added model
Browse files- checkpoint-3750/config.json +60 -0
- checkpoint-3750/optimizer.pt +3 -0
- checkpoint-3750/pytorch_model.bin +3 -0
- checkpoint-3750/rng_state.pth +3 -0
- checkpoint-3750/scheduler.pt +3 -0
- checkpoint-3750/special_tokens_map.json +107 -0
- checkpoint-3750/tokenizer.json +0 -0
- checkpoint-3750/tokenizer_config.json +112 -0
- checkpoint-3750/trainer_state.json +2330 -0
- checkpoint-3750/training_args.bin +3 -0
- checkpoint-4000/config.json +60 -0
- checkpoint-4000/optimizer.pt +3 -0
- checkpoint-4000/pytorch_model.bin +3 -0
- checkpoint-4000/rng_state.pth +3 -0
- checkpoint-4000/scheduler.pt +3 -0
- checkpoint-4000/special_tokens_map.json +107 -0
- checkpoint-4000/tokenizer.json +0 -0
- checkpoint-4000/tokenizer_config.json +112 -0
- checkpoint-4000/trainer_state.json +2488 -0
- checkpoint-4000/training_args.bin +3 -0
- config.json +60 -0
- outputs.json +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +107 -0
- tokenizer.json +0 -0
- tokenizer_config.json +112 -0
- training_args.bin +3 -0
checkpoint-3750/config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "t5-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 4096,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 1024,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dense_act_fn": "relu",
|
| 11 |
+
"dropout_rate": 0.1,
|
| 12 |
+
"eos_token_id": 1,
|
| 13 |
+
"feed_forward_proj": "relu",
|
| 14 |
+
"initializer_factor": 1.0,
|
| 15 |
+
"is_encoder_decoder": true,
|
| 16 |
+
"is_gated_act": false,
|
| 17 |
+
"layer_norm_epsilon": 1e-06,
|
| 18 |
+
"model_type": "t5",
|
| 19 |
+
"n_positions": 512,
|
| 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 |
+
"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 |
+
"torch_dtype": "float32",
|
| 57 |
+
"transformers_version": "4.21.1",
|
| 58 |
+
"use_cache": true,
|
| 59 |
+
"vocab_size": 32128
|
| 60 |
+
}
|
checkpoint-3750/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5669dd7ce5ace51bac0029a9d08ac5a9816282c0be8c02df0984e09aae4e976d
|
| 3 |
+
size 5901650259
|
checkpoint-3750/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:15aa9beb5f66acb237e9b7d0f753aa4d0dbf01a107d0bc6131b782e0b7c8c77b
|
| 3 |
+
size 2950844807
|
checkpoint-3750/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb2cf15a9c125b4e5069961eaf06cf372f98b626cffb8dcabf8a2e164237d224
|
| 3 |
+
size 14503
|
checkpoint-3750/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c514d21168c417eaf3fb06d80c273b6fa48ae486e6698a9aa69f5565b809351
|
| 3 |
+
size 623
|
checkpoint-3750/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": "<unk>"
|
| 107 |
+
}
|
checkpoint-3750/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-3750/tokenizer_config.json
ADDED
|
@@ -0,0 +1,112 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"extra_ids": 100,
|
| 106 |
+
"model_max_length": 512,
|
| 107 |
+
"name_or_path": "t5-large",
|
| 108 |
+
"pad_token": "<pad>",
|
| 109 |
+
"special_tokens_map_file": null,
|
| 110 |
+
"tokenizer_class": "T5Tokenizer",
|
| 111 |
+
"unk_token": "<unk>"
|
| 112 |
+
}
|
checkpoint-3750/trainer_state.json
ADDED
|
@@ -0,0 +1,2330 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 8.84433962264151,
|
| 5 |
+
"global_step": 3750,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 1e-05,
|
| 13 |
+
"loss": 0.0022,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 2e-05,
|
| 19 |
+
"loss": 0.0023,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 3e-05,
|
| 25 |
+
"loss": 0.0024,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.09,
|
| 30 |
+
"learning_rate": 4e-05,
|
| 31 |
+
"loss": 0.0036,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 5e-05,
|
| 37 |
+
"loss": 0.0067,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.14,
|
| 42 |
+
"learning_rate": 6e-05,
|
| 43 |
+
"loss": 0.0108,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.17,
|
| 48 |
+
"learning_rate": 7e-05,
|
| 49 |
+
"loss": 0.0199,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"learning_rate": 8e-05,
|
| 55 |
+
"loss": 0.0332,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.21,
|
| 60 |
+
"learning_rate": 9e-05,
|
| 61 |
+
"loss": 0.0461,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.24,
|
| 66 |
+
"learning_rate": 0.0001,
|
| 67 |
+
"loss": 0.0791,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26,
|
| 72 |
+
"learning_rate": 9.97584541062802e-05,
|
| 73 |
+
"loss": 0.0862,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.28,
|
| 78 |
+
"learning_rate": 9.951690821256039e-05,
|
| 79 |
+
"loss": 0.0989,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.31,
|
| 84 |
+
"learning_rate": 9.927536231884058e-05,
|
| 85 |
+
"loss": 0.1003,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.33,
|
| 90 |
+
"learning_rate": 9.903381642512077e-05,
|
| 91 |
+
"loss": 0.1084,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.35,
|
| 96 |
+
"learning_rate": 9.879227053140096e-05,
|
| 97 |
+
"loss": 0.1114,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.38,
|
| 102 |
+
"learning_rate": 9.855072463768117e-05,
|
| 103 |
+
"loss": 0.1322,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.4,
|
| 108 |
+
"learning_rate": 9.830917874396136e-05,
|
| 109 |
+
"loss": 0.1178,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.42,
|
| 114 |
+
"learning_rate": 9.806763285024155e-05,
|
| 115 |
+
"loss": 0.0989,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.45,
|
| 120 |
+
"learning_rate": 9.782608695652174e-05,
|
| 121 |
+
"loss": 0.1154,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.47,
|
| 126 |
+
"learning_rate": 9.758454106280194e-05,
|
| 127 |
+
"loss": 0.1135,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.5,
|
| 132 |
+
"learning_rate": 9.734299516908213e-05,
|
| 133 |
+
"loss": 0.1299,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.52,
|
| 138 |
+
"learning_rate": 9.710144927536232e-05,
|
| 139 |
+
"loss": 0.1257,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.54,
|
| 144 |
+
"learning_rate": 9.685990338164251e-05,
|
| 145 |
+
"loss": 0.1217,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.57,
|
| 150 |
+
"learning_rate": 9.66183574879227e-05,
|
| 151 |
+
"loss": 0.1191,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.59,
|
| 156 |
+
"learning_rate": 9.63768115942029e-05,
|
| 157 |
+
"loss": 0.1159,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.61,
|
| 162 |
+
"learning_rate": 9.61352657004831e-05,
|
| 163 |
+
"loss": 0.1251,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.64,
|
| 168 |
+
"learning_rate": 9.589371980676329e-05,
|
| 169 |
+
"loss": 0.114,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.66,
|
| 174 |
+
"learning_rate": 9.565217391304348e-05,
|
| 175 |
+
"loss": 0.1302,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.68,
|
| 180 |
+
"learning_rate": 9.541062801932368e-05,
|
| 181 |
+
"loss": 0.1198,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.71,
|
| 186 |
+
"learning_rate": 9.516908212560387e-05,
|
| 187 |
+
"loss": 0.1275,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.73,
|
| 192 |
+
"learning_rate": 9.492753623188406e-05,
|
| 193 |
+
"loss": 0.1331,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.75,
|
| 198 |
+
"learning_rate": 9.468599033816425e-05,
|
| 199 |
+
"loss": 0.1252,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.78,
|
| 204 |
+
"learning_rate": 9.444444444444444e-05,
|
| 205 |
+
"loss": 0.1206,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.8,
|
| 210 |
+
"learning_rate": 9.420289855072463e-05,
|
| 211 |
+
"loss": 0.1356,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.83,
|
| 216 |
+
"learning_rate": 9.396135265700483e-05,
|
| 217 |
+
"loss": 0.1154,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.85,
|
| 222 |
+
"learning_rate": 9.371980676328503e-05,
|
| 223 |
+
"loss": 0.1306,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.87,
|
| 228 |
+
"learning_rate": 9.347826086956522e-05,
|
| 229 |
+
"loss": 0.1258,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.9,
|
| 234 |
+
"learning_rate": 9.323671497584541e-05,
|
| 235 |
+
"loss": 0.1342,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.92,
|
| 240 |
+
"learning_rate": 9.29951690821256e-05,
|
| 241 |
+
"loss": 0.133,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.94,
|
| 246 |
+
"learning_rate": 9.27536231884058e-05,
|
| 247 |
+
"loss": 0.1258,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.97,
|
| 252 |
+
"learning_rate": 9.251207729468599e-05,
|
| 253 |
+
"loss": 0.1414,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.99,
|
| 258 |
+
"learning_rate": 9.227053140096618e-05,
|
| 259 |
+
"loss": 0.1428,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.0,
|
| 264 |
+
"eval_loss": 1.8589301109313965,
|
| 265 |
+
"eval_runtime": 81.3061,
|
| 266 |
+
"eval_samples_per_second": 9.261,
|
| 267 |
+
"eval_steps_per_second": 0.59,
|
| 268 |
+
"step": 424
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.01,
|
| 272 |
+
"learning_rate": 9.202898550724637e-05,
|
| 273 |
+
"loss": 0.1159,
|
| 274 |
+
"step": 430
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 1.04,
|
| 278 |
+
"learning_rate": 9.178743961352657e-05,
|
| 279 |
+
"loss": 0.1101,
|
| 280 |
+
"step": 440
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.06,
|
| 284 |
+
"learning_rate": 9.154589371980676e-05,
|
| 285 |
+
"loss": 0.0938,
|
| 286 |
+
"step": 450
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 1.08,
|
| 290 |
+
"learning_rate": 9.130434782608696e-05,
|
| 291 |
+
"loss": 0.1014,
|
| 292 |
+
"step": 460
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 1.11,
|
| 296 |
+
"learning_rate": 9.106280193236715e-05,
|
| 297 |
+
"loss": 0.1137,
|
| 298 |
+
"step": 470
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.13,
|
| 302 |
+
"learning_rate": 9.082125603864735e-05,
|
| 303 |
+
"loss": 0.1279,
|
| 304 |
+
"step": 480
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 1.16,
|
| 308 |
+
"learning_rate": 9.057971014492754e-05,
|
| 309 |
+
"loss": 0.114,
|
| 310 |
+
"step": 490
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.18,
|
| 314 |
+
"learning_rate": 9.033816425120773e-05,
|
| 315 |
+
"loss": 0.1188,
|
| 316 |
+
"step": 500
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 1.2,
|
| 320 |
+
"learning_rate": 9.009661835748792e-05,
|
| 321 |
+
"loss": 0.1163,
|
| 322 |
+
"step": 510
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 1.23,
|
| 326 |
+
"learning_rate": 8.985507246376813e-05,
|
| 327 |
+
"loss": 0.1006,
|
| 328 |
+
"step": 520
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 1.25,
|
| 332 |
+
"learning_rate": 8.961352657004832e-05,
|
| 333 |
+
"loss": 0.1064,
|
| 334 |
+
"step": 530
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 1.27,
|
| 338 |
+
"learning_rate": 8.937198067632851e-05,
|
| 339 |
+
"loss": 0.1064,
|
| 340 |
+
"step": 540
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.3,
|
| 344 |
+
"learning_rate": 8.91304347826087e-05,
|
| 345 |
+
"loss": 0.11,
|
| 346 |
+
"step": 550
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 1.32,
|
| 350 |
+
"learning_rate": 8.888888888888889e-05,
|
| 351 |
+
"loss": 0.1193,
|
| 352 |
+
"step": 560
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.34,
|
| 356 |
+
"learning_rate": 8.864734299516908e-05,
|
| 357 |
+
"loss": 0.1274,
|
| 358 |
+
"step": 570
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 1.37,
|
| 362 |
+
"learning_rate": 8.840579710144929e-05,
|
| 363 |
+
"loss": 0.1181,
|
| 364 |
+
"step": 580
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 1.39,
|
| 368 |
+
"learning_rate": 8.816425120772948e-05,
|
| 369 |
+
"loss": 0.1038,
|
| 370 |
+
"step": 590
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 1.42,
|
| 374 |
+
"learning_rate": 8.792270531400967e-05,
|
| 375 |
+
"loss": 0.1156,
|
| 376 |
+
"step": 600
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 1.44,
|
| 380 |
+
"learning_rate": 8.768115942028987e-05,
|
| 381 |
+
"loss": 0.1176,
|
| 382 |
+
"step": 610
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 1.46,
|
| 386 |
+
"learning_rate": 8.743961352657006e-05,
|
| 387 |
+
"loss": 0.1154,
|
| 388 |
+
"step": 620
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.49,
|
| 392 |
+
"learning_rate": 8.719806763285025e-05,
|
| 393 |
+
"loss": 0.1328,
|
| 394 |
+
"step": 630
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 1.51,
|
| 398 |
+
"learning_rate": 8.695652173913044e-05,
|
| 399 |
+
"loss": 0.1039,
|
| 400 |
+
"step": 640
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 1.53,
|
| 404 |
+
"learning_rate": 8.671497584541063e-05,
|
| 405 |
+
"loss": 0.1311,
|
| 406 |
+
"step": 650
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 1.56,
|
| 410 |
+
"learning_rate": 8.647342995169082e-05,
|
| 411 |
+
"loss": 0.1172,
|
| 412 |
+
"step": 660
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.58,
|
| 416 |
+
"learning_rate": 8.623188405797103e-05,
|
| 417 |
+
"loss": 0.1266,
|
| 418 |
+
"step": 670
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 1.6,
|
| 422 |
+
"learning_rate": 8.599033816425122e-05,
|
| 423 |
+
"loss": 0.1052,
|
| 424 |
+
"step": 680
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 1.63,
|
| 428 |
+
"learning_rate": 8.574879227053141e-05,
|
| 429 |
+
"loss": 0.1311,
|
| 430 |
+
"step": 690
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.65,
|
| 434 |
+
"learning_rate": 8.55072463768116e-05,
|
| 435 |
+
"loss": 0.1162,
|
| 436 |
+
"step": 700
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 1.67,
|
| 440 |
+
"learning_rate": 8.52657004830918e-05,
|
| 441 |
+
"loss": 0.1131,
|
| 442 |
+
"step": 710
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.7,
|
| 446 |
+
"learning_rate": 8.502415458937199e-05,
|
| 447 |
+
"loss": 0.0934,
|
| 448 |
+
"step": 720
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 1.72,
|
| 452 |
+
"learning_rate": 8.478260869565218e-05,
|
| 453 |
+
"loss": 0.1117,
|
| 454 |
+
"step": 730
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 1.75,
|
| 458 |
+
"learning_rate": 8.454106280193237e-05,
|
| 459 |
+
"loss": 0.116,
|
| 460 |
+
"step": 740
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 1.77,
|
| 464 |
+
"learning_rate": 8.429951690821256e-05,
|
| 465 |
+
"loss": 0.1256,
|
| 466 |
+
"step": 750
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 1.79,
|
| 470 |
+
"learning_rate": 8.405797101449276e-05,
|
| 471 |
+
"loss": 0.0895,
|
| 472 |
+
"step": 760
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.82,
|
| 476 |
+
"learning_rate": 8.381642512077295e-05,
|
| 477 |
+
"loss": 0.1266,
|
| 478 |
+
"step": 770
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.84,
|
| 482 |
+
"learning_rate": 8.357487922705315e-05,
|
| 483 |
+
"loss": 0.1245,
|
| 484 |
+
"step": 780
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 1.86,
|
| 488 |
+
"learning_rate": 8.333333333333334e-05,
|
| 489 |
+
"loss": 0.11,
|
| 490 |
+
"step": 790
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 1.89,
|
| 494 |
+
"learning_rate": 8.309178743961354e-05,
|
| 495 |
+
"loss": 0.1081,
|
| 496 |
+
"step": 800
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.91,
|
| 500 |
+
"learning_rate": 8.285024154589373e-05,
|
| 501 |
+
"loss": 0.127,
|
| 502 |
+
"step": 810
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 1.93,
|
| 506 |
+
"learning_rate": 8.260869565217392e-05,
|
| 507 |
+
"loss": 0.115,
|
| 508 |
+
"step": 820
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 1.96,
|
| 512 |
+
"learning_rate": 8.236714975845411e-05,
|
| 513 |
+
"loss": 0.1197,
|
| 514 |
+
"step": 830
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.98,
|
| 518 |
+
"learning_rate": 8.21256038647343e-05,
|
| 519 |
+
"loss": 0.1224,
|
| 520 |
+
"step": 840
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 2.0,
|
| 524 |
+
"eval_loss": 1.8221625089645386,
|
| 525 |
+
"eval_runtime": 73.5866,
|
| 526 |
+
"eval_samples_per_second": 10.233,
|
| 527 |
+
"eval_steps_per_second": 0.652,
|
| 528 |
+
"step": 848
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 2.0,
|
| 532 |
+
"learning_rate": 8.18840579710145e-05,
|
| 533 |
+
"loss": 0.1095,
|
| 534 |
+
"step": 850
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 2.03,
|
| 538 |
+
"learning_rate": 8.164251207729469e-05,
|
| 539 |
+
"loss": 0.056,
|
| 540 |
+
"step": 860
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 2.05,
|
| 544 |
+
"learning_rate": 8.140096618357488e-05,
|
| 545 |
+
"loss": 0.0497,
|
| 546 |
+
"step": 870
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 2.08,
|
| 550 |
+
"learning_rate": 8.115942028985508e-05,
|
| 551 |
+
"loss": 0.0566,
|
| 552 |
+
"step": 880
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 2.1,
|
| 556 |
+
"learning_rate": 8.091787439613528e-05,
|
| 557 |
+
"loss": 0.0595,
|
| 558 |
+
"step": 890
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 2.12,
|
| 562 |
+
"learning_rate": 8.067632850241547e-05,
|
| 563 |
+
"loss": 0.0681,
|
| 564 |
+
"step": 900
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 2.15,
|
| 568 |
+
"learning_rate": 8.043478260869566e-05,
|
| 569 |
+
"loss": 0.0881,
|
| 570 |
+
"step": 910
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 2.17,
|
| 574 |
+
"learning_rate": 8.019323671497585e-05,
|
| 575 |
+
"loss": 0.0604,
