Upload 11 files
Browse files- config.json +32 -0
- generation_config.json +8 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +107 -0
- tokenizer.json +0 -0
- tokenizer_config.json +111 -0
- trainer_state.json +2768 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "lcw99/t5-large-korean-text-summary",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 2816,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 1024,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dense_act_fn": "gelu_new",
|
| 11 |
+
"dropout_rate": 0.1,
|
| 12 |
+
"eos_token_id": 1,
|
| 13 |
+
"feed_forward_proj": "gated-gelu",
|
| 14 |
+
"initializer_factor": 1.0,
|
| 15 |
+
"is_encoder_decoder": true,
|
| 16 |
+
"is_gated_act": true,
|
| 17 |
+
"layer_norm_epsilon": 1e-06,
|
| 18 |
+
"max_length": 128,
|
| 19 |
+
"model_type": "t5",
|
| 20 |
+
"num_decoder_layers": 24,
|
| 21 |
+
"num_heads": 16,
|
| 22 |
+
"num_layers": 24,
|
| 23 |
+
"output_past": true,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"relative_attention_max_distance": 128,
|
| 26 |
+
"relative_attention_num_buckets": 32,
|
| 27 |
+
"tie_word_embeddings": false,
|
| 28 |
+
"torch_dtype": "float32",
|
| 29 |
+
"transformers_version": "4.28.1",
|
| 30 |
+
"use_cache": true,
|
| 31 |
+
"vocab_size": 50358
|
| 32 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 1,
|
| 5 |
+
"max_length": 128,
|
| 6 |
+
"pad_token_id": 0,
|
| 7 |
+
"transformers_version": "4.28.1"
|
| 8 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:458e7397681d76a9af52de6efa53a64d2f9b9f1a73c481b7597d127444ca5ab0
|
| 3 |
+
size 6564217929
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:42ff2699ae8ab43df44f5839be2c507c8c4bab0fce65475102c4778e3fdf84c5
|
| 3 |
+
size 3282133829
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:38133f3a5f47d59d484cd2ddefed4d1215613077b6b404b43147860faa36caf5
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:951ba7654c359733f025001e368bd54635389c9128285a5879c624e1f8f533ee
|
| 3 |
+
size 627
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<extra_id_0>",
|
| 4 |
+
"<extra_id_1>",
|
| 5 |
+
"<extra_id_2>",
|
| 6 |
+
"<extra_id_3>",
|
| 7 |
+
"<extra_id_4>",
|
| 8 |
+
"<extra_id_5>",
|
| 9 |
+
"<extra_id_6>",
|
| 10 |
+
"<extra_id_7>",
|
| 11 |
+
"<extra_id_8>",
|
| 12 |
+
"<extra_id_9>",
|
| 13 |
+
"<extra_id_10>",
|
| 14 |
+
"<extra_id_11>",
|
| 15 |
+
"<extra_id_12>",
|
| 16 |
+
"<extra_id_13>",
|
| 17 |
+
"<extra_id_14>",
|
| 18 |
+
"<extra_id_15>",
|
| 19 |
+
"<extra_id_16>",
|
| 20 |
+
"<extra_id_17>",
|
| 21 |
+
"<extra_id_18>",
|
| 22 |
+
"<extra_id_19>",
|
| 23 |
+
"<extra_id_20>",
|
| 24 |
+
"<extra_id_21>",
|
| 25 |
+
"<extra_id_22>",
|
| 26 |
+
"<extra_id_23>",
|
| 27 |
+
"<extra_id_24>",
|
| 28 |
+
"<extra_id_25>",
|
| 29 |
+
"<extra_id_26>",
|
| 30 |
+
"<extra_id_27>",
|
| 31 |
+
"<extra_id_28>",
|
| 32 |
+
"<extra_id_29>",
|
| 33 |
+
"<extra_id_30>",
|
| 34 |
+
"<extra_id_31>",
|
| 35 |
+
"<extra_id_32>",
|
| 36 |
+
"<extra_id_33>",
|
| 37 |
+
"<extra_id_34>",
|
| 38 |
+
"<extra_id_35>",
|
| 39 |
+
"<extra_id_36>",
|
| 40 |
+
"<extra_id_37>",
|
| 41 |
+
"<extra_id_38>",
|
| 42 |
+
"<extra_id_39>",
|
| 43 |
+
"<extra_id_40>",
|
| 44 |
+
"<extra_id_41>",
|
| 45 |
+
"<extra_id_42>",
|
| 46 |
+
"<extra_id_43>",
|
| 47 |
+
"<extra_id_44>",
|
| 48 |
+
"<extra_id_45>",
|
| 49 |
+
"<extra_id_46>",
|
| 50 |
+
"<extra_id_47>",
|
| 51 |
+
"<extra_id_48>",
|
| 52 |
+
"<extra_id_49>",
|
| 53 |
+
"<extra_id_50>",
|
| 54 |
+
"<extra_id_51>",
|
| 55 |
+
"<extra_id_52>",
|
| 56 |
+
"<extra_id_53>",
|
| 57 |
+
"<extra_id_54>",
|
| 58 |
+
"<extra_id_55>",
|
| 59 |
+
"<extra_id_56>",
|
| 60 |
+
"<extra_id_57>",
|
| 61 |
+
"<extra_id_58>",
|
| 62 |
+
"<extra_id_59>",
|
| 63 |
+
"<extra_id_60>",
|
| 64 |
+
"<extra_id_61>",
|
| 65 |
+
"<extra_id_62>",
|
| 66 |
+
"<extra_id_63>",
|
| 67 |
+
"<extra_id_64>",
|
| 68 |
+
"<extra_id_65>",
|
| 69 |
+
"<extra_id_66>",
|
| 70 |
+
"<extra_id_67>",
|
| 71 |
+
"<extra_id_68>",
|
| 72 |
+
"<extra_id_69>",
|
| 73 |
+
"<extra_id_70>",
|
| 74 |
+
"<extra_id_71>",
|
| 75 |
+
"<extra_id_72>",
|
| 76 |
+
"<extra_id_73>",
|
| 77 |
+
"<extra_id_74>",
|
| 78 |
+
"<extra_id_75>",
|
| 79 |
+
"<extra_id_76>",
|
| 80 |
+
"<extra_id_77>",
|
| 81 |
+
"<extra_id_78>",
|
| 82 |
+
"<extra_id_79>",
|
| 83 |
+
"<extra_id_80>",
|
| 84 |
+
"<extra_id_81>",
|
| 85 |
+
"<extra_id_82>",
|
| 86 |
+
"<extra_id_83>",
|
| 87 |
+
"<extra_id_84>",
|
| 88 |
+
"<extra_id_85>",
|
| 89 |
+
"<extra_id_86>",
|
| 90 |
+
"<extra_id_87>",
|
| 91 |
+
"<extra_id_88>",
|
| 92 |
+
"<extra_id_89>",
|
| 93 |
+
"<extra_id_90>",
|
| 94 |
+
"<extra_id_91>",
|
| 95 |
+
"<extra_id_92>",
|
| 96 |
+
"<extra_id_93>",
|
| 97 |
+
"<extra_id_94>",
|
| 98 |
+
"<extra_id_95>",
|
| 99 |
+
"<extra_id_96>",
|
| 100 |
+
"<extra_id_97>",
|
| 101 |
+
"<extra_id_98>",
|
| 102 |
+
"<extra_id_99>"
|
| 103 |
+
],
|
| 104 |
+
"eos_token": "</s>",
|
| 105 |
+
"pad_token": "<pad>",
|
| 106 |
+
"unk_token": "<pad>"
|
| 107 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,111 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<extra_id_0>",
|
| 4 |
+
"<extra_id_1>",
|
| 5 |
+
"<extra_id_2>",
|
| 6 |
+
"<extra_id_3>",
|
| 7 |
+
"<extra_id_4>",
|
| 8 |
+
"<extra_id_5>",
|
| 9 |
+
"<extra_id_6>",
|
| 10 |
+
"<extra_id_7>",
|
| 11 |
+
"<extra_id_8>",
|
| 12 |
+
"<extra_id_9>",
|
| 13 |
+
"<extra_id_10>",
|
| 14 |
+
"<extra_id_11>",
|
| 15 |
+
"<extra_id_12>",
|
| 16 |
+
"<extra_id_13>",
|
| 17 |
+
"<extra_id_14>",
|
| 18 |
+
"<extra_id_15>",
|
| 19 |
+
"<extra_id_16>",
|
| 20 |
+
"<extra_id_17>",
|
| 21 |
+
"<extra_id_18>",
|
| 22 |
+
"<extra_id_19>",
|
| 23 |
+
"<extra_id_20>",
|
| 24 |
+
"<extra_id_21>",
|
| 25 |
+
"<extra_id_22>",
|
| 26 |
+
"<extra_id_23>",
|
| 27 |
+
"<extra_id_24>",
|
| 28 |
+
"<extra_id_25>",
|
| 29 |
+
"<extra_id_26>",
|
| 30 |
+
"<extra_id_27>",
|
| 31 |
+
"<extra_id_28>",
|
| 32 |
+
"<extra_id_29>",
|
| 33 |
+
"<extra_id_30>",
|
| 34 |
+
"<extra_id_31>",
|
| 35 |
+
"<extra_id_32>",
|
| 36 |
+
"<extra_id_33>",
|
| 37 |
+
"<extra_id_34>",
|
| 38 |
+
"<extra_id_35>",
|
| 39 |
+
"<extra_id_36>",
|
| 40 |
+
"<extra_id_37>",
|
| 41 |
+
"<extra_id_38>",
|
| 42 |
+
"<extra_id_39>",
|
| 43 |
+
"<extra_id_40>",
|
| 44 |
+
"<extra_id_41>",
|
| 45 |
+
"<extra_id_42>",
|
| 46 |
+
"<extra_id_43>",
|
| 47 |
+
"<extra_id_44>",
|
| 48 |
+
"<extra_id_45>",
|
| 49 |
+
"<extra_id_46>",
|
| 50 |
+
"<extra_id_47>",
|
| 51 |
+
"<extra_id_48>",
|
| 52 |
+
"<extra_id_49>",
|
| 53 |
+
"<extra_id_50>",
|
| 54 |
+
"<extra_id_51>",
|
| 55 |
+
"<extra_id_52>",
|
| 56 |
+
"<extra_id_53>",
|
| 57 |
+
"<extra_id_54>",
|
| 58 |
+
"<extra_id_55>",
|
| 59 |
+
"<extra_id_56>",
|
| 60 |
+
"<extra_id_57>",
|
| 61 |
+
"<extra_id_58>",
|
| 62 |
+
"<extra_id_59>",
|
| 63 |
+
"<extra_id_60>",
|
| 64 |
+
"<extra_id_61>",
|
| 65 |
+
"<extra_id_62>",
|
| 66 |
+
"<extra_id_63>",
|
| 67 |
+
"<extra_id_64>",
|
| 68 |
+
"<extra_id_65>",
|
| 69 |
+
"<extra_id_66>",
|
| 70 |
+
"<extra_id_67>",
|
| 71 |
+
"<extra_id_68>",
|
| 72 |
+
"<extra_id_69>",
|
| 73 |
+
"<extra_id_70>",
|
| 74 |
+
"<extra_id_71>",
|
| 75 |
+
"<extra_id_72>",
|
| 76 |
+
"<extra_id_73>",
|
| 77 |
+
"<extra_id_74>",
|
| 78 |
+
"<extra_id_75>",
|
| 79 |
+
"<extra_id_76>",
|
| 80 |
+
"<extra_id_77>",
|
| 81 |
+
"<extra_id_78>",
|
| 82 |
+
"<extra_id_79>",
|
| 83 |
+
"<extra_id_80>",
|
| 84 |
+
"<extra_id_81>",
|
| 85 |
+
"<extra_id_82>",
|
| 86 |
+
"<extra_id_83>",
|
| 87 |
+
"<extra_id_84>",
|
| 88 |
+
"<extra_id_85>",
|
| 89 |
+
"<extra_id_86>",
|
| 90 |
+
"<extra_id_87>",
|
| 91 |
+
"<extra_id_88>",
|
| 92 |
+
"<extra_id_89>",
|
| 93 |
+
"<extra_id_90>",
|
| 94 |
+
"<extra_id_91>",
|
| 95 |
+
"<extra_id_92>",
|
| 96 |
+
"<extra_id_93>",
|
| 97 |
+
"<extra_id_94>",
|
| 98 |
+
"<extra_id_95>",
|
| 99 |
+
"<extra_id_96>",
|
| 100 |
+
"<extra_id_97>",
|
| 101 |
+
"<extra_id_98>",
|
| 102 |
+
"<extra_id_99>"
|
| 103 |
+
],
|
| 104 |
+
"clean_up_tokenization_spaces": true,
|
| 105 |
+
"eos_token": "</s>",
|
| 106 |
+
"extra_ids": 100,
|
| 107 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 108 |
+
"pad_token": "<pad>",
|
| 109 |
+
"tokenizer_class": "T5Tokenizer",
|
| 110 |
+
"unk_token": "<pad>"
|
| 111 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2768 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 22840,
|
| 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": 1.9956217162872154e-05,
|
| 13 |
+
"loss": 13.2845,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 1.991243432574431e-05,
|
| 19 |
+
"loss": 5.8932,
|
| 20 |
+
"step": 100
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 1.9868651488616462e-05,
|
| 25 |
+
"loss": 1.2995,
|
| 26 |
+
"step": 150
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.02,
|
| 30 |
+
"learning_rate": 1.9824868651488618e-05,
|
| 31 |
+
"loss": 0.6063,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.02,
|
| 36 |
+
"learning_rate": 1.978108581436077e-05,
|
| 37 |
+
"loss": 0.5446,
|
| 38 |
+
"step": 250
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.03,
|
| 42 |
+
"learning_rate": 1.9737302977232926e-05,
|
| 43 |
+
"loss": 0.469,
|
| 44 |
+
"step": 300
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03,
|
| 48 |
+
"learning_rate": 1.969352014010508e-05,
|
| 49 |
+
"loss": 0.4239,
|
| 50 |
+
"step": 350
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.04,
|
| 54 |
+
"learning_rate": 1.9649737302977235e-05,
|
| 55 |
+
"loss": 0.3874,
|
| 56 |
+
"step": 400
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.04,
|
| 60 |
+
"learning_rate": 1.9605954465849387e-05,
|
| 61 |
+
"loss": 0.3522,
|
| 62 |
+
"step": 450
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.04,
|
| 66 |
+
"learning_rate": 1.9562171628721543e-05,
|
| 67 |
+
"loss": 0.3294,
|
| 68 |
+
"step": 500
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.05,
|
| 72 |
+
"learning_rate": 1.9518388791593695e-05,
|
| 73 |
+
"loss": 0.3258,
|
| 74 |
+
"step": 550
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.05,
|
| 78 |
+
"learning_rate": 1.947460595446585e-05,
|
| 79 |
+
"loss": 0.3032,
|
| 80 |
+
"step": 600
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.06,
|
| 84 |
+
"learning_rate": 1.9430823117338004e-05,
|
| 85 |
+
"loss": 0.3083,
|
| 86 |
+
"step": 650
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.06,
|
| 90 |
+
"learning_rate": 1.938704028021016e-05,
|
| 91 |
+
"loss": 0.298,
|
| 92 |
+
"step": 700
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.07,
|
| 96 |
+
"learning_rate": 1.9343257443082312e-05,
|
| 97 |
+
"loss": 0.2915,
|
| 98 |
+
"step": 750
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.07,
|
| 102 |
+
"learning_rate": 1.9299474605954468e-05,
