Hugging Face's logo

nllg
/
poetry-bygpt5-medium-de

Fill-Mask
Transformers
PyTorch
bygpt5
Inference Endpoints
Model card Files Community
potamides commited on
Commit
4e30fbc
•
1 Parent(s): 00a8baa

add model files

Browse files
Files changed (6) hide show
  1. config.json +34 -0
  2. pytorch_model.bin +3 -0
  3. special_tokens_map.json +151 -0
  4. tokenizer_config.json +160 -0
  5. trainer_state.json +2929 -0
  6. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "models/bygpt5-medium/de",
3
+ "architectures": [
4
+ "ByGPT5LMHeadModel"
5
+ ],
6
+ "d_ff": 3840,
7
+ "d_kv": 64,
8
+ "d_model": 1536,
9
+ "decoder_start_token_id": 0,
10
+ "dense_act_fn": "gelu_new",
11
+ "dropout_rate": 0.1,
12
+ "eos_token_id": 1,
13
+ "feed_forward_proj": "gated-gelu",
14
+ "gradient_checkpointing": false,
15
+ "initializer_factor": 1.0,
16
+ "is_decoder": true,
17
+ "is_encoder_decoder": false,
18
+ "is_gated_act": true,
19
+ "layer_norm_epsilon": 1e-06,
20
+ "model_type": "bygpt5",
21
+ "num_decoder_layers": 12,
22
+ "num_heads": 16,
23
+ "num_layers": 12,
24
+ "output_past": true,
25
+ "pad_token_id": 0,
26
+ "relative_attention_max_distance": 128,
27
+ "relative_attention_num_buckets": 32,
28
+ "tie_word_embeddings": false,
29
+ "tokenizer_class": "ByT5Tokenizer",
30
+ "torch_dtype": "float32",
31
+ "transformers_version": "4.20.0",
32
+ "use_cache": true,
33
+ "vocab_size": 384
34
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b3e5b5b05d9d1145defca722bc6f63b60c4e50451585c176bdf7247a62407168
3
+ size 1156247841
special_tokens_map.json ADDED
@@ -0,0 +1,151 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "<extra_id_100>",
104
+ "<extra_id_101>",
105
+ "<extra_id_102>",
106
+ "<extra_id_103>",
107
+ "<extra_id_104>",
108
+ "<extra_id_105>",
109
+ "<extra_id_106>",
110
+ "<extra_id_107>",
111
+ "<extra_id_108>",
112
+ "<extra_id_109>",
113
+ "<extra_id_110>",
114
+ "<extra_id_111>",
115
+ "<extra_id_112>",
116
+ "<extra_id_113>",
117
+ "<extra_id_114>",
118
+ "<extra_id_115>",
119
+ "<extra_id_116>",
120
+ "<extra_id_117>",
121
+ "<extra_id_118>",
122
+ "<extra_id_119>",
123
+ "<extra_id_120>",
124
+ "<extra_id_121>",
125
+ "<extra_id_122>",
126
+ "<extra_id_123>",
127
+ "<extra_id_124>"
128
+ ],
129
+ "bos_token": "</s>",
130
+ "eos_token": {
131
+ "content": "</s>",
132
+ "lstrip": false,
133
+ "normalized": true,
134
+ "rstrip": false,
135
+ "single_word": false
136
+ },
137
+ "pad_token": {
138
+ "content": "<pad>",
139
+ "lstrip": false,
140
+ "normalized": true,
141
+ "rstrip": false,
142
+ "single_word": false
143
+ },
144
+ "unk_token": {
145
+ "content": "<unk>",
146
+ "lstrip": false,
147
+ "normalized": true,
148
+ "rstrip": false,
149
+ "single_word": false
150
+ }
151
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,160 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": false,
5
+ "additional_special_tokens": [
6
+ "<extra_id_0>",
7
+ "<extra_id_1>",
8
+ "<extra_id_2>",
9
+ "<extra_id_3>",
10
+ "<extra_id_4>",
11
+ "<extra_id_5>",
12
+ "<extra_id_6>",
13
+ "<extra_id_7>",
14
+ "<extra_id_8>",
15
+ "<extra_id_9>",
16
+ "<extra_id_10>",
17
+ "<extra_id_11>",
18
+ "<extra_id_12>",
19
+ "<extra_id_13>",
20
+ "<extra_id_14>",
21
+ "<extra_id_15>",
22
+ "<extra_id_16>",
23
+ "<extra_id_17>",
24
+ "<extra_id_18>",
25
+ "<extra_id_19>",
26
+ "<extra_id_20>",
27
+ "<extra_id_21>",
28
+ "<extra_id_22>",
29
+ "<extra_id_23>",
30
+ "<extra_id_24>",
31
+ "<extra_id_25>",
32
+ "<extra_id_26>",
33
+ "<extra_id_27>",
34
+ "<extra_id_28>",