|
| 576 |
+
"step": 920
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 2.19,
|
| 580 |
+
"learning_rate": 7.995169082125604e-05,
|
| 581 |
+
"loss": 0.0632,
|
| 582 |
+
"step": 930
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 2.22,
|
| 586 |
+
"learning_rate": 7.971014492753623e-05,
|
| 587 |
+
"loss": 0.0599,
|
| 588 |
+
"step": 940
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 2.24,
|
| 592 |
+
"learning_rate": 7.946859903381643e-05,
|
| 593 |
+
"loss": 0.0684,
|
| 594 |
+
"step": 950
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 2.26,
|
| 598 |
+
"learning_rate": 7.922705314009662e-05,
|
| 599 |
+
"loss": 0.0713,
|
| 600 |
+
"step": 960
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 2.29,
|
| 604 |
+
"learning_rate": 7.898550724637681e-05,
|
| 605 |
+
"loss": 0.0595,
|
| 606 |
+
"step": 970
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 2.31,
|
| 610 |
+
"learning_rate": 7.874396135265701e-05,
|
| 611 |
+
"loss": 0.0687,
|
| 612 |
+
"step": 980
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 2.33,
|
| 616 |
+
"learning_rate": 7.85024154589372e-05,
|
| 617 |
+
"loss": 0.0732,
|
| 618 |
+
"step": 990
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 2.36,
|
| 622 |
+
"learning_rate": 7.82608695652174e-05,
|
| 623 |
+
"loss": 0.0615,
|
| 624 |
+
"step": 1000
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 2.38,
|
| 628 |
+
"learning_rate": 7.801932367149759e-05,
|
| 629 |
+
"loss": 0.0636,
|
| 630 |
+
"step": 1010
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 2.41,
|
| 634 |
+
"learning_rate": 7.777777777777778e-05,
|
| 635 |
+
"loss": 0.0617,
|
| 636 |
+
"step": 1020
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.43,
|
| 640 |
+
"learning_rate": 7.753623188405797e-05,
|
| 641 |
+
"loss": 0.0776,
|
| 642 |
+
"step": 1030
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 2.45,
|
| 646 |
+
"learning_rate": 7.729468599033817e-05,
|
| 647 |
+
"loss": 0.0671,
|
| 648 |
+
"step": 1040
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 2.48,
|
| 652 |
+
"learning_rate": 7.705314009661836e-05,
|
| 653 |
+
"loss": 0.073,
|
| 654 |
+
"step": 1050
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 2.5,
|
| 658 |
+
"learning_rate": 7.681159420289855e-05,
|
| 659 |
+
"loss": 0.0724,
|
| 660 |
+
"step": 1060
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 2.52,
|
| 664 |
+
"learning_rate": 7.657004830917874e-05,
|
| 665 |
+
"loss": 0.0607,
|
| 666 |
+
"step": 1070
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 2.55,
|
| 670 |
+
"learning_rate": 7.632850241545893e-05,
|
| 671 |
+
"loss": 0.0645,
|
| 672 |
+
"step": 1080
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 2.57,
|
| 676 |
+
"learning_rate": 7.608695652173914e-05,
|
| 677 |
+
"loss": 0.0749,
|
| 678 |
+
"step": 1090
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 2.59,
|
| 682 |
+
"learning_rate": 7.584541062801933e-05,
|
| 683 |
+
"loss": 0.0594,
|
| 684 |
+
"step": 1100
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 2.62,
|
| 688 |
+
"learning_rate": 7.560386473429952e-05,
|
| 689 |
+
"loss": 0.0699,
|
| 690 |
+
"step": 1110
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 2.64,
|
| 694 |
+
"learning_rate": 7.536231884057971e-05,
|
| 695 |
+
"loss": 0.0827,
|
| 696 |
+
"step": 1120
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 2.67,
|
| 700 |
+
"learning_rate": 7.51207729468599e-05,
|
| 701 |
+
"loss": 0.0596,
|
| 702 |
+
"step": 1130
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 2.69,
|
| 706 |
+
"learning_rate": 7.48792270531401e-05,
|
| 707 |
+
"loss": 0.0754,
|
| 708 |
+
"step": 1140
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 2.71,
|
| 712 |
+
"learning_rate": 7.463768115942029e-05,
|
| 713 |
+
"loss": 0.0706,
|
| 714 |
+
"step": 1150
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 2.74,
|
| 718 |
+
"learning_rate": 7.439613526570048e-05,
|
| 719 |
+
"loss": 0.063,
|
| 720 |
+
"step": 1160
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 2.76,
|
| 724 |
+
"learning_rate": 7.415458937198067e-05,
|
| 725 |
+
"loss": 0.0665,
|
| 726 |
+
"step": 1170
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 2.78,
|
| 730 |
+
"learning_rate": 7.391304347826086e-05,
|
| 731 |
+
"loss": 0.0748,
|
| 732 |
+
"step": 1180
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 2.81,
|
| 736 |
+
"learning_rate": 7.367149758454107e-05,
|
| 737 |
+
"loss": 0.0594,
|
| 738 |
+
"step": 1190
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 2.83,
|
| 742 |
+
"learning_rate": 7.342995169082126e-05,
|
| 743 |
+
"loss": 0.0839,
|
| 744 |
+
"step": 1200
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.85,
|
| 748 |
+
"learning_rate": 7.318840579710145e-05,
|
| 749 |
+
"loss": 0.0693,
|
| 750 |
+
"step": 1210
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 2.88,
|
| 754 |
+
"learning_rate": 7.294685990338164e-05,
|
| 755 |
+
"loss": 0.0663,
|
| 756 |
+
"step": 1220
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 2.9,
|
| 760 |
+
"learning_rate": 7.270531400966184e-05,
|
| 761 |
+
"loss": 0.0752,
|
| 762 |
+
"step": 1230
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 2.92,
|
| 766 |
+
"learning_rate": 7.246376811594203e-05,
|
| 767 |
+
"loss": 0.092,
|
| 768 |
+
"step": 1240
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 2.95,
|
| 772 |
+
"learning_rate": 7.222222222222222e-05,
|
| 773 |
+
"loss": 0.0739,
|
| 774 |
+
"step": 1250
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.97,
|
| 778 |
+
"learning_rate": 7.198067632850241e-05,
|
| 779 |
+
"loss": 0.067,
|
| 780 |
+
"step": 1260
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 3.0,
|
| 784 |
+
"learning_rate": 7.17391304347826e-05,
|
| 785 |
+
"loss": 0.0768,
|
| 786 |
+
"step": 1270
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 3.0,
|
| 790 |
+
"eval_loss": 1.9414550065994263,
|
| 791 |
+
"eval_runtime": 73.8083,
|
| 792 |
+
"eval_samples_per_second": 10.202,
|
| 793 |
+
"eval_steps_per_second": 0.65,
|
| 794 |
+
"step": 1272
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 3.02,
|
| 798 |
+
"learning_rate": 7.14975845410628e-05,
|
| 799 |
+
"loss": 0.0424,
|
| 800 |
+
"step": 1280
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 3.04,
|
| 804 |
+
"learning_rate": 7.1256038647343e-05,
|
| 805 |
+
"loss": 0.0284,
|
| 806 |
+
"step": 1290
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.07,
|
| 810 |
+
"learning_rate": 7.101449275362319e-05,
|
| 811 |
+
"loss": 0.0286,
|
| 812 |
+
"step": 1300
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 3.09,
|
| 816 |
+
"learning_rate": 7.077294685990338e-05,
|
| 817 |
+
"loss": 0.0368,
|
| 818 |
+
"step": 1310
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 3.11,
|
| 822 |
+
"learning_rate": 7.053140096618357e-05,
|
| 823 |
+
"loss": 0.0286,
|
| 824 |
+
"step": 1320
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 3.14,
|
| 828 |
+
"learning_rate": 7.028985507246377e-05,
|
| 829 |
+
"loss": 0.0363,
|
| 830 |
+
"step": 1330
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.16,
|
| 834 |
+
"learning_rate": 7.004830917874396e-05,
|
| 835 |
+
"loss": 0.0414,
|
| 836 |
+
"step": 1340
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 3.18,
|
| 840 |
+
"learning_rate": 6.980676328502416e-05,
|
| 841 |
+
"loss": 0.0498,
|
| 842 |
+
"step": 1350
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 3.21,
|
| 846 |
+
"learning_rate": 6.956521739130436e-05,
|
| 847 |
+
"loss": 0.0288,
|
| 848 |
+
"step": 1360
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 3.23,
|
| 852 |
+
"learning_rate": 6.932367149758455e-05,
|
| 853 |
+
"loss": 0.0423,
|
| 854 |
+
"step": 1370
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 3.25,
|
| 858 |
+
"learning_rate": 6.908212560386474e-05,
|
| 859 |
+
"loss": 0.0372,
|
| 860 |
+
"step": 1380
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 3.28,
|
| 864 |
+
"learning_rate": 6.884057971014493e-05,
|
| 865 |
+
"loss": 0.0395,
|
| 866 |
+
"step": 1390
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 3.3,
|
| 870 |
+
"learning_rate": 6.859903381642512e-05,
|
| 871 |
+
"loss": 0.038,
|
| 872 |
+
"step": 1400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 3.33,
|
| 876 |
+
"learning_rate": 6.835748792270533e-05,
|
| 877 |
+
"loss": 0.0381,
|
| 878 |
+
"step": 1410
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 3.35,
|
| 882 |
+
"learning_rate": 6.811594202898552e-05,
|
| 883 |
+
"loss": 0.0379,
|
| 884 |
+
"step": 1420
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 3.37,
|
| 888 |
+
"learning_rate": 6.787439613526571e-05,
|
| 889 |
+
"loss": 0.0484,
|
| 890 |
+
"step": 1430
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.4,
|
| 894 |
+
"learning_rate": 6.76328502415459e-05,
|
| 895 |
+
"loss": 0.0459,
|
| 896 |
+
"step": 1440
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 3.42,
|
| 900 |
+
"learning_rate": 6.73913043478261e-05,
|
| 901 |
+
"loss": 0.0434,
|
| 902 |
+
"step": 1450
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 3.44,
|
| 906 |
+
"learning_rate": 6.714975845410629e-05,
|
| 907 |
+
"loss": 0.0448,
|
| 908 |
+
"step": 1460
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 3.47,
|
| 912 |
+
"learning_rate": 6.690821256038648e-05,
|
| 913 |
+
"loss": 0.0502,
|
| 914 |
+
"step": 1470
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 3.49,
|
| 918 |
+
"learning_rate": 6.666666666666667e-05,
|
| 919 |
+
"loss": 0.04,
|
| 920 |
+
"step": 1480
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 3.51,
|
| 924 |
+
"learning_rate": 6.642512077294686e-05,
|
| 925 |
+
"loss": 0.0383,
|
| 926 |
+
"step": 1490
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 3.54,
|
| 930 |
+
"learning_rate": 6.618357487922707e-05,
|
| 931 |
+
"loss": 0.0448,
|
| 932 |
+
"step": 1500
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 3.56,
|
| 936 |
+
"learning_rate": 6.594202898550726e-05,
|
| 937 |
+
"loss": 0.0378,
|
| 938 |
+
"step": 1510
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 3.58,
|
| 942 |
+
"learning_rate": 6.570048309178745e-05,
|
| 943 |
+
"loss": 0.0316,
|
| 944 |
+
"step": 1520
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 3.61,
|
| 948 |
+
"learning_rate": 6.545893719806764e-05,
|
| 949 |
+
"loss": 0.0457,
|
| 950 |
+
"step": 1530
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 3.63,
|
| 954 |
+
"learning_rate": 6.521739130434783e-05,
|
| 955 |
+
"loss": 0.0395,
|
| 956 |
+
"step": 1540
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 3.66,
|
| 960 |
+
"learning_rate": 6.497584541062803e-05,
|
| 961 |
+
"loss": 0.0409,
|
| 962 |
+
"step": 1550
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 3.68,
|
| 966 |
+
"learning_rate": 6.473429951690822e-05,
|
| 967 |
+
"loss": 0.045,
|
| 968 |
+
"step": 1560
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 3.7,
|
| 972 |
+
"learning_rate": 6.449275362318841e-05,
|
| 973 |
+
"loss": 0.0389,
|
| 974 |
+
"step": 1570
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 3.73,
|
| 978 |
+
"learning_rate": 6.42512077294686e-05,
|
| 979 |
+
"loss": 0.0396,
|
| 980 |
+
"step": 1580
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 3.75,
|
| 984 |
+
"learning_rate": 6.400966183574879e-05,
|
| 985 |
+
"loss": 0.0396,
|
| 986 |
+
"step": 1590
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 3.77,
|
| 990 |
+
"learning_rate": 6.376811594202898e-05,
|
| 991 |
+
"loss": 0.0388,
|
| 992 |
+
"step": 1600
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 3.8,
|
| 996 |
+
"learning_rate": 6.352657004830919e-05,
|
| 997 |
+
"loss": 0.0401,
|
| 998 |
+
"step": 1610
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 3.82,
|
| 1002 |
+
"learning_rate": 6.328502415458938e-05,
|
| 1003 |
+
"loss": 0.0441,
|
| 1004 |
+
"step": 1620
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 3.84,
|
| 1008 |
+
"learning_rate": 6.304347826086957e-05,
|
| 1009 |
+
"loss": 0.037,
|
| 1010 |
+
"step": 1630
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 3.87,
|
| 1014 |
+
"learning_rate": 6.280193236714976e-05,
|
| 1015 |
+
"loss": 0.0395,
|
| 1016 |
+
"step": 1640
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 3.89,
|
| 1020 |
+
"learning_rate": 6.256038647342996e-05,
|
| 1021 |
+
"loss": 0.0463,
|
| 1022 |
+
"step": 1650
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 3.92,
|
| 1026 |
+
"learning_rate": 6.231884057971015e-05,
|
| 1027 |
+
"loss": 0.0571,
|
| 1028 |
+
"step": 1660
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 3.94,
|
| 1032 |
+
"learning_rate": 6.207729468599034e-05,
|
| 1033 |
+
"loss": 0.0291,
|
| 1034 |
+
"step": 1670
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 3.96,
|
| 1038 |
+
"learning_rate": 6.183574879227053e-05,
|
| 1039 |
+
"loss": 0.0396,
|
| 1040 |
+
"step": 1680
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 3.99,
|
| 1044 |
+
"learning_rate": 6.159420289855072e-05,
|
| 1045 |
+
"loss": 0.0406,
|
| 1046 |
+
"step": 1690
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 4.0,
|
| 1050 |
+
"eval_loss": 1.9886054992675781,
|
| 1051 |
+
"eval_runtime": 74.4486,
|
| 1052 |
+
"eval_samples_per_second": 10.114,
|
| 1053 |
+
"eval_steps_per_second": 0.645,
|
| 1054 |
+
"step": 1696
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 4.01,
|
| 1058 |
+
"learning_rate": 6.135265700483092e-05,
|
| 1059 |
+
"loss": 0.0336,
|
| 1060 |
+
"step": 1700
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 4.03,
|
| 1064 |
+
"learning_rate": 6.111111111111112e-05,
|
| 1065 |
+
"loss": 0.0238,
|
| 1066 |
+
"step": 1710
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 4.06,
|
| 1070 |
+
"learning_rate": 6.086956521739131e-05,
|
| 1071 |
+
"loss": 0.0255,
|
| 1072 |
+
"step": 1720
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 4.08,
|
| 1076 |
+
"learning_rate": 6.0628019323671504e-05,
|
| 1077 |
+
"loss": 0.0217,
|
| 1078 |
+
"step": 1730
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 4.1,
|
| 1082 |
+
"learning_rate": 6.0386473429951696e-05,
|
| 1083 |
+
"loss": 0.024,
|
| 1084 |
+
"step": 1740
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 4.13,
|
| 1088 |
+
"learning_rate": 6.014492753623189e-05,
|
| 1089 |
+
"loss": 0.0245,
|
| 1090 |
+
"step": 1750
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 4.15,
|
| 1094 |
+
"learning_rate": 5.990338164251208e-05,
|
| 1095 |
+
"loss": 0.0203,
|
| 1096 |
+
"step": 1760
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 4.17,
|
| 1100 |
+
"learning_rate": 5.966183574879227e-05,
|
| 1101 |
+
"loss": 0.0192,
|
| 1102 |
+
"step": 1770
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 4.2,
|
| 1106 |
+
"learning_rate": 5.942028985507246e-05,
|
| 1107 |
+
"loss": 0.0238,
|
| 1108 |
+
"step": 1780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 4.22,
|
| 1112 |
+
"learning_rate": 5.9178743961352655e-05,
|
| 1113 |
+
"loss": 0.0213,
|
| 1114 |
+
"step": 1790
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 4.25,
|
| 1118 |
+
"learning_rate": 5.8937198067632847e-05,
|
| 1119 |
+
"loss": 0.024,
|
| 1120 |
+
"step": 1800
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 4.27,
|
| 1124 |
+
"learning_rate": 5.869565217391305e-05,
|
| 1125 |
+
"loss": 0.0137,
|
| 1126 |
+
"step": 1810
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 4.29,
|
| 1130 |
+
"learning_rate": 5.8454106280193244e-05,
|
| 1131 |
+
"loss": 0.0179,
|
| 1132 |
+
"step": 1820
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 4.32,
|
| 1136 |
+
"learning_rate": 5.8212560386473435e-05,
|
| 1137 |
+
"loss": 0.018,
|
| 1138 |
+
"step": 1830
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 4.34,
|
| 1142 |
+
"learning_rate": 5.797101449275363e-05,
|
| 1143 |
+
"loss": 0.0136,
|
| 1144 |
+
"step": 1840
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 4.36,
|
| 1148 |
+
"learning_rate": 5.772946859903382e-05,
|
| 1149 |
+
"loss": 0.0156,
|
| 1150 |
+
"step": 1850
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 4.39,
|
| 1154 |
+
"learning_rate": 5.748792270531401e-05,
|
| 1155 |
+
"loss": 0.0192,
|
| 1156 |
+
"step": 1860
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 4.41,
|
| 1160 |
+
"learning_rate": 5.72463768115942e-05,
|
| 1161 |
+
"loss": 0.0154,
|
| 1162 |
+
"step": 1870
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 4.43,
|
| 1166 |
+
"learning_rate": 5.7004830917874394e-05,
|
| 1167 |
+
"loss": 0.0197,
|
| 1168 |
+
"step": 1880
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 4.46,
|
| 1172 |
+
"learning_rate": 5.6763285024154586e-05,
|
| 1173 |
+
"loss": 0.0229,
|
| 1174 |
+
"step": 1890
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 4.48,
|
| 1178 |
+
"learning_rate": 5.652173913043478e-05,
|
| 1179 |
+
"loss": 0.0199,
|
| 1180 |
+
"step": 1900
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 4.5,
|
| 1184 |
+
"learning_rate": 5.6280193236714976e-05,
|
| 1185 |
+
"loss": 0.0222,
|
| 1186 |
+
"step": 1910
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 4.53,
|
| 1190 |
+
"learning_rate": 5.6038647342995175e-05,
|
| 1191 |
+
"loss": 0.0277,
|
| 1192 |
+
"step": 1920
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 4.55,
|
| 1196 |
+
"learning_rate": 5.5797101449275366e-05,
|
| 1197 |
+
"loss": 0.0246,
|
| 1198 |
+
"step": 1930
