udpate model
Browse files- .gitattributes +1 -0
- added_tokens.json +3 -0
- config.json +61 -0
- generation_config.json +7 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +108 -0
- spiece.model +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +113 -0
- trainer_state.json +2281 -0
- training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<s>": 32100
|
| 3 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/flan-t5-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 2048,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 768,
|
| 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 |
+
"model_type": "t5",
|
| 19 |
+
"n_positions": 512,
|
| 20 |
+
"num_decoder_layers": 12,
|
| 21 |
+
"num_heads": 12,
|
| 22 |
+
"num_layers": 12,
|
| 23 |
+
"output_past": true,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"relative_attention_max_distance": 128,
|
| 26 |
+
"relative_attention_num_buckets": 32,
|
| 27 |
+
"task_specific_params": {
|
| 28 |
+
"summarization": {
|
| 29 |
+
"early_stopping": true,
|
| 30 |
+
"length_penalty": 2.0,
|
| 31 |
+
"max_length": 200,
|
| 32 |
+
"min_length": 30,
|
| 33 |
+
"no_repeat_ngram_size": 3,
|
| 34 |
+
"num_beams": 4,
|
| 35 |
+
"prefix": "summarize: "
|
| 36 |
+
},
|
| 37 |
+
"translation_en_to_de": {
|
| 38 |
+
"early_stopping": true,
|
| 39 |
+
"max_length": 300,
|
| 40 |
+
"num_beams": 4,
|
| 41 |
+
"prefix": "translate English to German: "
|
| 42 |
+
},
|
| 43 |
+
"translation_en_to_fr": {
|
| 44 |
+
"early_stopping": true,
|
| 45 |
+
"max_length": 300,
|
| 46 |
+
"num_beams": 4,
|
| 47 |
+
"prefix": "translate English to French: "
|
| 48 |
+
},
|
| 49 |
+
"translation_en_to_ro": {
|
| 50 |
+
"early_stopping": true,
|
| 51 |
+
"max_length": 300,
|
| 52 |
+
"num_beams": 4,
|
| 53 |
+
"prefix": "translate English to Romanian: "
|
| 54 |
+
}
|
| 55 |
+
},
|
| 56 |
+
"tie_word_embeddings": false,
|
| 57 |
+
"torch_dtype": "float32",
|
| 58 |
+
"transformers_version": "4.29.1",
|
| 59 |
+
"use_cache": true,
|
| 60 |
+
"vocab_size": 32101
|
| 61 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 1,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.29.1"
|
| 7 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:31a98c4ed9949814ffc96ffb5fea857e335782f2c94f7eab1d311ccae7ccd8ed
|
| 3 |
+
size 990242997
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,108 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"bos_token": "<s>",
|
| 105 |
+
"eos_token": "</s>",
|
| 106 |
+
"pad_token": "<pad>",
|
| 107 |
+
"unk_token": "<unk>"
|
| 108 |
+
}
|
spiece.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d60acb128cf7b7f2536e8f38a5b18a05535c9e14c7a355904270e15b0945ea86
|
| 3 |
+
size 791656
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,113 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": 512,
|
| 108 |
+
"pad_token": "<pad>",
|
| 109 |
+
"padding_side": "right",
|
| 110 |
+
"sp_model_kwargs": {},
|
| 111 |
+
"tokenizer_class": "T5Tokenizer",
|
| 112 |
+
"unk_token": "<unk>"
|
| 113 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2281 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 40.0,
|
| 5 |
+
"global_step": 3760,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.11,
|
| 12 |
+
"learning_rate": 1.3297872340425533e-06,
|
| 13 |
+
"loss": 29.6158,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.21,
|
| 18 |
+
"learning_rate": 2.6595744680851065e-06,
|
| 19 |
+
"loss": 27.1775,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.32,
|
| 24 |
+
"learning_rate": 3.98936170212766e-06,
|
| 25 |
+
"loss": 24.7793,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.43,
|
| 30 |
+
"learning_rate": 5.319148936170213e-06,
|
| 31 |
+
"loss": 23.8977,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.53,
|
| 36 |
+
"learning_rate": 6.648936170212766e-06,
|
| 37 |
+
"loss": 20.9254,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.64,
|
| 42 |
+
"learning_rate": 7.97872340425532e-06,
|
| 43 |
+
"loss": 20.9818,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.74,
|
| 48 |
+
"learning_rate": 9.308510638297872e-06,
|
| 49 |
+
"loss": 16.5909,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.85,
|
| 54 |
+
"learning_rate": 1.0638297872340426e-05,
|
| 55 |
+
"loss": 13.8147,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.96,
|
| 60 |
+
"learning_rate": 1.196808510638298e-05,
|
| 61 |
+
"loss": 10.0983,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 1.06,
|
| 66 |
+
"learning_rate": 1.3297872340425532e-05,
|
| 67 |
+
"loss": 6.1833,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.17,
|
| 72 |
+
"learning_rate": 1.4627659574468085e-05,
|
| 73 |
+
"loss": 3.6727,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.28,
|
| 78 |
+
"learning_rate": 1.595744680851064e-05,
|
| 79 |
+
"loss": 2.9481,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.38,
|
| 84 |
+
"learning_rate": 1.7287234042553193e-05,
|
| 85 |
+
"loss": 2.4666,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.49,
|
| 90 |
+
"learning_rate": 1.8617021276595745e-05,
|
| 91 |
+
"loss": 2.104,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.6,
|
| 96 |
+
"learning_rate": 1.9946808510638297e-05,
|
| 97 |
+
"loss": 1.8897,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7,
|
| 102 |
+
"learning_rate": 2.1276595744680852e-05,
|
| 103 |
+
"loss": 1.5951,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.81,
|
| 108 |
+
"learning_rate": 2.2606382978723408e-05,
|
| 109 |
+
"loss": 1.305,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.91,
|
| 114 |
+
"learning_rate": 2.393617021276596e-05,
|
| 115 |
+
"loss": 0.974,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.02,
|
| 120 |
+
"learning_rate": 2.5265957446808515e-05,
|
| 121 |
+
"loss": 0.7627,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 2.13,
|
| 126 |
+
"learning_rate": 2.6595744680851064e-05,
|
| 127 |
+
"loss": 0.6068,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.23,
|
| 132 |
+
"learning_rate": 2.7925531914893616e-05,
|
| 133 |
+
"loss": 0.4438,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 2.34,
|
| 138 |
+
"learning_rate": 2.925531914893617e-05,
|
| 139 |
+
"loss": 0.3635,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 2.45,
|
| 144 |
+
"learning_rate": 3.0585106382978726e-05,
|
| 145 |
+
"loss": 0.2884,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.55,
|
| 150 |
+
"learning_rate": 3.191489361702128e-05,
|
| 151 |
+
"loss": 0.2461,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.66,
|
| 156 |
+
"learning_rate": 3.324468085106383e-05,
|
| 157 |
+
"loss": 0.2003,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.77,
|
| 162 |
+
"learning_rate": 3.4574468085106386e-05,
|
| 163 |
+
"loss": 0.2406,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.87,
|
| 168 |
+
"learning_rate": 3.590425531914894e-05,
|
| 169 |
+
"loss": 0.1561,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.98,
|
| 174 |
+
"learning_rate": 3.723404255319149e-05,
|
| 175 |
+
"loss": 0.1476,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.09,
|
| 180 |
+
"learning_rate": 3.8563829787234045e-05,
|
| 181 |
+
"loss": 0.1365,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.19,
|
| 186 |
+
"learning_rate": 3.9893617021276594e-05,
|
| 187 |
+
"loss": 0.1181,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.3,
|
| 192 |
+
"learning_rate": 4.1223404255319156e-05,
|
| 193 |
+
"loss": 0.1274,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.4,
|
| 198 |
+
"learning_rate": 4.2553191489361704e-05,
|
| 199 |
+
"loss": 0.1283,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.51,
|
| 204 |
+
"learning_rate": 4.388297872340425e-05,
|
| 205 |
+
"loss": 0.1117,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 4.5212765957446815e-05,
|
| 211 |
+
"loss": 0.1104,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.72,
|
| 216 |
+
"learning_rate": 4.6542553191489364e-05,
|
| 217 |
+
"loss": 0.1041,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.83,
|
| 222 |
+
"learning_rate": 4.787234042553192e-05,
|
| 223 |
+
"loss": 0.1026,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 3.94,
|
| 228 |
+
"learning_rate": 4.920212765957447e-05,
|
| 229 |
+
