1.3205,
1664
+ "step": 1380
1665
+ },
1666
+ {
1667
+ "epoch": 1.38,
1668
+ "learning_rate": 0.0010392144063027574,
1669
+ "loss": 1.1656,
1670
+ "step": 1385
1671
+ },
1672
+ {
1673
+ "epoch": 1.39,
1674
+ "learning_rate": 0.0010312571750140686,
1675
+ "loss": 1.2727,
1676
+ "step": 1390
1677
+ },
1678
+ {
1679
+ "epoch": 1.39,
1680
+ "learning_rate": 0.0010232999437253798,
1681
+ "loss": 1.2887,
1682
+ "step": 1395
1683
+ },
1684
+ {
1685
+ "epoch": 1.4,
1686
+ "learning_rate": 0.001015342712436691,
1687
+ "loss": 1.2759,
1688
+ "step": 1400
1689
+ },
1690
+ {
1691
+ "epoch": 1.4,
1692
+ "learning_rate": 0.0010073854811480023,
1693
+ "loss": 1.2698,
1694
+ "step": 1405
1695
+ },
1696
+ {
1697
+ "epoch": 1.41,
1698
+ "learning_rate": 0.0009994282498593135,
1699
+ "loss": 1.3167,
1700
+ "step": 1410
1701
+ },
1702
+ {
1703
+ "epoch": 1.41,
1704
+ "learning_rate": 0.0009914710185706245,
1705
+ "loss": 1.2659,
1706
+ "step": 1415
1707
+ },
1708
+ {
1709
+ "epoch": 1.42,
1710
+ "learning_rate": 0.0009835137872819357,
1711
+ "loss": 1.2209,
1712
+ "step": 1420
1713
+ },
1714
+ {
1715
+ "epoch": 1.42,
1716
+ "learning_rate": 0.000975556555993247,
1717
+ "loss": 1.2778,
1718
+ "step": 1425
1719
+ },
1720
+ {
1721
+ "epoch": 1.43,
1722
+ "learning_rate": 0.0009675993247045582,
1723
+ "loss": 1.2276,
1724
+ "step": 1430
1725
+ },
1726
+ {
1727
+ "epoch": 1.43,
1728
+ "learning_rate": 0.0009596420934158695,
1729
+ "loss": 1.2704,
1730
+ "step": 1435
1731
+ },
1732
+ {
1733
+ "epoch": 1.44,
1734
+ "learning_rate": 0.0009516848621271805,
1735
+ "loss": 1.2859,
1736
+ "step": 1440
1737
+ },
1738
+ {
1739
+ "epoch": 1.44,
1740
+ "learning_rate": 0.0009437276308384918,
1741
+ "loss": 1.3544,
1742
+ "step": 1445
1743
+ },
1744
+ {
1745
+ "epoch": 1.45,
1746
+ "learning_rate": 0.000935770399549803,
1747
+ "loss": 1.2112,
1748
+ "step": 1450
1749
+ },
1750
+ {
1751
+ "epoch": 1.45,
1752
+ "learning_rate": 0.0009278131682611142,
1753
+ "loss": 1.2632,
1754
+ "step": 1455
1755
+ },
1756
+ {
1757
+ "epoch": 1.46,
1758
+ "learning_rate": 0.0009198559369724254,
1759
+ "loss": 1.2354,
1760
+ "step": 1460
1761
+ },
1762
+ {
1763
+ "epoch": 1.46,
1764
+ "learning_rate": 0.0009118987056837366,
1765
+ "loss": 1.2159,
1766
+ "step": 1465
1767
+ },
1768
+ {
1769
+ "epoch": 1.47,
1770
+ "learning_rate": 0.0009039414743950478,
1771
+ "loss": 1.2772,
1772
+ "step": 1470
1773
+ },
1774
+ {
1775
+ "epoch": 1.47,
1776
+ "learning_rate": 0.000895984243106359,
1777
+ "loss": 1.2689,
1778
+ "step": 1475
1779
+ },
1780
+ {
1781
+ "epoch": 1.48,
1782