|
| 103 |
+
"loss": 0.2927,
|
| 104 |
+
"step": 800
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.07,
|
| 108 |
+
"learning_rate": 1.925569176882662e-05,
|
| 109 |
+
"loss": 0.2854,
|
| 110 |
+
"step": 850
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.08,
|
| 114 |
+
"learning_rate": 1.9211908931698776e-05,
|
| 115 |
+
"loss": 0.2821,
|
| 116 |
+
"step": 900
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.08,
|
| 120 |
+
"learning_rate": 1.916812609457093e-05,
|
| 121 |
+
"loss": 0.2876,
|
| 122 |
+
"step": 950
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.09,
|
| 126 |
+
"learning_rate": 1.9124343257443084e-05,
|
| 127 |
+
"loss": 0.2695,
|
| 128 |
+
"step": 1000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.09,
|
| 132 |
+
"learning_rate": 1.9080560420315237e-05,
|
| 133 |
+
"loss": 0.2801,
|
| 134 |
+
"step": 1050
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.1,
|
| 138 |
+
"learning_rate": 1.9036777583187393e-05,
|
| 139 |
+
"loss": 0.2772,
|
| 140 |
+
"step": 1100
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.1,
|
| 144 |
+
"learning_rate": 1.8992994746059545e-05,
|
| 145 |
+
"loss": 0.2702,
|
| 146 |
+
"step": 1150
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.11,
|
| 150 |
+
"learning_rate": 1.89492119089317e-05,
|
| 151 |
+
"loss": 0.277,
|
| 152 |
+
"step": 1200
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.11,
|
| 156 |
+
"learning_rate": 1.8905429071803853e-05,
|
| 157 |
+
"loss": 0.2749,
|
| 158 |
+
"step": 1250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.11,
|
| 162 |
+
"learning_rate": 1.886164623467601e-05,
|
| 163 |
+
"loss": 0.2751,
|
| 164 |
+
"step": 1300
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.12,
|
| 168 |
+
"learning_rate": 1.881786339754816e-05,
|
| 169 |
+
"loss": 0.2747,
|
| 170 |
+
"step": 1350
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.12,
|
| 174 |
+
"learning_rate": 1.8774080560420317e-05,
|
| 175 |
+
"loss": 0.2768,
|
| 176 |
+
"step": 1400
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.13,
|
| 180 |
+
"learning_rate": 1.873029772329247e-05,
|
| 181 |
+
"loss": 0.269,
|
| 182 |
+
"step": 1450
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.13,
|
| 186 |
+
"learning_rate": 1.8686514886164622e-05,
|
| 187 |
+
"loss": 0.2711,
|
| 188 |
+
"step": 1500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.14,
|
| 192 |
+
"learning_rate": 1.8642732049036778e-05,
|
| 193 |
+
"loss": 0.2669,
|
| 194 |
+
"step": 1550
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.14,
|
| 198 |
+
"learning_rate": 1.8598949211908934e-05,
|
| 199 |
+
"loss": 0.2617,
|
| 200 |
+
"step": 1600
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.14,
|
| 204 |
+
"learning_rate": 1.855516637478109e-05,
|
| 205 |
+
"loss": 0.269,
|
| 206 |
+
"step": 1650
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.15,
|
| 210 |
+
"learning_rate": 1.8511383537653242e-05,
|
| 211 |
+
"loss": 0.2629,
|
| 212 |
+
"step": 1700
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.15,
|
| 216 |
+
"learning_rate": 1.8467600700525398e-05,
|
| 217 |
+
"loss": 0.2606,
|
| 218 |
+
"step": 1750
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.16,
|
| 222 |
+
"learning_rate": 1.842381786339755e-05,
|
| 223 |
+
"loss": 0.2623,
|
| 224 |
+
"step": 1800
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.16,
|
| 228 |
+
"learning_rate": 1.8380035026269706e-05,
|
| 229 |
+
"loss": 0.2638,
|
| 230 |
+
"step": 1850
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.17,
|
| 234 |
+
"learning_rate": 1.833625218914186e-05,
|
| 235 |
+
"loss": 0.2617,
|
| 236 |
+
"step": 1900
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.17,
|
| 240 |
+
"learning_rate": 1.829246935201401e-05,
|
| 241 |
+
"loss": 0.2547,
|
| 242 |
+
"step": 1950
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.18,
|
| 246 |
+
"learning_rate": 1.8248686514886167e-05,
|
| 247 |
+
"loss": 0.2656,
|
| 248 |
+
"step": 2000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.18,
|
| 252 |
+
"learning_rate": 1.820490367775832e-05,
|
| 253 |
+
"loss": 0.2635,
|
| 254 |
+
"step": 2050
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.18,
|
| 258 |
+
"learning_rate": 1.8161120840630475e-05,
|
| 259 |
+
"loss": 0.2598,
|
| 260 |
+
"step": 2100
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.19,
|
| 264 |
+
"learning_rate": 1.8117338003502628e-05,
|
| 265 |
+
"loss": 0.2927,
|
| 266 |
+
"step": 2150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.19,
|
| 270 |
+
"learning_rate": 1.8073555166374784e-05,
|
| 271 |
+
"loss": 0.2635,
|
| 272 |
+
"step": 2200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.2,
|
| 276 |
+
"learning_rate": 1.8029772329246936e-05,
|
| 277 |
+
"loss": 0.2532,
|
| 278 |
+
"step": 2250
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.2,
|
| 282 |
+
"learning_rate": 1.7985989492119092e-05,
|
| 283 |
+
"loss": 0.2665,
|
| 284 |
+
"step": 2300
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.21,
|
| 288 |
+
"learning_rate": 1.7942206654991244e-05,
|
| 289 |
+
"loss": 0.2518,
|
| 290 |
+
"step": 2350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.21,
|
| 294 |
+
"learning_rate": 1.78984238178634e-05,
|
| 295 |
+
"loss": 0.2622,
|
| 296 |
+
"step": 2400
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.21,
|
| 300 |
+
"learning_rate": 1.7854640980735553e-05,
|
| 301 |
+
"loss": 0.2599,
|
| 302 |
+
"step": 2450
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.22,
|
| 306 |
+
"learning_rate": 1.781085814360771e-05,
|
| 307 |
+
"loss": 0.2571,
|
| 308 |
+
"step": 2500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.22,
|
| 312 |
+
"learning_rate": 1.776707530647986e-05,
|
| 313 |
+
"loss": 0.2593,
|
| 314 |
+
"step": 2550
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.23,
|
| 318 |
+
"learning_rate": 1.7723292469352017e-05,
|
| 319 |
+
"loss": 0.2659,
|
| 320 |
+
"step": 2600
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.23,
|
| 324 |
+
"learning_rate": 1.767950963222417e-05,
|
| 325 |
+
"loss": 0.2544,
|
| 326 |
+
"step": 2650
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.24,
|
| 330 |
+
"learning_rate": 1.7635726795096325e-05,
|
| 331 |
+
"loss": 0.2523,
|
| 332 |
+
"step": 2700
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.24,
|
| 336 |
+
"learning_rate": 1.7591943957968477e-05,
|
| 337 |
+
"loss": 0.2618,
|
| 338 |
+
"step": 2750
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.25,
|
| 342 |
+
"learning_rate": 1.7548161120840633e-05,
|
| 343 |
+
"loss": 0.2539,
|
| 344 |
+
"step": 2800
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.25,
|
| 348 |
+
"learning_rate": 1.7504378283712786e-05,
|
| 349 |
+
"loss": 0.2617,
|
| 350 |
+
"step": 2850
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.25,
|
| 354 |
+
"learning_rate": 1.746059544658494e-05,
|
| 355 |
+
"loss": 0.2503,
|
| 356 |
+
"step": 2900
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.26,
|
| 360 |
+
"learning_rate": 1.7416812609457094e-05,
|
| 361 |
+
"loss": 0.2548,
|
| 362 |
+
"step": 2950
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.26,
|
| 366 |
+
"learning_rate": 1.737302977232925e-05,
|
| 367 |
+
"loss": 0.2547,
|
| 368 |
+
"step": 3000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.27,
|
| 372 |
+
"learning_rate": 1.7329246935201402e-05,
|
| 373 |
+
"loss": 0.2516,
|
| 374 |
+
"step": 3050
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.27,
|
| 378 |
+
"learning_rate": 1.7285464098073558e-05,
|
| 379 |
+
"loss": 0.2525,
|
| 380 |
+
"step": 3100
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.28,
|
| 384 |
+
"learning_rate": 1.724168126094571e-05,
|
| 385 |
+
"loss": 0.2549,
|
| 386 |
+
"step": 3150
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.28,
|
| 390 |
+
"learning_rate": 1.7197898423817866e-05,
|
| 391 |
+
"loss": 0.2495,
|
| 392 |
+
"step": 3200
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.28,
|
| 396 |
+
"learning_rate": 1.715411558669002e-05,
|
| 397 |
+
"loss": 0.262,
|
| 398 |
+
"step": 3250
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.29,
|
| 402 |
+
"learning_rate": 1.7110332749562174e-05,
|
| 403 |
+
"loss": 0.2557,
|
| 404 |
+
"step": 3300
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.29,
|
| 408 |
+
"learning_rate": 1.7066549912434327e-05,
|
| 409 |
+
"loss": 0.2493,
|
| 410 |
+
"step": 3350
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.3,
|
| 414 |
+
"learning_rate": 1.7022767075306483e-05,
|
| 415 |
+
"loss": 0.2561,
|
| 416 |
+
"step": 3400
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.3,
|
| 420 |
+
"learning_rate": 1.6978984238178635e-05,
|
| 421 |
+
"loss": 0.2473,
|
| 422 |
+
"step": 3450
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.31,
|
| 426 |
+
"learning_rate": 1.6935201401050788e-05,
|
| 427 |
+
"loss": 0.2507,
|
| 428 |
+
"step": 3500
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.31,
|
| 432 |
+
"learning_rate": 1.6891418563922943e-05,
|
| 433 |
+
"loss": 0.253,
|
| 434 |
+
"step": 3550
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.32,
|
| 438 |
+
"learning_rate": 1.6847635726795096e-05,
|
| 439 |
+
"loss": 0.2442,
|
| 440 |
+
"step": 3600
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.32,
|
| 444 |
+
"learning_rate": 1.6803852889667252e-05,
|
| 445 |
+
"loss": 0.2447,
|
| 446 |
+
"step": 3650
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.32,
|
| 450 |
+
"learning_rate": 1.6760070052539404e-05,
|
| 451 |
+
"loss": 0.2544,
|
| 452 |
+
"step": 3700
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.33,
|
| 456 |
+
"learning_rate": 1.671628721541156e-05,
|
| 457 |
+
"loss": 0.2554,
|
| 458 |
+
"step": 3750
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.33,
|
| 462 |
+
"learning_rate": 1.6672504378283712e-05,
|
| 463 |
+
"loss": 0.2571,
|
| 464 |
+
"step": 3800
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.34,
|
| 468 |
+
"learning_rate": 1.6628721541155868e-05,
|
| 469 |
+
"loss": 0.2503,
|
| 470 |
+
"step": 3850
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.34,
|
| 474 |
+
"learning_rate": 1.658493870402802e-05,
|
| 475 |
+
"loss": 0.2483,
|
| 476 |
+
"step": 3900
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.35,
|
| 480 |
+
"learning_rate": 1.6541155866900177e-05,
|
| 481 |
+
"loss": 0.2543,
|
| 482 |
+
"step": 3950
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.35,
|
| 486 |
+
"learning_rate": 1.649737302977233e-05,
|
| 487 |
+
"loss": 0.2366,
|
| 488 |
+
"step": 4000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.35,
|
| 492 |
+
"learning_rate": 1.6453590192644485e-05,