"loss": 0.1016,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.04,
|
| 234 |
+
"learning_rate": 4.999982762728059e-05,
|
| 235 |
+
"loss": 0.11,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 4.15,
|
| 240 |
+
"learning_rate": 4.99978884614854e-05,
|
| 241 |
+
"loss": 0.0975,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 4.26,
|
| 246 |
+
"learning_rate": 4.999379483168148e-05,
|
| 247 |
+
"loss": 0.0873,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 4.36,
|
| 252 |
+
"learning_rate": 4.9987547090681766e-05,
|
| 253 |
+
"loss": 0.0864,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 4.47,
|
| 258 |
+
"learning_rate": 4.997914577695306e-05,
|
| 259 |
+
"loss": 0.0955,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.57,
|
| 264 |
+
"learning_rate": 4.996859161456965e-05,
|
| 265 |
+
"loss": 0.0933,
|
| 266 |
+
"step": 430
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 4.68,
|
| 270 |
+
"learning_rate": 4.995588551315087e-05,
|
| 271 |
+
"loss": 0.1315,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 4.79,
|
| 276 |
+
"learning_rate": 4.994102856778274e-05,
|
| 277 |
+
"loss": 0.0739,
|
| 278 |
+
"step": 450
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.89,
|
| 282 |
+
"learning_rate": 4.992402205892358e-05,
|
| 283 |
+
"loss": 0.07,
|
| 284 |
+
"step": 460
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 5.0,
|
| 288 |
+
"learning_rate": 4.990486745229364e-05,
|
| 289 |
+
"loss": 0.0766,
|
| 290 |
+
"step": 470
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 5.11,
|
| 294 |
+
"learning_rate": 4.9883566398748774e-05,
|
| 295 |
+
"loss": 0.0718,
|
| 296 |
+
"step": 480
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 5.21,
|
| 300 |
+
"learning_rate": 4.986012073413816e-05,
|
| 301 |
+
"loss": 0.0726,
|
| 302 |
+
"step": 490
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 5.32,
|
| 306 |
+
"learning_rate": 4.98345324791461e-05,
|
| 307 |
+
"loss": 0.0725,
|
| 308 |
+
"step": 500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 5.43,
|
| 312 |
+
"learning_rate": 4.9806803839117845e-05,
|
| 313 |
+
"loss": 0.0813,
|
| 314 |
+
"step": 510
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 5.53,
|
| 318 |
+
"learning_rate": 4.97769372038695e-05,
|
| 319 |
+
"loss": 0.074,
|
| 320 |
+
"step": 520
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.64,
|
| 324 |
+
"learning_rate": 4.974493514748213e-05,
|
| 325 |
+
"loss": 0.0686,
|
| 326 |
+
"step": 530
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 5.74,
|
| 330 |
+
"learning_rate": 4.971080042807981e-05,
|
| 331 |
+
"loss": 0.0754,
|
| 332 |
+
"step": 540
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 5.85,
|
| 336 |
+
"learning_rate": 4.9674535987592016e-05,
|
| 337 |
+
"loss": 0.0781,
|
| 338 |
+
"step": 550
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.96,
|
| 342 |
+
"learning_rate": 4.96361449515e-05,
|
| 343 |
+
"loss": 0.0683,
|
| 344 |
+
"step": 560
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 6.06,
|
| 348 |
+
"learning_rate": 4.959563062856744e-05,
|
| 349 |
+
"loss": 0.0724,
|
| 350 |
+
"step": 570
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 6.17,
|
| 354 |
+
"learning_rate": 4.955299651055528e-05,
|
| 355 |
+
"loss": 0.075,
|
| 356 |
+
"step": 580
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.28,
|
| 360 |
+
"learning_rate": 4.950824627192079e-05,
|
| 361 |
+
"loss": 0.0618,
|
| 362 |
+
"step": 590
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.38,
|
| 366 |
+
"learning_rate": 4.946138376950087e-05,
|
| 367 |
+
"loss": 0.0687,
|
| 368 |
+
"step": 600
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.49,
|
| 372 |
+
"learning_rate": 4.9412413042179615e-05,
|
| 373 |
+
"loss": 0.0625,
|
| 374 |
+
"step": 610
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.6,
|
| 378 |
+
"learning_rate": 4.936133831054032e-05,
|
| 379 |
+
"loss": 0.0635,
|
| 380 |
+
"step": 620
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.7,
|
| 384 |
+
"learning_rate": 4.930816397650159e-05,
|
| 385 |
+
"loss": 0.0677,
|
| 386 |
+
"step": 630
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.81,
|
| 390 |
+
"learning_rate": 4.925289462293807e-05,
|
| 391 |
+
"loss": 0.0571,
|
| 392 |
+
"step": 640
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.91,
|
| 396 |
+
"learning_rate": 4.9195535013285396e-05,
|
| 397 |
+
"loss": 0.0683,
|
| 398 |
+
"step": 650
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.02,
|
| 402 |
+
"learning_rate": 4.9136090091129725e-05,
|
| 403 |
+
"loss": 0.0701,
|
| 404 |
+
"step": 660
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.13,
|
| 408 |
+
"learning_rate": 4.907456497978156e-05,
|
| 409 |
+
"loss": 0.0594,
|
| 410 |
+
"step": 670
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 7.23,
|
| 414 |
+
"learning_rate": 4.901096498183429e-05,
|
| 415 |
+
"loss": 0.0546,
|
| 416 |
+
"step": 680
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.34,
|
| 420 |
+
"learning_rate": 4.894529557870716e-05,
|
| 421 |
+
"loss": 0.0629,
|
| 422 |
+
"step": 690
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 7.45,
|
| 426 |
+
"learning_rate": 4.8877562430172815e-05,
|
| 427 |
+
"loss": 0.0717,
|
| 428 |
+
"step": 700
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 7.55,
|
| 432 |
+
"learning_rate": 4.880777137386954e-05,
|
| 433 |
+
"loss": 0.0611,
|
| 434 |
+
"step": 710
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 7.66,
|
| 438 |
+
"learning_rate": 4.873592842479813e-05,
|
| 439 |
+
"loss": 0.0532,
|
| 440 |
+
"step": 720
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.77,
|
| 444 |
+
"learning_rate": 4.8662039774803476e-05,
|
| 445 |
+
"loss": 0.0586,
|
| 446 |
+
"step": 730
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 7.87,
|
| 450 |
+
"learning_rate": 4.8586111792040935e-05,
|
| 451 |
+
"loss": 0.0589,
|
| 452 |
+
"step": 740
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 7.98,
|
| 456 |
+
"learning_rate": 4.850815102042744e-05,
|
| 457 |
+
"loss": 0.0553,
|
| 458 |
+
"step": 750
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 8.09,
|
| 462 |
+
"learning_rate": 4.842816417907759e-05,
|
| 463 |
+
"loss": 0.0535,
|
| 464 |
+
"step": 760
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 8.19,
|
| 468 |
+
"learning_rate": 4.8346158161724454e-05,
|
| 469 |
+
"loss": 0.0624,
|
| 470 |
+
"step": 770
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 8.3,
|
| 474 |
+
"learning_rate": 4.82621400361255e-05,
|
| 475 |
+
"loss": 0.0536,
|
| 476 |
+
"step": 780
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 8.4,
|
| 480 |
+
"learning_rate": 4.817611704345344e-05,
|
| 481 |
+
"loss": 0.0612,
|
| 482 |
+
"step": 790
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 8.51,
|
| 486 |
+
"learning_rate": 4.808809659767214e-05,
|
| 487 |
+
"loss": 0.0535,
|
| 488 |
+
"step": 800
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 8.62,
|
| 492 |
+
"learning_rate": 4.7998086284897626e-05,
|
| 493 |
+
"loss": 0.0511,
|
| 494 |
+
"step": 810
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 8.72,
|
| 498 |
+
"learning_rate": 4.7906093862744297e-05,
|
| 499 |
+
"loss": 0.0542,
|
| 500 |
+
"step": 820
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 8.83,
|
| 504 |
+
"learning_rate": 4.781212725965629e-05,
|
| 505 |
+
"loss": 0.0614,
|
| 506 |
+
"step": 830
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 8.94,
|
| 510 |
+
"learning_rate": 4.7716194574224214e-05,
|
| 511 |
+
"loss": 0.0561,
|
| 512 |
+
"step": 840
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 9.04,
|
| 516 |
+
"learning_rate": 4.7618304074487124e-05,
|
| 517 |
+
"loss": 0.0535,
|
| 518 |
+
"step": 850
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 9.15,
|
| 522 |
+
"learning_rate": 4.751846419721994e-05,
|
| 523 |
+
"loss": 0.0506,
|
| 524 |
+
"step": 860
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 9.26,
|
| 528 |
+
"learning_rate": 4.7416683547206306e-05,
|
| 529 |
+
"loss": 0.0522,
|
| 530 |
+
"step": 870
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.36,
|
| 534 |
+
"learning_rate": 4.731297089649703e-05,
|
| 535 |
+
"loss": 0.0495,
|
| 536 |
+
"step": 880