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 4.58,
|
| 1202 |
+
"learning_rate": 5.555555555555556e-05,
|
| 1203 |
+
"loss": 0.0252,
|
| 1204 |
+
"step": 1940
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 4.6,
|
| 1208 |
+
"learning_rate": 5.531400966183575e-05,
|
| 1209 |
+
"loss": 0.0269,
|
| 1210 |
+
"step": 1950
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 4.62,
|
| 1214 |
+
"learning_rate": 5.507246376811594e-05,
|
| 1215 |
+
"loss": 0.0183,
|
| 1216 |
+
"step": 1960
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 4.65,
|
| 1220 |
+
"learning_rate": 5.483091787439614e-05,
|
| 1221 |
+
"loss": 0.025,
|
| 1222 |
+
"step": 1970
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 4.67,
|
| 1226 |
+
"learning_rate": 5.458937198067633e-05,
|
| 1227 |
+
"loss": 0.0208,
|
| 1228 |
+
"step": 1980
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 4.69,
|
| 1232 |
+
"learning_rate": 5.4347826086956524e-05,
|
| 1233 |
+
"loss": 0.0198,
|
| 1234 |
+
"step": 1990
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 4.72,
|
| 1238 |
+
"learning_rate": 5.4106280193236716e-05,
|
| 1239 |
+
"loss": 0.0135,
|
| 1240 |
+
"step": 2000
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 4.74,
|
| 1244 |
+
"learning_rate": 5.386473429951691e-05,
|
| 1245 |
+
"loss": 0.0204,
|
| 1246 |
+
"step": 2010
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 4.76,
|
| 1250 |
+
"learning_rate": 5.3623188405797106e-05,
|
| 1251 |
+
"loss": 0.0226,
|
| 1252 |
+
"step": 2020
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 4.79,
|
| 1256 |
+
"learning_rate": 5.3381642512077304e-05,
|
| 1257 |
+
"loss": 0.019,
|
| 1258 |
+
"step": 2030
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 4.81,
|
| 1262 |
+
"learning_rate": 5.3140096618357496e-05,
|
| 1263 |
+
"loss": 0.0162,
|
| 1264 |
+
"step": 2040
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 4.83,
|
| 1268 |
+
"learning_rate": 5.289855072463769e-05,
|
| 1269 |
+
"loss": 0.0181,
|
| 1270 |
+
"step": 2050
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 4.86,
|
| 1274 |
+
"learning_rate": 5.265700483091788e-05,
|
| 1275 |
+
"loss": 0.0187,
|
| 1276 |
+
"step": 2060
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 4.88,
|
| 1280 |
+
"learning_rate": 5.241545893719807e-05,
|
| 1281 |
+
"loss": 0.0192,
|
| 1282 |
+
"step": 2070
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 4.91,
|
| 1286 |
+
"learning_rate": 5.217391304347826e-05,
|
| 1287 |
+
"loss": 0.0255,
|
| 1288 |
+
"step": 2080
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 4.93,
|
| 1292 |
+
"learning_rate": 5.1932367149758455e-05,
|
| 1293 |
+
"loss": 0.0258,
|
| 1294 |
+
"step": 2090
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 4.95,
|
| 1298 |
+
"learning_rate": 5.1690821256038647e-05,
|
| 1299 |
+
"loss": 0.0274,
|
| 1300 |
+
"step": 2100
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 4.98,
|
| 1304 |
+
"learning_rate": 5.144927536231884e-05,
|
| 1305 |
+
"loss": 0.0291,
|
| 1306 |
+
"step": 2110
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 5.0,
|
| 1310 |
+
"learning_rate": 5.1207729468599044e-05,
|
| 1311 |
+
"loss": 0.0422,
|
| 1312 |
+
"step": 2120
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 5.0,
|
| 1316 |
+
"eval_loss": 1.9903086423873901,
|
| 1317 |
+
"eval_runtime": 72.5407,
|
| 1318 |
+
"eval_samples_per_second": 10.38,
|
| 1319 |
+
"eval_steps_per_second": 0.662,
|
| 1320 |
+
"step": 2120
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 5.02,
|
| 1324 |
+
"learning_rate": 5.0966183574879235e-05,
|
| 1325 |
+
"loss": 0.0099,
|
| 1326 |
+
"step": 2130
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 5.05,
|
| 1330 |
+
"learning_rate": 5.072463768115943e-05,
|
| 1331 |
+
"loss": 0.0121,
|
| 1332 |
+
"step": 2140
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 5.07,
|
| 1336 |
+
"learning_rate": 5.048309178743962e-05,
|
| 1337 |
+
"loss": 0.008,
|
| 1338 |
+
"step": 2150
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 5.09,
|
| 1342 |
+
"learning_rate": 5.024154589371981e-05,
|
| 1343 |
+
"loss": 0.01,
|
| 1344 |
+
"step": 2160
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 5.12,
|
| 1348 |
+
"learning_rate": 5e-05,
|
| 1349 |
+
"loss": 0.0065,
|
| 1350 |
+
"step": 2170
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 5.14,
|
| 1354 |
+
"learning_rate": 4.9758454106280194e-05,
|
| 1355 |
+
"loss": 0.0096,
|
| 1356 |
+
"step": 2180
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 5.17,
|
| 1360 |
+
"learning_rate": 4.9516908212560386e-05,
|
| 1361 |
+
"loss": 0.0147,
|
| 1362 |
+
"step": 2190
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 5.19,
|
| 1366 |
+
"learning_rate": 4.9275362318840584e-05,
|
| 1367 |
+
"loss": 0.0094,
|
| 1368 |
+
"step": 2200
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 5.21,
|
| 1372 |
+
"learning_rate": 4.9033816425120776e-05,
|
| 1373 |
+
"loss": 0.0123,
|
| 1374 |
+
"step": 2210
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 5.24,
|
| 1378 |
+
"learning_rate": 4.879227053140097e-05,
|
| 1379 |
+
"loss": 0.007,
|
| 1380 |
+
"step": 2220
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 5.26,
|
| 1384 |
+
"learning_rate": 4.855072463768116e-05,
|
| 1385 |
+
"loss": 0.0136,
|
| 1386 |
+
"step": 2230
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 5.28,
|
| 1390 |
+
"learning_rate": 4.830917874396135e-05,
|
| 1391 |
+
"loss": 0.017,
|
| 1392 |
+
"step": 2240
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 5.31,
|
| 1396 |
+
"learning_rate": 4.806763285024155e-05,
|
| 1397 |
+
"loss": 0.0078,
|
| 1398 |
+
"step": 2250
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 5.33,
|
| 1402 |
+
"learning_rate": 4.782608695652174e-05,
|
| 1403 |
+
"loss": 0.011,
|
| 1404 |
+
"step": 2260
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 5.35,
|
| 1408 |
+
"learning_rate": 4.7584541062801933e-05,
|
| 1409 |
+
"loss": 0.0078,
|
| 1410 |
+
"step": 2270
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 5.38,
|
| 1414 |
+
"learning_rate": 4.7342995169082125e-05,
|
| 1415 |
+
"loss": 0.0118,
|
| 1416 |
+
"step": 2280
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 5.4,
|
| 1420 |
+
"learning_rate": 4.710144927536232e-05,
|
| 1421 |
+
"loss": 0.0072,
|
| 1422 |
+
"step": 2290
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 5.42,
|
| 1426 |
+
"learning_rate": 4.6859903381642516e-05,
|
| 1427 |
+
"loss": 0.0122,
|
| 1428 |
+
"step": 2300
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 5.45,
|
| 1432 |
+
"learning_rate": 4.661835748792271e-05,
|
| 1433 |
+
"loss": 0.0145,
|
| 1434 |
+
"step": 2310
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 5.47,
|
| 1438 |
+
"learning_rate": 4.63768115942029e-05,
|
| 1439 |
+
"loss": 0.0094,
|
| 1440 |
+
"step": 2320
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 5.5,
|
| 1444 |
+
"learning_rate": 4.613526570048309e-05,
|
| 1445 |
+
"loss": 0.0087,
|
| 1446 |
+
"step": 2330
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 5.52,
|
| 1450 |
+
"learning_rate": 4.589371980676328e-05,
|
| 1451 |
+
"loss": 0.0127,
|
| 1452 |
+
"step": 2340
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 5.54,
|
| 1456 |
+
"learning_rate": 4.565217391304348e-05,
|
| 1457 |
+
"loss": 0.0076,
|
| 1458 |
+
"step": 2350
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 5.57,
|
| 1462 |
+
"learning_rate": 4.541062801932367e-05,
|
| 1463 |
+
"loss": 0.0076,
|
| 1464 |
+
"step": 2360
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 5.59,
|
| 1468 |
+
"learning_rate": 4.5169082125603865e-05,
|
| 1469 |
+
"loss": 0.0074,
|
| 1470 |
+
"step": 2370
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 5.61,
|
| 1474 |
+
"learning_rate": 4.492753623188406e-05,
|
| 1475 |
+
"loss": 0.0093,
|
| 1476 |
+
"step": 2380
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 5.64,
|
| 1480 |
+
"learning_rate": 4.4685990338164255e-05,
|
| 1481 |
+
"loss": 0.0074,
|
| 1482 |
+
"step": 2390
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 5.66,
|
| 1486 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 1487 |
+
"loss": 0.0162,
|
| 1488 |
+
"step": 2400
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 5.68,
|
| 1492 |
+
"learning_rate": 4.4202898550724645e-05,
|
| 1493 |
+
"loss": 0.0146,
|
| 1494 |
+
"step": 2410
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 5.71,
|
| 1498 |
+
"learning_rate": 4.396135265700484e-05,
|
| 1499 |
+
"loss": 0.007,
|
| 1500 |
+
"step": 2420
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 5.73,
|
| 1504 |
+
"learning_rate": 4.371980676328503e-05,
|
| 1505 |
+
"loss": 0.01,
|
| 1506 |
+
"step": 2430
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 5.75,
|
| 1510 |
+
"learning_rate": 4.347826086956522e-05,
|
| 1511 |
+
"loss": 0.0098,
|
| 1512 |
+
"step": 2440
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 5.78,
|
| 1516 |
+
"learning_rate": 4.323671497584541e-05,
|
| 1517 |
+
"loss": 0.009,
|
| 1518 |
+
"step": 2450
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 5.8,
|
| 1522 |
+
"learning_rate": 4.299516908212561e-05,
|
| 1523 |
+
"loss": 0.0071,
|
| 1524 |
+
"step": 2460
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 5.83,
|
| 1528 |
+
"learning_rate": 4.27536231884058e-05,
|
| 1529 |
+
"loss": 0.0112,
|
| 1530 |
+
"step": 2470
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 5.85,
|
| 1534 |
+
"learning_rate": 4.2512077294685994e-05,
|
| 1535 |
+
"loss": 0.0135,
|
| 1536 |
+
"step": 2480
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 5.87,
|
| 1540 |
+
"learning_rate": 4.2270531400966186e-05,
|
| 1541 |
+
"loss": 0.0104,
|
| 1542 |
+
"step": 2490
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 5.9,
|
| 1546 |
+
"learning_rate": 4.202898550724638e-05,
|
| 1547 |
+
"loss": 0.0084,
|
| 1548 |
+
"step": 2500
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 5.92,
|
| 1552 |
+
"learning_rate": 4.1787439613526576e-05,
|
| 1553 |
+
"loss": 0.0098,
|
| 1554 |
+
"step": 2510
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 5.94,
|
| 1558 |
+
"learning_rate": 4.154589371980677e-05,
|
| 1559 |
+
"loss": 0.0083,
|
| 1560 |
+
"step": 2520
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 5.97,
|
| 1564 |
+
"learning_rate": 4.130434782608696e-05,
|
| 1565 |
+
"loss": 0.0088,
|
| 1566 |
+
"step": 2530
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 5.99,
|
| 1570 |
+
"learning_rate": 4.106280193236715e-05,
|
| 1571 |
+
"loss": 0.0134,
|
| 1572 |
+
"step": 2540
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 6.0,
|
| 1576 |
+
"eval_loss": 2.0551583766937256,
|
| 1577 |
+
"eval_runtime": 73.0209,
|
| 1578 |
+
"eval_samples_per_second": 10.312,
|
| 1579 |
+
"eval_steps_per_second": 0.657,
|
| 1580 |
+
"step": 2544
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 6.01,
|
| 1584 |
+
"learning_rate": 4.082125603864734e-05,
|
| 1585 |
+
"loss": 0.005,
|
| 1586 |
+
"step": 2550
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 6.04,
|
| 1590 |
+
"learning_rate": 4.057971014492754e-05,
|
| 1591 |
+
"loss": 0.0043,
|
| 1592 |
+
"step": 2560
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 6.06,
|
| 1596 |
+
"learning_rate": 4.0338164251207733e-05,
|
| 1597 |
+
"loss": 0.005,
|
| 1598 |
+
"step": 2570
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 6.08,
|
| 1602 |
+
"learning_rate": 4.0096618357487925e-05,
|
| 1603 |
+
"loss": 0.0021,
|
| 1604 |
+
"step": 2580
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 6.11,
|
| 1608 |
+
"learning_rate": 3.985507246376812e-05,
|
| 1609 |
+
"loss": 0.0028,
|
| 1610 |
+
"step": 2590
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 6.13,
|
| 1614 |
+
"learning_rate": 3.961352657004831e-05,
|
| 1615 |
+
"loss": 0.0027,
|
| 1616 |
+
"step": 2600
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 6.16,
|
| 1620 |
+
"learning_rate": 3.937198067632851e-05,
|
| 1621 |
+
"loss": 0.004,
|
| 1622 |
+
"step": 2610
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 6.18,
|
| 1626 |
+
"learning_rate": 3.91304347826087e-05,
|
| 1627 |
+
"loss": 0.0047,
|
| 1628 |
+
"step": 2620
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 6.2,
|
| 1632 |
+
"learning_rate": 3.888888888888889e-05,
|
| 1633 |
+
"loss": 0.0031,
|
| 1634 |
+
"step": 2630
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 6.23,
|
| 1638 |
+
"learning_rate": 3.864734299516908e-05,
|
| 1639 |
+
"loss": 0.0042,
|
| 1640 |
+
"step": 2640
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 6.25,
|
| 1644 |
+
"learning_rate": 3.8405797101449274e-05,
|
| 1645 |
+
"loss": 0.0037,
|
| 1646 |
+
"step": 2650
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 6.27,
|
| 1650 |
+
"learning_rate": 3.8164251207729466e-05,
|
| 1651 |
+
"loss": 0.0029,
|
| 1652 |
+
"step": 2660
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 6.3,
|
| 1656 |
+
"learning_rate": 3.7922705314009665e-05,
|
| 1657 |
+
"loss": 0.0018,
|
| 1658 |
+
"step": 2670
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 6.32,
|
| 1662 |
+
"learning_rate": 3.7681159420289856e-05,
|
| 1663 |
+
"loss": 0.0042,
|
| 1664 |
+
"step": 2680
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 6.34,
|
| 1668 |
+
"learning_rate": 3.743961352657005e-05,
|
| 1669 |
+
"loss": 0.0054,
|
| 1670 |
+
"step": 2690
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 6.37,
|
| 1674 |
+
"learning_rate": 3.719806763285024e-05,
|
| 1675 |
+
"loss": 0.0051,
|
| 1676 |
+
"step": 2700
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 6.39,
|
| 1680 |
+
"learning_rate": 3.695652173913043e-05,
|
| 1681 |
+
"loss": 0.0054,
|
| 1682 |
+
"step": 2710
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 6.42,
|
| 1686 |
+
"learning_rate": 3.671497584541063e-05,
|
| 1687 |
+
"loss": 0.0026,
|
| 1688 |
+
"step": 2720
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 6.44,
|
| 1692 |
+
"learning_rate": 3.647342995169082e-05,
|
| 1693 |
+
"loss": 0.0038,
|
| 1694 |
+
"step": 2730
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 6.46,
|
| 1698 |
+
"learning_rate": 3.6231884057971014e-05,
|
| 1699 |
+
"loss": 0.0051,
|
| 1700 |
+
"step": 2740
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 6.49,
|
| 1704 |
+
"learning_rate": 3.5990338164251205e-05,
|
| 1705 |
+
"loss": 0.003,
|
| 1706 |
+
"step": 2750
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 6.51,
|
| 1710 |
+
"learning_rate": 3.57487922705314e-05,
|
| 1711 |
+
"loss": 0.0028,
|
| 1712 |
+
"step": 2760
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 6.53,
|
| 1716 |
+
"learning_rate": 3.5507246376811596e-05,
|
| 1717 |
+
"loss": 0.0036,
|
| 1718 |
+
"step": 2770
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 6.56,
|
| 1722 |
+
"learning_rate": 3.526570048309179e-05,
|
| 1723 |
+
"loss": 0.0033,
|
| 1724 |
+
"step": 2780
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 6.58,
|
| 1728 |
+
"learning_rate": 3.502415458937198e-05,
|
| 1729 |
+
"loss": 0.0028,
|
| 1730 |
+
"step": 2790
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 6.6,
|
| 1734 |
+
"learning_rate": 3.478260869565218e-05,
|
| 1735 |
+
"loss": 0.0048,
|
| 1736 |
+
"step": 2800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 6.63,
|
| 1740 |
+
"learning_rate": 3.454106280193237e-05,
|
| 1741 |
+
"loss": 0.002,
|
| 1742 |
+
"step": 2810
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 6.65,
|
| 1746 |
+
"learning_rate": 3.429951690821256e-05,
|
| 1747 |
+
"loss": 0.0017,
|
| 1748 |
+
"step": 2820
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 6.67,
|
| 1752 |
+
"learning_rate": 3.405797101449276e-05,
|
| 1753 |
+
"loss": 0.0036,
|
| 1754 |
+
"step": 2830
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 6.7,
|
| 1758 |
+
"learning_rate": 3.381642512077295e-05,
|
| 1759 |
+
"loss": 0.0035,
|
| 1760 |
+
"step": 2840
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 6.72,
|
| 1764 |
+
"learning_rate": 3.357487922705314e-05,
|
| 1765 |
+
"loss": 0.0047,
|
| 1766 |
+
"step": 2850
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 6.75,
|
| 1770 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 1771 |
+
"loss": 0.0047,
|
| 1772 |
+
"step": 2860
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 6.77,
|
| 1776 |
+
"learning_rate": 3.3091787439613533e-05,
|
| 1777 |
+
"loss": 0.0028,
|
| 1778 |
+
"step": 2870
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 6.79,
|
| 1782 |
+
"learning_rate": 3.2850241545893725e-05,
|
| 1783 |
+
"loss": 0.0045,
|
| 1784 |
+
"step": 2880
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 6.82,
|
| 1788 |
+
"learning_rate": 3.260869565217392e-05,
|
| 1789 |
+
"loss": 0.0027,
|
| 1790 |
+
"step": 2890
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 6.84,
|
| 1794 |
+
"learning_rate": 3.236714975845411e-05,
|
| 1795 |
+
"loss": 0.002,
|
| 1796 |
+
"step": 2900
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 6.86,
|
| 1800 |
+
"learning_rate": 3.21256038647343e-05,
|
| 1801 |
+
"loss": 0.004,
|
| 1802 |
+
"step": 2910
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 6.89,
|
| 1806 |
+
"learning_rate": 3.188405797101449e-05,
|
| 1807 |
+
"loss": 0.0033,
|
| 1808 |
+
"step": 2920
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 6.91,
|
| 1812 |
+
"learning_rate": 3.164251207729469e-05,