+ "learning_rate": 0.0008880270118176702,
1783
+ "loss": 1.2522,
1784
+ "step": 1480
1785
+ },
1786
+ {
1787
+ "epoch": 1.48,
1788
+ "learning_rate": 0.0008800697805289814,
1789
+ "loss": 1.2642,
1790
+ "step": 1485
1791
+ },
1792
+ {
1793
+ "epoch": 1.49,
1794
+ "learning_rate": 0.0008721125492402925,
1795
+ "loss": 1.3123,
1796
+ "step": 1490
1797
+ },
1798
+ {
1799
+ "epoch": 1.49,
1800
+ "learning_rate": 0.0008641553179516038,
1801
+ "loss": 1.2526,
1802
+ "step": 1495
1803
+ },
1804
+ {
1805
+ "epoch": 1.5,
1806
+ "learning_rate": 0.000856198086662915,
1807
+ "loss": 1.3556,
1808
+ "step": 1500
1809
+ },
1810
+ {
1811
+ "epoch": 1.5,
1812
+ "learning_rate": 0.0008482408553742262,
1813
+ "loss": 1.1951,
1814
+ "step": 1505
1815
+ },
1816
+ {
1817
+ "epoch": 1.51,
1818
+ "learning_rate": 0.0008402836240855374,
1819
+ "loss": 1.3322,
1820
+ "step": 1510
1821
+ },
1822
+ {
1823
+ "epoch": 1.51,
1824
+ "learning_rate": 0.0008323263927968485,
1825
+ "loss": 1.2582,
1826
+ "step": 1515
1827
+ },
1828
+ {
1829
+ "epoch": 1.52,
1830
+ "learning_rate": 0.0008243691615081597,
1831
+ "loss": 1.3049,
1832
+ "step": 1520
1833
+ },
1834
+ {
1835
+ "epoch": 1.52,
1836
+ "learning_rate": 0.000816411930219471,
1837
+ "loss": 1.1639,
1838
+ "step": 1525
1839
+ },
1840
+ {
1841
+ "epoch": 1.53,
1842
+ "learning_rate": 0.0008084546989307822,
1843
+ "loss": 1.2385,
1844
+ "step": 1530
1845
+ },
1846
+ {
1847
+ "epoch": 1.53,
1848
+ "learning_rate": 0.0008004974676420934,
1849
+ "loss": 1.268,
1850
+ "step": 1535
1851
+ },
1852
+ {
1853
+ "epoch": 1.54,
1854
+ "learning_rate": 0.0007925402363534045,
1855
+ "loss": 1.2466,
1856
+ "step": 1540
1857
+ },
1858
+ {
1859
+ "epoch": 1.54,
1860
+ "learning_rate": 0.0007845830050647157,
1861
+ "loss": 1.2585,
1862
+ "step": 1545
1863
+ },
1864
+ {
1865
+ "epoch": 1.55,
1866
+ "learning_rate": 0.000776625773776027,
1867
+ "loss": 1.2781,
1868
+ "step": 1550
1869
+ },
1870
+ {
1871
+ "epoch": 1.55,
1872
+ "learning_rate": 0.0007686685424873382,
1873
+ "loss": 1.19,
1874
+ "step": 1555
1875
+ },
1876
+ {
1877
+ "epoch": 1.56,
1878
+ "learning_rate": 0.0007607113111986494,
1879
+ "loss": 1.2718,
1880
+ "step": 1560
1881
+ },
1882
+ {
1883
+ "epoch": 1.56,
1884
+ "learning_rate": 0.0007527540799099605,
1885
+ "loss": 1.2066,
1886
+ "step": 1565
1887
+ },
1888
+ {
1889
+ "epoch": 1.57,
1890
+ "learning_rate": 0.0007447968486212717,
1891
+ "loss": 1.2605,
1892
+ "step": 1570
1893
+ },
1894
+ {
1895
+ "epoch": 1.57,
1896
+ "learning_rate": 0.0007368396173325829,
1897
+ "loss": 1.2403,
1898
+ "step": 1575
1899
+ },
1900
+ {