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.47,
|
| 540 |
+
"learning_rate": 4.7207335183654e-05,
|
| 541 |
+
"loss": 0.0538,
|
| 542 |
+
"step": 890
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 9.57,
|
| 546 |
+
"learning_rate": 4.709978551297984e-05,
|
| 547 |
+
"loss": 0.049,
|
| 548 |
+
"step": 900
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 9.68,
|
| 552 |
+
"learning_rate": 4.699033115373322e-05,
|
| 553 |
+
"loss": 0.0531,
|
| 554 |
+
"step": 910
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 9.79,
|
| 558 |
+
"learning_rate": 4.687898153933001e-05,
|
| 559 |
+
"loss": 0.0556,
|
| 560 |
+
"step": 920
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.89,
|
| 564 |
+
"learning_rate": 4.6765746266530244e-05,
|
| 565 |
+
"loss": 0.0573,
|
| 566 |
+
"step": 930
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 10.0,
|
| 570 |
+
"learning_rate": 4.665063509461097e-05,
|
| 571 |
+
"loss": 0.0494,
|
| 572 |
+
"step": 940
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 10.11,
|
| 576 |
+
"learning_rate": 4.653365794452521e-05,
|
| 577 |
+
"loss": 0.0491,
|
| 578 |
+
"step": 950
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.21,
|
| 582 |
+
"learning_rate": 4.6414824898046896e-05,
|
| 583 |
+
"loss": 0.0521,
|
| 584 |
+
"step": 960
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 10.32,
|
| 588 |
+
"learning_rate": 4.62941461969019e-05,
|
| 589 |
+
"loss": 0.0518,
|
| 590 |
+
"step": 970
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 10.43,
|
| 594 |
+
"learning_rate": 4.617163224188544e-05,
|
| 595 |
+
"loss": 0.0519,
|
| 596 |
+
"step": 980
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 10.53,
|
| 600 |
+
"learning_rate": 4.60472935919656e-05,
|
| 601 |
+
"loss": 0.0537,
|
| 602 |
+
"step": 990
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 10.64,
|
| 606 |
+
"learning_rate": 4.5921140963373335e-05,
|
| 607 |
+
"loss": 0.0481,
|
| 608 |
+
"step": 1000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 10.74,
|
| 612 |
+
"learning_rate": 4.579318522867886e-05,
|
| 613 |
+
"loss": 0.0452,
|
| 614 |
+
"step": 1010
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 10.85,
|
| 618 |
+
"learning_rate": 4.566343741585465e-05,
|
| 619 |
+
"loss": 0.0484,
|
| 620 |
+
"step": 1020
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 10.96,
|
| 624 |
+
"learning_rate": 4.553190870732488e-05,
|
| 625 |
+
"loss": 0.0479,
|
| 626 |
+
"step": 1030
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 11.06,
|
| 630 |
+
"learning_rate": 4.539861043900176e-05,
|
| 631 |
+
"loss": 0.0446,
|
| 632 |
+
"step": 1040
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 11.17,
|
| 636 |
+
"learning_rate": 4.526355409930847e-05,
|
| 637 |
+
"loss": 0.0431,
|
| 638 |
+
"step": 1050
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 11.28,
|
| 642 |
+
"learning_rate": 4.512675132818908e-05,
|
| 643 |
+
"loss": 0.0486,
|
| 644 |
+
"step": 1060
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 11.38,
|
| 648 |
+
"learning_rate": 4.498821391610527e-05,
|
| 649 |
+
"loss": 0.0485,
|
| 650 |
+
"step": 1070
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 11.49,
|
| 654 |
+
"learning_rate": 4.484795380302025e-05,
|
| 655 |
+
"loss": 0.0421,
|
| 656 |
+
"step": 1080
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 11.6,
|
| 660 |
+
"learning_rate": 4.470598307736964e-05,
|
| 661 |
+
"loss": 0.0512,
|
| 662 |
+
"step": 1090
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 11.7,
|
| 666 |
+
"learning_rate": 4.456231397501966e-05,
|
| 667 |
+
"loss": 0.0479,
|
| 668 |
+
"step": 1100
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 11.81,
|
| 672 |
+
"learning_rate": 4.441695887821253e-05,
|
| 673 |
+
"loss": 0.0467,
|
| 674 |
+
"step": 1110
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 11.91,
|
| 678 |
+
"learning_rate": 4.426993031449934e-05,
|
| 679 |
+
"loss": 0.0439,
|
| 680 |
+
"step": 1120
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 12.02,
|
| 684 |
+
"learning_rate": 4.4121240955660304e-05,
|
| 685 |
+
"loss": 0.043,
|
| 686 |
+
"step": 1130
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 12.13,
|
| 690 |
+
"learning_rate": 4.397090361661268e-05,
|
| 691 |
+
"loss": 0.0464,
|
| 692 |
+
"step": 1140
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 12.23,
|
| 696 |
+
"learning_rate": 4.381893125430628e-05,
|
| 697 |
+
"loss": 0.0485,
|
| 698 |
+
"step": 1150
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 12.34,
|
| 702 |
+
"learning_rate": 4.3665336966606774e-05,
|
| 703 |
+
"loss": 0.0423,
|
| 704 |
+
"step": 1160
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 12.45,
|
| 708 |
+
"learning_rate": 4.351013399116681e-05,
|
| 709 |
+
"loss": 0.0429,
|
| 710 |
+
"step": 1170
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 12.55,
|
| 714 |
+
"learning_rate": 4.335333570428516e-05,
|
| 715 |
+
"loss": 0.0478,
|
| 716 |
+
"step": 1180
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 12.66,
|
| 720 |
+
"learning_rate": 4.319495561975386e-05,
|
| 721 |
+
"loss": 0.0424,
|
| 722 |
+
"step": 1190
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 12.77,
|
| 726 |
+
"learning_rate": 4.303500738769348e-05,
|
| 727 |
+
"loss": 0.0467,
|
| 728 |
+
"step": 1200
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 12.87,
|
| 732 |
+
"learning_rate": 4.287350479337672e-05,
|
| 733 |
+
"loss": 0.0472,
|
| 734 |
+
"step": 1210
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 12.98,
|
| 738 |
+
"learning_rate": 4.271046175604029e-05,
|
| 739 |
+
"loss": 0.0443,
|
| 740 |
+
"step": 1220
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 13.09,
|
| 744 |
+
"learning_rate": 4.254589232768528e-05,
|
| 745 |
+
"loss": 0.0392,
|
| 746 |
+
"step": 1230
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 13.19,
|
| 750 |
+
"learning_rate": 4.2379810691866064e-05,
|
| 751 |
+
"loss": 0.0407,
|
| 752 |
+
"step": 1240
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 13.3,
|
| 756 |
+
"learning_rate": 4.2212231162467886e-05,
|
| 757 |
+
"loss": 0.0409,
|
| 758 |
+
"step": 1250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 13.4,
|
| 762 |
+
"learning_rate": 4.2043168182473215e-05,
|
| 763 |
+
"loss": 0.0435,
|
| 764 |
+
"step": 1260
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 13.51,
|
| 768 |
+
"learning_rate": 4.187263632271695e-05,
|
| 769 |
+
"loss": 0.0434,
|
| 770 |
+
"step": 1270
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 13.62,
|
| 774 |
+
"learning_rate": 4.170065028063062e-05,
|
| 775 |
+
"loss": 0.0441,
|
| 776 |
+
"step": 1280
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 13.72,
|
| 780 |
+
"learning_rate": 4.152722487897571e-05,
|
| 781 |
+
"loss": 0.044,
|
| 782 |
+
"step": 1290
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 13.83,
|
| 786 |
+
"learning_rate": 4.1352375064566054e-05,
|
| 787 |
+
"loss": 0.0446,
|
| 788 |
+
"step": 1300
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 13.94,
|
| 792 |
+
"learning_rate": 4.1176115906979774e-05,
|
| 793 |
+
"loss": 0.0464,
|
| 794 |
+
"step": 1310
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 14.04,
|
| 798 |
+
"learning_rate": 4.0998462597260346e-05,
|
| 799 |
+
"loss": 0.0412,
|
| 800 |
+
"step": 1320
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 14.15,
|
| 804 |
+
"learning_rate": 4.0819430446607465e-05,
|
| 805 |
+
"loss": 0.0382,
|
| 806 |
+
"step": 1330
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 14.26,
|
| 810 |
+
"learning_rate": 4.0639034885057376e-05,
|
| 811 |
+
"loss": 0.0414,
|
| 812 |
+
"step": 1340
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 14.36,
|
| 816 |
+
"learning_rate": 4.045729146015303e-05,
|
| 817 |
+
"loss": 0.0441,
|
| 818 |
+
"step": 1350
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 14.47,
|
| 822 |
+
"learning_rate": 4.027421583560414e-05,
|
| 823 |
+
"loss": 0.0435,
|
| 824 |
+
"step": 1360
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 14.57,
|
| 828 |
+
"learning_rate": 4.0089823789937106e-05,
|
| 829 |
+
"loss": 0.0424,
|
| 830 |
+
"step": 1370
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 14.68,
|
| 834 |
+