|
| 1813 |
+
"loss": 0.0032,
|
| 1814 |
+
"step": 2930
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 6.93,
|
| 1818 |
+
"learning_rate": 3.140096618357488e-05,
|
| 1819 |
+
"loss": 0.0106,
|
| 1820 |
+
"step": 2940
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 6.96,
|
| 1824 |
+
"learning_rate": 3.1159420289855074e-05,
|
| 1825 |
+
"loss": 0.0025,
|
| 1826 |
+
"step": 2950
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 6.98,
|
| 1830 |
+
"learning_rate": 3.0917874396135266e-05,
|
| 1831 |
+
"loss": 0.0026,
|
| 1832 |
+
"step": 2960
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 7.0,
|
| 1836 |
+
"eval_loss": 2.094109296798706,
|
| 1837 |
+
"eval_runtime": 73.2503,
|
| 1838 |
+
"eval_samples_per_second": 10.28,
|
| 1839 |
+
"eval_steps_per_second": 0.655,
|
| 1840 |
+
"step": 2968
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 7.0,
|
| 1844 |
+
"learning_rate": 3.067632850241546e-05,
|
| 1845 |
+
"loss": 0.0051,
|
| 1846 |
+
"step": 2970
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 7.03,
|
| 1850 |
+
"learning_rate": 3.0434782608695656e-05,
|
| 1851 |
+
"loss": 0.0017,
|
| 1852 |
+
"step": 2980
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 7.05,
|
| 1856 |
+
"learning_rate": 3.0193236714975848e-05,
|
| 1857 |
+
"loss": 0.001,
|
| 1858 |
+
"step": 2990
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 7.08,
|
| 1862 |
+
"learning_rate": 2.995169082125604e-05,
|
| 1863 |
+
"loss": 0.0035,
|
| 1864 |
+
"step": 3000
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 7.1,
|
| 1868 |
+
"learning_rate": 2.971014492753623e-05,
|
| 1869 |
+
"loss": 0.0011,
|
| 1870 |
+
"step": 3010
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 7.12,
|
| 1874 |
+
"learning_rate": 2.9468599033816423e-05,
|
| 1875 |
+
"loss": 0.0022,
|
| 1876 |
+
"step": 3020
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 7.15,
|
| 1880 |
+
"learning_rate": 2.9227053140096622e-05,
|
| 1881 |
+
"loss": 0.0013,
|
| 1882 |
+
"step": 3030
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 7.17,
|
| 1886 |
+
"learning_rate": 2.8985507246376814e-05,
|
| 1887 |
+
"loss": 0.0024,
|
| 1888 |
+
"step": 3040
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 7.19,
|
| 1892 |
+
"learning_rate": 2.8743961352657005e-05,
|
| 1893 |
+
"loss": 0.001,
|
| 1894 |
+
"step": 3050
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 7.22,
|
| 1898 |
+
"learning_rate": 2.8502415458937197e-05,
|
| 1899 |
+
"loss": 0.0012,
|
| 1900 |
+
"step": 3060
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 7.24,
|
| 1904 |
+
"learning_rate": 2.826086956521739e-05,
|
| 1905 |
+
"loss": 0.001,
|
| 1906 |
+
"step": 3070
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 7.26,
|
| 1910 |
+
"learning_rate": 2.8019323671497587e-05,
|
| 1911 |
+
"loss": 0.0023,
|
| 1912 |
+
"step": 3080
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 7.29,
|
| 1916 |
+
"learning_rate": 2.777777777777778e-05,
|
| 1917 |
+
"loss": 0.0013,
|
| 1918 |
+
"step": 3090
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 7.31,
|
| 1922 |
+
"learning_rate": 2.753623188405797e-05,
|
| 1923 |
+
"loss": 0.001,
|
| 1924 |
+
"step": 3100
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 7.33,
|
| 1928 |
+
"learning_rate": 2.7294685990338166e-05,
|
| 1929 |
+
"loss": 0.0022,
|
| 1930 |
+
"step": 3110
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 7.36,
|
| 1934 |
+
"learning_rate": 2.7053140096618358e-05,
|
| 1935 |
+
"loss": 0.0011,
|
| 1936 |
+
"step": 3120
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 7.38,
|
| 1940 |
+
"learning_rate": 2.6811594202898553e-05,
|
| 1941 |
+
"loss": 0.001,
|
| 1942 |
+
"step": 3130
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 7.41,
|
| 1946 |
+
"learning_rate": 2.6570048309178748e-05,
|
| 1947 |
+
"loss": 0.0016,
|
| 1948 |
+
"step": 3140
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 7.43,
|
| 1952 |
+
"learning_rate": 2.632850241545894e-05,
|
| 1953 |
+
"loss": 0.0009,
|
| 1954 |
+
"step": 3150
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 7.45,
|
| 1958 |
+
"learning_rate": 2.608695652173913e-05,
|
| 1959 |
+
"loss": 0.0019,
|
| 1960 |
+
"step": 3160
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 7.48,
|
| 1964 |
+
"learning_rate": 2.5845410628019323e-05,
|
| 1965 |
+
"loss": 0.0013,
|
| 1966 |
+
"step": 3170
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 7.5,
|
| 1970 |
+
"learning_rate": 2.5603864734299522e-05,
|
| 1971 |
+
"loss": 0.0009,
|
| 1972 |
+
"step": 3180
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 7.52,
|
| 1976 |
+
"learning_rate": 2.5362318840579714e-05,
|
| 1977 |
+
"loss": 0.0011,
|
| 1978 |
+
"step": 3190
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 7.55,
|
| 1982 |
+
"learning_rate": 2.5120772946859905e-05,
|
| 1983 |
+
"loss": 0.0024,
|
| 1984 |
+
"step": 3200
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 7.57,
|
| 1988 |
+
"learning_rate": 2.4879227053140097e-05,
|
| 1989 |
+
"loss": 0.0013,
|
| 1990 |
+
"step": 3210
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 7.59,
|
| 1994 |
+
"learning_rate": 2.4637681159420292e-05,
|
| 1995 |
+
"loss": 0.0012,
|
| 1996 |
+
"step": 3220
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 7.62,
|
| 2000 |
+
"learning_rate": 2.4396135265700484e-05,
|
| 2001 |
+
"loss": 0.0011,
|
| 2002 |
+
"step": 3230
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 7.64,
|
| 2006 |
+
"learning_rate": 2.4154589371980676e-05,
|
| 2007 |
+
"loss": 0.0022,
|
| 2008 |
+
"step": 3240
|
| 2009 |
+
},
|
| 2010 |
+
{
|
| 2011 |
+
"epoch": 7.67,
|
| 2012 |
+
"learning_rate": 2.391304347826087e-05,
|
| 2013 |
+
"loss": 0.0011,
|
| 2014 |
+
"step": 3250
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 7.69,
|
| 2018 |
+
"learning_rate": 2.3671497584541063e-05,
|
| 2019 |
+
"loss": 0.0074,
|
| 2020 |
+
"step": 3260
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 7.71,
|
| 2024 |
+
"learning_rate": 2.3429951690821258e-05,
|
| 2025 |
+
"loss": 0.001,
|
| 2026 |
+
"step": 3270
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 7.74,
|
| 2030 |
+
"learning_rate": 2.318840579710145e-05,
|
| 2031 |
+
"loss": 0.0016,
|
| 2032 |
+
"step": 3280
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 7.76,
|
| 2036 |
+
"learning_rate": 2.294685990338164e-05,
|
| 2037 |
+
"loss": 0.0028,
|
| 2038 |
+
"step": 3290
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 7.78,
|
| 2042 |
+
"learning_rate": 2.2705314009661836e-05,
|
| 2043 |
+
"loss": 0.0015,
|
| 2044 |
+
"step": 3300
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 7.81,
|
| 2048 |
+
"learning_rate": 2.246376811594203e-05,
|
| 2049 |
+
"loss": 0.0038,
|
| 2050 |
+
"step": 3310
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 7.83,
|
| 2054 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 2055 |
+
"loss": 0.0013,
|
| 2056 |
+
"step": 3320
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 7.85,
|
| 2060 |
+
"learning_rate": 2.198067632850242e-05,
|
| 2061 |
+
"loss": 0.0009,
|
| 2062 |
+
"step": 3330
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 7.88,
|
| 2066 |
+
"learning_rate": 2.173913043478261e-05,
|
| 2067 |
+
"loss": 0.001,
|
| 2068 |
+
"step": 3340
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 7.9,
|
| 2072 |
+
"learning_rate": 2.1497584541062805e-05,
|
| 2073 |
+
"loss": 0.001,
|
| 2074 |
+
"step": 3350
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 7.92,
|
| 2078 |
+
"learning_rate": 2.1256038647342997e-05,
|
| 2079 |
+
"loss": 0.001,
|
| 2080 |
+
"step": 3360
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 7.95,
|
| 2084 |
+
"learning_rate": 2.101449275362319e-05,
|
| 2085 |
+
"loss": 0.0018,
|
| 2086 |
+
"step": 3370
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 7.97,
|
| 2090 |
+
"learning_rate": 2.0772946859903384e-05,
|
| 2091 |
+
"loss": 0.0015,
|
| 2092 |
+
"step": 3380
|
| 2093 |
+
},
|
| 2094 |
+
{
|
| 2095 |
+
"epoch": 8.0,
|
| 2096 |
+
"learning_rate": 2.0531400966183576e-05,
|
| 2097 |
+
"loss": 0.001,
|
| 2098 |
+
"step": 3390
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 8.0,
|
| 2102 |
+
"eval_loss": 2.134007215499878,
|
| 2103 |
+
"eval_runtime": 73.9295,
|
| 2104 |
+
"eval_samples_per_second": 10.185,
|
| 2105 |
+
"eval_steps_per_second": 0.649,
|
| 2106 |
+
"step": 3392
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 8.02,
|
| 2110 |
+
"learning_rate": 2.028985507246377e-05,
|
| 2111 |
+
"loss": 0.0011,
|
| 2112 |
+
"step": 3400
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 8.04,
|
| 2116 |
+
"learning_rate": 2.0048309178743963e-05,
|
| 2117 |
+
"loss": 0.0014,
|
| 2118 |
+
"step": 3410
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 8.07,
|
| 2122 |
+
"learning_rate": 1.9806763285024154e-05,
|
| 2123 |
+
"loss": 0.0007,
|
| 2124 |
+
"step": 3420
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 8.09,
|
| 2128 |
+
"learning_rate": 1.956521739130435e-05,
|
| 2129 |
+
"loss": 0.0008,
|
| 2130 |
+
"step": 3430
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 8.11,
|
| 2134 |
+
"learning_rate": 1.932367149758454e-05,
|
| 2135 |
+
"loss": 0.0007,
|
| 2136 |
+
"step": 3440
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 8.14,
|
| 2140 |
+
"learning_rate": 1.9082125603864733e-05,
|
| 2141 |
+
"loss": 0.0008,
|
| 2142 |
+
"step": 3450
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 8.16,
|
| 2146 |
+
"learning_rate": 1.8840579710144928e-05,
|
| 2147 |
+
"loss": 0.0008,
|
| 2148 |
+
"step": 3460
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 8.18,
|
| 2152 |
+
"learning_rate": 1.859903381642512e-05,
|
| 2153 |
+
"loss": 0.0009,
|
| 2154 |
+
"step": 3470
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 8.21,
|
| 2158 |
+
"learning_rate": 1.8357487922705315e-05,
|
| 2159 |
+
"loss": 0.0007,
|
| 2160 |
+
"step": 3480
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 8.23,
|
| 2164 |
+
"learning_rate": 1.8115942028985507e-05,
|
| 2165 |
+
"loss": 0.0012,
|
| 2166 |
+
"step": 3490
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 8.25,
|
| 2170 |
+
"learning_rate": 1.78743961352657e-05,
|
| 2171 |
+
"loss": 0.0007,
|
| 2172 |
+
"step": 3500
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 8.28,
|
| 2176 |
+
"learning_rate": 1.7632850241545894e-05,
|
| 2177 |
+
"loss": 0.003,
|
| 2178 |
+
"step": 3510
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 8.3,
|
| 2182 |
+
"learning_rate": 1.739130434782609e-05,
|
| 2183 |
+
"loss": 0.0013,
|
| 2184 |
+
"step": 3520
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 8.33,
|
| 2188 |
+
"learning_rate": 1.714975845410628e-05,
|
| 2189 |
+
"loss": 0.0009,
|
| 2190 |
+
"step": 3530
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 8.35,
|
| 2194 |
+
"learning_rate": 1.6908212560386476e-05,
|
| 2195 |
+
"loss": 0.0011,
|
| 2196 |
+
"step": 3540
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 8.37,
|
| 2200 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 2201 |
+
"loss": 0.0009,
|
| 2202 |
+
"step": 3550
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 8.4,
|
| 2206 |
+
"learning_rate": 1.6425120772946863e-05,
|
| 2207 |
+
"loss": 0.0009,
|
| 2208 |
+
"step": 3560
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 8.42,
|
| 2212 |
+
"learning_rate": 1.6183574879227054e-05,
|
| 2213 |
+
"loss": 0.001,
|
| 2214 |
+
"step": 3570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 8.44,
|
| 2218 |
+
"learning_rate": 1.5942028985507246e-05,
|
| 2219 |
+
"loss": 0.0008,
|
| 2220 |
+
"step": 3580
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 8.47,
|
| 2224 |
+
"learning_rate": 1.570048309178744e-05,
|
| 2225 |
+
"loss": 0.0008,
|
| 2226 |
+
"step": 3590
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 8.49,
|
| 2230 |
+
"learning_rate": 1.5458937198067633e-05,
|
| 2231 |
+
"loss": 0.001,
|
| 2232 |
+
"step": 3600
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 8.51,
|
| 2236 |
+
"learning_rate": 1.5217391304347828e-05,
|
| 2237 |
+
"loss": 0.0007,
|
| 2238 |
+
"step": 3610
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 8.54,
|
| 2242 |
+
"learning_rate": 1.497584541062802e-05,
|
| 2243 |
+
"loss": 0.0016,
|
| 2244 |
+
"step": 3620
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 8.56,
|
| 2248 |
+
"learning_rate": 1.4734299516908212e-05,
|
| 2249 |
+
"loss": 0.0009,
|
| 2250 |
+
"step": 3630
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 8.58,
|
| 2254 |
+
"learning_rate": 1.4492753623188407e-05,
|
| 2255 |
+
"loss": 0.0007,
|
| 2256 |
+
"step": 3640
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 8.61,
|
| 2260 |
+
"learning_rate": 1.4251207729468599e-05,
|
| 2261 |
+
"loss": 0.0007,
|
| 2262 |
+
"step": 3650
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 8.63,
|
| 2266 |
+
"learning_rate": 1.4009661835748794e-05,
|
| 2267 |
+
"loss": 0.0007,
|
| 2268 |
+
"step": 3660
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 8.66,
|
| 2272 |
+
"learning_rate": 1.3768115942028985e-05,
|
| 2273 |
+
"loss": 0.0023,
|
| 2274 |
+
"step": 3670
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 8.68,
|
| 2278 |
+
"learning_rate": 1.3526570048309179e-05,
|
| 2279 |
+
"loss": 0.0008,
|
| 2280 |
+
"step": 3680
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 8.7,
|
| 2284 |
+
"learning_rate": 1.3285024154589374e-05,
|
| 2285 |
+
"loss": 0.0007,
|
| 2286 |
+
"step": 3690
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 8.73,
|
| 2290 |
+
"learning_rate": 1.3043478260869566e-05,
|
| 2291 |
+
"loss": 0.0007,
|
| 2292 |
+
"step": 3700
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 8.75,
|
| 2296 |
+
"learning_rate": 1.2801932367149761e-05,
|
| 2297 |
+
"loss": 0.0008,
|
| 2298 |
+
"step": 3710
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 8.77,
|
| 2302 |
+
"learning_rate": 1.2560386473429953e-05,
|
| 2303 |
+
"loss": 0.0009,
|
| 2304 |
+
"step": 3720
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 8.8,
|
| 2308 |
+
"learning_rate": 1.2318840579710146e-05,
|
| 2309 |
+
"loss": 0.0007,
|
| 2310 |
+
"step": 3730
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 8.82,
|
| 2314 |
+
"learning_rate": 1.2077294685990338e-05,
|
| 2315 |
+
"loss": 0.0008,
|
| 2316 |
+
"step": 3740
|
| 2317 |
+
},
|
| 2318 |
+
{
|
| 2319 |
+
"epoch": 8.84,
|
| 2320 |
+
"learning_rate": 1.1835748792270531e-05,
|
| 2321 |
+
"loss": 0.0012,
|
| 2322 |
+
"step": 3750
|
| 2323 |
+
}
|
| 2324 |
+
],
|
| 2325 |
+
"max_steps": 4240,
|
| 2326 |
+
"num_train_epochs": 10,
|
| 2327 |
+
"total_flos": 0.0,
|
| 2328 |
+
"trial_name": null,
|
| 2329 |
+
"trial_params": null
|
| 2330 |
+
}
|
checkpoint-3750/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7b27dc19e266bc07c05772a12a4c0ba666e04d9c5ca7a7c6a3ae1c9c6c2ce46
|
| 3 |
+
size 3759
|
checkpoint-4000/config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "t5-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 4096,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 1024,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dense_act_fn": "relu",
|
| 11 |
+
"dropout_rate": 0.1,
|
| 12 |
+
"eos_token_id": 1,
|
| 13 |
+
"feed_forward_proj": "relu",
|
| 14 |
+
"initializer_factor": 1.0,
|
| 15 |
+
"is_encoder_decoder": true,
|
| 16 |
+
"is_gated_act": false,
|
| 17 |
+
"layer_norm_epsilon": 1e-06,
|
| 18 |
+
"model_type": "t5",
|
| 19 |
+
"n_positions": 512,
|
| 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 |
+
"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 |
+
"torch_dtype": "float32",
|
| 57 |
+
"transformers_version": "4.21.1",
|
| 58 |
+
"use_cache": true,
|
| 59 |
+
"vocab_size": 32128
|
| 60 |
+
}
|
checkpoint-4000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a59a46ca32af31a1a244bc07c78c6987317a87983ec4b037d83ad0349e6a70da
|
| 3 |
+
size 5901650259
|
checkpoint-4000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dca6f05cb19950f195ccfc427e2e232054560f7487a5ddd7292625e973c82194
|
| 3 |
+
size 2950844807
|
checkpoint-4000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9885b98c4ec2860b7f3ea8dae7a7ec0e71b3769e68ea2821165e5d2e5c4e057c
|
| 3 |
+
size 14503
|
checkpoint-4000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:456c7042a6e7dcd2efa81e79cd4ec8a30a285830b62fa0dca6d8f7c3c79c077c
|
| 3 |
+
size 623
|
checkpoint-4000/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": "<unk>"
|
| 107 |
+
}
|
checkpoint-4000/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-4000/tokenizer_config.json
ADDED
|
@@ -0,0 +1,112 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"extra_ids": 100,
|
| 106 |
+
"model_max_length": 512,
|
| 107 |
+
"name_or_path": "t5-large",
|
| 108 |
+
"pad_token": "<pad>",
|
| 109 |
+
"special_tokens_map_file": null,
|
| 110 |
+
"tokenizer_class": "T5Tokenizer",
|
| 111 |
+
"unk_token": "<unk>"
|
| 112 |
+
}
|
checkpoint-4000/trainer_state.json
ADDED
|
@@ -0,0 +1,2488 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.433962264150944,
|
| 5 |
+
"global_step": 4000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 1e-05,
|
| 13 |
+
"loss": 0.0022,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 2e-05,
|
| 19 |
+
"loss": 0.0023,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 3e-05,
|
| 25 |
+
"loss": 0.0024,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.09,