|
| 493 |
+
"loss": 0.2476,
|
| 494 |
+
"step": 4050
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.36,
|
| 498 |
+
"learning_rate": 1.6409807355516637e-05,
|
| 499 |
+
"loss": 0.248,
|
| 500 |
+
"step": 4100
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.36,
|
| 504 |
+
"learning_rate": 1.6366024518388793e-05,
|
| 505 |
+
"loss": 0.2474,
|
| 506 |
+
"step": 4150
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.37,
|
| 510 |
+
"learning_rate": 1.6322241681260946e-05,
|
| 511 |
+
"loss": 0.2493,
|
| 512 |
+
"step": 4200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.37,
|
| 516 |
+
"learning_rate": 1.62784588441331e-05,
|
| 517 |
+
"loss": 0.2383,
|
| 518 |
+
"step": 4250
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.38,
|
| 522 |
+
"learning_rate": 1.6234676007005254e-05,
|
| 523 |
+
"loss": 0.2493,
|
| 524 |
+
"step": 4300
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.38,
|
| 528 |
+
"learning_rate": 1.619089316987741e-05,
|
| 529 |
+
"loss": 0.251,
|
| 530 |
+
"step": 4350
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.39,
|
| 534 |
+
"learning_rate": 1.6147110332749562e-05,
|
| 535 |
+
"loss": 0.2558,
|
| 536 |
+
"step": 4400
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.39,
|
| 540 |
+
"learning_rate": 1.6103327495621718e-05,
|
| 541 |
+
"loss": 0.2448,
|
| 542 |
+
"step": 4450
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.39,
|
| 546 |
+
"learning_rate": 1.605954465849387e-05,
|
| 547 |
+
"loss": 0.2502,
|
| 548 |
+
"step": 4500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.4,
|
| 552 |
+
"learning_rate": 1.6015761821366026e-05,
|
| 553 |
+
"loss": 0.2507,
|
| 554 |
+
"step": 4550
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.4,
|
| 558 |
+
"learning_rate": 1.597197898423818e-05,
|
| 559 |
+
"loss": 0.2468,
|
| 560 |
+
"step": 4600
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.41,
|
| 564 |
+
"learning_rate": 1.5928196147110334e-05,
|
| 565 |
+
"loss": 0.24,
|
| 566 |
+
"step": 4650
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.41,
|
| 570 |
+
"learning_rate": 1.5884413309982487e-05,
|
| 571 |
+
"loss": 0.2485,
|
| 572 |
+
"step": 4700
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.42,
|
| 576 |
+
"learning_rate": 1.5840630472854643e-05,
|
| 577 |
+
"loss": 0.2521,
|
| 578 |
+
"step": 4750
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.42,
|
| 582 |
+
"learning_rate": 1.57968476357268e-05,
|
| 583 |
+
"loss": 0.2504,
|
| 584 |
+
"step": 4800
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.42,
|
| 588 |
+
"learning_rate": 1.575306479859895e-05,
|
| 589 |
+
"loss": 0.2481,
|
| 590 |
+
"step": 4850
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.43,
|
| 594 |
+
"learning_rate": 1.5709281961471107e-05,
|
| 595 |
+
"loss": 0.2469,
|
| 596 |
+
"step": 4900
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.43,
|
| 600 |
+
"learning_rate": 1.566549912434326e-05,
|
| 601 |
+
"loss": 0.2446,
|
| 602 |
+
"step": 4950
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.44,
|
| 606 |
+
"learning_rate": 1.5621716287215415e-05,
|
| 607 |
+
"loss": 0.2494,
|
| 608 |
+
"step": 5000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.44,
|
| 612 |
+
"learning_rate": 1.5577933450087568e-05,
|
| 613 |
+
"loss": 0.2442,
|
| 614 |
+
"step": 5050
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.45,
|
| 618 |
+
"learning_rate": 1.5534150612959723e-05,
|
| 619 |
+
"loss": 0.2458,
|
| 620 |
+
"step": 5100
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.45,
|
| 624 |
+
"learning_rate": 1.5490367775831876e-05,
|
| 625 |
+
"loss": 0.2579,
|
| 626 |
+
"step": 5150
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.46,
|
| 630 |
+
"learning_rate": 1.544658493870403e-05,
|
| 631 |
+
"loss": 0.2456,
|
| 632 |
+
"step": 5200
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.46,
|
| 636 |
+
"learning_rate": 1.5402802101576184e-05,
|
| 637 |
+
"loss": 0.2478,
|
| 638 |
+
"step": 5250
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.46,
|
| 642 |
+
"learning_rate": 1.535901926444834e-05,
|
| 643 |
+
"loss": 0.2491,
|
| 644 |
+
"step": 5300
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.47,
|
| 648 |
+
"learning_rate": 1.5315236427320492e-05,
|
| 649 |
+
"loss": 0.2434,
|
| 650 |
+
"step": 5350
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.47,
|
| 654 |
+
"learning_rate": 1.5271453590192645e-05,
|
| 655 |
+
"loss": 0.2438,
|
| 656 |
+
"step": 5400
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.48,
|
| 660 |
+
"learning_rate": 1.52276707530648e-05,
|
| 661 |
+
"loss": 0.2435,
|
| 662 |
+
"step": 5450
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.48,
|
| 666 |
+
"learning_rate": 1.5183887915936955e-05,
|
| 667 |
+
"loss": 0.2524,
|
| 668 |
+
"step": 5500
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.49,
|
| 672 |
+
"learning_rate": 1.5140105078809109e-05,
|
| 673 |
+
"loss": 0.2511,
|
| 674 |
+
"step": 5550
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.49,
|
| 678 |
+
"learning_rate": 1.5096322241681263e-05,
|
| 679 |
+
"loss": 0.2534,
|
| 680 |
+
"step": 5600
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.49,
|
| 684 |
+
"learning_rate": 1.5052539404553417e-05,
|
| 685 |
+
"loss": 0.2477,
|
| 686 |
+
"step": 5650
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.5,
|
| 690 |
+
"learning_rate": 1.5008756567425571e-05,
|
| 691 |
+
"loss": 0.255,
|
| 692 |
+
"step": 5700
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.5,
|
| 696 |
+
"learning_rate": 1.4964973730297725e-05,
|
| 697 |
+
"loss": 0.2446,
|
| 698 |
+
"step": 5750
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.51,
|
| 702 |
+
"learning_rate": 1.492119089316988e-05,
|
| 703 |
+
"loss": 0.2492,
|
| 704 |
+
"step": 5800
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.51,
|
| 708 |
+
"learning_rate": 1.4877408056042034e-05,
|
| 709 |
+
"loss": 0.25,
|
| 710 |
+
"step": 5850
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.52,
|
| 714 |
+
"learning_rate": 1.4833625218914188e-05,
|
| 715 |
+
"loss": 0.241,
|
| 716 |
+
"step": 5900
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.52,
|
| 720 |
+
"learning_rate": 1.4789842381786342e-05,
|
| 721 |
+
"loss": 0.2517,
|
| 722 |
+
"step": 5950
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.53,
|
| 726 |
+
"learning_rate": 1.4746059544658496e-05,
|
| 727 |
+
"loss": 0.2453,
|
| 728 |
+
"step": 6000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.53,
|
| 732 |
+
"learning_rate": 1.470227670753065e-05,
|
| 733 |
+
"loss": 0.2485,
|
| 734 |
+
"step": 6050
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.53,
|
| 738 |
+
"learning_rate": 1.4658493870402803e-05,
|
| 739 |
+
"loss": 0.2494,
|
| 740 |
+
"step": 6100
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.54,
|
| 744 |
+
"learning_rate": 1.4614711033274957e-05,
|
| 745 |
+
"loss": 0.2471,
|
| 746 |
+
"step": 6150
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.54,
|
| 750 |
+
"learning_rate": 1.4570928196147111e-05,
|
| 751 |
+
"loss": 0.2508,
|
| 752 |
+
"step": 6200
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.55,
|
| 756 |
+
"learning_rate": 1.4527145359019265e-05,
|
| 757 |
+
"loss": 0.2423,
|
| 758 |
+
"step": 6250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.55,
|
| 762 |
+
"learning_rate": 1.448336252189142e-05,
|
| 763 |
+
"loss": 0.2433,
|
| 764 |
+
"step": 6300
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.56,
|
| 768 |
+
"learning_rate": 1.4439579684763573e-05,
|
| 769 |
+
"loss": 0.2405,
|
| 770 |
+
"step": 6350
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.56,
|
| 774 |
+
"learning_rate": 1.4395796847635727e-05,
|
| 775 |
+
"loss": 0.2496,
|
| 776 |
+
"step": 6400
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.56,
|
| 780 |
+
"learning_rate": 1.4352014010507882e-05,
|
| 781 |
+
"loss": 0.2433,
|
| 782 |
+
"step": 6450
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.57,
|
| 786 |
+
"learning_rate": 1.4308231173380036e-05,
|
| 787 |
+
"loss": 0.2429,
|
| 788 |
+
"step": 6500
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.57,
|
| 792 |
+
"learning_rate": 1.426444833625219e-05,
|
| 793 |
+
"loss": 0.246,
|
| 794 |
+
"step": 6550
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.58,
|
| 798 |
+
"learning_rate": 1.4220665499124344e-05,
|
| 799 |
+
"loss": 0.2455,
|
| 800 |
+
"step": 6600
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.58,
|
| 804 |
+
"learning_rate": 1.4176882661996498e-05,
|
| 805 |
+
"loss": 0.2437,
|
| 806 |
+
"step": 6650
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.59,
|
| 810 |
+
"learning_rate": 1.4133099824868652e-05,
|
| 811 |
+
"loss": 0.2455,
|
| 812 |
+
"step": 6700
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.59,
|
| 816 |
+
"learning_rate": 1.4089316987740806e-05,
|
| 817 |
+
"loss": 0.2457,
|
| 818 |
+
"step": 6750
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.6,
|
| 822 |
+
"learning_rate": 1.404553415061296e-05,
|
| 823 |
+
"loss": 0.2304,
|
| 824 |
+
"step": 6800
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.6,
|
| 828 |
+
"learning_rate": 1.4001751313485115e-05,
|
| 829 |
+
"loss": 0.241,
|
| 830 |
+
"step": 6850
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.6,
|
| 834 |
+
"learning_rate": 1.3957968476357269e-05,
|
| 835 |
+
"loss": 0.2398,
|
| 836 |
+
"step": 6900
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.61,
|
| 840 |
+
"learning_rate": 1.3914185639229423e-05,
|
| 841 |
+
"loss": 0.2459,
|
| 842 |
+
"step": 6950
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.61,
|
| 846 |
+
"learning_rate": 1.3870402802101577e-05,
|
| 847 |
+
"loss": 0.2452,
|
| 848 |
+
"step": 7000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.62,
|
| 852 |
+
"learning_rate": 1.3826619964973731e-05,
|
| 853 |
+
"loss": 0.2416,
|
| 854 |
+
"step": 7050
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.62,
|
| 858 |
+
"learning_rate": 1.3782837127845885e-05,
|
| 859 |
+
"loss": 0.2365,
|
| 860 |
+
"step": 7100
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.63,
|
| 864 |
+
"learning_rate": 1.373905429071804e-05,
|
| 865 |
+
"loss": 0.245,
|
| 866 |
+
"step": 7150
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.63,
|
| 870 |
+
"learning_rate": 1.3695271453590194e-05,
|
| 871 |
+
"loss": 0.2472,
|
| 872 |
+
"step": 7200
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.63,
|
| 876 |
+
"learning_rate": 1.3651488616462348e-05,
|
| 877 |
+
"loss": 0.2462,
|
| 878 |
+
"step": 7250
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.64,
|
| 882 |
+
"learning_rate": 1.3607705779334502e-05,