"learning_rate": 3.990413121513526e-05,
|
| 835 |
+
"loss": 0.0422,
|
| 836 |
+
"step": 1380
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 14.79,
|
| 840 |
+
"learning_rate": 3.9717154115269076e-05,
|
| 841 |
+
"loss": 0.0389,
|
| 842 |
+
"step": 1390
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 14.89,
|
| 846 |
+
"learning_rate": 3.952890860511692e-05,
|
| 847 |
+
"loss": 0.0423,
|
| 848 |
+
"step": 1400
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 15.0,
|
| 852 |
+
"learning_rate": 3.933941090877615e-05,
|
| 853 |
+
"loss": 0.0419,
|
| 854 |
+
"step": 1410
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 15.11,
|
| 858 |
+
"learning_rate": 3.914867735826489e-05,
|
| 859 |
+
"loss": 0.0425,
|
| 860 |
+
"step": 1420
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 15.21,
|
| 864 |
+
"learning_rate": 3.895672439211431e-05,
|
| 865 |
+
"loss": 0.0383,
|
| 866 |
+
"step": 1430
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 15.32,
|
| 870 |
+
"learning_rate": 3.876356855395202e-05,
|
| 871 |
+
"loss": 0.0388,
|
| 872 |
+
"step": 1440
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 15.43,
|
| 876 |
+
"learning_rate": 3.856922649107613e-05,
|
| 877 |
+
"loss": 0.0406,
|
| 878 |
+
"step": 1450
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 15.53,
|
| 882 |
+
"learning_rate": 3.8373714953020506e-05,
|
| 883 |
+
"loss": 0.0415,
|
| 884 |
+
"step": 1460
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 15.64,
|
| 888 |
+
"learning_rate": 3.817705079011125e-05,
|
| 889 |
+
"loss": 0.0401,
|
| 890 |
+
"step": 1470
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 15.74,
|
| 894 |
+
"learning_rate": 3.797925095201438e-05,
|
| 895 |
+
"loss": 0.0358,
|
| 896 |
+
"step": 1480
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 15.85,
|
| 900 |
+
"learning_rate": 3.778033248627506e-05,
|
| 901 |
+
"loss": 0.0403,
|
| 902 |
+
"step": 1490
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 15.96,
|
| 906 |
+
"learning_rate": 3.758031253684827e-05,
|
| 907 |
+
"loss": 0.0431,
|
| 908 |
+
"step": 1500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 16.06,
|
| 912 |
+
"learning_rate": 3.737920834262134e-05,
|
| 913 |
+
"loss": 0.0392,
|
| 914 |
+
"step": 1510
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 16.17,
|
| 918 |
+
"learning_rate": 3.717703723592811e-05,
|
| 919 |
+
"loss": 0.041,
|
| 920 |
+
"step": 1520
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 16.28,
|
| 924 |
+
"learning_rate": 3.69738166410552e-05,
|
| 925 |
+
"loss": 0.0353,
|
| 926 |
+
"step": 1530
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 16.38,
|
| 930 |
+
"learning_rate": 3.676956407274021e-05,
|
| 931 |
+
"loss": 0.0388,
|
| 932 |
+
"step": 1540
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 16.49,
|
| 936 |
+
"learning_rate": 3.6564297134662264e-05,
|
| 937 |
+
"loss": 0.038,
|
| 938 |
+
"step": 1550
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 16.6,
|
| 942 |
+
"learning_rate": 3.63580335179248e-05,
|
| 943 |
+
"loss": 0.038,
|
| 944 |
+
"step": 1560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 16.7,
|
| 948 |
+
"learning_rate": 3.615079099953081e-05,
|
| 949 |
+
"loss": 0.0352,
|
| 950 |
+
"step": 1570
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 16.81,
|
| 954 |
+
"learning_rate": 3.59425874408508e-05,
|
| 955 |
+
"loss": 0.0404,
|
| 956 |
+
"step": 1580
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 16.91,
|
| 960 |
+
"learning_rate": 3.5733440786083304e-05,
|
| 961 |
+
"loss": 0.0378,
|
| 962 |
+
"step": 1590
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 17.02,
|
| 966 |
+
"learning_rate": 3.552336906070838e-05,
|
| 967 |
+
"loss": 0.0435,
|
| 968 |
+
"step": 1600
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 17.13,
|
| 972 |
+
"learning_rate": 3.531239036993409e-05,
|
| 973 |
+
"loss": 0.0352,
|
| 974 |
+
"step": 1610
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 17.23,
|
| 978 |
+
"learning_rate": 3.5100522897136065e-05,
|
| 979 |
+
"loss": 0.0416,
|
| 980 |
+
"step": 1620
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 17.34,
|
| 984 |
+
"learning_rate": 3.4887784902290366e-05,
|
| 985 |
+
"loss": 0.0336,
|
| 986 |
+
"step": 1630
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 17.45,
|
| 990 |
+
"learning_rate": 3.467419472039972e-05,
|
| 991 |
+
"loss": 0.0318,
|
| 992 |
+
"step": 1640
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 17.55,
|
| 996 |
+
"learning_rate": 3.4459770759913314e-05,
|
| 997 |
+
"loss": 0.035,
|
| 998 |
+
"step": 1650
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 17.66,
|
| 1002 |
+
"learning_rate": 3.424453150114023e-05,
|
| 1003 |
+
"loss": 0.0366,
|
| 1004 |
+
"step": 1660
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 17.77,
|
| 1008 |
+
"learning_rate": 3.4028495494656746e-05,
|
| 1009 |
+
"loss": 0.0458,
|
| 1010 |
+
"step": 1670
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 17.87,
|
| 1014 |
+
"learning_rate": 3.381168135970749e-05,
|
| 1015 |
+
"loss": 0.04,
|
| 1016 |
+
"step": 1680
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 17.98,
|
| 1020 |
+
"learning_rate": 3.3594107782600753e-05,
|
| 1021 |
+
"loss": 0.0381,
|
| 1022 |
+
"step": 1690
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 18.09,
|
| 1026 |
+
"learning_rate": 3.337579351509797e-05,
|
| 1027 |
+
"loss": 0.0345,
|
| 1028 |
+
"step": 1700
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 18.19,
|
| 1032 |
+
"learning_rate": 3.315675737279763e-05,
|
| 1033 |
+
"loss": 0.0358,
|
| 1034 |
+
"step": 1710
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 18.3,
|
| 1038 |
+
"learning_rate": 3.293701823351357e-05,
|
| 1039 |
+
"loss": 0.0346,
|
| 1040 |
+
"step": 1720
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 18.4,
|
| 1044 |
+
"learning_rate": 3.271659503564804e-05,
|
| 1045 |
+
"loss": 0.0383,
|
| 1046 |
+
"step": 1730
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 18.51,
|
| 1050 |
+
"learning_rate": 3.249550677655945e-05,
|
| 1051 |
+
"loss": 0.0332,
|
| 1052 |
+
"step": 1740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 18.62,
|
| 1056 |
+
"learning_rate": 3.227377251092507e-05,
|
| 1057 |
+
"loss": 0.0348,
|
| 1058 |
+
"step": 1750
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 18.72,
|
| 1062 |
+
"learning_rate": 3.205141134909878e-05,
|
| 1063 |
+
"loss": 0.0382,
|
| 1064 |
+
"step": 1760
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 18.83,
|
| 1068 |
+
"learning_rate": 3.182844245546403e-05,
|
| 1069 |
+
"loss": 0.0377,
|
| 1070 |
+
"step": 1770
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 18.94,
|
| 1074 |
+
"learning_rate": 3.160488504678216e-05,
|
| 1075 |
+
"loss": 0.0366,
|
| 1076 |
+
"step": 1780
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 19.04,
|
| 1080 |
+
"learning_rate": 3.138075839053615e-05,
|
| 1081 |
+
"loss": 0.0382,
|
| 1082 |
+
"step": 1790
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 19.15,
|
| 1086 |
+
"learning_rate": 3.11560818032701e-05,
|
| 1087 |
+
"loss": 0.0332,
|
| 1088 |
+
"step": 1800
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 19.26,
|
| 1092 |
+
"learning_rate": 3.093087464892434e-05,
|
| 1093 |
+
"loss": 0.0328,
|
| 1094 |
+
"step": 1810
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 19.36,
|
| 1098 |
+
"learning_rate": 3.070515633716658e-05,
|
| 1099 |
+
"loss": 0.0382,
|
| 1100 |
+
"step": 1820
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 19.47,
|
| 1104 |
+
"learning_rate": 3.047894632171907e-05,
|
| 1105 |
+
"loss": 0.0372,
|
| 1106 |
+
"step": 1830
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 19.57,
|
| 1110 |
+
"learning_rate": 3.0252264098681947e-05,
|
| 1111 |
+
"loss": 0.0347,
|
| 1112 |
+
"step": 1840
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 19.68,
|
| 1116 |
+
"learning_rate": 3.0025129204852964e-05,
|
| 1117 |
+
"loss": 0.0355,
|
| 1118 |
+
"step": 1850
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 19.79,
|
| 1122 |
+
"learning_rate": 2.979756121604366e-05,
|
| 1123 |
+
"loss": 0.0353,
|
| 1124 |
+
"step": 1860
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 19.89,
|
| 1128 |
+