1901
+ "epoch": 1.58,
1902
+ "learning_rate": 0.0007288823860438941,
1903
+ "loss": 1.3096,
1904
+ "step": 1580
1905
+ },
1906
+ {
1907
+ "epoch": 1.58,
1908
+ "learning_rate": 0.0007209251547552054,
1909
+ "loss": 1.152,
1910
+ "step": 1585
1911
+ },
1912
+ {
1913
+ "epoch": 1.59,
1914
+ "learning_rate": 0.0007129679234665165,
1915
+ "loss": 1.282,
1916
+ "step": 1590
1917
+ },
1918
+ {
1919
+ "epoch": 1.59,
1920
+ "learning_rate": 0.0007050106921778277,
1921
+ "loss": 1.2004,
1922
+ "step": 1595
1923
+ },
1924
+ {
1925
+ "epoch": 1.6,
1926
+ "learning_rate": 0.0006970534608891389,
1927
+ "loss": 1.2049,
1928
+ "step": 1600
1929
+ },
1930
+ {
1931
+ "epoch": 1.6,
1932
+ "learning_rate": 0.0006890962296004501,
1933
+ "loss": 1.1815,
1934
+ "step": 1605
1935
+ },
1936
+ {
1937
+ "epoch": 1.61,
1938
+ "learning_rate": 0.0006811389983117614,
1939
+ "loss": 1.1898,
1940
+ "step": 1610
1941
+ },
1942
+ {
1943
+ "epoch": 1.61,
1944
+ "learning_rate": 0.0006731817670230726,
1945
+ "loss": 1.244,
1946
+ "step": 1615
1947
+ },
1948
+ {
1949
+ "epoch": 1.62,
1950
+ "learning_rate": 0.0006652245357343837,
1951
+ "loss": 1.2229,
1952
+ "step": 1620
1953
+ },
1954
+ {
1955
+ "epoch": 1.62,
1956
+ "learning_rate": 0.0006572673044456949,
1957
+ "loss": 1.2084,
1958
+ "step": 1625
1959
+ },
1960
+ {
1961
+ "epoch": 1.63,
1962
+ "learning_rate": 0.0006493100731570061,
1963
+ "loss": 1.3129,
1964
+ "step": 1630
1965
+ },
1966
+ {
1967
+ "epoch": 1.63,
1968
+ "learning_rate": 0.0006413528418683173,
1969
+ "loss": 1.3901,
1970
+ "step": 1635
1971
+ },
1972
+ {
1973
+ "epoch": 1.64,
1974
+ "learning_rate": 0.0006333956105796286,
1975
+ "loss": 1.2896,
1976
+ "step": 1640
1977
+ },
1978
+ {
1979
+ "epoch": 1.64,
1980
+ "learning_rate": 0.0006254383792909397,
1981
+ "loss": 1.2866,
1982
+ "step": 1645
1983
+ },
1984
+ {
1985
+ "epoch": 1.65,
1986
+ "learning_rate": 0.0006174811480022509,
1987
+ "loss": 1.3251,
1988
+ "step": 1650
1989
+ },
1990
+ {
1991
+ "epoch": 1.65,
1992
+ "learning_rate": 0.0006095239167135622,
1993
+ "loss": 1.1864,
1994
+ "step": 1655
1995
+ },
1996
+ {
1997
+ "epoch": 1.66,
1998
+ "learning_rate": 0.0006015666854248733,
1999
+ "loss": 1.2127,
2000
+ "step": 1660
2001
+ },
2002
+ {
2003
+ "epoch": 1.66,
2004
+ "learning_rate": 0.0005936094541361845,
2005
+ "loss": 1.2944,
2006
+ "step": 1665
2007
+ },
2008
+ {
2009
+ "epoch": 1.67,
2010
+ "learning_rate": 0.0005856522228474956,
2011
+ "loss": 1.3106,
2012
+ "step": 1670
2013
+ },
2014
+ {
2015
+ "epoch": 1.67,
2016
+ "learning_rate": 0.000577694991558807,
2017
+ "loss": 1.3614,