|
| 30 |
+
"learning_rate": 4e-05,
|
| 31 |
+
"loss": 0.0036,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 5e-05,
|
| 37 |
+
"loss": 0.0067,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.14,
|
| 42 |
+
"learning_rate": 6e-05,
|
| 43 |
+
"loss": 0.0108,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.17,
|
| 48 |
+
"learning_rate": 7e-05,
|
| 49 |
+
"loss": 0.0199,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"learning_rate": 8e-05,
|
| 55 |
+
"loss": 0.0332,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.21,
|
| 60 |
+
"learning_rate": 9e-05,
|
| 61 |
+
"loss": 0.0461,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.24,
|
| 66 |
+
"learning_rate": 0.0001,
|
| 67 |
+
"loss": 0.0791,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26,
|
| 72 |
+
"learning_rate": 9.97584541062802e-05,
|
| 73 |
+
"loss": 0.0862,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.28,
|
| 78 |
+
"learning_rate": 9.951690821256039e-05,
|
| 79 |
+
"loss": 0.0989,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.31,
|
| 84 |
+
"learning_rate": 9.927536231884058e-05,
|
| 85 |
+
"loss": 0.1003,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.33,
|
| 90 |
+
"learning_rate": 9.903381642512077e-05,
|
| 91 |
+
"loss": 0.1084,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.35,
|
| 96 |
+
"learning_rate": 9.879227053140096e-05,
|
| 97 |
+
"loss": 0.1114,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.38,
|
| 102 |
+
"learning_rate": 9.855072463768117e-05,
|
| 103 |
+
"loss": 0.1322,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.4,
|
| 108 |
+
"learning_rate": 9.830917874396136e-05,
|
| 109 |
+
"loss": 0.1178,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.42,
|
| 114 |
+
"learning_rate": 9.806763285024155e-05,
|
| 115 |
+
"loss": 0.0989,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.45,
|
| 120 |
+
"learning_rate": 9.782608695652174e-05,
|
| 121 |
+
"loss": 0.1154,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.47,
|
| 126 |
+
"learning_rate": 9.758454106280194e-05,
|
| 127 |
+
"loss": 0.1135,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.5,
|
| 132 |
+
"learning_rate": 9.734299516908213e-05,
|
| 133 |
+
"loss": 0.1299,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.52,
|
| 138 |
+
"learning_rate": 9.710144927536232e-05,
|
| 139 |
+
"loss": 0.1257,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.54,
|
| 144 |
+
"learning_rate": 9.685990338164251e-05,
|
| 145 |
+
"loss": 0.1217,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.57,
|
| 150 |
+
"learning_rate": 9.66183574879227e-05,
|
| 151 |
+
"loss": 0.1191,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.59,
|
| 156 |
+
"learning_rate": 9.63768115942029e-05,
|
| 157 |
+
"loss": 0.1159,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.61,
|
| 162 |
+
"learning_rate": 9.61352657004831e-05,
|
| 163 |
+
"loss": 0.1251,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.64,
|
| 168 |
+
"learning_rate": 9.589371980676329e-05,
|
| 169 |
+
"loss": 0.114,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.66,
|
| 174 |
+
"learning_rate": 9.565217391304348e-05,
|
| 175 |
+
"loss": 0.1302,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.68,
|
| 180 |
+
"learning_rate": 9.541062801932368e-05,
|
| 181 |
+
"loss": 0.1198,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.71,
|
| 186 |
+
"learning_rate": 9.516908212560387e-05,
|
| 187 |
+
"loss": 0.1275,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.73,
|
| 192 |
+
"learning_rate": 9.492753623188406e-05,
|
| 193 |
+
"loss": 0.1331,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.75,
|
| 198 |
+
"learning_rate": 9.468599033816425e-05,
|
| 199 |
+
"loss": 0.1252,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.78,
|
| 204 |
+
"learning_rate": 9.444444444444444e-05,
|
| 205 |
+
"loss": 0.1206,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.8,
|
| 210 |
+
"learning_rate": 9.420289855072463e-05,
|
| 211 |
+
"loss": 0.1356,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.83,
|
| 216 |
+
"learning_rate": 9.396135265700483e-05,
|
| 217 |
+
"loss": 0.1154,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.85,
|
| 222 |
+
"learning_rate": 9.371980676328503e-05,
|
| 223 |
+
"loss": 0.1306,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.87,
|
| 228 |
+
"learning_rate": 9.347826086956522e-05,
|
| 229 |
+
"loss": 0.1258,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.9,
|
| 234 |
+
"learning_rate": 9.323671497584541e-05,
|
| 235 |
+
"loss": 0.1342,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.92,
|
| 240 |
+
"learning_rate": 9.29951690821256e-05,
|
| 241 |
+
"loss": 0.133,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.94,
|
| 246 |
+
"learning_rate": 9.27536231884058e-05,
|
| 247 |
+
"loss": 0.1258,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.97,
|
| 252 |
+
"learning_rate": 9.251207729468599e-05,
|
| 253 |
+
"loss": 0.1414,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.99,
|
| 258 |
+
"learning_rate": 9.227053140096618e-05,
|
| 259 |
+
"loss": 0.1428,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.0,
|
| 264 |
+
"eval_loss": 1.8589301109313965,
|
| 265 |
+
"eval_runtime": 81.3061,
|
| 266 |
+
"eval_samples_per_second": 9.261,
|
| 267 |
+
"eval_steps_per_second": 0.59,
|
| 268 |
+
"step": 424
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.01,
|
| 272 |
+
"learning_rate": 9.202898550724637e-05,
|
| 273 |
+
"loss": 0.1159,
|
| 274 |
+
"step": 430
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 1.04,
|
| 278 |
+
"learning_rate": 9.178743961352657e-05,
|
| 279 |
+
"loss": 0.1101,
|
| 280 |
+
"step": 440
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.06,
|
| 284 |
+
"learning_rate": 9.154589371980676e-05,
|
| 285 |
+
"loss": 0.0938,
|
| 286 |
+
"step": 450
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 1.08,
|
| 290 |
+
"learning_rate": 9.130434782608696e-05,
|
| 291 |
+
"loss": 0.1014,
|
| 292 |
+
"step": 460
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 1.11,
|
| 296 |
+
"learning_rate": 9.106280193236715e-05,
|
| 297 |
+
"loss": 0.1137,
|
| 298 |
+
"step": 470
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.13,
|
| 302 |
+
"learning_rate": 9.082125603864735e-05,
|
| 303 |
+
"loss": 0.1279,
|
| 304 |
+
"step": 480
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 1.16,
|
| 308 |
+
"learning_rate": 9.057971014492754e-05,
|
| 309 |
+
"loss": 0.114,
|
| 310 |
+
"step": 490
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.18,
|
| 314 |
+
"learning_rate": 9.033816425120773e-05,
|
| 315 |
+
"loss": 0.1188,
|
| 316 |
+
"step": 500
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 1.2,
|
| 320 |
+
"learning_rate": 9.009661835748792e-05,
|
| 321 |
+
"loss": 0.1163,
|
| 322 |
+
"step": 510
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 1.23,
|
| 326 |
+
"learning_rate": 8.985507246376813e-05,
|
| 327 |
+
"loss": 0.1006,
|
| 328 |
+
"step": 520
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 1.25,
|
| 332 |
+
"learning_rate": 8.961352657004832e-05,
|
| 333 |
+
"loss": 0.1064,
|
| 334 |
+
"step": 530
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 1.27,
|
| 338 |
+
"learning_rate": 8.937198067632851e-05,
|
| 339 |
+
"loss": 0.1064,
|
| 340 |
+
"step": 540
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.3,
|
| 344 |
+
"learning_rate": 8.91304347826087e-05,
|
| 345 |
+
"loss": 0.11,
|
| 346 |
+
"step": 550
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 1.32,
|
| 350 |
+
"learning_rate": 8.888888888888889e-05,
|
| 351 |
+
"loss": 0.1193,
|
| 352 |
+
"step": 560
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.34,
|
| 356 |
+
"learning_rate": 8.864734299516908e-05,
|
| 357 |
+
"loss": 0.1274,
|
| 358 |
+
"step": 570
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 1.37,
|
| 362 |
+
"learning_rate": 8.840579710144929e-05,
|
| 363 |
+
"loss": 0.1181,
|
| 364 |
+
"step": 580
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 1.39,
|
| 368 |
+
"learning_rate": 8.816425120772948e-05,
|
| 369 |
+
"loss": 0.1038,
|
| 370 |
+
"step": 590
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 1.42,
|
| 374 |
+
"learning_rate": 8.792270531400967e-05,
|
| 375 |
+
"loss": 0.1156,
|
| 376 |
+
"step": 600
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 1.44,
|
| 380 |
+
"learning_rate": 8.768115942028987e-05,
|
| 381 |
+
"loss": 0.1176,
|
| 382 |
+
"step": 610
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 1.46,
|
| 386 |
+
"learning_rate": 8.743961352657006e-05,
|
| 387 |
+
"loss": 0.1154,
|
| 388 |
+
"step": 620
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.49,
|
| 392 |
+
"learning_rate": 8.719806763285025e-05,
|
| 393 |
+
"loss": 0.1328,
|
| 394 |
+
"step": 630
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 1.51,
|
| 398 |
+
"learning_rate": 8.695652173913044e-05,
|
| 399 |
+
"loss": 0.1039,
|
| 400 |
+
"step": 640
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 1.53,
|
| 404 |
+
"learning_rate": 8.671497584541063e-05,
|
| 405 |
+
"loss": 0.1311,
|
| 406 |
+
"step": 650
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 1.56,
|
| 410 |
+
"learning_rate": 8.647342995169082e-05,
|
| 411 |
+
"loss": 0.1172,
|
| 412 |
+
"step": 660
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.58,
|
| 416 |
+
"learning_rate": 8.623188405797103e-05,
|
| 417 |
+
"loss": 0.1266,
|
| 418 |
+
"step": 670
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 1.6,
|
| 422 |
+
"learning_rate": 8.599033816425122e-05,
|
| 423 |
+
"loss": 0.1052,
|
| 424 |
+
"step": 680
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 1.63,
|
| 428 |
+
"learning_rate": 8.574879227053141e-05,
|
| 429 |
+
"loss": 0.1311,
|
| 430 |
+
"step": 690
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.65,
|
| 434 |
+
"learning_rate": 8.55072463768116e-05,
|
| 435 |
+
"loss": 0.1162,
|
| 436 |
+
"step": 700
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 1.67,
|
| 440 |
+
"learning_rate": 8.52657004830918e-05,
|
| 441 |
+
"loss": 0.1131,
|
| 442 |
+
"step": 710
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.7,
|
| 446 |
+
"learning_rate": 8.502415458937199e-05,
|
| 447 |
+
"loss": 0.0934,
|
| 448 |
+
"step": 720
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 1.72,
|
| 452 |
+
"learning_rate": 8.478260869565218e-05,
|
| 453 |
+
"loss": 0.1117,
|
| 454 |
+
"step": 730
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 1.75,
|
| 458 |
+
"learning_rate": 8.454106280193237e-05,
|
| 459 |
+
"loss": 0.116,
|
| 460 |
+
"step": 740
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 1.77,
|
| 464 |
+
"learning_rate": 8.429951690821256e-05,
|
| 465 |
+
"loss": 0.1256,
|
| 466 |
+
"step": 750
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 1.79,
|
| 470 |
+
"learning_rate": 8.405797101449276e-05,
|
| 471 |
+
"loss": 0.0895,
|
| 472 |
+
"step": 760
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.82,
|
| 476 |
+
"learning_rate": 8.381642512077295e-05,
|
| 477 |
+
"loss": 0.1266,
|
| 478 |
+
"step": 770
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.84,
|
| 482 |
+
"learning_rate": 8.357487922705315e-05,
|
| 483 |
+
"loss": 0.1245,
|
| 484 |
+
"step": 780
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 1.86,
|
| 488 |
+
"learning_rate": 8.333333333333334e-05,
|
| 489 |
+
"loss": 0.11,
|
| 490 |
+
"step": 790
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 1.89,
|
| 494 |
+
"learning_rate": 8.309178743961354e-05,
|
| 495 |
+
"loss": 0.1081,
|
| 496 |
+
"step": 800
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.91,
|
| 500 |
+
"learning_rate": 8.285024154589373e-05,
|
| 501 |
+
"loss": 0.127,
|
| 502 |
+
"step": 810
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 1.93,
|
| 506 |
+
"learning_rate": 8.260869565217392e-05,
|
| 507 |
+
"loss": 0.115,
|
| 508 |
+
"step": 820
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 1.96,
|
| 512 |
+
"learning_rate": 8.236714975845411e-05,
|
| 513 |
+
"loss": 0.1197,
|
| 514 |
+
"step": 830
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.98,
|
| 518 |
+
"learning_rate": 8.21256038647343e-05,
|
| 519 |
+
"loss": 0.1224,
|
| 520 |
+
"step": 840
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 2.0,
|
| 524 |
+
"eval_loss": 1.8221625089645386,
|
| 525 |
+
"eval_runtime": 73.5866,
|
| 526 |
+
"eval_samples_per_second": 10.233,
|
| 527 |
+
"eval_steps_per_second": 0.652,
|
| 528 |
+
"step": 848
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 2.0,
|
| 532 |
+
"learning_rate": 8.18840579710145e-05,
|
| 533 |
+
"loss": 0.1095,
|
| 534 |
+
"step": 850
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 2.03,
|
| 538 |
+
"learning_rate": 8.164251207729469e-05,
|
| 539 |
+
"loss": 0.056,
|
| 540 |
+
"step": 860
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 2.05,
|
| 544 |
+
"learning_rate": 8.140096618357488e-05,
|
| 545 |
+
"loss": 0.0497,
|
| 546 |
+
"step": 870
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 2.08,
|
| 550 |
+
"learning_rate": 8.115942028985508e-05,
|
| 551 |
+
"loss": 0.0566,
|
| 552 |
+
"step": 880
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 2.1,
|
| 556 |
+
"learning_rate": 8.091787439613528e-05,
|
| 557 |
+
"loss": 0.0595,
|
| 558 |
+
"step": 890
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 2.12,
|
| 562 |
+
"learning_rate": 8.067632850241547e-05,
|
| 563 |
+
"loss": 0.0681,
|
| 564 |
+
"step": 900
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 2.15,
|
| 568 |
+
"learning_rate": 8.043478260869566e-05,
|
| 569 |
+
"loss": 0.0881,
|
| 570 |
+
"step": 910
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 2.17,
|
| 574 |
+
"learning_rate": 8.019323671497585e-05,
|
| 575 |
+
"loss": 0.0604,
|
| 576 |
+
"step": 920
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 2.19,
|
| 580 |
+
"learning_rate": 7.995169082125604e-05,
|
| 581 |
+
"loss": 0.0632,
|
| 582 |
+
"step": 930
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 2.22,
|
| 586 |
+
"learning_rate": 7.971014492753623e-05,
|
| 587 |
+
"loss": 0.0599,
|
| 588 |
+
"step": 940
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 2.24,
|
| 592 |
+
"learning_rate": 7.946859903381643e-05,
|
| 593 |
+
"loss": 0.0684,
|
| 594 |
+
"step": 950
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 2.26,
|
| 598 |
+
"learning_rate": 7.922705314009662e-05,
|
| 599 |
+
"loss": 0.0713,
|
| 600 |
+
"step": 960
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 2.29,
|
| 604 |
+
"learning_rate": 7.898550724637681e-05,
|
| 605 |
+
"loss": 0.0595,
|
| 606 |
+
"step": 970
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 2.31,
|
| 610 |
+
"learning_rate": 7.874396135265701e-05,
|
| 611 |
+
"loss": 0.0687,
|
| 612 |
+
"step": 980
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 2.33,
|
| 616 |
+
"learning_rate": 7.85024154589372e-05,
|
| 617 |
+
"loss": 0.0732,
|
| 618 |
+
"step": 990
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 2.36,
|
| 622 |
+
"learning_rate": 7.82608695652174e-05,
|
| 623 |
+
"loss": 0.0615,
|
| 624 |
+
"step": 1000
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 2.38,
|
| 628 |
+
"learning_rate": 7.801932367149759e-05,
|
| 629 |
+
"loss": 0.0636,
|
| 630 |
+
"step": 1010
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 2.41,
|
| 634 |
+
"learning_rate": 7.777777777777778e-05,
|
| 635 |
+
"loss": 0.0617,
|
| 636 |
+
"step": 1020
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.43,
|
| 640 |
+
"learning_rate": 7.753623188405797e-05,
|
| 641 |
+
"loss": 0.0776,
|
| 642 |
+
"step": 1030
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 2.45,
|
| 646 |
+
"learning_rate": 7.729468599033817e-05,
|
| 647 |
+
"loss": 0.0671,
|
| 648 |
+
"step": 1040
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 2.48,
|
| 652 |
+
"learning_rate": 7.705314009661836e-05,
|
| 653 |
+
"loss": 0.073,
|
| 654 |
+
"step": 1050
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 2.5,
|
| 658 |
+
"learning_rate": 7.681159420289855e-05,
|
| 659 |
+
"loss": 0.0724,
|
| 660 |
+
"step": 1060
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 2.52,
|
| 664 |
+
"learning_rate": 7.657004830917874e-05,
|
| 665 |
+
"loss": 0.0607,
|
| 666 |
+
"step": 1070
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 2.55,
|
| 670 |
+
"learning_rate": 7.632850241545893e-05,
|
| 671 |
+
"loss": 0.0645,
|
| 672 |
+
"step": 1080
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 2.57,
|
| 676 |
+
"learning_rate": 7.608695652173914e-05,
|
| 677 |
+
"loss": 0.0749,
|
| 678 |
+
"step": 1090
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 2.59,
|
| 682 |
+
"learning_rate": 7.584541062801933e-05,
|
| 683 |
+
"loss": 0.0594,
|
| 684 |
+
"step": 1100
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 2.62,
|
| 688 |
+
"learning_rate": 7.560386473429952e-05,
|
| 689 |
+
"loss": 0.0699,
|
| 690 |
+
"step": 1110
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 2.64,
|
| 694 |
+
"learning_rate": 7.536231884057971e-05,
|
| 695 |
+
"loss": 0.0827,
|
| 696 |
+
"step": 1120
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 2.67,
|
| 700 |
+
"learning_rate": 7.51207729468599e-05,
|
| 701 |
+
"loss": 0.0596,
|
| 702 |
+
"step": 1130
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 2.69,