"learning_rate": 2.956957974539226e-05,
|
| 1129 |
+
"loss": 0.0352,
|
| 1130 |
+
"step": 1870
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 20.0,
|
| 1134 |
+
"learning_rate": 2.9341204441673266e-05,
|
| 1135 |
+
"loss": 0.0362,
|
| 1136 |
+
"step": 1880
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 20.11,
|
| 1140 |
+
"learning_rate": 2.911245498760398e-05,
|
| 1141 |
+
"loss": 0.0385,
|
| 1142 |
+
"step": 1890
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 20.21,
|
| 1146 |
+
"learning_rate": 2.8883351098148216e-05,
|
| 1147 |
+
"loss": 0.0308,
|
| 1148 |
+
"step": 1900
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 20.32,
|
| 1152 |
+
"learning_rate": 2.8653912518817055e-05,
|
| 1153 |
+
"loss": 0.0326,
|
| 1154 |
+
"step": 1910
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 20.43,
|
| 1158 |
+
"learning_rate": 2.8424159023967135e-05,
|
| 1159 |
+
"loss": 0.0359,
|
| 1160 |
+
"step": 1920
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 20.53,
|
| 1164 |
+
"learning_rate": 2.8194110415096335e-05,
|
| 1165 |
+
"loss": 0.0327,
|
| 1166 |
+
"step": 1930
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 20.64,
|
| 1170 |
+
"learning_rate": 2.7963786519137203e-05,
|
| 1171 |
+
"loss": 0.0325,
|
| 1172 |
+
"step": 1940
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 20.74,
|
| 1176 |
+
"learning_rate": 2.773320718674812e-05,
|
| 1177 |
+
"loss": 0.0358,
|
| 1178 |
+
"step": 1950
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 20.85,
|
| 1182 |
+
"learning_rate": 2.7502392290602463e-05,
|
| 1183 |
+
"loss": 0.031,
|
| 1184 |
+
"step": 1960
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 20.96,
|
| 1188 |
+
"learning_rate": 2.727136172367589e-05,
|
| 1189 |
+
"loss": 0.0357,
|
| 1190 |
+
"step": 1970
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 21.06,
|
| 1194 |
+
"learning_rate": 2.7040135397531808e-05,
|
| 1195 |
+
"loss": 0.033,
|
| 1196 |
+
"step": 1980
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 21.17,
|
| 1200 |
+
"learning_rate": 2.6808733240605323e-05,
|
| 1201 |
+
"loss": 0.0295,
|
| 1202 |
+
"step": 1990
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 21.28,
|
| 1206 |
+
"learning_rate": 2.657717519648562e-05,
|
| 1207 |
+
"loss": 0.0382,
|
| 1208 |
+
"step": 2000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 21.38,
|
| 1212 |
+
"learning_rate": 2.6345481222197205e-05,
|
| 1213 |
+
"loss": 0.0317,
|
| 1214 |
+
"step": 2010
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 21.49,
|
| 1218 |
+
"learning_rate": 2.6113671286479812e-05,
|
| 1219 |
+
"loss": 0.0314,
|
| 1220 |
+
"step": 2020
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 21.6,
|
| 1224 |
+
"learning_rate": 2.588176536806739e-05,
|
| 1225 |
+
"loss": 0.0328,
|
| 1226 |
+
"step": 2030
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 21.7,
|
| 1230 |
+
"learning_rate": 2.5649783453966274e-05,
|
| 1231 |
+
"loss": 0.0337,
|
| 1232 |
+
"step": 2040
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 21.81,
|
| 1236 |
+
"learning_rate": 2.5417745537732523e-05,
|
| 1237 |
+
"loss": 0.0347,
|
| 1238 |
+
"step": 2050
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 21.91,
|
| 1242 |
+
"learning_rate": 2.5185671617748812e-05,
|
| 1243 |
+
"loss": 0.0319,
|
| 1244 |
+
"step": 2060
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 22.02,
|
| 1248 |
+
"learning_rate": 2.4953581695500812e-05,
|
| 1249 |
+
"loss": 0.0343,
|
| 1250 |
+
"step": 2070
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 22.13,
|
| 1254 |
+
"learning_rate": 2.4721495773853366e-05,
|
| 1255 |
+
"loss": 0.0336,
|
| 1256 |
+
"step": 2080
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 22.23,
|
| 1260 |
+
"learning_rate": 2.4489433855326518e-05,
|
| 1261 |
+
"loss": 0.031,
|
| 1262 |
+
"step": 2090
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 22.34,
|
| 1266 |
+
"learning_rate": 2.4257415940371587e-05,
|
| 1267 |
+
"loss": 0.0349,
|
| 1268 |
+
"step": 2100
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 22.45,
|
| 1272 |
+
"learning_rate": 2.4025462025647407e-05,
|
| 1273 |
+
"loss": 0.0319,
|
| 1274 |
+
"step": 2110
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 22.55,
|
| 1278 |
+
"learning_rate": 2.3793592102296915e-05,
|
| 1279 |
+
"loss": 0.0311,
|
| 1280 |
+
"step": 2120
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 22.66,
|
| 1284 |
+
"learning_rate": 2.3561826154224144e-05,
|
| 1285 |
+
"loss": 0.0331,
|
| 1286 |
+
"step": 2130
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 22.77,
|
| 1290 |
+
"learning_rate": 2.333018415637196e-05,
|
| 1291 |
+
"loss": 0.0323,
|
| 1292 |
+
"step": 2140
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 22.87,
|
| 1296 |
+
"learning_rate": 2.309868607300048e-05,
|
| 1297 |
+
"loss": 0.0349,
|
| 1298 |
+
"step": 2150
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 22.98,
|
| 1302 |
+
"learning_rate": 2.2867351855966383e-05,
|
| 1303 |
+
"loss": 0.0331,
|
| 1304 |
+
"step": 2160
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 23.09,
|
| 1308 |
+
"learning_rate": 2.263620144300345e-05,
|
| 1309 |
+
"loss": 0.0302,
|
| 1310 |
+
"step": 2170
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 23.19,
|
| 1314 |
+
"learning_rate": 2.240525475600412e-05,
|
| 1315 |
+
"loss": 0.0353,
|
| 1316 |
+
"step": 2180
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 23.3,
|
| 1320 |
+
"learning_rate": 2.2174531699302562e-05,
|
| 1321 |
+
"loss": 0.0321,
|
| 1322 |
+
"step": 2190
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 23.4,
|
| 1326 |
+
"learning_rate": 2.1944052157959143e-05,
|
| 1327 |
+
"loss": 0.0336,
|
| 1328 |
+
"step": 2200
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 23.51,
|
| 1332 |
+
"learning_rate": 2.171383599604668e-05,
|
| 1333 |
+
"loss": 0.0309,
|
| 1334 |
+
"step": 2210
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 23.62,
|
| 1338 |
+
"learning_rate": 2.1483903054938426e-05,
|
| 1339 |
+
"loss": 0.0315,
|
| 1340 |
+
"step": 2220
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 23.72,
|
| 1344 |
+
"learning_rate": 2.1254273151597966e-05,
|
| 1345 |
+
"loss": 0.0303,
|
| 1346 |
+
"step": 2230
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 23.83,
|
| 1350 |
+
"learning_rate": 2.1024966076871364e-05,
|
| 1351 |
+
"loss": 0.0322,
|
| 1352 |
+
"step": 2240
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 23.94,
|
| 1356 |
+
"learning_rate": 2.0796001593781417e-05,
|
| 1357 |
+
"loss": 0.034,
|
| 1358 |
+
"step": 2250
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 24.04,
|
| 1362 |
+
"learning_rate": 2.0567399435824367e-05,
|
| 1363 |
+
"loss": 0.0313,
|
| 1364 |
+
"step": 2260
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 24.15,
|
| 1368 |
+
"learning_rate": 2.033917930526916e-05,
|
| 1369 |
+
"loss": 0.0311,
|
| 1370 |
+
"step": 2270
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 24.26,
|
| 1374 |
+
"learning_rate": 2.0111360871459388e-05,
|
| 1375 |
+
"loss": 0.0286,
|
| 1376 |
+
"step": 2280
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 24.36,
|
| 1380 |
+
"learning_rate": 1.9883963769118097e-05,
|
| 1381 |
+
"loss": 0.0307,
|
| 1382 |
+
"step": 2290
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 24.47,
|
| 1386 |
+
"learning_rate": 1.9657007596655477e-05,
|
| 1387 |
+
"loss": 0.0299,
|
| 1388 |
+
"step": 2300
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 24.57,
|
| 1392 |
+
"learning_rate": 1.9430511914479853e-05,
|
| 1393 |
+
"loss": 0.0319,
|
| 1394 |
+
"step": 2310
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 24.68,
|
| 1398 |
+
"learning_rate": 1.920449624331179e-05,
|
| 1399 |
+
"loss": 0.0328,
|
| 1400 |
+
"step": 2320
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 24.79,
|
| 1404 |
+
"learning_rate": 1.8978980062501712e-05,
|
| 1405 |
+
"loss": 0.0333,
|
| 1406 |
+
"step": 2330
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 24.89,
|
| 1410 |
+
"learning_rate": 1.8753982808351007e-05,
|
| 1411 |
+
"loss": 0.0299,
|
| 1412 |
+
"step": 2340
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 25.0,
|
| 1416 |
+
"learning_rate": 1.852952387243698e-05,
|
| 1417 |
+