|
| 883 |
+
"loss": 0.2426,
|
| 884 |
+
"step": 7300
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.64,
|
| 888 |
+
"learning_rate": 1.3563922942206656e-05,
|
| 889 |
+
"loss": 0.2463,
|
| 890 |
+
"step": 7350
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.65,
|
| 894 |
+
"learning_rate": 1.352014010507881e-05,
|
| 895 |
+
"loss": 0.2416,
|
| 896 |
+
"step": 7400
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.65,
|
| 900 |
+
"learning_rate": 1.3476357267950964e-05,
|
| 901 |
+
"loss": 0.2476,
|
| 902 |
+
"step": 7450
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.66,
|
| 906 |
+
"learning_rate": 1.3432574430823118e-05,
|
| 907 |
+
"loss": 0.2424,
|
| 908 |
+
"step": 7500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.66,
|
| 912 |
+
"learning_rate": 1.3388791593695273e-05,
|
| 913 |
+
"loss": 0.2354,
|
| 914 |
+
"step": 7550
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.67,
|
| 918 |
+
"learning_rate": 1.3345008756567425e-05,
|
| 919 |
+
"loss": 0.243,
|
| 920 |
+
"step": 7600
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.67,
|
| 924 |
+
"learning_rate": 1.3301225919439579e-05,
|
| 925 |
+
"loss": 0.2362,
|
| 926 |
+
"step": 7650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.67,
|
| 930 |
+
"learning_rate": 1.3257443082311733e-05,
|
| 931 |
+
"loss": 0.2374,
|
| 932 |
+
"step": 7700
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.68,
|
| 936 |
+
"learning_rate": 1.3213660245183887e-05,
|
| 937 |
+
"loss": 0.2379,
|
| 938 |
+
"step": 7750
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.68,
|
| 942 |
+
"learning_rate": 1.3169877408056041e-05,
|
| 943 |
+
"loss": 0.2348,
|
| 944 |
+
"step": 7800
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.69,
|
| 948 |
+
"learning_rate": 1.3126094570928196e-05,
|
| 949 |
+
"loss": 0.2432,
|
| 950 |
+
"step": 7850
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.69,
|
| 954 |
+
"learning_rate": 1.308231173380035e-05,
|
| 955 |
+
"loss": 0.2381,
|
| 956 |
+
"step": 7900
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.7,
|
| 960 |
+
"learning_rate": 1.3038528896672507e-05,
|
| 961 |
+
"loss": 0.2472,
|
| 962 |
+
"step": 7950
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.7,
|
| 966 |
+
"learning_rate": 1.2994746059544661e-05,
|
| 967 |
+
"loss": 0.2409,
|
| 968 |
+
"step": 8000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.7,
|
| 972 |
+
"learning_rate": 1.2950963222416814e-05,
|
| 973 |
+
"loss": 0.2448,
|
| 974 |
+
"step": 8050
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.71,
|
| 978 |
+
"learning_rate": 1.2907180385288968e-05,
|
| 979 |
+
"loss": 0.2421,
|
| 980 |
+
"step": 8100
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.71,
|
| 984 |
+
"learning_rate": 1.2863397548161122e-05,
|
| 985 |
+
"loss": 0.2392,
|
| 986 |
+
"step": 8150
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.72,
|
| 990 |
+
"learning_rate": 1.2819614711033276e-05,
|
| 991 |
+
"loss": 0.2432,
|
| 992 |
+
"step": 8200
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.72,
|
| 996 |
+
"learning_rate": 1.277583187390543e-05,
|
| 997 |
+
"loss": 0.2336,
|
| 998 |
+
"step": 8250
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.73,
|
| 1002 |
+
"learning_rate": 1.2732049036777585e-05,
|
| 1003 |
+
"loss": 0.2416,
|
| 1004 |
+
"step": 8300
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.73,
|
| 1008 |
+
"learning_rate": 1.2688266199649739e-05,
|
| 1009 |
+
"loss": 0.236,
|
| 1010 |
+
"step": 8350
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.74,
|
| 1014 |
+
"learning_rate": 1.2644483362521893e-05,
|
| 1015 |
+
"loss": 0.2387,
|
| 1016 |
+
"step": 8400
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.74,
|
| 1020 |
+
"learning_rate": 1.2600700525394047e-05,
|
| 1021 |
+
"loss": 0.2398,
|
| 1022 |
+
"step": 8450
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.74,
|
| 1026 |
+
"learning_rate": 1.2556917688266201e-05,
|
| 1027 |
+
"loss": 0.2413,
|
| 1028 |
+
"step": 8500
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.75,
|
| 1032 |
+
"learning_rate": 1.2513134851138355e-05,
|
| 1033 |
+
"loss": 0.2445,
|
| 1034 |
+
"step": 8550
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.75,
|
| 1038 |
+
"learning_rate": 1.246935201401051e-05,
|
| 1039 |
+
"loss": 0.2405,
|
| 1040 |
+
"step": 8600
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.76,
|
| 1044 |
+
"learning_rate": 1.2425569176882663e-05,
|
| 1045 |
+
"loss": 0.2389,
|
| 1046 |
+
"step": 8650
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.76,
|
| 1050 |
+
"learning_rate": 1.2381786339754818e-05,
|
| 1051 |
+
"loss": 0.2348,
|
| 1052 |
+
"step": 8700
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.77,
|
| 1056 |
+
"learning_rate": 1.2338003502626972e-05,
|
| 1057 |
+
"loss": 0.2397,
|
| 1058 |
+
"step": 8750
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.77,
|
| 1062 |
+
"learning_rate": 1.2294220665499126e-05,
|
| 1063 |
+
"loss": 0.233,
|
| 1064 |
+
"step": 8800
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.77,
|
| 1068 |
+
"learning_rate": 1.225043782837128e-05,
|
| 1069 |
+
"loss": 0.2395,
|
| 1070 |
+
"step": 8850
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.78,
|
| 1074 |
+
"learning_rate": 1.2206654991243434e-05,
|
| 1075 |
+
"loss": 0.2366,
|
| 1076 |
+
"step": 8900
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.78,
|
| 1080 |
+
"learning_rate": 1.2162872154115588e-05,
|
| 1081 |
+
"loss": 0.234,
|
| 1082 |
+
"step": 8950
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.79,
|
| 1086 |
+
"learning_rate": 1.2119089316987742e-05,
|
| 1087 |
+
"loss": 0.2391,
|
| 1088 |
+
"step": 9000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.79,
|
| 1092 |
+
"learning_rate": 1.2075306479859897e-05,
|
| 1093 |
+
"loss": 0.2382,
|
| 1094 |
+
"step": 9050
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.8,
|
| 1098 |
+
"learning_rate": 1.203152364273205e-05,
|
| 1099 |
+
"loss": 0.231,
|
| 1100 |
+
"step": 9100
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.8,
|
| 1104 |
+
"learning_rate": 1.1987740805604205e-05,
|
| 1105 |
+
"loss": 0.2383,
|
| 1106 |
+
"step": 9150
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.81,
|
| 1110 |
+
"learning_rate": 1.1943957968476359e-05,
|
| 1111 |
+
"loss": 0.2443,
|
| 1112 |
+
"step": 9200
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.81,
|
| 1116 |
+
"learning_rate": 1.1900175131348513e-05,
|
| 1117 |
+
"loss": 0.2446,
|
| 1118 |
+
"step": 9250
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.81,
|
| 1122 |
+
"learning_rate": 1.1856392294220667e-05,
|
| 1123 |
+
"loss": 0.2404,
|
| 1124 |
+
"step": 9300
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.82,
|
| 1128 |
+
"learning_rate": 1.1812609457092821e-05,
|
| 1129 |
+
"loss": 0.2371,
|
| 1130 |
+
"step": 9350
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.82,
|
| 1134 |
+
"learning_rate": 1.1768826619964975e-05,
|
| 1135 |
+
"loss": 0.2427,
|
| 1136 |
+
"step": 9400
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.83,
|
| 1140 |
+
"learning_rate": 1.172504378283713e-05,
|
| 1141 |
+
"loss": 0.2354,
|
| 1142 |
+
"step": 9450
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.83,
|
| 1146 |
+
"learning_rate": 1.1681260945709284e-05,
|
| 1147 |
+
"loss": 0.2332,
|
| 1148 |
+
"step": 9500
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.84,
|
| 1152 |
+
"learning_rate": 1.1637478108581436e-05,
|
| 1153 |
+
"loss": 0.2394,
|
| 1154 |
+
"step": 9550
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.84,
|
| 1158 |
+
"learning_rate": 1.159369527145359e-05,
|
| 1159 |
+
"loss": 0.238,
|
| 1160 |
+
"step": 9600
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.85,
|
| 1164 |
+
"learning_rate": 1.1549912434325744e-05,
|
| 1165 |
+
"loss": 0.2365,
|
| 1166 |
+
"step": 9650
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.85,
|
| 1170 |
+
"learning_rate": 1.1506129597197899e-05,
|
| 1171 |
+
"loss": 0.235,
|
| 1172 |
+
"step": 9700
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.85,
|
| 1176 |
+
"learning_rate": 1.1462346760070053e-05,
|
| 1177 |
+
"loss": 0.2374,
|
| 1178 |
+
"step": 9750
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.86,
|
| 1182 |
+
"learning_rate": 1.1418563922942207e-05,
|
| 1183 |
+
"loss": 0.2291,
|
| 1184 |
+
"step": 9800
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.86,
|
| 1188 |
+
"learning_rate": 1.1374781085814361e-05,
|
| 1189 |
+
"loss": 0.2404,
|
| 1190 |
+
"step": 9850
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.87,
|
| 1194 |
+
"learning_rate": 1.1330998248686515e-05,
|
| 1195 |
+
"loss": 0.2373,
|
| 1196 |
+
"step": 9900
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.87,
|
| 1200 |
+
"learning_rate": 1.128721541155867e-05,
|
| 1201 |
+
"loss": 0.2426,
|
| 1202 |
+
"step": 9950
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.88,
|
| 1206 |
+
"learning_rate": 1.1243432574430823e-05,
|
| 1207 |
+
"loss": 0.2476,
|
| 1208 |
+
"step": 10000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.88,
|
| 1212 |
+
"learning_rate": 1.1199649737302978e-05,
|
| 1213 |
+
"loss": 0.2404,
|
| 1214 |
+
"step": 10050
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.88,
|
| 1218 |
+
"learning_rate": 1.1155866900175132e-05,
|
| 1219 |
+
"loss": 0.2401,
|
| 1220 |
+
"step": 10100
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.89,
|
| 1224 |
+
"learning_rate": 1.1112084063047286e-05,
|
| 1225 |
+
"loss": 0.2443,
|
| 1226 |
+
"step": 10150
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.89,
|
| 1230 |
+
"learning_rate": 1.106830122591944e-05,
|
| 1231 |
+
"loss": 0.2349,
|
| 1232 |
+
"step": 10200
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 0.9,
|
| 1236 |
+
"learning_rate": 1.1024518388791594e-05,
|
| 1237 |
+
"loss": 0.2408,
|
| 1238 |
+
"step": 10250
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.9,
|
| 1242 |
+
"learning_rate": 1.0980735551663748e-05,
|
| 1243 |
+
"loss": 0.2393,
|
| 1244 |
+
"step": 10300
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.91,
|
| 1248 |
+
"learning_rate": 1.0936952714535902e-05,
|
| 1249 |
+
"loss": 0.2358,
|
| 1250 |
+
"step": 10350
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.91,
|
| 1254 |
+
"learning_rate": 1.0893169877408056e-05,
|
| 1255 |
+
"loss": 0.2412,
|
| 1256 |
+
"step": 10400
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.92,
|
| 1260 |
+
"learning_rate": 1.084938704028021e-05,
|
| 1261 |
+
"loss": 0.2536,
|
| 1262 |
+
"step": 10450
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.92,
|
| 1266 |
+
"learning_rate": 1.0805604203152365e-05,