|
| 706 |
+
"learning_rate": 7.48792270531401e-05,
|
| 707 |
+
"loss": 0.0754,
|
| 708 |
+
"step": 1140
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 2.71,
|
| 712 |
+
"learning_rate": 7.463768115942029e-05,
|
| 713 |
+
"loss": 0.0706,
|
| 714 |
+
"step": 1150
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 2.74,
|
| 718 |
+
"learning_rate": 7.439613526570048e-05,
|
| 719 |
+
"loss": 0.063,
|
| 720 |
+
"step": 1160
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 2.76,
|
| 724 |
+
"learning_rate": 7.415458937198067e-05,
|
| 725 |
+
"loss": 0.0665,
|
| 726 |
+
"step": 1170
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 2.78,
|
| 730 |
+
"learning_rate": 7.391304347826086e-05,
|
| 731 |
+
"loss": 0.0748,
|
| 732 |
+
"step": 1180
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 2.81,
|
| 736 |
+
"learning_rate": 7.367149758454107e-05,
|
| 737 |
+
"loss": 0.0594,
|
| 738 |
+
"step": 1190
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 2.83,
|
| 742 |
+
"learning_rate": 7.342995169082126e-05,
|
| 743 |
+
"loss": 0.0839,
|
| 744 |
+
"step": 1200
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.85,
|
| 748 |
+
"learning_rate": 7.318840579710145e-05,
|
| 749 |
+
"loss": 0.0693,
|
| 750 |
+
"step": 1210
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 2.88,
|
| 754 |
+
"learning_rate": 7.294685990338164e-05,
|
| 755 |
+
"loss": 0.0663,
|
| 756 |
+
"step": 1220
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 2.9,
|
| 760 |
+
"learning_rate": 7.270531400966184e-05,
|
| 761 |
+
"loss": 0.0752,
|
| 762 |
+
"step": 1230
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 2.92,
|
| 766 |
+
"learning_rate": 7.246376811594203e-05,
|
| 767 |
+
"loss": 0.092,
|
| 768 |
+
"step": 1240
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 2.95,
|
| 772 |
+
"learning_rate": 7.222222222222222e-05,
|
| 773 |
+
"loss": 0.0739,
|
| 774 |
+
"step": 1250
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.97,
|
| 778 |
+
"learning_rate": 7.198067632850241e-05,
|
| 779 |
+
"loss": 0.067,
|
| 780 |
+
"step": 1260
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 3.0,
|
| 784 |
+
"learning_rate": 7.17391304347826e-05,
|
| 785 |
+
"loss": 0.0768,
|
| 786 |
+
"step": 1270
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 3.0,
|
| 790 |
+
"eval_loss": 1.9414550065994263,
|
| 791 |
+
"eval_runtime": 73.8083,
|
| 792 |
+
"eval_samples_per_second": 10.202,
|
| 793 |
+
"eval_steps_per_second": 0.65,
|
| 794 |
+
"step": 1272
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 3.02,
|
| 798 |
+
"learning_rate": 7.14975845410628e-05,
|
| 799 |
+
"loss": 0.0424,
|
| 800 |
+
"step": 1280
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 3.04,
|
| 804 |
+
"learning_rate": 7.1256038647343e-05,
|
| 805 |
+
"loss": 0.0284,
|
| 806 |
+
"step": 1290
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.07,
|
| 810 |
+
"learning_rate": 7.101449275362319e-05,
|
| 811 |
+
"loss": 0.0286,
|
| 812 |
+
"step": 1300
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 3.09,
|
| 816 |
+
"learning_rate": 7.077294685990338e-05,
|
| 817 |
+
"loss": 0.0368,
|
| 818 |
+
"step": 1310
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 3.11,
|
| 822 |
+
"learning_rate": 7.053140096618357e-05,
|
| 823 |
+
"loss": 0.0286,
|
| 824 |
+
"step": 1320
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 3.14,
|
| 828 |
+
"learning_rate": 7.028985507246377e-05,
|
| 829 |
+
"loss": 0.0363,
|
| 830 |
+
"step": 1330
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.16,
|
| 834 |
+
"learning_rate": 7.004830917874396e-05,
|
| 835 |
+
"loss": 0.0414,
|
| 836 |
+
"step": 1340
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 3.18,
|
| 840 |
+
"learning_rate": 6.980676328502416e-05,
|
| 841 |
+
"loss": 0.0498,
|
| 842 |
+
"step": 1350
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 3.21,
|
| 846 |
+
"learning_rate": 6.956521739130436e-05,
|
| 847 |
+
"loss": 0.0288,
|
| 848 |
+
"step": 1360
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 3.23,
|
| 852 |
+
"learning_rate": 6.932367149758455e-05,
|
| 853 |
+
"loss": 0.0423,
|
| 854 |
+
"step": 1370
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 3.25,
|
| 858 |
+
"learning_rate": 6.908212560386474e-05,
|
| 859 |
+
"loss": 0.0372,
|
| 860 |
+
"step": 1380
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 3.28,
|
| 864 |
+
"learning_rate": 6.884057971014493e-05,
|
| 865 |
+
"loss": 0.0395,
|
| 866 |
+
"step": 1390
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 3.3,
|
| 870 |
+
"learning_rate": 6.859903381642512e-05,
|
| 871 |
+
"loss": 0.038,
|
| 872 |
+
"step": 1400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 3.33,
|
| 876 |
+
"learning_rate": 6.835748792270533e-05,
|
| 877 |
+
"loss": 0.0381,
|
| 878 |
+
"step": 1410
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 3.35,
|
| 882 |
+
"learning_rate": 6.811594202898552e-05,
|
| 883 |
+
"loss": 0.0379,
|
| 884 |
+
"step": 1420
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 3.37,
|
| 888 |
+
"learning_rate": 6.787439613526571e-05,
|
| 889 |
+
"loss": 0.0484,
|
| 890 |
+
"step": 1430
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.4,
|
| 894 |
+
"learning_rate": 6.76328502415459e-05,
|
| 895 |
+
"loss": 0.0459,
|
| 896 |
+
"step": 1440
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 3.42,
|
| 900 |
+
"learning_rate": 6.73913043478261e-05,
|
| 901 |
+
"loss": 0.0434,
|
| 902 |
+
"step": 1450
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 3.44,
|
| 906 |
+
"learning_rate": 6.714975845410629e-05,
|
| 907 |
+
"loss": 0.0448,
|
| 908 |
+
"step": 1460
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 3.47,
|
| 912 |
+
"learning_rate": 6.690821256038648e-05,
|
| 913 |
+
"loss": 0.0502,
|
| 914 |
+
"step": 1470
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 3.49,
|
| 918 |
+
"learning_rate": 6.666666666666667e-05,
|
| 919 |
+
"loss": 0.04,
|
| 920 |
+
"step": 1480
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 3.51,
|
| 924 |
+
"learning_rate": 6.642512077294686e-05,
|
| 925 |
+
"loss": 0.0383,
|
| 926 |
+
"step": 1490
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 3.54,
|
| 930 |
+
"learning_rate": 6.618357487922707e-05,
|
| 931 |
+
"loss": 0.0448,
|
| 932 |
+
"step": 1500
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 3.56,
|
| 936 |
+
"learning_rate": 6.594202898550726e-05,
|
| 937 |
+
"loss": 0.0378,
|
| 938 |
+
"step": 1510
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 3.58,
|
| 942 |
+
"learning_rate": 6.570048309178745e-05,
|
| 943 |
+
"loss": 0.0316,
|
| 944 |
+
"step": 1520
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 3.61,
|
| 948 |
+
"learning_rate": 6.545893719806764e-05,
|
| 949 |
+
"loss": 0.0457,
|
| 950 |
+
"step": 1530
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 3.63,
|
| 954 |
+
"learning_rate": 6.521739130434783e-05,
|
| 955 |
+
"loss": 0.0395,
|
| 956 |
+
"step": 1540
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 3.66,
|
| 960 |
+
"learning_rate": 6.497584541062803e-05,
|
| 961 |
+
"loss": 0.0409,
|
| 962 |
+
"step": 1550
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 3.68,
|
| 966 |
+
"learning_rate": 6.473429951690822e-05,
|
| 967 |
+
"loss": 0.045,
|
| 968 |
+
"step": 1560
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 3.7,
|
| 972 |
+
"learning_rate": 6.449275362318841e-05,
|
| 973 |
+
"loss": 0.0389,
|
| 974 |
+
"step": 1570
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 3.73,
|
| 978 |
+
"learning_rate": 6.42512077294686e-05,
|
| 979 |
+
"loss": 0.0396,
|
| 980 |
+
"step": 1580
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 3.75,
|
| 984 |
+
"learning_rate": 6.400966183574879e-05,
|
| 985 |
+
"loss": 0.0396,
|
| 986 |
+
"step": 1590
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 3.77,
|
| 990 |
+
"learning_rate": 6.376811594202898e-05,
|
| 991 |
+
"loss": 0.0388,
|
| 992 |
+
"step": 1600
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 3.8,
|
| 996 |
+
"learning_rate": 6.352657004830919e-05,
|
| 997 |
+
"loss": 0.0401,
|
| 998 |
+
"step": 1610
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 3.82,
|
| 1002 |
+
"learning_rate": 6.328502415458938e-05,
|
| 1003 |
+
"loss": 0.0441,
|
| 1004 |
+
"step": 1620
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 3.84,
|
| 1008 |
+
"learning_rate": 6.304347826086957e-05,
|
| 1009 |
+
"loss": 0.037,
|
| 1010 |
+
"step": 1630
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 3.87,
|
| 1014 |
+
"learning_rate": 6.280193236714976e-05,
|
| 1015 |
+
"loss": 0.0395,
|
| 1016 |
+
"step": 1640
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 3.89,
|
| 1020 |
+
"learning_rate": 6.256038647342996e-05,
|
| 1021 |
+
"loss": 0.0463,
|
| 1022 |
+
"step": 1650
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 3.92,
|
| 1026 |
+
"learning_rate": 6.231884057971015e-05,
|
| 1027 |
+
"loss": 0.0571,
|
| 1028 |
+
"step": 1660
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 3.94,
|
| 1032 |
+
"learning_rate": 6.207729468599034e-05,
|
| 1033 |
+
"loss": 0.0291,
|
| 1034 |
+
"step": 1670
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 3.96,
|
| 1038 |
+
"learning_rate": 6.183574879227053e-05,
|
| 1039 |
+
"loss": 0.0396,
|
| 1040 |
+
"step": 1680
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 3.99,
|
| 1044 |
+
"learning_rate": 6.159420289855072e-05,
|
| 1045 |
+
"loss": 0.0406,
|
| 1046 |
+
"step": 1690
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 4.0,
|
| 1050 |
+
"eval_loss": 1.9886054992675781,
|
| 1051 |
+
"eval_runtime": 74.4486,
|
| 1052 |
+
"eval_samples_per_second": 10.114,
|
| 1053 |
+
"eval_steps_per_second": 0.645,
|
| 1054 |
+
"step": 1696
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 4.01,
|
| 1058 |
+
"learning_rate": 6.135265700483092e-05,
|
| 1059 |
+
"loss": 0.0336,
|
| 1060 |
+
"step": 1700
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 4.03,
|
| 1064 |
+
"learning_rate": 6.111111111111112e-05,
|
| 1065 |
+
"loss": 0.0238,
|
| 1066 |
+
"step": 1710
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 4.06,
|
| 1070 |
+
"learning_rate": 6.086956521739131e-05,
|
| 1071 |
+
"loss": 0.0255,
|
| 1072 |
+
"step": 1720
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 4.08,
|
| 1076 |
+
"learning_rate": 6.0628019323671504e-05,
|
| 1077 |
+
"loss": 0.0217,
|
| 1078 |
+
"step": 1730
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 4.1,
|
| 1082 |
+
"learning_rate": 6.0386473429951696e-05,
|
| 1083 |
+
"loss": 0.024,
|
| 1084 |
+
"step": 1740
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 4.13,
|
| 1088 |
+
"learning_rate": 6.014492753623189e-05,
|
| 1089 |
+
"loss": 0.0245,
|
| 1090 |
+
"step": 1750
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 4.15,
|
| 1094 |
+
"learning_rate": 5.990338164251208e-05,
|
| 1095 |
+
"loss": 0.0203,
|
| 1096 |
+
"step": 1760
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 4.17,
|
| 1100 |
+
"learning_rate": 5.966183574879227e-05,
|
| 1101 |
+
"loss": 0.0192,
|
| 1102 |
+
"step": 1770
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 4.2,
|
| 1106 |
+
"learning_rate": 5.942028985507246e-05,
|
| 1107 |
+
"loss": 0.0238,
|
| 1108 |
+
"step": 1780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 4.22,
|
| 1112 |
+
"learning_rate": 5.9178743961352655e-05,
|
| 1113 |
+
"loss": 0.0213,
|
| 1114 |
+
"step": 1790
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 4.25,
|
| 1118 |
+
"learning_rate": 5.8937198067632847e-05,
|
| 1119 |
+
"loss": 0.024,
|
| 1120 |
+
"step": 1800
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 4.27,
|
| 1124 |
+
"learning_rate": 5.869565217391305e-05,
|
| 1125 |
+
"loss": 0.0137,
|
| 1126 |
+
"step": 1810
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 4.29,
|
| 1130 |
+
"learning_rate": 5.8454106280193244e-05,
|
| 1131 |
+
"loss": 0.0179,
|
| 1132 |
+
"step": 1820
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 4.32,
|
| 1136 |
+
"learning_rate": 5.8212560386473435e-05,
|
| 1137 |
+
"loss": 0.018,
|
| 1138 |
+
"step": 1830
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 4.34,
|
| 1142 |
+
"learning_rate": 5.797101449275363e-05,
|
| 1143 |
+
"loss": 0.0136,
|
| 1144 |
+
"step": 1840
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 4.36,
|
| 1148 |
+
"learning_rate": 5.772946859903382e-05,
|
| 1149 |
+
"loss": 0.0156,
|
| 1150 |
+
"step": 1850
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 4.39,
|
| 1154 |
+
"learning_rate": 5.748792270531401e-05,
|
| 1155 |
+
"loss": 0.0192,
|
| 1156 |
+
"step": 1860
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 4.41,
|
| 1160 |
+
"learning_rate": 5.72463768115942e-05,
|
| 1161 |
+
"loss": 0.0154,
|
| 1162 |
+
"step": 1870
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 4.43,
|
| 1166 |
+
"learning_rate": 5.7004830917874394e-05,
|
| 1167 |
+
"loss": 0.0197,
|
| 1168 |
+
"step": 1880
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 4.46,
|
| 1172 |
+
"learning_rate": 5.6763285024154586e-05,
|
| 1173 |
+
"loss": 0.0229,
|
| 1174 |
+
"step": 1890
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 4.48,
|
| 1178 |
+
"learning_rate": 5.652173913043478e-05,
|
| 1179 |
+
"loss": 0.0199,
|
| 1180 |
+
"step": 1900
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 4.5,
|
| 1184 |
+
"learning_rate": 5.6280193236714976e-05,
|
| 1185 |
+
"loss": 0.0222,
|
| 1186 |
+
"step": 1910
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 4.53,
|
| 1190 |
+
"learning_rate": 5.6038647342995175e-05,
|
| 1191 |
+
"loss": 0.0277,
|
| 1192 |
+
"step": 1920
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 4.55,
|
| 1196 |
+
"learning_rate": 5.5797101449275366e-05,
|
| 1197 |
+
"loss": 0.0246,
|
| 1198 |
+
"step": 1930
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 4.58,
|
| 1202 |
+
"learning_rate": 5.555555555555556e-05,
|
| 1203 |
+
"loss": 0.0252,
|
| 1204 |
+
"step": 1940
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 4.6,
|
| 1208 |
+
"learning_rate": 5.531400966183575e-05,
|
| 1209 |
+
"loss": 0.0269,
|
| 1210 |
+
"step": 1950
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 4.62,
|
| 1214 |
+
"learning_rate": 5.507246376811594e-05,
|
| 1215 |
+
"loss": 0.0183,
|
| 1216 |
+
"step": 1960
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 4.65,
|
| 1220 |
+
"learning_rate": 5.483091787439614e-05,
|
| 1221 |
+
"loss": 0.025,
|
| 1222 |
+
"step": 1970
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 4.67,
|
| 1226 |
+
"learning_rate": 5.458937198067633e-05,
|
| 1227 |
+
"loss": 0.0208,
|
| 1228 |
+
"step": 1980
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 4.69,
|
| 1232 |
+
"learning_rate": 5.4347826086956524e-05,
|
| 1233 |
+
"loss": 0.0198,
|
| 1234 |
+
"step": 1990
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 4.72,
|
| 1238 |
+
"learning_rate": 5.4106280193236716e-05,
|
| 1239 |
+
"loss": 0.0135,
|
| 1240 |
+
"step": 2000
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 4.74,
|
| 1244 |
+
"learning_rate": 5.386473429951691e-05,
|
| 1245 |
+
"loss": 0.0204,
|
| 1246 |
+
"step": 2010
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 4.76,
|
| 1250 |
+
"learning_rate": 5.3623188405797106e-05,
|
| 1251 |
+
"loss": 0.0226,
|
| 1252 |
+
"step": 2020
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 4.79,
|
| 1256 |
+
"learning_rate": 5.3381642512077304e-05,
|
| 1257 |
+
"loss": 0.019,
|
| 1258 |
+
"step": 2030
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 4.81,
|
| 1262 |
+
"learning_rate": 5.3140096618357496e-05,
|
| 1263 |
+
"loss": 0.0162,
|
| 1264 |
+
"step": 2040
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 4.83,
|
| 1268 |
+
"learning_rate": 5.289855072463769e-05,
|
| 1269 |
+
"loss": 0.0181,
|
| 1270 |
+
"step": 2050
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 4.86,
|
| 1274 |
+
"learning_rate": 5.265700483091788e-05,
|
| 1275 |
+
"loss": 0.0187,
|
| 1276 |
+
"step": 2060
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 4.88,
|
| 1280 |
+
"learning_rate": 5.241545893719807e-05,
|
| 1281 |
+
"loss": 0.0192,
|
| 1282 |
+
"step": 2070
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 4.91,
|
| 1286 |
+
"learning_rate": 5.217391304347826e-05,
|
| 1287 |
+
"loss": 0.0255,
|
| 1288 |
+
"step": 2080
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 4.93,
|
| 1292 |
+
"learning_rate": 5.1932367149758455e-05,
|
| 1293 |
+
"loss": 0.0258,
|
| 1294 |
+
"step": 2090
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 4.95,
|
| 1298 |
+
"learning_rate": 5.1690821256038647e-05,
|
| 1299 |
+
"loss": 0.0274,
|
| 1300 |
+
"step": 2100
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 4.98,
|
| 1304 |
+
"learning_rate": 5.144927536231884e-05,
|
| 1305 |
+
"loss": 0.0291,
|
| 1306 |
+
"step": 2110
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 5.0,
|
| 1310 |
+
"learning_rate": 5.1207729468599044e-05,
|
| 1311 |
+
"loss": 0.0422,
|
| 1312 |
+
"step": 2120
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 5.0,
|
| 1316 |
+
"eval_loss": 1.9903086423873901,
|
| 1317 |
+
"eval_runtime": 72.5407,
|
| 1318 |
+
"eval_samples_per_second": 10.38,
|
| 1319 |
+
"eval_steps_per_second": 0.662,
|
| 1320 |
+
"step": 2120
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 5.02,
|
| 1324 |
+