"loss": 0.0284,
|
| 1418 |
+
"step": 2350
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 25.11,
|
| 1422 |
+
"learning_rate": 1.830562259994152e-05,
|
| 1423 |
+
"loss": 0.0286,
|
| 1424 |
+
"step": 2360
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 25.21,
|
| 1428 |
+
"learning_rate": 1.808229828798379e-05,
|
| 1429 |
+
"loss": 0.0331,
|
| 1430 |
+
"step": 2370
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 25.32,
|
| 1434 |
+
"learning_rate": 1.7859570183957174e-05,
|
| 1435 |
+
"loss": 0.0272,
|
| 1436 |
+
"step": 2380
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 25.43,
|
| 1440 |
+
"learning_rate": 1.7637457483870333e-05,
|
| 1441 |
+
"loss": 0.0273,
|
| 1442 |
+
"step": 2390
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 25.53,
|
| 1446 |
+
"learning_rate": 1.7415979330692857e-05,
|
| 1447 |
+
"loss": 0.0319,
|
| 1448 |
+
"step": 2400
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 25.64,
|
| 1452 |
+
"learning_rate": 1.7195154812705343e-05,
|
| 1453 |
+
"loss": 0.0272,
|
| 1454 |
+
"step": 2410
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 25.74,
|
| 1458 |
+
"learning_rate": 1.697500296185432e-05,
|
| 1459 |
+
"loss": 0.0351,
|
| 1460 |
+
"step": 2420
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 25.85,
|
| 1464 |
+
"learning_rate": 1.6755542752111925e-05,
|
| 1465 |
+
"loss": 0.0303,
|
| 1466 |
+
"step": 2430
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 25.96,
|
| 1470 |
+
"learning_rate": 1.6536793097840615e-05,
|
| 1471 |
+
"loss": 0.0332,
|
| 1472 |
+
"step": 2440
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 26.06,
|
| 1476 |
+
"learning_rate": 1.631877285216304e-05,
|
| 1477 |
+
"loss": 0.0306,
|
| 1478 |
+
"step": 2450
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 26.17,
|
| 1482 |
+
"learning_rate": 1.610150080533716e-05,
|
| 1483 |
+
"loss": 0.0311,
|
| 1484 |
+
"step": 2460
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 26.28,
|
| 1488 |
+
"learning_rate": 1.5884995683136765e-05,
|
| 1489 |
+
"loss": 0.0321,
|
| 1490 |
+
"step": 2470
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 26.38,
|
| 1494 |
+
"learning_rate": 1.5669276145237633e-05,
|
| 1495 |
+
"loss": 0.0351,
|
| 1496 |
+
"step": 2480
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 26.49,
|
| 1500 |
+
"learning_rate": 1.5454360783609278e-05,
|
| 1501 |
+
"loss": 0.026,
|
| 1502 |
+
"step": 2490
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 26.6,
|
| 1506 |
+
"learning_rate": 1.5240268120912631e-05,
|
| 1507 |
+
"loss": 0.0273,
|
| 1508 |
+
"step": 2500
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 26.7,
|
| 1512 |
+
"learning_rate": 1.5027016608903594e-05,
|
| 1513 |
+
"loss": 0.0325,
|
| 1514 |
+
"step": 2510
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 26.81,
|
| 1518 |
+
"learning_rate": 1.4814624626842798e-05,
|
| 1519 |
+
"loss": 0.029,
|
| 1520 |
+
"step": 2520
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 26.91,
|
| 1524 |
+
"learning_rate": 1.4603110479911592e-05,
|
| 1525 |
+
"loss": 0.0304,
|
| 1526 |
+
"step": 2530
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 27.02,
|
| 1530 |
+
"learning_rate": 1.4392492397634321e-05,
|
| 1531 |
+
"loss": 0.0297,
|
| 1532 |
+
"step": 2540
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 27.13,
|
| 1536 |
+
"learning_rate": 1.418278853230729e-05,
|
| 1537 |
+
"loss": 0.0271,
|
| 1538 |
+
"step": 2550
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 27.23,
|
| 1542 |
+
"learning_rate": 1.3974016957434208e-05,
|
| 1543 |
+
"loss": 0.0327,
|
| 1544 |
+
"step": 2560
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 27.34,
|
| 1548 |
+
"learning_rate": 1.3766195666168603e-05,
|
| 1549 |
+
"loss": 0.0275,
|
| 1550 |
+
"step": 2570
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 27.45,
|
| 1554 |
+
"learning_rate": 1.3559342569762933e-05,
|
| 1555 |
+
"loss": 0.0319,
|
| 1556 |
+
"step": 2580
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 27.55,
|
| 1560 |
+
"learning_rate": 1.3353475496025047e-05,
|
| 1561 |
+
"loss": 0.0286,
|
| 1562 |
+
"step": 2590
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 27.66,
|
| 1566 |
+
"learning_rate": 1.314861218778156e-05,
|
| 1567 |
+
"loss": 0.0326,
|
| 1568 |
+
"step": 2600
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 27.77,
|
| 1572 |
+
"learning_rate": 1.2944770301348727e-05,
|
| 1573 |
+
"loss": 0.0289,
|
| 1574 |
+
"step": 2610
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 27.87,
|
| 1578 |
+
"learning_rate": 1.2741967405010698e-05,
|
| 1579 |
+
"loss": 0.0294,
|
| 1580 |
+
"step": 2620
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 27.98,
|
| 1584 |
+
"learning_rate": 1.254022097750539e-05,
|
| 1585 |
+
"loss": 0.0308,
|
| 1586 |
+
"step": 2630
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 28.09,
|
| 1590 |
+
"learning_rate": 1.23395484065181e-05,
|
| 1591 |
+
"loss": 0.0334,
|
| 1592 |
+
"step": 2640
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 28.19,
|
| 1596 |
+
"learning_rate": 1.213996698718283e-05,
|
| 1597 |
+
"loss": 0.0279,
|
| 1598 |
+
"step": 2650
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 28.3,
|
| 1602 |
+
"learning_rate": 1.1941493920591828e-05,
|
| 1603 |
+
"loss": 0.0293,
|
| 1604 |
+
"step": 2660
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 28.4,
|
| 1608 |
+
"learning_rate": 1.1744146312313012e-05,
|
| 1609 |
+
"loss": 0.0274,
|
| 1610 |
+
"step": 2670
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 28.51,
|
| 1614 |
+
"learning_rate": 1.1547941170915686e-05,
|
| 1615 |
+
"loss": 0.0303,
|
| 1616 |
+
"step": 2680
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 28.62,
|
| 1620 |
+
"learning_rate": 1.1352895406504751e-05,
|
| 1621 |
+
"loss": 0.0302,
|
| 1622 |
+
"step": 2690
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 28.72,
|
| 1626 |
+
"learning_rate": 1.1159025829263183e-05,
|
| 1627 |
+
"loss": 0.0283,
|
| 1628 |
+
"step": 2700
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 28.83,
|
| 1632 |
+
"learning_rate": 1.0966349148003325e-05,
|
| 1633 |
+
"loss": 0.0313,
|
| 1634 |
+
"step": 2710
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 28.94,
|
| 1638 |
+
"learning_rate": 1.0774881968726722e-05,
|
| 1639 |
+
"loss": 0.0269,
|
| 1640 |
+
"step": 2720
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 29.04,
|
| 1644 |
+
"learning_rate": 1.0584640793192995e-05,
|
| 1645 |
+
"loss": 0.0336,
|
| 1646 |
+
"step": 2730
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 29.15,
|
| 1650 |
+
"learning_rate": 1.0395642017497648e-05,
|
| 1651 |
+
"loss": 0.0293,
|
| 1652 |
+
"step": 2740
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 29.26,
|
| 1656 |
+
"learning_rate": 1.020790193065883e-05,
|
| 1657 |
+
"loss": 0.027,
|
| 1658 |
+
"step": 2750
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 29.36,
|
| 1662 |
+
"learning_rate": 1.0021436713213606e-05,
|
| 1663 |
+
"loss": 0.0281,
|
| 1664 |
+
"step": 2760
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 29.47,
|
| 1668 |
+
"learning_rate": 9.836262435823316e-06,
|
| 1669 |
+
"loss": 0.031,
|
| 1670 |
+
"step": 2770
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 29.57,
|
| 1674 |
+
"learning_rate": 9.652395057888555e-06,
|
| 1675 |
+
"loss": 0.0284,
|
| 1676 |
+
"step": 2780
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 29.68,
|
| 1680 |
+
"learning_rate": 9.469850426173688e-06,
|
| 1681 |
+
"loss": 0.0292,
|
| 1682 |
+
"step": 2790
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 29.79,
|
| 1686 |
+
"learning_rate": 9.288644273441083e-06,
|
| 1687 |
+
"loss": 0.0284,
|
| 1688 |
+
"step": 2800
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 29.89,
|
| 1692 |
+
"learning_rate": 9.108792217095203e-06,
|
| 1693 |
+
"loss": 0.0285,
|
| 1694 |
+
"step": 2810
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 30.0,
|
| 1698 |
+
"learning_rate": 8.930309757836517e-06,
|
| 1699 |
+
"loss": 0.0257,
|
| 1700 |
+
"step": 2820
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 30.11,
|
| 1704 |
+
"learning_rate": 8.753212278325684e-06,
|
| 1705 |
+
"loss": 0.0309,
|
| 1706 |
+
"step": 2830