|
| 1267 |
+
"loss": 0.2344,
|
| 1268 |
+
"step": 10500
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.92,
|
| 1272 |
+
"learning_rate": 1.0761821366024519e-05,
|
| 1273 |
+
"loss": 0.2413,
|
| 1274 |
+
"step": 10550
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 0.93,
|
| 1278 |
+
"learning_rate": 1.0718038528896673e-05,
|
| 1279 |
+
"loss": 0.2487,
|
| 1280 |
+
"step": 10600
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.93,
|
| 1284 |
+
"learning_rate": 1.0674255691768827e-05,
|
| 1285 |
+
"loss": 0.2374,
|
| 1286 |
+
"step": 10650
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.94,
|
| 1290 |
+
"learning_rate": 1.0630472854640981e-05,
|
| 1291 |
+
"loss": 0.2374,
|
| 1292 |
+
"step": 10700
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.94,
|
| 1296 |
+
"learning_rate": 1.0586690017513135e-05,
|
| 1297 |
+
"loss": 0.2396,
|
| 1298 |
+
"step": 10750
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.95,
|
| 1302 |
+
"learning_rate": 1.054290718038529e-05,
|
| 1303 |
+
"loss": 0.236,
|
| 1304 |
+
"step": 10800
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.95,
|
| 1308 |
+
"learning_rate": 1.0499124343257444e-05,
|
| 1309 |
+
"loss": 0.2393,
|
| 1310 |
+
"step": 10850
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 0.95,
|
| 1314 |
+
"learning_rate": 1.0455341506129598e-05,
|
| 1315 |
+
"loss": 0.238,
|
| 1316 |
+
"step": 10900
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 0.96,
|
| 1320 |
+
"learning_rate": 1.0411558669001752e-05,
|
| 1321 |
+
"loss": 0.2385,
|
| 1322 |
+
"step": 10950
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.96,
|
| 1326 |
+
"learning_rate": 1.0367775831873904e-05,
|
| 1327 |
+
"loss": 0.2376,
|
| 1328 |
+
"step": 11000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.97,
|
| 1332 |
+
"learning_rate": 1.0323992994746059e-05,
|
| 1333 |
+
"loss": 0.2347,
|
| 1334 |
+
"step": 11050
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.97,
|
| 1338 |
+
"learning_rate": 1.0280210157618213e-05,
|
| 1339 |
+
"loss": 0.2337,
|
| 1340 |
+
"step": 11100
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.98,
|
| 1344 |
+
"learning_rate": 1.023642732049037e-05,
|
| 1345 |
+
"loss": 0.2423,
|
| 1346 |
+
"step": 11150
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.98,
|
| 1350 |
+
"learning_rate": 1.0192644483362524e-05,
|
| 1351 |
+
"loss": 0.2347,
|
| 1352 |
+
"step": 11200
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.99,
|
| 1356 |
+
"learning_rate": 1.0148861646234678e-05,
|
| 1357 |
+
"loss": 0.2387,
|
| 1358 |
+
"step": 11250
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.99,
|
| 1362 |
+
"learning_rate": 1.0105078809106833e-05,
|
| 1363 |
+
"loss": 0.2427,
|
| 1364 |
+
"step": 11300
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.99,
|
| 1368 |
+
"learning_rate": 1.0061295971978987e-05,
|
| 1369 |
+
"loss": 0.2243,
|
| 1370 |
+
"step": 11350
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 1.0,
|
| 1374 |
+
"learning_rate": 1.0017513134851141e-05,
|
| 1375 |
+
"loss": 0.2386,
|
| 1376 |
+
"step": 11400
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 1.0,
|
| 1380 |
+
"eval_loss": 0.2167460024356842,
|
| 1381 |
+
"eval_runtime": 72.1234,
|
| 1382 |
+
"eval_samples_per_second": 63.336,
|
| 1383 |
+
"eval_steps_per_second": 15.834,
|
| 1384 |
+
"step": 11420
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 1.0,
|
| 1388 |
+
"learning_rate": 9.973730297723293e-06,
|
| 1389 |
+
"loss": 0.2285,
|
| 1390 |
+
"step": 11450
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 1.01,
|
| 1394 |
+
"learning_rate": 9.929947460595447e-06,
|
| 1395 |
+
"loss": 0.2315,
|
| 1396 |
+
"step": 11500
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.01,
|
| 1400 |
+
"learning_rate": 9.886164623467602e-06,
|
| 1401 |
+
"loss": 0.2317,
|
| 1402 |
+
"step": 11550
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.02,
|
| 1406 |
+
"learning_rate": 9.842381786339756e-06,
|
| 1407 |
+
"loss": 0.2314,
|
| 1408 |
+
"step": 11600
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 1.02,
|
| 1412 |
+
"learning_rate": 9.79859894921191e-06,
|
| 1413 |
+
"loss": 0.2392,
|
| 1414 |
+
"step": 11650
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 1.02,
|
| 1418 |
+
"learning_rate": 9.754816112084064e-06,
|
| 1419 |
+
"loss": 0.2311,
|
| 1420 |
+
"step": 11700
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 1.03,
|
| 1424 |
+
"learning_rate": 9.711033274956218e-06,
|
| 1425 |
+
"loss": 0.223,
|
| 1426 |
+
"step": 11750
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 1.03,
|
| 1430 |
+
"learning_rate": 9.667250437828372e-06,
|
| 1431 |
+
"loss": 0.2176,
|
| 1432 |
+
"step": 11800
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 1.04,
|
| 1436 |
+
"learning_rate": 9.623467600700526e-06,
|
| 1437 |
+
"loss": 0.2311,
|
| 1438 |
+
"step": 11850
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.04,
|
| 1442 |
+
"learning_rate": 9.57968476357268e-06,
|
| 1443 |
+
"loss": 0.2242,
|
| 1444 |
+
"step": 11900
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.05,
|
| 1448 |
+
"learning_rate": 9.535901926444835e-06,
|
| 1449 |
+
"loss": 0.2203,
|
| 1450 |
+
"step": 11950
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 1.05,
|
| 1454 |
+
"learning_rate": 9.492119089316989e-06,
|
| 1455 |
+
"loss": 0.2301,
|
| 1456 |
+
"step": 12000
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 1.06,
|
| 1460 |
+
"learning_rate": 9.448336252189143e-06,
|
| 1461 |
+
"loss": 0.2356,
|
| 1462 |
+
"step": 12050
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 1.06,
|
| 1466 |
+
"learning_rate": 9.404553415061297e-06,
|
| 1467 |
+
"loss": 0.2261,
|
| 1468 |
+
"step": 12100
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 1.06,
|
| 1472 |
+
"learning_rate": 9.360770577933451e-06,
|
| 1473 |
+
"loss": 0.2269,
|
| 1474 |
+
"step": 12150
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 1.07,
|
| 1478 |
+
"learning_rate": 9.316987740805605e-06,
|
| 1479 |
+
"loss": 0.2224,
|
| 1480 |
+
"step": 12200
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.07,
|
| 1484 |
+
"learning_rate": 9.27320490367776e-06,
|
| 1485 |
+
"loss": 0.2214,
|
| 1486 |
+
"step": 12250
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.08,
|
| 1490 |
+
"learning_rate": 9.229422066549914e-06,
|
| 1491 |
+
"loss": 0.2327,
|
| 1492 |
+
"step": 12300
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 1.08,
|
| 1496 |
+
"learning_rate": 9.185639229422068e-06,
|
| 1497 |
+
"loss": 0.2245,
|
| 1498 |
+
"step": 12350
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 1.09,
|
| 1502 |
+
"learning_rate": 9.141856392294222e-06,
|
| 1503 |
+
"loss": 0.2304,
|
| 1504 |
+
"step": 12400
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 1.09,
|
| 1508 |
+
"learning_rate": 9.098073555166376e-06,
|
| 1509 |
+
"loss": 0.2387,
|
| 1510 |
+
"step": 12450
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 1.09,
|
| 1514 |
+
"learning_rate": 9.05429071803853e-06,
|
| 1515 |
+
"loss": 0.2285,
|
| 1516 |
+
"step": 12500
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 1.1,
|
| 1520 |
+
"learning_rate": 9.010507880910684e-06,
|
| 1521 |
+
"loss": 0.2274,
|
| 1522 |
+
"step": 12550
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.1,
|
| 1526 |
+
"learning_rate": 8.966725043782838e-06,
|
| 1527 |
+
"loss": 0.2278,
|
| 1528 |
+
"step": 12600
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.11,
|
| 1532 |
+
"learning_rate": 8.922942206654993e-06,
|
| 1533 |
+
"loss": 0.2277,
|
| 1534 |
+
"step": 12650
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 1.11,
|
| 1538 |
+
"learning_rate": 8.879159369527147e-06,
|
| 1539 |
+
"loss": 0.2261,
|
| 1540 |
+
"step": 12700
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.12,
|
| 1544 |
+
"learning_rate": 8.8353765323993e-06,
|
| 1545 |
+
"loss": 0.2274,
|
| 1546 |
+
"step": 12750
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 1.12,
|
| 1550 |
+
"learning_rate": 8.791593695271455e-06,
|
| 1551 |
+
"loss": 0.227,
|
| 1552 |
+
"step": 12800
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 1.13,
|
| 1556 |
+
"learning_rate": 8.747810858143609e-06,
|
| 1557 |
+
"loss": 0.2363,
|
| 1558 |
+
"step": 12850
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 1.13,
|
| 1562 |
+
"learning_rate": 8.704028021015763e-06,
|
| 1563 |
+
"loss": 0.2244,
|
| 1564 |
+
"step": 12900
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.13,
|
| 1568 |
+
"learning_rate": 8.660245183887916e-06,
|
| 1569 |
+
"loss": 0.2272,
|
| 1570 |
+
"step": 12950
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.14,
|
| 1574 |
+
"learning_rate": 8.61646234676007e-06,
|
| 1575 |
+
"loss": 0.2277,
|
| 1576 |
+
"step": 13000
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 1.14,
|
| 1580 |
+
"learning_rate": 8.572679509632224e-06,
|
| 1581 |
+
"loss": 0.232,
|
| 1582 |
+
"step": 13050
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 1.15,
|
| 1586 |
+
"learning_rate": 8.528896672504378e-06,
|
| 1587 |
+
"loss": 0.2309,
|
| 1588 |
+
"step": 13100
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 1.15,
|
| 1592 |
+
"learning_rate": 8.485113835376532e-06,
|
| 1593 |
+
"loss": 0.2292,
|
| 1594 |
+
"step": 13150
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 1.16,
|
| 1598 |
+
"learning_rate": 8.441330998248686e-06,
|
| 1599 |
+
"loss": 0.2308,
|
| 1600 |
+
"step": 13200
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 1.16,
|
| 1604 |
+
"learning_rate": 8.39754816112084e-06,
|
| 1605 |
+
"loss": 0.2291,
|
| 1606 |
+
"step": 13250
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.16,
|
| 1610 |
+
"learning_rate": 8.353765323992995e-06,
|
| 1611 |
+
"loss": 0.2326,
|
| 1612 |
+
"step": 13300
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 1.17,
|
| 1616 |
+
"learning_rate": 8.309982486865149e-06,
|
| 1617 |
+
"loss": 0.2215,
|
| 1618 |
+
"step": 13350
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 1.17,
|
| 1622 |
+
"learning_rate": 8.266199649737303e-06,
|
| 1623 |
+
"loss": 0.2312,
|
| 1624 |
+
"step": 13400
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 1.18,
|
| 1628 |
+
"learning_rate": 8.222416812609457e-06,
|
| 1629 |
+
"loss": 0.2216,
|
| 1630 |
+
"step": 13450
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 1.18,
|
| 1634 |
+
"learning_rate": 8.178633975481613e-06,
|
| 1635 |
+
"loss": 0.2269,
|
| 1636 |
+
"step": 13500
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 1.19,
|
| 1640 |
+
"learning_rate": 8.134851138353767e-06,
|
| 1641 |
+
"loss": 0.229,
|
| 1642 |
+
"step": 13550