2018
+ "step": 1675
2019
+ },
2020
+ {
2021
+ "epoch": 1.68,
2022
+ "learning_rate": 0.0005697377602701182,
2023
+ "loss": 1.2786,
2024
+ "step": 1680
2025
+ },
2026
+ {
2027
+ "epoch": 1.68,
2028
+ "learning_rate": 0.0005617805289814293,
2029
+ "loss": 1.2758,
2030
+ "step": 1685
2031
+ },
2032
+ {
2033
+ "epoch": 1.69,
2034
+ "learning_rate": 0.0005538232976927405,
2035
+ "loss": 1.31,
2036
+ "step": 1690
2037
+ },
2038
+ {
2039
+ "epoch": 1.69,
2040
+ "learning_rate": 0.0005458660664040517,
2041
+ "loss": 1.2246,
2042
+ "step": 1695
2043
+ },
2044
+ {
2045
+ "epoch": 1.7,
2046
+ "learning_rate": 0.000537908835115363,
2047
+ "loss": 1.2295,
2048
+ "step": 1700
2049
+ },
2050
+ {
2051
+ "epoch": 1.7,
2052
+ "learning_rate": 0.0005299516038266742,
2053
+ "loss": 1.2822,
2054
+ "step": 1705
2055
+ },
2056
+ {
2057
+ "epoch": 1.71,
2058
+ "learning_rate": 0.0005219943725379853,
2059
+ "loss": 1.2582,
2060
+ "step": 1710
2061
+ },
2062
+ {
2063
+ "epoch": 1.71,
2064
+ "learning_rate": 0.0005140371412492965,
2065
+ "loss": 1.1898,
2066
+ "step": 1715
2067
+ },
2068
+ {
2069
+ "epoch": 1.72,
2070
+ "learning_rate": 0.0005060799099606077,
2071
+ "loss": 1.2697,
2072
+ "step": 1720
2073
+ },
2074
+ {
2075
+ "epoch": 1.72,
2076
+ "learning_rate": 0.0004981226786719189,
2077
+ "loss": 1.2696,
2078
+ "step": 1725
2079
+ },
2080
+ {
2081
+ "epoch": 1.73,
2082
+ "learning_rate": 0.0004901654473832302,
2083
+ "loss": 1.2383,
2084
+ "step": 1730
2085
+ },
2086
+ {
2087
+ "epoch": 1.73,
2088
+ "learning_rate": 0.0004822082160945413,
2089
+ "loss": 1.2665,
2090
+ "step": 1735
2091
+ },
2092
+ {
2093
+ "epoch": 1.74,
2094
+ "learning_rate": 0.00047425098480585254,
2095
+ "loss": 1.3193,
2096
+ "step": 1740
2097
+ },
2098
+ {
2099
+ "epoch": 1.74,
2100
+ "learning_rate": 0.0004662937535171637,
2101
+ "loss": 1.2431,
2102
+ "step": 1745
2103
+ },
2104
+ {
2105
+ "epoch": 1.75,
2106
+ "learning_rate": 0.0004583365222284749,
2107
+ "loss": 1.2243,
2108
+ "step": 1750
2109
+ },
2110
+ {
2111
+ "epoch": 1.75,
2112
+ "learning_rate": 0.00045037929093978614,
2113
+ "loss": 1.2323,
2114
+ "step": 1755
2115
+ },
2116
+ {
2117
+ "epoch": 1.76,
2118
+ "learning_rate": 0.0004424220596510973,
2119
+ "loss": 1.3318,
2120
+ "step": 1760
2121
+ },
2122
+ {
2123
+ "epoch": 1.76,
2124
+ "learning_rate": 0.0004344648283624085,
2125
+ "loss": 1.2339,
2126
+ "step": 1765
2127
+ },
2128
+ {
2129
+ "epoch": 1.77,
2130
+ "learning_rate": 0.0004265075970737197,
2131
+ "loss": 1.3316,
2132
+ "step": 1770
2133
+ },
2134
+ {
2135
+ "epoch": 1.77,
2136