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 30.21,
|
| 1710 |
+
"learning_rate": 8.577515041857681e-06,
|
| 1711 |
+
"loss": 0.0274,
|
| 1712 |
+
"step": 2840
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 30.32,
|
| 1716 |
+
"learning_rate": 8.40323319104637e-06,
|
| 1717 |
+
"loss": 0.0309,
|
| 1718 |
+
"step": 2850
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 30.43,
|
| 1722 |
+
"learning_rate": 8.23038174651942e-06,
|
| 1723 |
+
"loss": 0.0274,
|
| 1724 |
+
"step": 2860
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 30.53,
|
| 1728 |
+
"learning_rate": 8.05897560562371e-06,
|
| 1729 |
+
"loss": 0.0263,
|
| 1730 |
+
"step": 2870
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 30.64,
|
| 1734 |
+
"learning_rate": 7.889029541141465e-06,
|
| 1735 |
+
"loss": 0.0293,
|
| 1736 |
+
"step": 2880
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 30.74,
|
| 1740 |
+
"learning_rate": 7.720558200016922e-06,
|
| 1741 |
+
"loss": 0.0294,
|
| 1742 |
+
"step": 2890
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 30.85,
|
| 1746 |
+
"learning_rate": 7.553576102094115e-06,
|
| 1747 |
+
"loss": 0.0281,
|
| 1748 |
+
"step": 2900
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 30.96,
|
| 1752 |
+
"learning_rate": 7.3880976388653625e-06,
|
| 1753 |
+
"loss": 0.0303,
|
| 1754 |
+
"step": 2910
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 31.06,
|
| 1758 |
+
"learning_rate": 7.224137072230982e-06,
|
| 1759 |
+
"loss": 0.028,
|
| 1760 |
+
"step": 2920
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 31.17,
|
| 1764 |
+
"learning_rate": 7.061708533270089e-06,
|
| 1765 |
+
"loss": 0.0284,
|
| 1766 |
+
"step": 2930
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 31.28,
|
| 1770 |
+
"learning_rate": 6.900826021022708e-06,
|
| 1771 |
+
"loss": 0.0294,
|
| 1772 |
+
"step": 2940
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 31.38,
|
| 1776 |
+
"learning_rate": 6.741503401283272e-06,
|
| 1777 |
+
"loss": 0.0293,
|
| 1778 |
+
"step": 2950
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 31.49,
|
| 1782 |
+
"learning_rate": 6.583754405405529e-06,
|
| 1783 |
+
"loss": 0.0247,
|
| 1784 |
+
"step": 2960
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 31.6,
|
| 1788 |
+
"learning_rate": 6.427592629119172e-06,
|
| 1789 |
+
"loss": 0.0257,
|
| 1790 |
+
"step": 2970
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 31.7,
|
| 1794 |
+
"learning_rate": 6.273031531358034e-06,
|
| 1795 |
+
"loss": 0.0275,
|
| 1796 |
+
"step": 2980
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 31.81,
|
| 1800 |
+
"learning_rate": 6.120084433100084e-06,
|
| 1801 |
+
"loss": 0.0302,
|
| 1802 |
+
"step": 2990
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 31.91,
|
| 1806 |
+
"learning_rate": 5.968764516219452e-06,
|
| 1807 |
+
"loss": 0.0291,
|
| 1808 |
+
"step": 3000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 32.02,
|
| 1812 |
+
"learning_rate": 5.819084822350224e-06,
|
| 1813 |
+
"loss": 0.0302,
|
| 1814 |
+
"step": 3010
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 32.13,
|
| 1818 |
+
"learning_rate": 5.671058251762529e-06,
|
| 1819 |
+
"loss": 0.03,
|
| 1820 |
+
"step": 3020
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 32.23,
|
| 1824 |
+
"learning_rate": 5.5246975622506354e-06,
|
| 1825 |
+
"loss": 0.0304,
|
| 1826 |
+
"step": 3030
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 32.34,
|
| 1830 |
+
"learning_rate": 5.380015368033476e-06,
|
| 1831 |
+
"loss": 0.0243,
|
| 1832 |
+
"step": 3040
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 32.45,
|
| 1836 |
+
"learning_rate": 5.2370241386674705e-06,
|
| 1837 |
+
"loss": 0.0292,
|
| 1838 |
+
"step": 3050
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 32.55,
|
| 1842 |
+
"learning_rate": 5.0957361979717764e-06,
|
| 1843 |
+
"loss": 0.0295,
|
| 1844 |
+
"step": 3060
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 32.66,
|
| 1848 |
+
"learning_rate": 4.956163722966223e-06,
|
| 1849 |
+
"loss": 0.0266,
|
| 1850 |
+
"step": 3070
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 32.77,
|
| 1854 |
+
"learning_rate": 4.818318742821765e-06,
|
| 1855 |
+
"loss": 0.0271,
|
| 1856 |
+
"step": 3080
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 32.87,
|
| 1860 |
+
"learning_rate": 4.682213137823774e-06,
|
| 1861 |
+
"loss": 0.0287,
|
| 1862 |
+
"step": 3090
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 32.98,
|
| 1866 |
+
"learning_rate": 4.547858638348107e-06,
|
| 1867 |
+
"loss": 0.0282,
|
| 1868 |
+
"step": 3100
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 33.09,
|
| 1872 |
+
"learning_rate": 4.415266823850117e-06,
|
| 1873 |
+
"loss": 0.029,
|
| 1874 |
+
"step": 3110
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 33.19,
|
| 1878 |
+
"learning_rate": 4.284449121866699e-06,
|
| 1879 |
+
"loss": 0.0273,
|
| 1880 |
+
"step": 3120
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 33.3,
|
| 1884 |
+
"learning_rate": 4.1554168070313255e-06,
|
| 1885 |
+
"loss": 0.0315,
|
| 1886 |
+
"step": 3130
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 33.4,
|
| 1890 |
+
"learning_rate": 4.028181000102424e-06,
|
| 1891 |
+
"loss": 0.0242,
|
| 1892 |
+
"step": 3140
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 33.51,
|
| 1896 |
+
"learning_rate": 3.90275266700485e-06,
|
| 1897 |
+
"loss": 0.0247,
|
| 1898 |
+
"step": 3150
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 33.62,
|
| 1902 |
+
"learning_rate": 3.779142617884823e-06,
|
| 1903 |
+
"loss": 0.0271,
|
| 1904 |
+
"step": 3160
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 33.72,
|
| 1908 |
+
"learning_rate": 3.65736150617822e-06,
|
| 1909 |
+
"loss": 0.0247,
|
| 1910 |
+
"step": 3170
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 33.83,
|
| 1914 |
+
"learning_rate": 3.5374198276924175e-06,
|
| 1915 |
+
"loss": 0.0273,
|
| 1916 |
+
"step": 3180
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 33.94,
|
| 1920 |
+
"learning_rate": 3.4193279197017094e-06,
|
| 1921 |
+
"loss": 0.0288,
|
| 1922 |
+
"step": 3190
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 34.04,
|
| 1926 |
+
"learning_rate": 3.3030959600563323e-06,
|
| 1927 |
+
"loss": 0.0313,
|
| 1928 |
+
"step": 3200
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 34.15,
|
| 1932 |
+
"learning_rate": 3.1887339663053463e-06,
|
| 1933 |
+
"loss": 0.0266,
|
| 1934 |
+
"step": 3210
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 34.26,
|
| 1938 |
+
"learning_rate": 3.0762517948332127e-06,
|
| 1939 |
+
"loss": 0.0283,
|
| 1940 |
+
"step": 3220
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 34.36,
|
| 1944 |
+
"learning_rate": 2.965659140010335e-06,
|
| 1945 |
+
"loss": 0.0263,
|
| 1946 |
+
"step": 3230
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 34.47,
|
| 1950 |
+
"learning_rate": 2.856965533357539e-06,
|
| 1951 |
+
"loss": 0.0282,
|
| 1952 |
+
"step": 3240
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 34.57,
|
| 1956 |
+
"learning_rate": 2.7501803427245816e-06,
|
| 1957 |
+
"loss": 0.0257,
|
| 1958 |
+
"step": 3250
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 34.68,
|
| 1962 |
+
"learning_rate": 2.645312771482805e-06,
|
| 1963 |
+
"loss": 0.0293,
|
| 1964 |
+
"step": 3260
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 34.79,
|
| 1968 |
+
"learning_rate": 2.542371857731879e-06,
|
| 1969 |
+
"loss": 0.0295,
|
| 1970 |
+
"step": 3270
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 34.89,
|
| 1974 |
+
"learning_rate": 2.441366473520909e-06,
|
| 1975 |
+
"loss": 0.026,
|
| 1976 |
+
"step": 3280
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 35.0,
|
| 1980 |
+
"learning_rate": 2.3423053240837515e-06,
|
| 1981 |
+
"loss": 0.0295,
|
| 1982 |
+
"step": 3290
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 35.11,
|
| 1986 |
+
"learning_rate": 2.245196947088754e-06,
|
| 1987 |
+
"loss": 0.0311,
|
| 1988 |
+
"step": 3300
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 35.21,
|
| 1992 |
+
"learning_rate": 2.1500497119029322e-06,
|
| 1993 |
+
"loss": 0.0279,
|
| 1994 |
+
"step": 3310
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 35.32,
|
| 1998 |
+