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 1.19,
|
| 1646 |
+
"learning_rate": 8.091068301225921e-06,
|
| 1647 |
+
"loss": 0.2319,
|
| 1648 |
+
"step": 13600
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.2,
|
| 1652 |
+
"learning_rate": 8.047285464098075e-06,
|
| 1653 |
+
"loss": 0.226,
|
| 1654 |
+
"step": 13650
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 1.2,
|
| 1658 |
+
"learning_rate": 8.00350262697023e-06,
|
| 1659 |
+
"loss": 0.2294,
|
| 1660 |
+
"step": 13700
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 1.2,
|
| 1664 |
+
"learning_rate": 7.959719789842383e-06,
|
| 1665 |
+
"loss": 0.2347,
|
| 1666 |
+
"step": 13750
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 1.21,
|
| 1670 |
+
"learning_rate": 7.915936952714538e-06,
|
| 1671 |
+
"loss": 0.2315,
|
| 1672 |
+
"step": 13800
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 1.21,
|
| 1676 |
+
"learning_rate": 7.872154115586692e-06,
|
| 1677 |
+
"loss": 0.2209,
|
| 1678 |
+
"step": 13850
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 1.22,
|
| 1682 |
+
"learning_rate": 7.828371278458846e-06,
|
| 1683 |
+
"loss": 0.2248,
|
| 1684 |
+
"step": 13900
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 1.22,
|
| 1688 |
+
"learning_rate": 7.784588441330998e-06,
|
| 1689 |
+
"loss": 0.2305,
|
| 1690 |
+
"step": 13950
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.23,
|
| 1694 |
+
"learning_rate": 7.740805604203152e-06,
|
| 1695 |
+
"loss": 0.2288,
|
| 1696 |
+
"step": 14000
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 1.23,
|
| 1700 |
+
"learning_rate": 7.697022767075307e-06,
|
| 1701 |
+
"loss": 0.2286,
|
| 1702 |
+
"step": 14050
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 1.23,
|
| 1706 |
+
"learning_rate": 7.65323992994746e-06,
|
| 1707 |
+
"loss": 0.2274,
|
| 1708 |
+
"step": 14100
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 1.24,
|
| 1712 |
+
"learning_rate": 7.609457092819616e-06,
|
| 1713 |
+
"loss": 0.2235,
|
| 1714 |
+
"step": 14150
|
| 1715 |
+
},
|
| 1716 |
+
{
|
| 1717 |
+
"epoch": 1.24,
|
| 1718 |
+
"learning_rate": 7.56567425569177e-06,
|
| 1719 |
+
"loss": 0.2328,
|
| 1720 |
+
"step": 14200
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 1.25,
|
| 1724 |
+
"learning_rate": 7.521891418563924e-06,
|
| 1725 |
+
"loss": 0.2267,
|
| 1726 |
+
"step": 14250
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 1.25,
|
| 1730 |
+
"learning_rate": 7.478108581436077e-06,
|
| 1731 |
+
"loss": 0.2231,
|
| 1732 |
+
"step": 14300
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.26,
|
| 1736 |
+
"learning_rate": 7.434325744308231e-06,
|
| 1737 |
+
"loss": 0.2256,
|
| 1738 |
+
"step": 14350
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.26,
|
| 1742 |
+
"learning_rate": 7.3905429071803855e-06,
|
| 1743 |
+
"loss": 0.2276,
|
| 1744 |
+
"step": 14400
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 1.27,
|
| 1748 |
+
"learning_rate": 7.34676007005254e-06,
|
| 1749 |
+
"loss": 0.2312,
|
| 1750 |
+
"step": 14450
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 1.27,
|
| 1754 |
+
"learning_rate": 7.302977232924694e-06,
|
| 1755 |
+
"loss": 0.2307,
|
| 1756 |
+
"step": 14500
|
| 1757 |
+
},
|
| 1758 |
+
{
|
| 1759 |
+
"epoch": 1.27,
|
| 1760 |
+
"learning_rate": 7.259194395796848e-06,
|
| 1761 |
+
"loss": 0.2265,
|
| 1762 |
+
"step": 14550
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 1.28,
|
| 1766 |
+
"learning_rate": 7.215411558669002e-06,
|
| 1767 |
+
"loss": 0.2351,
|
| 1768 |
+
"step": 14600
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 1.28,
|
| 1772 |
+
"learning_rate": 7.171628721541156e-06,
|
| 1773 |
+
"loss": 0.2264,
|
| 1774 |
+
"step": 14650
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.29,
|
| 1778 |
+
"learning_rate": 7.12784588441331e-06,
|
| 1779 |
+
"loss": 0.2309,
|
| 1780 |
+
"step": 14700
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 1.29,
|
| 1784 |
+
"learning_rate": 7.0840630472854645e-06,
|
| 1785 |
+
"loss": 0.2325,
|
| 1786 |
+
"step": 14750
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 1.3,
|
| 1790 |
+
"learning_rate": 7.040280210157619e-06,
|
| 1791 |
+
"loss": 0.2329,
|
| 1792 |
+
"step": 14800
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 1.3,
|
| 1796 |
+
"learning_rate": 6.996497373029773e-06,
|
| 1797 |
+
"loss": 0.2301,
|
| 1798 |
+
"step": 14850
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 1.3,
|
| 1802 |
+
"learning_rate": 6.952714535901927e-06,
|
| 1803 |
+
"loss": 0.2308,
|
| 1804 |
+
"step": 14900
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 1.31,
|
| 1808 |
+
"learning_rate": 6.908931698774081e-06,
|
| 1809 |
+
"loss": 0.2273,
|
| 1810 |
+
"step": 14950
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 1.31,
|
| 1814 |
+
"learning_rate": 6.865148861646235e-06,
|
| 1815 |
+
"loss": 0.2347,
|
| 1816 |
+
"step": 15000
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.32,
|
| 1820 |
+
"learning_rate": 6.8213660245183884e-06,
|
| 1821 |
+
"loss": 0.2303,
|
| 1822 |
+
"step": 15050
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 1.32,
|
| 1826 |
+
"learning_rate": 6.777583187390544e-06,
|
| 1827 |
+
"loss": 0.2293,
|
| 1828 |
+
"step": 15100
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 1.33,
|
| 1832 |
+
"learning_rate": 6.733800350262698e-06,
|
| 1833 |
+
"loss": 0.2299,
|
| 1834 |
+
"step": 15150
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 1.33,
|
| 1838 |
+
"learning_rate": 6.6900175131348525e-06,
|
| 1839 |
+
"loss": 0.2271,
|
| 1840 |
+
"step": 15200
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 1.34,
|
| 1844 |
+
"learning_rate": 6.646234676007006e-06,
|
| 1845 |
+
"loss": 0.2246,
|
| 1846 |
+
"step": 15250
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 1.34,
|
| 1850 |
+
"learning_rate": 6.60245183887916e-06,
|
| 1851 |
+
"loss": 0.2262,
|
| 1852 |
+
"step": 15300
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 1.34,
|
| 1856 |
+
"learning_rate": 6.558669001751314e-06,
|
| 1857 |
+
"loss": 0.2176,
|
| 1858 |
+
"step": 15350
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.35,
|
| 1862 |
+
"learning_rate": 6.514886164623468e-06,
|
| 1863 |
+
"loss": 0.2277,
|
| 1864 |
+
"step": 15400
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 1.35,
|
| 1868 |
+
"learning_rate": 6.471103327495622e-06,
|
| 1869 |
+
"loss": 0.2184,
|
| 1870 |
+
"step": 15450
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 1.36,
|
| 1874 |
+
"learning_rate": 6.4273204903677765e-06,
|
| 1875 |
+
"loss": 0.2365,
|
| 1876 |
+
"step": 15500
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 1.36,
|
| 1880 |
+
"learning_rate": 6.383537653239931e-06,
|
| 1881 |
+
"loss": 0.2288,
|
| 1882 |
+
"step": 15550
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 1.37,
|
| 1886 |
+
"learning_rate": 6.339754816112085e-06,
|
| 1887 |
+
"loss": 0.2252,
|
| 1888 |
+
"step": 15600
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 1.37,
|
| 1892 |
+
"learning_rate": 6.295971978984239e-06,
|
| 1893 |
+
"loss": 0.2293,
|
| 1894 |
+
"step": 15650
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 1.37,
|
| 1898 |
+
"learning_rate": 6.252189141856393e-06,
|
| 1899 |
+
"loss": 0.2252,
|
| 1900 |
+
"step": 15700
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 1.38,
|
| 1904 |
+
"learning_rate": 6.208406304728547e-06,
|
| 1905 |
+
"loss": 0.2382,
|
| 1906 |
+
"step": 15750
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 1.38,
|
| 1910 |
+
"learning_rate": 6.164623467600701e-06,
|
| 1911 |
+
"loss": 0.2293,
|
| 1912 |
+
"step": 15800
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 1.39,
|
| 1916 |
+
"learning_rate": 6.120840630472855e-06,
|
| 1917 |
+
"loss": 0.2274,
|
| 1918 |
+
"step": 15850
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 1.39,
|
| 1922 |
+
"learning_rate": 6.0770577933450096e-06,
|
| 1923 |
+
"loss": 0.2227,
|
| 1924 |
+
"step": 15900
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 1.4,
|
| 1928 |
+
"learning_rate": 6.033274956217164e-06,
|
| 1929 |
+
"loss": 0.2311,
|
| 1930 |
+
"step": 15950
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 1.4,
|
| 1934 |
+
"learning_rate": 5.989492119089317e-06,
|
| 1935 |
+
"loss": 0.2227,
|
| 1936 |
+
"step": 16000
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 1.41,
|
| 1940 |
+
"learning_rate": 5.945709281961471e-06,
|
| 1941 |
+
"loss": 0.2242,
|
| 1942 |
+
"step": 16050
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 1.41,
|
| 1946 |
+
"learning_rate": 5.901926444833625e-06,
|
| 1947 |
+
"loss": 0.2341,
|
| 1948 |
+
"step": 16100
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 1.41,
|
| 1952 |
+
"learning_rate": 5.858143607705779e-06,
|
| 1953 |
+
"loss": 0.2341,
|
| 1954 |
+
"step": 16150
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 1.42,
|
| 1958 |
+
"learning_rate": 5.8143607705779335e-06,
|
| 1959 |
+
"loss": 0.2306,
|
| 1960 |
+
"step": 16200
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 1.42,
|
| 1964 |
+
"learning_rate": 5.770577933450088e-06,
|
| 1965 |
+
"loss": 0.224,
|
| 1966 |
+
"step": 16250
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 1.43,
|
| 1970 |
+
"learning_rate": 5.726795096322242e-06,
|
| 1971 |
+
"loss": 0.2283,
|
| 1972 |
+
"step": 16300
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 1.43,
|
| 1976 |
+
"learning_rate": 5.683012259194396e-06,
|
| 1977 |
+
"loss": 0.2266,
|
| 1978 |
+
"step": 16350
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 1.44,
|
| 1982 |
+
"learning_rate": 5.63922942206655e-06,
|
| 1983 |
+
"loss": 0.2283,
|
| 1984 |
+
"step": 16400
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 1.44,
|
| 1988 |
+
"learning_rate": 5.595446584938704e-06,
|
| 1989 |
+
"loss": 0.2296,
|
| 1990 |
+
"step": 16450
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 1.44,
|
| 1994 |
+
"learning_rate": 5.551663747810858e-06,
|
| 1995 |
+
"loss": 0.2214,
|
| 1996 |
+
"step": 16500
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 1.45,
|
| 2000 |
+
"learning_rate": 5.5078809106830125e-06,
|
| 2001 |
+
"loss": 0.2229,
|
| 2002 |
+
"step": 16550
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 1.45,
|
| 2006 |
+
"learning_rate": 5.464098073555167e-06,
|
| 2007 |
+
"loss": 0.2333,
|
| 2008 |
+
"step": 16600
|
| 2009 |
+
},
|
| 2010 |
+
{
|
| 2011 |
+
"epoch": 1.46,
|
| 2012 |
+
"learning_rate": 5.420315236427321e-06,
|
| 2013 |
+
"loss": 0.2276,
|
| 2014 |
+
"step": 16650
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 1.46,
|
| 2018 |
+
"learning_rate": 5.376532399299476e-06,
|
| 2019 |
+
"loss": 0.2261,
|
| 2020 |
+
"step": 16700