+ "learning_rate": 0.0004185503657850309,
2137
+ "loss": 1.2981,
2138
+ "step": 1775
2139
+ },
2140
+ {
2141
+ "epoch": 1.78,
2142
+ "learning_rate": 0.00041059313449634213,
2143
+ "loss": 1.1703,
2144
+ "step": 1780
2145
+ },
2146
+ {
2147
+ "epoch": 1.78,
2148
+ "learning_rate": 0.0004026359032076533,
2149
+ "loss": 1.2283,
2150
+ "step": 1785
2151
+ },
2152
+ {
2153
+ "epoch": 1.79,
2154
+ "learning_rate": 0.0003946786719189645,
2155
+ "loss": 1.2554,
2156
+ "step": 1790
2157
+ },
2158
+ {
2159
+ "epoch": 1.79,
2160
+ "learning_rate": 0.00038672144063027573,
2161
+ "loss": 1.3264,
2162
+ "step": 1795
2163
+ },
2164
+ {
2165
+ "epoch": 1.8,
2166
+ "learning_rate": 0.0003787642093415869,
2167
+ "loss": 1.2297,
2168
+ "step": 1800
2169
+ },
2170
+ {
2171
+ "epoch": 1.8,
2172
+ "learning_rate": 0.0003708069780528981,
2173
+ "loss": 1.3238,
2174
+ "step": 1805
2175
+ },
2176
+ {
2177
+ "epoch": 1.81,
2178
+ "learning_rate": 0.0003628497467642093,
2179
+ "loss": 1.2793,
2180
+ "step": 1810
2181
+ },
2182
+ {
2183
+ "epoch": 1.81,
2184
+ "learning_rate": 0.0003548925154755205,
2185
+ "loss": 1.2891,
2186
+ "step": 1815
2187
+ },
2188
+ {
2189
+ "epoch": 1.82,
2190
+ "learning_rate": 0.0003469352841868317,
2191
+ "loss": 1.2713,
2192
+ "step": 1820
2193
+ },
2194
+ {
2195
+ "epoch": 1.82,
2196
+ "learning_rate": 0.0003389780528981429,
2197
+ "loss": 1.218,
2198
+ "step": 1825
2199
+ },
2200
+ {
2201
+ "epoch": 1.83,
2202
+ "learning_rate": 0.0003310208216094541,
2203
+ "loss": 1.29,
2204
+ "step": 1830
2205
+ },
2206
+ {
2207
+ "epoch": 1.83,
2208
+ "learning_rate": 0.0003230635903207653,
2209
+ "loss": 1.2781,
2210
+ "step": 1835
2211
+ },
2212
+ {
2213
+ "epoch": 1.84,
2214
+ "learning_rate": 0.00031510635903207653,
2215
+ "loss": 1.2553,
2216
+ "step": 1840
2217
+ },
2218
+ {
2219
+ "epoch": 1.84,
2220
+ "learning_rate": 0.0003071491277433877,
2221
+ "loss": 1.3065,
2222
+ "step": 1845
2223
+ },
2224
+ {
2225
+ "epoch": 1.85,
2226
+ "learning_rate": 0.0002991918964546989,
2227
+ "loss": 1.2249,
2228
+ "step": 1850
2229
+ },
2230
+ {
2231
+ "epoch": 1.85,
2232
+ "learning_rate": 0.0002912346651660101,
2233
+ "loss": 1.1943,
2234
+ "step": 1855
2235
+ },
2236
+ {
2237
+ "epoch": 1.86,
2238
+ "learning_rate": 0.0002832774338773213,
2239
+ "loss": 1.3517,
2240
+ "step": 1860
2241
+ },
2242
+ {
2243
+ "epoch": 1.86,
2244
+ "learning_rate": 0.0002753202025886325,
2245
+ "loss": 1.2667,
2246
+ "step": 1865
2247
+ },
2248
+ {
2249
+ "epoch": 1.87,
2250
+ "learning_rate": 0.0002673629712999437,
2251
+ "loss": 1.2597,
2252
+ "step": 1870
2253
+ },
2254