"learning_rate": 2.056871818870648e-06,
|
| 1999 |
+
"loss": 0.0261,
|
| 2000 |
+
"step": 3320
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 35.43,
|
| 2004 |
+
"learning_rate": 1.9656712986068664e-06,
|
| 2005 |
+
"loss": 0.0285,
|
| 2006 |
+
"step": 3330
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 35.53,
|
| 2010 |
+
"learning_rate": 1.8764560113050027e-06,
|
| 2011 |
+
"loss": 0.0282,
|
| 2012 |
+
"step": 3340
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 35.64,
|
| 2016 |
+
"learning_rate": 1.7892336460595043e-06,
|
| 2017 |
+
"loss": 0.0258,
|
| 2018 |
+
"step": 3350
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 35.74,
|
| 2022 |
+
"learning_rate": 1.7040117202031774e-06,
|
| 2023 |
+
"loss": 0.0256,
|
| 2024 |
+
"step": 3360
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 35.85,
|
| 2028 |
+
"learning_rate": 1.620797578659261e-06,
|
| 2029 |
+
"loss": 0.0287,
|
| 2030 |
+
"step": 3370
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 35.96,
|
| 2034 |
+
"learning_rate": 1.5395983933084196e-06,
|
| 2035 |
+
"loss": 0.0281,
|
| 2036 |
+
"step": 3380
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 36.06,
|
| 2040 |
+
"learning_rate": 1.4604211623706315e-06,
|
| 2041 |
+
"loss": 0.0263,
|
| 2042 |
+
"step": 3390
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 36.17,
|
| 2046 |
+
"learning_rate": 1.3832727098020332e-06,
|
| 2047 |
+
"loss": 0.0282,
|
| 2048 |
+
"step": 3400
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 36.28,
|
| 2052 |
+
"learning_rate": 1.308159684706789e-06,
|
| 2053 |
+
"loss": 0.029,
|
| 2054 |
+
"step": 3410
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 36.38,
|
| 2058 |
+
"learning_rate": 1.2350885607640445e-06,
|
| 2059 |
+
"loss": 0.0242,
|
| 2060 |
+
"step": 3420
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 36.49,
|
| 2064 |
+
"learning_rate": 1.1640656356699808e-06,
|
| 2065 |
+
"loss": 0.0262,
|
| 2066 |
+
"step": 3430
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 36.6,
|
| 2070 |
+
"learning_rate": 1.0950970305950353e-06,
|
| 2071 |
+
"loss": 0.0247,
|
| 2072 |
+
"step": 3440
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 36.7,
|
| 2076 |
+
"learning_rate": 1.028188689656351e-06,
|
| 2077 |
+
"loss": 0.0285,
|
| 2078 |
+
"step": 3450
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 36.81,
|
| 2082 |
+
"learning_rate": 9.633463794054776e-07,
|
| 2083 |
+
"loss": 0.0274,
|
| 2084 |
+
"step": 3460
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 36.91,
|
| 2088 |
+
"learning_rate": 9.005756883313759e-07,
|
| 2089 |
+
"loss": 0.0301,
|
| 2090 |
+
"step": 3470
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 37.02,
|
| 2094 |
+
"learning_rate": 8.39882026378766e-07,
|
| 2095 |
+
"loss": 0.0277,
|
| 2096 |
+
"step": 3480
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 37.13,
|
| 2100 |
+
"learning_rate": 7.812706244818669e-07,
|
| 2101 |
+
"loss": 0.0295,
|
| 2102 |
+
"step": 3490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 37.23,
|
| 2106 |
+
"learning_rate": 7.247465341135795e-07,
|
| 2107 |
+
"loss": 0.0279,
|
| 2108 |
+
"step": 3500
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 37.34,
|
| 2112 |
+
"learning_rate": 6.703146268500827e-07,
|
| 2113 |
+
"loss": 0.0255,
|
| 2114 |
+
"step": 3510
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 37.45,
|
| 2118 |
+
"learning_rate": 6.179795939510263e-07,
|
| 2119 |
+
"loss": 0.0286,
|
| 2120 |
+
"step": 3520
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 37.55,
|
| 2124 |
+
"learning_rate": 5.677459459551587e-07,
|
| 2125 |
+
"loss": 0.024,
|
| 2126 |
+
"step": 3530
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 37.66,
|
| 2130 |
+
"learning_rate": 5.1961801229162e-07,
|
| 2131 |
+
"loss": 0.0267,
|
| 2132 |
+
"step": 3540
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 37.77,
|
| 2136 |
+
"learning_rate": 4.7359994090678816e-07,
|
| 2137 |
+
"loss": 0.0289,
|
| 2138 |
+
"step": 3550
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 37.87,
|
| 2142 |
+
"learning_rate": 4.2969569790678456e-07,
|
| 2143 |
+
"loss": 0.0278,
|
| 2144 |
+
"step": 3560
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 37.98,
|
| 2148 |
+
"learning_rate": 3.8790906721566935e-07,
|
| 2149 |
+
"loss": 0.0259,
|
| 2150 |
+
"step": 3570
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 38.09,
|
| 2154 |
+
"learning_rate": 3.482436502492858e-07,
|
| 2155 |
+
"loss": 0.03,
|
| 2156 |
+
"step": 3580
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 38.19,
|
| 2160 |
+
"learning_rate": 3.1070286560491436e-07,
|
| 2161 |
+
"loss": 0.0289,
|
| 2162 |
+
"step": 3590
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 38.3,
|
| 2166 |
+
"learning_rate": 2.7528994876659663e-07,
|
| 2167 |
+
"loss": 0.0271,
|
| 2168 |
+
"step": 3600
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 38.4,
|
| 2172 |
+
"learning_rate": 2.420079518263119e-07,
|
| 2173 |
+
"loss": 0.0278,
|
| 2174 |
+
"step": 3610
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 38.51,
|
| 2178 |
+
"learning_rate": 2.1085974322091473e-07,
|
| 2179 |
+
"loss": 0.0258,
|
| 2180 |
+
"step": 3620
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 38.62,
|
| 2184 |
+
"learning_rate": 1.8184800748491937e-07,
|
| 2185 |
+
"loss": 0.0283,
|
| 2186 |
+
"step": 3630
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 38.72,
|
| 2190 |
+
"learning_rate": 1.5497524501913163e-07,
|
| 2191 |
+
"loss": 0.0249,
|
| 2192 |
+
"step": 3640
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 38.83,
|
| 2196 |
+
"learning_rate": 1.3024377187515203e-07,
|
| 2197 |
+
"loss": 0.0275,
|
| 2198 |
+
"step": 3650
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 38.94,
|
| 2202 |
+
"learning_rate": 1.076557195557576e-07,
|
| 2203 |
+
"loss": 0.0293,
|
| 2204 |
+
"step": 3660
|
| 2205 |
+
},
|
| 2206 |
+
{
|
| 2207 |
+
"epoch": 39.04,
|
| 2208 |
+
"learning_rate": 8.721303483121002e-08,
|
| 2209 |
+
"loss": 0.0278,
|
| 2210 |
+
"step": 3670
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 39.15,
|
| 2214 |
+
"learning_rate": 6.891747957145922e-08,
|
| 2215 |
+
"loss": 0.0286,
|
| 2216 |
+
"step": 3680
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 39.26,
|
| 2220 |
+
"learning_rate": 5.277063059429543e-08,
|
| 2221 |
+
"loss": 0.0295,
|
| 2222 |
+
"step": 3690
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 39.36,
|
| 2226 |
+
"learning_rate": 3.877387952945788e-08,
|
| 2227 |
+
"loss": 0.0272,
|
| 2228 |
+
"step": 3700
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 39.47,
|
| 2232 |
+
"learning_rate": 2.692843269869183e-08,
|
| 2233 |
+
"loss": 0.0277,
|
| 2234 |
+
"step": 3710
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 39.57,
|
| 2238 |
+
"learning_rate": 1.7235311011778997e-08,
|
| 2239 |
+
"loss": 0.0259,
|
| 2240 |
+
"step": 3720
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 39.68,
|
| 2244 |
+
"learning_rate": 9.695349878549586e-09,
|
| 2245 |
+
"loss": 0.0247,
|
| 2246 |
+
"step": 3730
|
| 2247 |
+
},
|
| 2248 |
+
{
|
| 2249 |
+
"epoch": 39.79,
|
| 2250 |
+
"learning_rate": 4.309199136889874e-09,
|
| 2251 |
+
"loss": 0.0317,
|
| 2252 |
+
"step": 3740
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 39.89,
|
| 2256 |
+
"learning_rate": 1.077322996717589e-09,
|
| 2257 |
+
"loss": 0.0285,
|
| 2258 |
+
"step": 3750
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 40.0,
|
| 2262 |
+
"learning_rate": 0.0,
|
| 2263 |
+
"loss": 0.0255,
|
| 2264 |
+
"step": 3760
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 40.0,
|
| 2268 |
+
"step": 3760,
|
| 2269 |
+
"total_flos": 4.579924393259827e+16,
|
| 2270 |
+
"train_loss": 0.608423255047107,
|
| 2271 |
+
"train_runtime": 971.1571,
|
| 2272 |
+
"train_samples_per_second": 247.457,
|
| 2273 |
+
"train_steps_per_second": 3.872
|
| 2274 |
+
}
|
| 2275 |
+
],
|
| 2276 |
+
"max_steps": 3760,
|
| 2277 |
+
"num_train_epochs": 40,
|
| 2278 |
+
"total_flos": 4.579924393259827e+16,
|
| 2279 |
+
"trial_name": null,
|
| 2280 |
+
"trial_params": null
|
| 2281 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e172c9f7b82d5d56a4ce5ed6d04a65d71d6a8dc5a7ed2595954b8b9f7aea0a12
|
| 3 |
+
size 3963
|