"learning_rate": 5.0966183574879235e-05,
|
| 1325 |
+
"loss": 0.0099,
|
| 1326 |
+
"step": 2130
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 5.05,
|
| 1330 |
+
"learning_rate": 5.072463768115943e-05,
|
| 1331 |
+
"loss": 0.0121,
|
| 1332 |
+
"step": 2140
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 5.07,
|
| 1336 |
+
"learning_rate": 5.048309178743962e-05,
|
| 1337 |
+
"loss": 0.008,
|
| 1338 |
+
"step": 2150
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 5.09,
|
| 1342 |
+
"learning_rate": 5.024154589371981e-05,
|
| 1343 |
+
"loss": 0.01,
|
| 1344 |
+
"step": 2160
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 5.12,
|
| 1348 |
+
"learning_rate": 5e-05,
|
| 1349 |
+
"loss": 0.0065,
|
| 1350 |
+
"step": 2170
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 5.14,
|
| 1354 |
+
"learning_rate": 4.9758454106280194e-05,
|
| 1355 |
+
"loss": 0.0096,
|
| 1356 |
+
"step": 2180
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 5.17,
|
| 1360 |
+
"learning_rate": 4.9516908212560386e-05,
|
| 1361 |
+
"loss": 0.0147,
|
| 1362 |
+
"step": 2190
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 5.19,
|
| 1366 |
+
"learning_rate": 4.9275362318840584e-05,
|
| 1367 |
+
"loss": 0.0094,
|
| 1368 |
+
"step": 2200
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 5.21,
|
| 1372 |
+
"learning_rate": 4.9033816425120776e-05,
|
| 1373 |
+
"loss": 0.0123,
|
| 1374 |
+
"step": 2210
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 5.24,
|
| 1378 |
+
"learning_rate": 4.879227053140097e-05,
|
| 1379 |
+
"loss": 0.007,
|
| 1380 |
+
"step": 2220
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 5.26,
|
| 1384 |
+
"learning_rate": 4.855072463768116e-05,
|
| 1385 |
+
"loss": 0.0136,
|
| 1386 |
+
"step": 2230
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 5.28,
|
| 1390 |
+
"learning_rate": 4.830917874396135e-05,
|
| 1391 |
+
"loss": 0.017,
|
| 1392 |
+
"step": 2240
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 5.31,
|
| 1396 |
+
"learning_rate": 4.806763285024155e-05,
|
| 1397 |
+
"loss": 0.0078,
|
| 1398 |
+
"step": 2250
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 5.33,
|
| 1402 |
+
"learning_rate": 4.782608695652174e-05,
|
| 1403 |
+
"loss": 0.011,
|
| 1404 |
+
"step": 2260
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 5.35,
|
| 1408 |
+
"learning_rate": 4.7584541062801933e-05,
|
| 1409 |
+
"loss": 0.0078,
|
| 1410 |
+
"step": 2270
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 5.38,
|
| 1414 |
+
"learning_rate": 4.7342995169082125e-05,
|
| 1415 |
+
"loss": 0.0118,
|
| 1416 |
+
"step": 2280
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 5.4,
|
| 1420 |
+
"learning_rate": 4.710144927536232e-05,
|
| 1421 |
+
"loss": 0.0072,
|
| 1422 |
+
"step": 2290
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 5.42,
|
| 1426 |
+
"learning_rate": 4.6859903381642516e-05,
|
| 1427 |
+
"loss": 0.0122,
|
| 1428 |
+
"step": 2300
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 5.45,
|
| 1432 |
+
"learning_rate": 4.661835748792271e-05,
|
| 1433 |
+
"loss": 0.0145,
|
| 1434 |
+
"step": 2310
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 5.47,
|
| 1438 |
+
"learning_rate": 4.63768115942029e-05,
|
| 1439 |
+
"loss": 0.0094,
|
| 1440 |
+
"step": 2320
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 5.5,
|
| 1444 |
+
"learning_rate": 4.613526570048309e-05,
|
| 1445 |
+
"loss": 0.0087,
|
| 1446 |
+
"step": 2330
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 5.52,
|
| 1450 |
+
"learning_rate": 4.589371980676328e-05,
|
| 1451 |
+
"loss": 0.0127,
|
| 1452 |
+
"step": 2340
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 5.54,
|
| 1456 |
+
"learning_rate": 4.565217391304348e-05,
|
| 1457 |
+
"loss": 0.0076,
|
| 1458 |
+
"step": 2350
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 5.57,
|
| 1462 |
+
"learning_rate": 4.541062801932367e-05,
|
| 1463 |
+
"loss": 0.0076,
|
| 1464 |
+
"step": 2360
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 5.59,
|
| 1468 |
+
"learning_rate": 4.5169082125603865e-05,
|
| 1469 |
+
"loss": 0.0074,
|
| 1470 |
+
"step": 2370
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 5.61,
|
| 1474 |
+
"learning_rate": 4.492753623188406e-05,
|
| 1475 |
+
"loss": 0.0093,
|
| 1476 |
+
"step": 2380
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 5.64,
|
| 1480 |
+
"learning_rate": 4.4685990338164255e-05,
|
| 1481 |
+
"loss": 0.0074,
|
| 1482 |
+
"step": 2390
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 5.66,
|
| 1486 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 1487 |
+
"loss": 0.0162,
|
| 1488 |
+
"step": 2400
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 5.68,
|
| 1492 |
+
"learning_rate": 4.4202898550724645e-05,
|
| 1493 |
+
"loss": 0.0146,
|
| 1494 |
+
"step": 2410
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 5.71,
|
| 1498 |
+
"learning_rate": 4.396135265700484e-05,
|
| 1499 |
+
"loss": 0.007,
|
| 1500 |
+
"step": 2420
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 5.73,
|
| 1504 |
+
"learning_rate": 4.371980676328503e-05,
|
| 1505 |
+
"loss": 0.01,
|
| 1506 |
+
"step": 2430
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 5.75,
|
| 1510 |
+
"learning_rate": 4.347826086956522e-05,
|
| 1511 |
+
"loss": 0.0098,
|
| 1512 |
+
"step": 2440
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 5.78,
|
| 1516 |
+
"learning_rate": 4.323671497584541e-05,
|
| 1517 |
+
"loss": 0.009,
|
| 1518 |
+
"step": 2450
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 5.8,
|
| 1522 |
+
"learning_rate": 4.299516908212561e-05,
|
| 1523 |
+
"loss": 0.0071,
|
| 1524 |
+
"step": 2460
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 5.83,
|
| 1528 |
+
"learning_rate": 4.27536231884058e-05,
|
| 1529 |
+
"loss": 0.0112,
|
| 1530 |
+
"step": 2470
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 5.85,
|
| 1534 |
+
"learning_rate": 4.2512077294685994e-05,
|
| 1535 |
+
"loss": 0.0135,
|
| 1536 |
+
"step": 2480
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 5.87,
|
| 1540 |
+
"learning_rate": 4.2270531400966186e-05,
|
| 1541 |
+
"loss": 0.0104,
|
| 1542 |
+
"step": 2490
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 5.9,
|
| 1546 |
+
"learning_rate": 4.202898550724638e-05,
|
| 1547 |
+
"loss": 0.0084,
|
| 1548 |
+
"step": 2500
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 5.92,
|
| 1552 |
+
"learning_rate": 4.1787439613526576e-05,
|
| 1553 |
+
"loss": 0.0098,
|
| 1554 |
+
"step": 2510
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 5.94,
|
| 1558 |
+
"learning_rate": 4.154589371980677e-05,
|
| 1559 |
+
"loss": 0.0083,
|
| 1560 |
+
"step": 2520
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 5.97,
|
| 1564 |
+
"learning_rate": 4.130434782608696e-05,
|
| 1565 |
+
"loss": 0.0088,
|
| 1566 |
+
"step": 2530
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 5.99,
|
| 1570 |
+
"learning_rate": 4.106280193236715e-05,
|
| 1571 |
+
"loss": 0.0134,
|
| 1572 |
+
"step": 2540
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 6.0,
|
| 1576 |
+
"eval_loss": 2.0551583766937256,
|
| 1577 |
+
"eval_runtime": 73.0209,
|
| 1578 |
+
"eval_samples_per_second": 10.312,
|
| 1579 |
+
"eval_steps_per_second": 0.657,
|
| 1580 |
+
"step": 2544
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 6.01,
|
| 1584 |
+
"learning_rate": 4.082125603864734e-05,
|
| 1585 |
+
"loss": 0.005,
|
| 1586 |
+
"step": 2550
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 6.04,
|
| 1590 |
+
"learning_rate": 4.057971014492754e-05,
|
| 1591 |
+
"loss": 0.0043,
|
| 1592 |
+
"step": 2560
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 6.06,
|
| 1596 |
+
"learning_rate": 4.0338164251207733e-05,
|
| 1597 |
+
"loss": 0.005,
|
| 1598 |
+
"step": 2570
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 6.08,
|
| 1602 |
+
"learning_rate": 4.0096618357487925e-05,
|
| 1603 |
+
"loss": 0.0021,
|
| 1604 |
+
"step": 2580
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 6.11,
|
| 1608 |
+
"learning_rate": 3.985507246376812e-05,
|
| 1609 |
+
"loss": 0.0028,
|
| 1610 |
+
"step": 2590
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 6.13,
|
| 1614 |
+
"learning_rate": 3.961352657004831e-05,
|
| 1615 |
+
"loss": 0.0027,
|
| 1616 |
+
"step": 2600
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 6.16,
|
| 1620 |
+
"learning_rate": 3.937198067632851e-05,
|
| 1621 |
+
"loss": 0.004,
|
| 1622 |
+
"step": 2610
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 6.18,
|
| 1626 |
+
"learning_rate": 3.91304347826087e-05,
|
| 1627 |
+
"loss": 0.0047,
|
| 1628 |
+
"step": 2620
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 6.2,
|
| 1632 |
+
"learning_rate": 3.888888888888889e-05,
|
| 1633 |
+
"loss": 0.0031,
|
| 1634 |
+
"step": 2630
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 6.23,
|
| 1638 |
+
"learning_rate": 3.864734299516908e-05,
|
| 1639 |
+
"loss": 0.0042,
|
| 1640 |
+
"step": 2640
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 6.25,
|
| 1644 |
+
"learning_rate": 3.8405797101449274e-05,
|
| 1645 |
+
"loss": 0.0037,
|
| 1646 |
+
"step": 2650
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 6.27,
|
| 1650 |
+
"learning_rate": 3.8164251207729466e-05,
|
| 1651 |
+
"loss": 0.0029,
|
| 1652 |
+
"step": 2660
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 6.3,
|
| 1656 |
+
"learning_rate": 3.7922705314009665e-05,
|
| 1657 |
+
"loss": 0.0018,
|
| 1658 |
+
"step": 2670
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 6.32,
|
| 1662 |
+
"learning_rate": 3.7681159420289856e-05,
|
| 1663 |
+
"loss": 0.0042,
|
| 1664 |
+
"step": 2680
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 6.34,
|
| 1668 |
+
"learning_rate": 3.743961352657005e-05,
|
| 1669 |
+
"loss": 0.0054,
|
| 1670 |
+
"step": 2690
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 6.37,
|
| 1674 |
+
"learning_rate": 3.719806763285024e-05,
|
| 1675 |
+
"loss": 0.0051,
|
| 1676 |
+
"step": 2700
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 6.39,
|
| 1680 |
+
"learning_rate": 3.695652173913043e-05,
|
| 1681 |
+
"loss": 0.0054,
|
| 1682 |
+
"step": 2710
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 6.42,
|
| 1686 |
+
"learning_rate": 3.671497584541063e-05,
|
| 1687 |
+
"loss": 0.0026,
|
| 1688 |
+
"step": 2720
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 6.44,
|
| 1692 |
+
"learning_rate": 3.647342995169082e-05,
|
| 1693 |
+
"loss": 0.0038,
|
| 1694 |
+
"step": 2730
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 6.46,
|
| 1698 |
+
"learning_rate": 3.6231884057971014e-05,
|
| 1699 |
+
"loss": 0.0051,
|
| 1700 |
+
"step": 2740
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 6.49,
|
| 1704 |
+
"learning_rate": 3.5990338164251205e-05,
|
| 1705 |
+
"loss": 0.003,
|
| 1706 |
+
"step": 2750
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 6.51,
|
| 1710 |
+
"learning_rate": 3.57487922705314e-05,
|
| 1711 |
+
"loss": 0.0028,
|
| 1712 |
+
"step": 2760
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 6.53,
|
| 1716 |
+
"learning_rate": 3.5507246376811596e-05,
|
| 1717 |
+
"loss": 0.0036,
|
| 1718 |
+
"step": 2770
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 6.56,
|
| 1722 |
+
"learning_rate": 3.526570048309179e-05,
|
| 1723 |
+
"loss": 0.0033,
|
| 1724 |
+
"step": 2780
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 6.58,
|
| 1728 |
+
"learning_rate": 3.502415458937198e-05,
|
| 1729 |
+
"loss": 0.0028,
|
| 1730 |
+
"step": 2790
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 6.6,
|
| 1734 |
+
"learning_rate": 3.478260869565218e-05,
|
| 1735 |
+
"loss": 0.0048,
|
| 1736 |
+
"step": 2800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 6.63,
|
| 1740 |
+
"learning_rate": 3.454106280193237e-05,
|
| 1741 |
+
"loss": 0.002,
|
| 1742 |
+
"step": 2810
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 6.65,
|
| 1746 |
+
"learning_rate": 3.429951690821256e-05,
|
| 1747 |
+
"loss": 0.0017,
|
| 1748 |
+
"step": 2820
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 6.67,
|
| 1752 |
+
"learning_rate": 3.405797101449276e-05,
|
| 1753 |
+
"loss": 0.0036,
|
| 1754 |
+
"step": 2830
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 6.7,
|
| 1758 |
+
"learning_rate": 3.381642512077295e-05,
|
| 1759 |
+
"loss": 0.0035,
|
| 1760 |
+
"step": 2840
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 6.72,
|
| 1764 |
+
"learning_rate": 3.357487922705314e-05,
|
| 1765 |
+
"loss": 0.0047,
|
| 1766 |
+
"step": 2850
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 6.75,
|
| 1770 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 1771 |
+
"loss": 0.0047,
|
| 1772 |
+
"step": 2860
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 6.77,
|
| 1776 |
+
"learning_rate": 3.3091787439613533e-05,
|
| 1777 |
+
"loss": 0.0028,
|
| 1778 |
+
"step": 2870
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 6.79,
|
| 1782 |
+
"learning_rate": 3.2850241545893725e-05,
|
| 1783 |
+
"loss": 0.0045,
|
| 1784 |
+
"step": 2880
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 6.82,
|
| 1788 |
+
"learning_rate": 3.260869565217392e-05,
|
| 1789 |
+
"loss": 0.0027,
|
| 1790 |
+
"step": 2890
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 6.84,
|
| 1794 |
+
"learning_rate": 3.236714975845411e-05,
|
| 1795 |
+
"loss": 0.002,
|
| 1796 |
+
"step": 2900
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 6.86,
|
| 1800 |
+
"learning_rate": 3.21256038647343e-05,
|
| 1801 |
+
"loss": 0.004,
|
| 1802 |
+
"step": 2910
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 6.89,
|
| 1806 |
+
"learning_rate": 3.188405797101449e-05,
|
| 1807 |
+
"loss": 0.0033,
|
| 1808 |
+
"step": 2920
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 6.91,
|
| 1812 |
+
"learning_rate": 3.164251207729469e-05,
|
| 1813 |
+
"loss": 0.0032,
|
| 1814 |
+
"step": 2930
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 6.93,
|
| 1818 |
+
"learning_rate": 3.140096618357488e-05,
|
| 1819 |
+
"loss": 0.0106,
|
| 1820 |
+
"step": 2940
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 6.96,
|
| 1824 |
+
"learning_rate": 3.1159420289855074e-05,
|
| 1825 |
+
"loss": 0.0025,
|
| 1826 |
+
"step": 2950
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 6.98,
|
| 1830 |
+
"learning_rate": 3.0917874396135266e-05,
|
| 1831 |
+
"loss": 0.0026,
|
| 1832 |
+
"step": 2960
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 7.0,
|
| 1836 |
+
"eval_loss": 2.094109296798706,
|
| 1837 |
+
"eval_runtime": 73.2503,
|
| 1838 |
+
"eval_samples_per_second": 10.28,
|
| 1839 |
+
"eval_steps_per_second": 0.655,
|
| 1840 |
+
"step": 2968
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 7.0,
|
| 1844 |
+
"learning_rate": 3.067632850241546e-05,
|
| 1845 |
+
"loss": 0.0051,
|
| 1846 |
+
"step": 2970
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 7.03,
|
| 1850 |
+
"learning_rate": 3.0434782608695656e-05,
|
| 1851 |
+
"loss": 0.0017,
|
| 1852 |
+
"step": 2980
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 7.05,
|
| 1856 |
+
"learning_rate": 3.0193236714975848e-05,
|
| 1857 |
+
"loss": 0.001,
|
| 1858 |
+
"step": 2990
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 7.08,
|
| 1862 |
+
"learning_rate": 2.995169082125604e-05,
|
| 1863 |
+
"loss": 0.0035,
|
| 1864 |
+
"step": 3000
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 7.1,
|
| 1868 |
+
"learning_rate": 2.971014492753623e-05,
|
| 1869 |
+
"loss": 0.0011,
|
| 1870 |
+
"step": 3010
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 7.12,
|
| 1874 |
+
"learning_rate": 2.9468599033816423e-05,
|
| 1875 |
+
"loss": 0.0022,
|
| 1876 |
+
"step": 3020
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 7.15,
|
| 1880 |
+
"learning_rate": 2.9227053140096622e-05,
|
| 1881 |
+
"loss": 0.0013,
|
| 1882 |
+
"step": 3030
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 7.17,
|
| 1886 |
+
"learning_rate": 2.8985507246376814e-05,
|
| 1887 |
+
"loss": 0.0024,
|
| 1888 |
+
"step": 3040
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 7.19,
|
| 1892 |
+
"learning_rate": 2.8743961352657005e-05,
|
| 1893 |
+
"loss": 0.001,
|
| 1894 |
+
"step": 3050
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 7.22,
|
| 1898 |
+
"learning_rate": 2.8502415458937197e-05,
|
| 1899 |
+
"loss": 0.0012,
|
| 1900 |
+
"step": 3060
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 7.24,
|
| 1904 |
+
"learning_rate": 2.826086956521739e-05,
|
| 1905 |
+
"loss": 0.001,
|
| 1906 |
+
"step": 3070
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 7.26,
|
| 1910 |
+
"learning_rate": 2.8019323671497587e-05,
|
| 1911 |
+
"loss": 0.0023,
|
| 1912 |
+
"step": 3080
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 7.29,
|
| 1916 |
+
"learning_rate": 2.777777777777778e-05,
|
| 1917 |
+
"loss": 0.0013,
|
| 1918 |
+
"step": 3090
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 7.31,
|
| 1922 |
+
"learning_rate": 2.753623188405797e-05,
|
| 1923 |
+
"loss": 0.001,
|
| 1924 |
+
"step": 3100
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 7.33,
|
| 1928 |
+
"learning_rate": 2.7294685990338166e-05,
|
| 1929 |
+
"loss": 0.0022,
|
| 1930 |
+
"step": 3110
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 7.36,
|
| 1934 |
+
"learning_rate": 2.7053140096618358e-05,
|
| 1935 |
+
"loss": 0.0011,
|
| 1936 |
+
"step": 3120