+ {
2255
+ "epoch": 1.87,
2256
+ "learning_rate": 0.0002594057400112549,
2257
+ "loss": 1.2336,
2258
+ "step": 1875
2259
+ },
2260
+ {
2261
+ "epoch": 1.88,
2262
+ "learning_rate": 0.0002514485087225661,
2263
+ "loss": 1.1898,
2264
+ "step": 1880
2265
+ },
2266
+ {
2267
+ "epoch": 1.88,
2268
+ "learning_rate": 0.0002434912774338773,
2269
+ "loss": 1.1498,
2270
+ "step": 1885
2271
+ },
2272
+ {
2273
+ "epoch": 1.89,
2274
+ "learning_rate": 0.0002355340461451885,
2275
+ "loss": 1.2345,
2276
+ "step": 1890
2277
+ },
2278
+ {
2279
+ "epoch": 1.89,
2280
+ "learning_rate": 0.0002275768148564997,
2281
+ "loss": 1.2426,
2282
+ "step": 1895
2283
+ },
2284
+ {
2285
+ "epoch": 1.9,
2286
+ "learning_rate": 0.00021961958356781088,
2287
+ "loss": 1.256,
2288
+ "step": 1900
2289
+ },
2290
+ {
2291
+ "epoch": 1.9,
2292
+ "learning_rate": 0.00021166235227912213,
2293
+ "loss": 1.2133,
2294
+ "step": 1905
2295
+ },
2296
+ {
2297
+ "epoch": 1.91,
2298
+ "learning_rate": 0.00020370512099043332,
2299
+ "loss": 1.1982,
2300
+ "step": 1910
2301
+ },
2302
+ {
2303
+ "epoch": 1.91,
2304
+ "learning_rate": 0.0001957478897017445,
2305
+ "loss": 1.2487,
2306
+ "step": 1915
2307
+ },
2308
+ {
2309
+ "epoch": 1.92,
2310
+ "learning_rate": 0.0001877906584130557,
2311
+ "loss": 1.3181,
2312
+ "step": 1920
2313
+ },
2314
+ {
2315
+ "epoch": 1.92,
2316
+ "learning_rate": 0.0001798334271243669,
2317
+ "loss": 1.2821,
2318
+ "step": 1925
2319
+ },
2320
+ {
2321
+ "epoch": 1.93,
2322
+ "learning_rate": 0.00017187619583567808,
2323
+ "loss": 1.2304,
2324
+ "step": 1930
2325
+ },
2326
+ {
2327
+ "epoch": 1.93,
2328
+ "learning_rate": 0.0001639189645469893,
2329
+ "loss": 1.2659,
2330
+ "step": 1935
2331
+ },
2332
+ {
2333
+ "epoch": 1.94,
2334
+ "learning_rate": 0.0001559617332583005,
2335
+ "loss": 1.2992,
2336
+ "step": 1940
2337
+ },
2338
+ {
2339
+ "epoch": 1.94,
2340
+ "learning_rate": 0.00014800450196961168,
2341
+ "loss": 1.2244,
2342
+ "step": 1945
2343
+ },
2344
+ {
2345
+ "epoch": 1.95,
2346
+ "learning_rate": 0.0001400472706809229,
2347
+ "loss": 1.2843,
2348
+ "step": 1950
2349
+ },
2350
+ {
2351
+ "epoch": 1.95,
2352
+ "learning_rate": 0.0001320900393922341,
2353
+ "loss": 1.3562,
2354
+ "step": 1955
2355
+ },
2356
+ {
2357
+ "epoch": 1.96,
2358
+ "learning_rate": 0.0001241328081035453,
2359
+ "loss": 1.2034,
2360
+ "step": 1960
2361
+ },
2362
+ {
2363
+ "epoch": 1.96,
2364
+ "learning_rate": 0.00011617557681485649,
2365
+ "loss": 1.1652,
2366
+ "step": 1965
2367
+ },
2368
+ {
2369
+ "epoch": 1.97,
2370
+ "learning_rate": 0.00010821834552616768,
2371