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 1.47,
|
| 2024 |
+
"learning_rate": 5.33274956217163e-06,
|
| 2025 |
+
"loss": 0.2305,
|
| 2026 |
+
"step": 16750
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.47,
|
| 2030 |
+
"learning_rate": 5.288966725043784e-06,
|
| 2031 |
+
"loss": 0.2314,
|
| 2032 |
+
"step": 16800
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 1.48,
|
| 2036 |
+
"learning_rate": 5.245183887915938e-06,
|
| 2037 |
+
"loss": 0.2269,
|
| 2038 |
+
"step": 16850
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 1.48,
|
| 2042 |
+
"learning_rate": 5.201401050788092e-06,
|
| 2043 |
+
"loss": 0.2313,
|
| 2044 |
+
"step": 16900
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 1.48,
|
| 2048 |
+
"learning_rate": 5.157618213660246e-06,
|
| 2049 |
+
"loss": 0.2341,
|
| 2050 |
+
"step": 16950
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 1.49,
|
| 2054 |
+
"learning_rate": 5.1138353765324e-06,
|
| 2055 |
+
"loss": 0.225,
|
| 2056 |
+
"step": 17000
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 1.49,
|
| 2060 |
+
"learning_rate": 5.070052539404554e-06,
|
| 2061 |
+
"loss": 0.2189,
|
| 2062 |
+
"step": 17050
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 1.5,
|
| 2066 |
+
"learning_rate": 5.026269702276708e-06,
|
| 2067 |
+
"loss": 0.231,
|
| 2068 |
+
"step": 17100
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 1.5,
|
| 2072 |
+
"learning_rate": 4.982486865148862e-06,
|
| 2073 |
+
"loss": 0.229,
|
| 2074 |
+
"step": 17150
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 1.51,
|
| 2078 |
+
"learning_rate": 4.938704028021016e-06,
|
| 2079 |
+
"loss": 0.2324,
|
| 2080 |
+
"step": 17200
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 1.51,
|
| 2084 |
+
"learning_rate": 4.89492119089317e-06,
|
| 2085 |
+
"loss": 0.2303,
|
| 2086 |
+
"step": 17250
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 1.51,
|
| 2090 |
+
"learning_rate": 4.8511383537653245e-06,
|
| 2091 |
+
"loss": 0.2274,
|
| 2092 |
+
"step": 17300
|
| 2093 |
+
},
|
| 2094 |
+
{
|
| 2095 |
+
"epoch": 1.52,
|
| 2096 |
+
"learning_rate": 4.807355516637479e-06,
|
| 2097 |
+
"loss": 0.2244,
|
| 2098 |
+
"step": 17350
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 1.52,
|
| 2102 |
+
"learning_rate": 4.763572679509633e-06,
|
| 2103 |
+
"loss": 0.2273,
|
| 2104 |
+
"step": 17400
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 1.53,
|
| 2108 |
+
"learning_rate": 4.719789842381787e-06,
|
| 2109 |
+
"loss": 0.2265,
|
| 2110 |
+
"step": 17450
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 1.53,
|
| 2114 |
+
"learning_rate": 4.676007005253941e-06,
|
| 2115 |
+
"loss": 0.2253,
|
| 2116 |
+
"step": 17500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 1.54,
|
| 2120 |
+
"learning_rate": 4.632224168126095e-06,
|
| 2121 |
+
"loss": 0.2349,
|
| 2122 |
+
"step": 17550
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 1.54,
|
| 2126 |
+
"learning_rate": 4.588441330998249e-06,
|
| 2127 |
+
"loss": 0.2202,
|
| 2128 |
+
"step": 17600
|
| 2129 |
+
},
|
| 2130 |
+
{
|
| 2131 |
+
"epoch": 1.55,
|
| 2132 |
+
"learning_rate": 4.544658493870403e-06,
|
| 2133 |
+
"loss": 0.2304,
|
| 2134 |
+
"step": 17650
|
| 2135 |
+
},
|
| 2136 |
+
{
|
| 2137 |
+
"epoch": 1.55,
|
| 2138 |
+
"learning_rate": 4.500875656742557e-06,
|
| 2139 |
+
"loss": 0.2321,
|
| 2140 |
+
"step": 17700
|
| 2141 |
+
},
|
| 2142 |
+
{
|
| 2143 |
+
"epoch": 1.55,
|
| 2144 |
+
"learning_rate": 4.457092819614711e-06,
|
| 2145 |
+
"loss": 0.2194,
|
| 2146 |
+
"step": 17750
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 1.56,
|
| 2150 |
+
"learning_rate": 4.413309982486865e-06,
|
| 2151 |
+
"loss": 0.2278,
|
| 2152 |
+
"step": 17800
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 1.56,
|
| 2156 |
+
"learning_rate": 4.36952714535902e-06,
|
| 2157 |
+
"loss": 0.2237,
|
| 2158 |
+
"step": 17850
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 1.57,
|
| 2162 |
+
"learning_rate": 4.325744308231174e-06,
|
| 2163 |
+
"loss": 0.2241,
|
| 2164 |
+
"step": 17900
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 1.57,
|
| 2168 |
+
"learning_rate": 4.281961471103328e-06,
|
| 2169 |
+
"loss": 0.2244,
|
| 2170 |
+
"step": 17950
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 1.58,
|
| 2174 |
+
"learning_rate": 4.238178633975482e-06,
|
| 2175 |
+
"loss": 0.2247,
|
| 2176 |
+
"step": 18000
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 1.58,
|
| 2180 |
+
"learning_rate": 4.1943957968476365e-06,
|
| 2181 |
+
"loss": 0.2307,
|
| 2182 |
+
"step": 18050
|
| 2183 |
+
},
|
| 2184 |
+
{
|
| 2185 |
+
"epoch": 1.58,
|
| 2186 |
+
"learning_rate": 4.150612959719791e-06,
|
| 2187 |
+
"loss": 0.2212,
|
| 2188 |
+
"step": 18100
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 1.59,
|
| 2192 |
+
"learning_rate": 4.106830122591945e-06,
|
| 2193 |
+
"loss": 0.224,
|
| 2194 |
+
"step": 18150
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 1.59,
|
| 2198 |
+
"learning_rate": 4.063047285464098e-06,
|
| 2199 |
+
"loss": 0.221,
|
| 2200 |
+
"step": 18200
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 1.6,
|
| 2204 |
+
"learning_rate": 4.019264448336252e-06,
|
| 2205 |
+
"loss": 0.2237,
|
| 2206 |
+
"step": 18250
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 1.6,
|
| 2210 |
+
"learning_rate": 3.975481611208406e-06,
|
| 2211 |
+
"loss": 0.2294,
|
| 2212 |
+
"step": 18300
|
| 2213 |
+
},
|
| 2214 |
+
{
|
| 2215 |
+
"epoch": 1.61,
|
| 2216 |
+
"learning_rate": 3.9316987740805604e-06,
|
| 2217 |
+
"loss": 0.222,
|
| 2218 |
+
"step": 18350
|
| 2219 |
+
},
|
| 2220 |
+
{
|
| 2221 |
+
"epoch": 1.61,
|
| 2222 |
+
"learning_rate": 3.8879159369527146e-06,
|
| 2223 |
+
"loss": 0.2235,
|
| 2224 |
+
"step": 18400
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 1.62,
|
| 2228 |
+
"learning_rate": 3.844133099824869e-06,
|
| 2229 |
+
"loss": 0.2299,
|
| 2230 |
+
"step": 18450
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 1.62,
|
| 2234 |
+
"learning_rate": 3.800350262697023e-06,
|
| 2235 |
+
"loss": 0.2308,
|
| 2236 |
+
"step": 18500
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 1.62,
|
| 2240 |
+
"learning_rate": 3.756567425569177e-06,
|
| 2241 |
+
"loss": 0.2279,
|
| 2242 |
+
"step": 18550
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 1.63,
|
| 2246 |
+
"learning_rate": 3.712784588441331e-06,
|
| 2247 |
+
"loss": 0.2293,
|
| 2248 |
+
"step": 18600
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 1.63,
|
| 2252 |
+
"learning_rate": 3.6690017513134857e-06,
|
| 2253 |
+
"loss": 0.2279,
|
| 2254 |
+
"step": 18650
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 1.64,
|
| 2258 |
+
"learning_rate": 3.62521891418564e-06,
|
| 2259 |
+
"loss": 0.2292,
|
| 2260 |
+
"step": 18700
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 1.64,
|
| 2264 |
+
"learning_rate": 3.581436077057794e-06,
|
| 2265 |
+
"loss": 0.2276,
|
| 2266 |
+
"step": 18750
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 1.65,
|
| 2270 |
+
"learning_rate": 3.537653239929948e-06,
|
| 2271 |
+
"loss": 0.2276,
|
| 2272 |
+
"step": 18800
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 1.65,
|
| 2276 |
+
"learning_rate": 3.4938704028021018e-06,
|
| 2277 |
+
"loss": 0.2209,
|
| 2278 |
+
"step": 18850
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 1.65,
|
| 2282 |
+
"learning_rate": 3.450087565674256e-06,
|
| 2283 |
+
"loss": 0.2251,
|
| 2284 |
+
"step": 18900
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 1.66,
|
| 2288 |
+
"learning_rate": 3.40630472854641e-06,
|
| 2289 |
+
"loss": 0.2266,
|
| 2290 |
+
"step": 18950
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 1.66,
|
| 2294 |
+
"learning_rate": 3.362521891418564e-06,
|
| 2295 |
+
"loss": 0.228,
|
| 2296 |
+
"step": 19000
|
| 2297 |
+
},
|
| 2298 |
+
{
|
| 2299 |
+
"epoch": 1.67,
|
| 2300 |
+
"learning_rate": 3.3187390542907183e-06,
|
| 2301 |
+
"loss": 0.2274,
|
| 2302 |
+
"step": 19050
|
| 2303 |
+
},
|
| 2304 |
+
{
|
| 2305 |
+
"epoch": 1.67,
|
| 2306 |
+
"learning_rate": 3.2749562171628725e-06,
|
| 2307 |
+
"loss": 0.2256,
|
| 2308 |
+
"step": 19100
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 1.68,
|
| 2312 |
+
"learning_rate": 3.2311733800350266e-06,
|
| 2313 |
+
"loss": 0.2279,
|
| 2314 |
+
"step": 19150
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 1.68,
|
| 2318 |
+
"learning_rate": 3.1873905429071807e-06,
|
| 2319 |
+
"loss": 0.2254,
|
| 2320 |
+
"step": 19200
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 1.69,
|
| 2324 |
+
"learning_rate": 3.1436077057793344e-06,
|
| 2325 |
+
"loss": 0.2273,
|
| 2326 |
+
"step": 19250
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 1.69,
|
| 2330 |
+
"learning_rate": 3.0998248686514886e-06,
|
| 2331 |
+
"loss": 0.2277,
|
| 2332 |
+
"step": 19300
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 1.69,
|
| 2336 |
+
"learning_rate": 3.0560420315236427e-06,
|
| 2337 |
+
"loss": 0.2291,
|
| 2338 |
+
"step": 19350
|
| 2339 |
+
},
|
| 2340 |
+
{
|
| 2341 |
+
"epoch": 1.7,
|
| 2342 |
+
"learning_rate": 3.012259194395797e-06,
|
| 2343 |
+
"loss": 0.2258,
|
| 2344 |
+
"step": 19400
|
| 2345 |
+
},
|
| 2346 |
+
{
|
| 2347 |
+
"epoch": 1.7,
|
| 2348 |
+
"learning_rate": 2.9684763572679514e-06,
|
| 2349 |
+
"loss": 0.2247,
|
| 2350 |
+
"step": 19450
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 1.71,
|
| 2354 |
+
"learning_rate": 2.9246935201401055e-06,
|
| 2355 |
+
"loss": 0.2246,
|
| 2356 |
+
"step": 19500
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 1.71,
|
| 2360 |
+
"learning_rate": 2.8809106830122597e-06,
|
| 2361 |
+
"loss": 0.2228,
|
| 2362 |
+
"step": 19550
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 1.72,
|
| 2366 |
+
"learning_rate": 2.837127845884414e-06,
|
| 2367 |
+
"loss": 0.2281,
|
| 2368 |
+
"step": 19600
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 1.72,
|
| 2372 |
+
"learning_rate": 2.793345008756568e-06,
|
| 2373 |
+
"loss": 0.2396,
|
| 2374 |
+
"step": 19650
|
| 2375 |
+
},
|
| 2376 |
+
{
|
| 2377 |
+
"epoch": 1.73,
|
| 2378 |
+
"learning_rate": 2.749562171628722e-06,
|
| 2379 |
+
"loss": 0.2214,
|
| 2380 |
+
"step": 19700
|
| 2381 |
+
},
|
| 2382 |
+
{
|
| 2383 |
+
"epoch": 1.73,
|
| 2384 |
+
"learning_rate": 2.7057793345008758e-06,
|
| 2385 |
+
"loss": 0.2286,
|
| 2386 |
+
"step": 19750
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"epoch": 1.73,
|
| 2390 |
+
"learning_rate": 2.66199649737303e-06,
|
| 2391 |
+
"loss": 0.2289,
|
| 2392 |
+
"step": 19800
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 1.74,
|
| 2396 |
+
"learning_rate": 2.618213660245184e-06,
|
| 2397 |
+
"loss": 0.2282,
|
| 2398 |