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 7.38,
|
| 1940 |
+
"learning_rate": 2.6811594202898553e-05,
|
| 1941 |
+
"loss": 0.001,
|
| 1942 |
+
"step": 3130
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 7.41,
|
| 1946 |
+
"learning_rate": 2.6570048309178748e-05,
|
| 1947 |
+
"loss": 0.0016,
|
| 1948 |
+
"step": 3140
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 7.43,
|
| 1952 |
+
"learning_rate": 2.632850241545894e-05,
|
| 1953 |
+
"loss": 0.0009,
|
| 1954 |
+
"step": 3150
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 7.45,
|
| 1958 |
+
"learning_rate": 2.608695652173913e-05,
|
| 1959 |
+
"loss": 0.0019,
|
| 1960 |
+
"step": 3160
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 7.48,
|
| 1964 |
+
"learning_rate": 2.5845410628019323e-05,
|
| 1965 |
+
"loss": 0.0013,
|
| 1966 |
+
"step": 3170
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 7.5,
|
| 1970 |
+
"learning_rate": 2.5603864734299522e-05,
|
| 1971 |
+
"loss": 0.0009,
|
| 1972 |
+
"step": 3180
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 7.52,
|
| 1976 |
+
"learning_rate": 2.5362318840579714e-05,
|
| 1977 |
+
"loss": 0.0011,
|
| 1978 |
+
"step": 3190
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 7.55,
|
| 1982 |
+
"learning_rate": 2.5120772946859905e-05,
|
| 1983 |
+
"loss": 0.0024,
|
| 1984 |
+
"step": 3200
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 7.57,
|
| 1988 |
+
"learning_rate": 2.4879227053140097e-05,
|
| 1989 |
+
"loss": 0.0013,
|
| 1990 |
+
"step": 3210
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 7.59,
|
| 1994 |
+
"learning_rate": 2.4637681159420292e-05,
|
| 1995 |
+
"loss": 0.0012,
|
| 1996 |
+
"step": 3220
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 7.62,
|
| 2000 |
+
"learning_rate": 2.4396135265700484e-05,
|
| 2001 |
+
"loss": 0.0011,
|
| 2002 |
+
"step": 3230
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 7.64,
|
| 2006 |
+
"learning_rate": 2.4154589371980676e-05,
|
| 2007 |
+
"loss": 0.0022,
|
| 2008 |
+
"step": 3240
|
| 2009 |
+
},
|
| 2010 |
+
{
|
| 2011 |
+
"epoch": 7.67,
|
| 2012 |
+
"learning_rate": 2.391304347826087e-05,
|
| 2013 |
+
"loss": 0.0011,
|
| 2014 |
+
"step": 3250
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 7.69,
|
| 2018 |
+
"learning_rate": 2.3671497584541063e-05,
|
| 2019 |
+
"loss": 0.0074,
|
| 2020 |
+
"step": 3260
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 7.71,
|
| 2024 |
+
"learning_rate": 2.3429951690821258e-05,
|
| 2025 |
+
"loss": 0.001,
|
| 2026 |
+
"step": 3270
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 7.74,
|
| 2030 |
+
"learning_rate": 2.318840579710145e-05,
|
| 2031 |
+
"loss": 0.0016,
|
| 2032 |
+
"step": 3280
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 7.76,
|
| 2036 |
+
"learning_rate": 2.294685990338164e-05,
|
| 2037 |
+
"loss": 0.0028,
|
| 2038 |
+
"step": 3290
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 7.78,
|
| 2042 |
+
"learning_rate": 2.2705314009661836e-05,
|
| 2043 |
+
"loss": 0.0015,
|
| 2044 |
+
"step": 3300
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 7.81,
|
| 2048 |
+
"learning_rate": 2.246376811594203e-05,
|
| 2049 |
+
"loss": 0.0038,
|
| 2050 |
+
"step": 3310
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 7.83,
|
| 2054 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 2055 |
+
"loss": 0.0013,
|
| 2056 |
+
"step": 3320
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 7.85,
|
| 2060 |
+
"learning_rate": 2.198067632850242e-05,
|
| 2061 |
+
"loss": 0.0009,
|
| 2062 |
+
"step": 3330
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 7.88,
|
| 2066 |
+
"learning_rate": 2.173913043478261e-05,
|
| 2067 |
+
"loss": 0.001,
|
| 2068 |
+
"step": 3340
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 7.9,
|
| 2072 |
+
"learning_rate": 2.1497584541062805e-05,
|
| 2073 |
+
"loss": 0.001,
|
| 2074 |
+
"step": 3350
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 7.92,
|
| 2078 |
+
"learning_rate": 2.1256038647342997e-05,
|
| 2079 |
+
"loss": 0.001,
|
| 2080 |
+
"step": 3360
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 7.95,
|
| 2084 |
+
"learning_rate": 2.101449275362319e-05,
|
| 2085 |
+
"loss": 0.0018,
|
| 2086 |
+
"step": 3370
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 7.97,
|
| 2090 |
+
"learning_rate": 2.0772946859903384e-05,
|
| 2091 |
+
"loss": 0.0015,
|
| 2092 |
+
"step": 3380
|
| 2093 |
+
},
|
| 2094 |
+
{
|
| 2095 |
+
"epoch": 8.0,
|
| 2096 |
+
"learning_rate": 2.0531400966183576e-05,
|
| 2097 |
+
"loss": 0.001,
|
| 2098 |
+
"step": 3390
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 8.0,
|
| 2102 |
+
"eval_loss": 2.134007215499878,
|
| 2103 |
+
"eval_runtime": 73.9295,
|
| 2104 |
+
"eval_samples_per_second": 10.185,
|
| 2105 |
+
"eval_steps_per_second": 0.649,
|
| 2106 |
+
"step": 3392
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 8.02,
|
| 2110 |
+
"learning_rate": 2.028985507246377e-05,
|
| 2111 |
+
"loss": 0.0011,
|
| 2112 |
+
"step": 3400
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 8.04,
|
| 2116 |
+
"learning_rate": 2.0048309178743963e-05,
|
| 2117 |
+
"loss": 0.0014,
|
| 2118 |
+
"step": 3410
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 8.07,
|
| 2122 |
+
"learning_rate": 1.9806763285024154e-05,
|
| 2123 |
+
"loss": 0.0007,
|
| 2124 |
+
"step": 3420
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 8.09,
|
| 2128 |
+
"learning_rate": 1.956521739130435e-05,
|
| 2129 |
+
"loss": 0.0008,
|
| 2130 |
+
"step": 3430
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 8.11,
|
| 2134 |
+
"learning_rate": 1.932367149758454e-05,
|
| 2135 |
+
"loss": 0.0007,
|
| 2136 |
+
"step": 3440
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 8.14,
|
| 2140 |
+
"learning_rate": 1.9082125603864733e-05,
|
| 2141 |
+
"loss": 0.0008,
|
| 2142 |
+
"step": 3450
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 8.16,
|
| 2146 |
+
"learning_rate": 1.8840579710144928e-05,
|
| 2147 |
+
"loss": 0.0008,
|
| 2148 |
+
"step": 3460
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 8.18,
|
| 2152 |
+
"learning_rate": 1.859903381642512e-05,
|
| 2153 |
+
"loss": 0.0009,
|
| 2154 |
+
"step": 3470
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 8.21,
|
| 2158 |
+
"learning_rate": 1.8357487922705315e-05,
|
| 2159 |
+
"loss": 0.0007,
|
| 2160 |
+
"step": 3480
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 8.23,
|
| 2164 |
+
"learning_rate": 1.8115942028985507e-05,
|
| 2165 |
+
"loss": 0.0012,
|
| 2166 |
+
"step": 3490
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 8.25,
|
| 2170 |
+
"learning_rate": 1.78743961352657e-05,
|
| 2171 |
+
"loss": 0.0007,
|
| 2172 |
+
"step": 3500
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 8.28,
|
| 2176 |
+
"learning_rate": 1.7632850241545894e-05,
|
| 2177 |
+
"loss": 0.003,
|
| 2178 |
+
"step": 3510
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 8.3,
|
| 2182 |
+
"learning_rate": 1.739130434782609e-05,
|
| 2183 |
+
"loss": 0.0013,
|
| 2184 |
+
"step": 3520
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 8.33,
|
| 2188 |
+
"learning_rate": 1.714975845410628e-05,
|
| 2189 |
+
"loss": 0.0009,
|
| 2190 |
+
"step": 3530
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 8.35,
|
| 2194 |
+
"learning_rate": 1.6908212560386476e-05,
|
| 2195 |
+
"loss": 0.0011,
|
| 2196 |
+
"step": 3540
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 8.37,
|
| 2200 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 2201 |
+
"loss": 0.0009,
|
| 2202 |
+
"step": 3550
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 8.4,
|
| 2206 |
+
"learning_rate": 1.6425120772946863e-05,
|
| 2207 |
+
"loss": 0.0009,
|
| 2208 |
+
"step": 3560
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 8.42,
|
| 2212 |
+
"learning_rate": 1.6183574879227054e-05,
|
| 2213 |
+
"loss": 0.001,
|
| 2214 |
+
"step": 3570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 8.44,
|
| 2218 |
+
"learning_rate": 1.5942028985507246e-05,
|
| 2219 |
+
"loss": 0.0008,
|
| 2220 |
+
"step": 3580
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 8.47,
|
| 2224 |
+
"learning_rate": 1.570048309178744e-05,
|
| 2225 |
+
"loss": 0.0008,
|
| 2226 |
+
"step": 3590
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 8.49,
|
| 2230 |
+
"learning_rate": 1.5458937198067633e-05,
|
| 2231 |
+
"loss": 0.001,
|
| 2232 |
+
"step": 3600
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 8.51,
|
| 2236 |
+
"learning_rate": 1.5217391304347828e-05,
|
| 2237 |
+
"loss": 0.0007,
|
| 2238 |
+
"step": 3610
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 8.54,
|
| 2242 |
+
"learning_rate": 1.497584541062802e-05,
|
| 2243 |
+
"loss": 0.0016,
|
| 2244 |
+
"step": 3620
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 8.56,
|
| 2248 |
+
"learning_rate": 1.4734299516908212e-05,
|
| 2249 |
+
"loss": 0.0009,
|
| 2250 |
+
"step": 3630
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 8.58,
|
| 2254 |
+
"learning_rate": 1.4492753623188407e-05,
|
| 2255 |
+
"loss": 0.0007,
|
| 2256 |
+
"step": 3640
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 8.61,
|
| 2260 |
+
"learning_rate": 1.4251207729468599e-05,
|
| 2261 |
+
"loss": 0.0007,
|
| 2262 |
+
"step": 3650
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 8.63,
|
| 2266 |
+
"learning_rate": 1.4009661835748794e-05,
|
| 2267 |
+
"loss": 0.0007,
|
| 2268 |
+
"step": 3660
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 8.66,
|
| 2272 |
+
"learning_rate": 1.3768115942028985e-05,
|
| 2273 |
+
"loss": 0.0023,
|
| 2274 |
+
"step": 3670
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 8.68,
|
| 2278 |
+
"learning_rate": 1.3526570048309179e-05,
|
| 2279 |
+
"loss": 0.0008,
|
| 2280 |
+
"step": 3680
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 8.7,
|
| 2284 |
+
"learning_rate": 1.3285024154589374e-05,
|
| 2285 |
+
"loss": 0.0007,
|
| 2286 |
+
"step": 3690
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 8.73,
|
| 2290 |
+
"learning_rate": 1.3043478260869566e-05,
|
| 2291 |
+
"loss": 0.0007,
|
| 2292 |
+
"step": 3700
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 8.75,
|
| 2296 |
+
"learning_rate": 1.2801932367149761e-05,
|
| 2297 |
+
"loss": 0.0008,
|
| 2298 |
+
"step": 3710
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 8.77,
|
| 2302 |
+
"learning_rate": 1.2560386473429953e-05,
|
| 2303 |
+
"loss": 0.0009,
|
| 2304 |
+
"step": 3720
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 8.8,
|
| 2308 |
+
"learning_rate": 1.2318840579710146e-05,
|
| 2309 |
+
"loss": 0.0007,
|
| 2310 |
+
"step": 3730
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 8.82,
|
| 2314 |
+
"learning_rate": 1.2077294685990338e-05,
|
| 2315 |
+
"loss": 0.0008,
|
| 2316 |
+
"step": 3740
|
| 2317 |
+
},
|
| 2318 |
+
{
|
| 2319 |
+
"epoch": 8.84,
|
| 2320 |
+
"learning_rate": 1.1835748792270531e-05,
|
| 2321 |
+
"loss": 0.0012,
|
| 2322 |
+
"step": 3750
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 8.87,
|
| 2326 |
+
"learning_rate": 1.1594202898550725e-05,
|
| 2327 |
+
"loss": 0.0007,
|
| 2328 |
+
"step": 3760
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 8.89,
|
| 2332 |
+
"learning_rate": 1.1352657004830918e-05,
|
| 2333 |
+
"loss": 0.0008,
|
| 2334 |
+
"step": 3770
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 8.92,
|
| 2338 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 2339 |
+
"loss": 0.0013,
|
| 2340 |
+
"step": 3780
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 8.94,
|
| 2344 |
+
"learning_rate": 1.0869565217391305e-05,
|
| 2345 |
+
"loss": 0.0007,
|
| 2346 |
+
"step": 3790
|
| 2347 |
+
},
|
| 2348 |
+
{
|
| 2349 |
+
"epoch": 8.96,
|
| 2350 |
+
"learning_rate": 1.0628019323671499e-05,
|
| 2351 |
+
"loss": 0.0007,
|
| 2352 |
+
"step": 3800
|
| 2353 |
+
},
|
| 2354 |
+
{
|
| 2355 |
+
"epoch": 8.99,
|
| 2356 |
+
"learning_rate": 1.0386473429951692e-05,
|
| 2357 |
+
"loss": 0.0007,
|
| 2358 |
+
"step": 3810
|
| 2359 |
+
},
|
| 2360 |
+
{
|
| 2361 |
+
"epoch": 9.0,
|
| 2362 |
+
"eval_loss": 2.1589605808258057,
|
| 2363 |
+
"eval_runtime": 74.7303,
|
| 2364 |
+
"eval_samples_per_second": 10.076,
|
| 2365 |
+
"eval_steps_per_second": 0.642,
|
| 2366 |
+
"step": 3816
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 9.01,
|
| 2370 |
+
"learning_rate": 1.0144927536231885e-05,
|
| 2371 |
+
"loss": 0.0008,
|
| 2372 |
+
"step": 3820
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 9.03,
|
| 2376 |
+
"learning_rate": 9.903381642512077e-06,
|
| 2377 |
+
"loss": 0.0006,
|
| 2378 |
+
"step": 3830
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 9.06,
|
| 2382 |
+
"learning_rate": 9.66183574879227e-06,
|
| 2383 |
+
"loss": 0.0007,
|
| 2384 |
+
"step": 3840
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 9.08,
|
| 2388 |
+
"learning_rate": 9.420289855072464e-06,
|
| 2389 |
+
"loss": 0.0006,
|
| 2390 |
+
"step": 3850
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 9.1,
|
| 2394 |
+
"learning_rate": 9.178743961352658e-06,
|
| 2395 |
+
"loss": 0.0006,
|
| 2396 |
+
"step": 3860
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 9.13,
|
| 2400 |
+
"learning_rate": 8.93719806763285e-06,
|
| 2401 |
+
"loss": 0.0006,
|
| 2402 |
+
"step": 3870
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 9.15,
|
| 2406 |
+
"learning_rate": 8.695652173913044e-06,
|
| 2407 |
+
"loss": 0.0006,
|
| 2408 |
+
"step": 3880
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 9.17,
|
| 2412 |
+
"learning_rate": 8.454106280193238e-06,
|
| 2413 |
+
"loss": 0.0009,
|
| 2414 |
+
"step": 3890
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 9.2,
|
| 2418 |
+
"learning_rate": 8.212560386473431e-06,
|
| 2419 |
+
"loss": 0.0009,
|
| 2420 |
+
"step": 3900
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 9.22,
|
| 2424 |
+
"learning_rate": 7.971014492753623e-06,
|
| 2425 |
+
"loss": 0.0006,
|
| 2426 |
+
"step": 3910
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 9.25,
|
| 2430 |
+
"learning_rate": 7.729468599033817e-06,
|
| 2431 |
+
"loss": 0.0006,
|
| 2432 |
+
"step": 3920
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 9.27,
|
| 2436 |
+
"learning_rate": 7.48792270531401e-06,
|
| 2437 |
+
"loss": 0.0007,
|
| 2438 |
+
"step": 3930
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 9.29,
|
| 2442 |
+
"learning_rate": 7.246376811594203e-06,
|
| 2443 |
+
"loss": 0.0008,
|
| 2444 |
+
"step": 3940
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 9.32,
|
| 2448 |
+
"learning_rate": 7.004830917874397e-06,
|
| 2449 |
+
"loss": 0.0006,
|
| 2450 |
+
"step": 3950
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 9.34,
|
| 2454 |
+
"learning_rate": 6.7632850241545894e-06,
|
| 2455 |
+
"loss": 0.0006,
|
| 2456 |
+
"step": 3960
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 9.36,
|
| 2460 |
+
"learning_rate": 6.521739130434783e-06,
|
| 2461 |
+
"loss": 0.0006,
|
| 2462 |
+
"step": 3970
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 9.39,
|
| 2466 |
+
"learning_rate": 6.280193236714976e-06,
|
| 2467 |
+
"loss": 0.0007,
|
| 2468 |
+
"step": 3980
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 9.41,
|
| 2472 |
+
"learning_rate": 6.038647342995169e-06,
|
| 2473 |
+
"loss": 0.0007,
|
| 2474 |
+
"step": 3990
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 9.43,
|
| 2478 |
+
"learning_rate": 5.797101449275362e-06,
|
| 2479 |
+
"loss": 0.0006,
|
| 2480 |
+
"step": 4000
|
| 2481 |
+
}
|
| 2482 |
+
],
|
| 2483 |
+
"max_steps": 4240,
|
| 2484 |
+
"num_train_epochs": 10,
|
| 2485 |
+
"total_flos": 0.0,
|
| 2486 |
+
"trial_name": null,
|
| 2487 |
+
"trial_params": null
|
| 2488 |
+
}
|
checkpoint-4000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7b27dc19e266bc07c05772a12a4c0ba666e04d9c5ca7a7c6a3ae1c9c6c2ce46
|
| 3 |
+
size 3759
|
config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "t5-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 4096,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 1024,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dense_act_fn": "relu",
|
| 11 |
+
"dropout_rate": 0.1,
|
| 12 |
+
"eos_token_id": 1,
|
| 13 |
+
"feed_forward_proj": "relu",
|
| 14 |
+
"initializer_factor": 1.0,
|
| 15 |
+
"is_encoder_decoder": true,
|
| 16 |
+
"is_gated_act": false,
|
| 17 |
+
"layer_norm_epsilon": 1e-06,
|
| 18 |
+
"model_type": "t5",
|
| 19 |
+
"n_positions": 512,
|
| 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 |
+
"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 |
+
"torch_dtype": "float32",
|
| 57 |
+
"transformers_version": "4.21.1",
|
| 58 |
+
"use_cache": true,
|
| 59 |
+
"vocab_size": 32128
|
| 60 |
+
}
|
outputs.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2bec7ee4be19d7c0e2a23371a3b17176c6012e207f7b3e69d371f753ec45bb64
|
| 3 |
+
size 2950844807
|
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": "<unk>"
|
| 107 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,112 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"extra_ids": 100,
|
| 106 |
+
"model_max_length": 512,
|
| 107 |
+
"name_or_path": "t5-large",
|
| 108 |
+
"pad_token": "<pad>",
|
| 109 |
+
"special_tokens_map_file": null,
|
| 110 |
+
"tokenizer_class": "T5Tokenizer",
|
| 111 |
+
"unk_token": "<unk>"
|
| 112 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7b27dc19e266bc07c05772a12a4c0ba666e04d9c5ca7a7c6a3ae1c9c6c2ce46
|
| 3 |
+
size 3759
|