+ "loss": 1.3189,
2372
+ "step": 1970
2373
+ },
2374
+ {
2375
+ "epoch": 1.97,
2376
+ "learning_rate": 0.00010026111423747889,
2377
+ "loss": 1.2199,
2378
+ "step": 1975
2379
+ },
2380
+ {
2381
+ "epoch": 1.98,
2382
+ "learning_rate": 9.230388294879008e-05,
2383
+ "loss": 1.2195,
2384
+ "step": 1980
2385
+ },
2386
+ {
2387
+ "epoch": 1.98,
2388
+ "learning_rate": 8.434665166010128e-05,
2389
+ "loss": 1.2418,
2390
+ "step": 1985
2391
+ },
2392
+ {
2393
+ "epoch": 1.99,
2394
+ "learning_rate": 7.638942037141249e-05,
2395
+ "loss": 1.2555,
2396
+ "step": 1990
2397
+ },
2398
+ {
2399
+ "epoch": 1.99,
2400
+ "learning_rate": 6.843218908272369e-05,
2401
+ "loss": 1.3079,
2402
+ "step": 1995
2403
+ },
2404
+ {
2405
+ "epoch": 2.0,
2406
+ "learning_rate": 6.047495779403489e-05,
2407
+ "loss": 1.2959,
2408
+ "step": 2000
2409
+ },
2410
+ {
2411
+ "epoch": 2.0,
2412
+ "learning_rate": 5.251772650534609e-05,
2413
+ "loss": 1.3329,
2414
+ "step": 2005
2415
+ },
2416
+ {
2417
+ "epoch": 2.01,
2418
+ "learning_rate": 4.456049521665728e-05,
2419
+ "loss": 1.2211,
2420
+ "step": 2010
2421
+ },
2422
+ {
2423
+ "epoch": 2.01,
2424
+ "learning_rate": 3.660326392796848e-05,
2425
+ "loss": 1.1615,
2426
+ "step": 2015
2427
+ },
2428
+ {
2429
+ "epoch": 2.02,
2430
+ "learning_rate": 2.8646032639279683e-05,
2431
+ "loss": 1.1695,
2432
+ "step": 2020
2433
+ },
2434
+ {
2435
+ "epoch": 2.02,
2436
+ "learning_rate": 2.068880135059088e-05,
2437
+ "loss": 1.1895,
2438
+ "step": 2025
2439
+ },
2440
+ {
2441
+ "epoch": 2.03,
2442
+ "learning_rate": 1.2731570061902082e-05,
2443
+ "loss": 1.2908,
2444
+ "step": 2030
2445
+ },
2446
+ {
2447
+ "epoch": 2.03,
2448
+ "learning_rate": 4.7743387732132805e-06,
2449
+ "loss": 1.2253,
2450
+ "step": 2035
2451
+ },
2452
+ {
2453
+ "epoch": 2.04,
2454
+ "step": 2038,
2455
+ "total_flos": 2.5582587482196332e+23,
2456
+ "train_loss": 0.02294661615032911,
2457
+ "train_runtime": 3034.1773,
2458
+ "train_samples_per_second": 11004.826,
2459
+ "train_steps_per_second": 0.672
2460
+ }
2461
+ ],
2462
+ "max_steps": 2038,
2463
+ "num_train_epochs": 3,
2464
+ "total_flos": 2.5582587482196332e+23,
2465
+ "trial_name": null,
2466
+ "trial_params": null
2467
+ }
training_state.pt → training_args.bin RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8e170da99eae1aeb46476dfb799e440557c0a07e40043abe92e1e339048452f2
3
- size 5167
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9c14f22879a7195a66db722671812d8fd8dd94ea48f7cb5a76242f115aba6dc
3
+ size 2799
vocab.txt ADDED
The diff for this file is too large to render. See raw diff