+
"step": 19850
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 1.74,
|
| 2402 |
+
"learning_rate": 2.574430823117338e-06,
|
| 2403 |
+
"loss": 0.2299,
|
| 2404 |
+
"step": 19900
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 1.75,
|
| 2408 |
+
"learning_rate": 2.5306479859894923e-06,
|
| 2409 |
+
"loss": 0.2276,
|
| 2410 |
+
"step": 19950
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 1.75,
|
| 2414 |
+
"learning_rate": 2.4868651488616464e-06,
|
| 2415 |
+
"loss": 0.2257,
|
| 2416 |
+
"step": 20000
|
| 2417 |
+
},
|
| 2418 |
+
{
|
| 2419 |
+
"epoch": 1.76,
|
| 2420 |
+
"learning_rate": 2.4430823117338006e-06,
|
| 2421 |
+
"loss": 0.228,
|
| 2422 |
+
"step": 20050
|
| 2423 |
+
},
|
| 2424 |
+
{
|
| 2425 |
+
"epoch": 1.76,
|
| 2426 |
+
"learning_rate": 2.3992994746059547e-06,
|
| 2427 |
+
"loss": 0.2245,
|
| 2428 |
+
"step": 20100
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 1.76,
|
| 2432 |
+
"learning_rate": 2.355516637478109e-06,
|
| 2433 |
+
"loss": 0.2199,
|
| 2434 |
+
"step": 20150
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 1.77,
|
| 2438 |
+
"learning_rate": 2.311733800350263e-06,
|
| 2439 |
+
"loss": 0.2264,
|
| 2440 |
+
"step": 20200
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 1.77,
|
| 2444 |
+
"learning_rate": 2.267950963222417e-06,
|
| 2445 |
+
"loss": 0.228,
|
| 2446 |
+
"step": 20250
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 1.78,
|
| 2450 |
+
"learning_rate": 2.2241681260945713e-06,
|
| 2451 |
+
"loss": 0.2311,
|
| 2452 |
+
"step": 20300
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 1.78,
|
| 2456 |
+
"learning_rate": 2.180385288966725e-06,
|
| 2457 |
+
"loss": 0.2252,
|
| 2458 |
+
"step": 20350
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 1.79,
|
| 2462 |
+
"learning_rate": 2.136602451838879e-06,
|
| 2463 |
+
"loss": 0.2259,
|
| 2464 |
+
"step": 20400
|
| 2465 |
+
},
|
| 2466 |
+
{
|
| 2467 |
+
"epoch": 1.79,
|
| 2468 |
+
"learning_rate": 2.0928196147110337e-06,
|
| 2469 |
+
"loss": 0.2254,
|
| 2470 |
+
"step": 20450
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 1.8,
|
| 2474 |
+
"learning_rate": 2.0490367775831878e-06,
|
| 2475 |
+
"loss": 0.2273,
|
| 2476 |
+
"step": 20500
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 1.8,
|
| 2480 |
+
"learning_rate": 2.005253940455342e-06,
|
| 2481 |
+
"loss": 0.2188,
|
| 2482 |
+
"step": 20550
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 1.8,
|
| 2486 |
+
"learning_rate": 1.9614711033274956e-06,
|
| 2487 |
+
"loss": 0.2295,
|
| 2488 |
+
"step": 20600
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 1.81,
|
| 2492 |
+
"learning_rate": 1.9176882661996498e-06,
|
| 2493 |
+
"loss": 0.2257,
|
| 2494 |
+
"step": 20650
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 1.81,
|
| 2498 |
+
"learning_rate": 1.873905429071804e-06,
|
| 2499 |
+
"loss": 0.2296,
|
| 2500 |
+
"step": 20700
|
| 2501 |
+
},
|
| 2502 |
+
{
|
| 2503 |
+
"epoch": 1.82,
|
| 2504 |
+
"learning_rate": 1.830122591943958e-06,
|
| 2505 |
+
"loss": 0.2304,
|
| 2506 |
+
"step": 20750
|
| 2507 |
+
},
|
| 2508 |
+
{
|
| 2509 |
+
"epoch": 1.82,
|
| 2510 |
+
"learning_rate": 1.7863397548161122e-06,
|
| 2511 |
+
"loss": 0.2316,
|
| 2512 |
+
"step": 20800
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 1.83,
|
| 2516 |
+
"learning_rate": 1.7425569176882665e-06,
|
| 2517 |
+
"loss": 0.2311,
|
| 2518 |
+
"step": 20850
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"epoch": 1.83,
|
| 2522 |
+
"learning_rate": 1.6987740805604204e-06,
|
| 2523 |
+
"loss": 0.2254,
|
| 2524 |
+
"step": 20900
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 1.83,
|
| 2528 |
+
"learning_rate": 1.6549912434325746e-06,
|
| 2529 |
+
"loss": 0.2234,
|
| 2530 |
+
"step": 20950
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 1.84,
|
| 2534 |
+
"learning_rate": 1.6112084063047287e-06,
|
| 2535 |
+
"loss": 0.2193,
|
| 2536 |
+
"step": 21000
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 1.84,
|
| 2540 |
+
"learning_rate": 1.5674255691768828e-06,
|
| 2541 |
+
"loss": 0.236,
|
| 2542 |
+
"step": 21050
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 1.85,
|
| 2546 |
+
"learning_rate": 1.5236427320490368e-06,
|
| 2547 |
+
"loss": 0.2282,
|
| 2548 |
+
"step": 21100
|
| 2549 |
+
},
|
| 2550 |
+
{
|
| 2551 |
+
"epoch": 1.85,
|
| 2552 |
+
"learning_rate": 1.479859894921191e-06,
|
| 2553 |
+
"loss": 0.23,
|
| 2554 |
+
"step": 21150
|
| 2555 |
+
},
|
| 2556 |
+
{
|
| 2557 |
+
"epoch": 1.86,
|
| 2558 |
+
"learning_rate": 1.436077057793345e-06,
|
| 2559 |
+
"loss": 0.232,
|
| 2560 |
+
"step": 21200
|
| 2561 |
+
},
|
| 2562 |
+
{
|
| 2563 |
+
"epoch": 1.86,
|
| 2564 |
+
"learning_rate": 1.3922942206654994e-06,
|
| 2565 |
+
"loss": 0.2283,
|
| 2566 |
+
"step": 21250
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 1.87,
|
| 2570 |
+
"learning_rate": 1.3485113835376535e-06,
|
| 2571 |
+
"loss": 0.2334,
|
| 2572 |
+
"step": 21300
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 1.87,
|
| 2576 |
+
"learning_rate": 1.3047285464098074e-06,
|
| 2577 |
+
"loss": 0.2297,
|
| 2578 |
+
"step": 21350
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 1.87,
|
| 2582 |
+
"learning_rate": 1.2609457092819616e-06,
|
| 2583 |
+
"loss": 0.223,
|
| 2584 |
+
"step": 21400
|
| 2585 |
+
},
|
| 2586 |
+
{
|
| 2587 |
+
"epoch": 1.88,
|
| 2588 |
+
"learning_rate": 1.2171628721541157e-06,
|
| 2589 |
+
"loss": 0.223,
|
| 2590 |
+
"step": 21450
|
| 2591 |
+
},
|
| 2592 |
+
{
|
| 2593 |
+
"epoch": 1.88,
|
| 2594 |
+
"learning_rate": 1.1733800350262698e-06,
|
| 2595 |
+
"loss": 0.2252,
|
| 2596 |
+
"step": 21500
|
| 2597 |
+
},
|
| 2598 |
+
{
|
| 2599 |
+
"epoch": 1.89,
|
| 2600 |
+
"learning_rate": 1.129597197898424e-06,
|
| 2601 |
+
"loss": 0.2256,
|
| 2602 |
+
"step": 21550
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 1.89,
|
| 2606 |
+
"learning_rate": 1.085814360770578e-06,
|
| 2607 |
+
"loss": 0.2228,
|
| 2608 |
+
"step": 21600
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 1.9,
|
| 2612 |
+
"learning_rate": 1.042031523642732e-06,
|
| 2613 |
+
"loss": 0.2288,
|
| 2614 |
+
"step": 21650
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 1.9,
|
| 2618 |
+
"learning_rate": 9.982486865148862e-07,
|
| 2619 |
+
"loss": 0.2342,
|
| 2620 |
+
"step": 21700
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 1.9,
|
| 2624 |
+
"learning_rate": 9.544658493870403e-07,
|
| 2625 |
+
"loss": 0.2267,
|
| 2626 |
+
"step": 21750
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 1.91,
|
| 2630 |
+
"learning_rate": 9.106830122591944e-07,
|
| 2631 |
+
"loss": 0.2234,
|
| 2632 |
+
"step": 21800
|
| 2633 |
+
},
|
| 2634 |
+
{
|
| 2635 |
+
"epoch": 1.91,
|
| 2636 |
+
"learning_rate": 8.669001751313486e-07,
|
| 2637 |
+
"loss": 0.2191,
|
| 2638 |
+
"step": 21850
|
| 2639 |
+
},
|
| 2640 |
+
{
|
| 2641 |
+
"epoch": 1.92,
|
| 2642 |
+
"learning_rate": 8.231173380035026e-07,
|
| 2643 |
+
"loss": 0.2219,
|
| 2644 |
+
"step": 21900
|
| 2645 |
+
},
|
| 2646 |
+
{
|
| 2647 |
+
"epoch": 1.92,
|
| 2648 |
+
"learning_rate": 7.793345008756568e-07,
|
| 2649 |
+
"loss": 0.2239,
|
| 2650 |
+
"step": 21950
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 1.93,
|
| 2654 |
+
"learning_rate": 7.35551663747811e-07,
|
| 2655 |
+
"loss": 0.2262,
|
| 2656 |
+
"step": 22000
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 1.93,
|
| 2660 |
+
"learning_rate": 6.91768826619965e-07,
|
| 2661 |
+
"loss": 0.2185,
|
| 2662 |
+
"step": 22050
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 1.94,
|
| 2666 |
+
"learning_rate": 6.47985989492119e-07,
|
| 2667 |
+
"loss": 0.2239,
|
| 2668 |
+
"step": 22100
|
| 2669 |
+
},
|
| 2670 |
+
{
|
| 2671 |
+
"epoch": 1.94,
|
| 2672 |
+
"learning_rate": 6.042031523642733e-07,
|
| 2673 |
+
"loss": 0.2286,
|
| 2674 |
+
"step": 22150
|
| 2675 |
+
},
|
| 2676 |
+
{
|
| 2677 |
+
"epoch": 1.94,
|
| 2678 |
+
"learning_rate": 5.604203152364274e-07,
|
| 2679 |
+
"loss": 0.2253,
|
| 2680 |
+
"step": 22200
|
| 2681 |
+
},
|
| 2682 |
+
{
|
| 2683 |
+
"epoch": 1.95,
|
| 2684 |
+
"learning_rate": 5.166374781085814e-07,
|
| 2685 |
+
"loss": 0.2252,
|
| 2686 |
+
"step": 22250
|
| 2687 |
+
},
|
| 2688 |
+
{
|
| 2689 |
+
"epoch": 1.95,
|
| 2690 |
+
"learning_rate": 4.728546409807356e-07,
|
| 2691 |
+
"loss": 0.2245,
|
| 2692 |
+
"step": 22300
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 1.96,
|
| 2696 |
+
"learning_rate": 4.290718038528897e-07,
|
| 2697 |
+
"loss": 0.2285,
|
| 2698 |
+
"step": 22350
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 1.96,
|
| 2702 |
+
"learning_rate": 3.8528896672504383e-07,
|
| 2703 |
+
"loss": 0.2256,
|
| 2704 |
+
"step": 22400
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 1.97,
|
| 2708 |
+
"learning_rate": 3.415061295971979e-07,
|
| 2709 |
+
"loss": 0.2334,
|
| 2710 |
+
"step": 22450
|
| 2711 |
+
},
|
| 2712 |
+
{
|
| 2713 |
+
"epoch": 1.97,
|
| 2714 |
+
"learning_rate": 2.9772329246935205e-07,
|
| 2715 |
+
"loss": 0.2313,
|
| 2716 |
+
"step": 22500
|
| 2717 |
+
},
|
| 2718 |
+
{
|
| 2719 |
+
"epoch": 1.97,
|
| 2720 |
+
"learning_rate": 2.5394045534150613e-07,
|
| 2721 |
+
"loss": 0.2202,
|
| 2722 |
+
"step": 22550
|
| 2723 |
+
},
|
| 2724 |
+
{
|
| 2725 |
+
"epoch": 1.98,
|
| 2726 |
+
"learning_rate": 2.1015761821366026e-07,
|
| 2727 |
+
"loss": 0.2235,
|
| 2728 |
+
"step": 22600
|
| 2729 |
+
},
|
| 2730 |
+
{
|
| 2731 |
+
"epoch": 1.98,
|
| 2732 |
+
"learning_rate": 1.6637478108581437e-07,
|
| 2733 |
+
"loss": 0.2311,
|
| 2734 |
+
"step": 22650
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 1.99,
|
| 2738 |
+
"learning_rate": 1.2259194395796848e-07,
|
| 2739 |
+
"loss": 0.2213,
|
| 2740 |
+
"step": 22700
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 1.99,
|
| 2744 |
+
"learning_rate": 7.88091068301226e-08,
|
| 2745 |
+
"loss": 0.2246,
|
| 2746 |
+
"step": 22750
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 2.0,
|
| 2750 |
+
"learning_rate": 3.502626970227671e-08,
|
| 2751 |
+
"loss": 0.2186,
|
| 2752 |
+
"step": 22800
|
| 2753 |
+
},
|
| 2754 |
+
{
|
| 2755 |
+
"epoch": 2.0,
|
| 2756 |
+
"eval_loss": 0.20982445776462555,
|
| 2757 |
+
"eval_runtime": 72.0375,
|
| 2758 |
+
"eval_samples_per_second": 63.411,
|
| 2759 |
+
"eval_steps_per_second": 15.853,
|
| 2760 |
+
"step": 22840
|
| 2761 |
+
}
|
| 2762 |
+
],
|
| 2763 |
+
"max_steps": 22840,
|
| 2764 |
+
"num_train_epochs": 2,
|
| 2765 |
+
"total_flos": 1.0789666943297126e+17,
|
| 2766 |
+
"trial_name": null,
|
| 2767 |
+
"trial_params": null
|
| 2768 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c50cadd5fbdd38e61af99ba2434211b0f93750595fee502ba0094fd4bb9990a
|
| 3 |
+
size 3707
|