35
+ "<extra_id_29>",
36
+ "<extra_id_30>",
37
+ "<extra_id_31>",
38
+ "<extra_id_32>",
39
+ "<extra_id_33>",
40
+ "<extra_id_34>",
41
+ "<extra_id_35>",
42
+ "<extra_id_36>",
43
+ "<extra_id_37>",
44
+ "<extra_id_38>",
45
+ "<extra_id_39>",
46
+ "<extra_id_40>",
47
+ "<extra_id_41>",
48
+ "<extra_id_42>",
49
+ "<extra_id_43>",
50
+ "<extra_id_44>",
51
+ "<extra_id_45>",
52
+ "<extra_id_46>",
53
+ "<extra_id_47>",
54
+ "<extra_id_48>",
55
+ "<extra_id_49>",
56
+ "<extra_id_50>",
57
+ "<extra_id_51>",
58
+ "<extra_id_52>",
59
+ "<extra_id_53>",
60
+ "<extra_id_54>",
61
+ "<extra_id_55>",
62
+ "<extra_id_56>",
63
+ "<extra_id_57>",
64
+ "<extra_id_58>",
65
+ "<extra_id_59>",
66
+ "<extra_id_60>",
67
+ "<extra_id_61>",
68
+ "<extra_id_62>",
69
+ "<extra_id_63>",
70
+ "<extra_id_64>",
71
+ "<extra_id_65>",
72
+ "<extra_id_66>",
73
+ "<extra_id_67>",
74
+ "<extra_id_68>",
75
+ "<extra_id_69>",
76
+ "<extra_id_70>",
77
+ "<extra_id_71>",
78
+ "<extra_id_72>",
79
+ "<extra_id_73>",
80
+ "<extra_id_74>",
81
+ "<extra_id_75>",
82
+ "<extra_id_76>",
83
+ "<extra_id_77>",
84
+ "<extra_id_78>",
85
+ "<extra_id_79>",
86
+ "<extra_id_80>",
87
+ "<extra_id_81>",
88
+ "<extra_id_82>",
89
+ "<extra_id_83>",
90
+ "<extra_id_84>",
91
+ "<extra_id_85>",
92
+ "<extra_id_86>",
93
+ "<extra_id_87>",
94
+ "<extra_id_88>",
95
+ "<extra_id_89>",
96
+ "<extra_id_90>",
97
+ "<extra_id_91>",
98
+ "<extra_id_92>",
99
+ "<extra_id_93>",
100
+ "<extra_id_94>",
101
+ "<extra_id_95>",
102
+ "<extra_id_96>",
103
+ "<extra_id_97>",
104
+ "<extra_id_98>",
105
+ "<extra_id_99>",
106
+ "<extra_id_100>",
107
+ "<extra_id_101>",
108
+ "<extra_id_102>",
109
+ "<extra_id_103>",
110
+ "<extra_id_104>",
111
+ "<extra_id_105>",
112
+ "<extra_id_106>",
113
+ "<extra_id_107>",
114
+ "<extra_id_108>",
115
+ "<extra_id_109>",
116
+ "<extra_id_110>",
117
+ "<extra_id_111>",
118
+ "<extra_id_112>",
119
+ "<extra_id_113>",
120
+ "<extra_id_114>",
121
+ "<extra_id_115>",
122
+ "<extra_id_116>",
123
+ "<extra_id_117>",
124
+ "<extra_id_118>",
125
+ "<extra_id_119>",
126
+ "<extra_id_120>",
127
+ "<extra_id_121>",
128
+ "<extra_id_122>",
129
+ "<extra_id_123>",
130
+ "<extra_id_124>"
131
+ ],
132
+ "eos_token": {
133
+ "__type": "AddedToken",
134
+ "content": "</s>",
135
+ "lstrip": false,
136
+ "normalized": true,
137
+ "rstrip": false,
138
+ "single_word": false
139
+ },
140
+ "extra_ids": 125,
141
+ "name_or_path": "models/bygpt5-medium/de",
142
+ "pad_token": {
143
+ "__type": "AddedToken",
144
+ "content": "<pad>",
145
+ "lstrip": false,
146
+ "normalized": true,
147
+ "rstrip": false,
148
+ "single_word": false
149
+ },
150
+ "special_tokens_map_file": "/storage/ukp/work/jbelouadi/.cache/huggingface/transformers/1d7794e7e6dcb9aff7f1b0f72b85d665041ce4cfc7c59cf78a0e83302f9ab8e3.063895353d5ef9e19a25220cb616c43abc5e84a2f11b1ffb71c29e097572a109",
151
+ "tokenizer_class": "ByGPT5Tokenizer",
152
+ "unk_token": {
153
+ "__type": "AddedToken",
154
+ "content": "<unk>",
155
+ "lstrip": false,
156
+ "normalized": true,
157
+ "rstrip": false,
158
+ "single_word": false
159
+ }
160
+ }
trainer_state.json ADDED
@@ -0,0 +1,2929 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 9.99995686494414,
5
+ "global_step": 115910,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 2.586206896551724e-07,
13
+ "loss": 1.9686,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.02,
18
+ "learning_rate": 6.465517241379309e-05,
19
+ "loss": 1.4946,
20
+ "step": 250
21
+ },
22
+ {
23
+ "epoch": 0.04,
24
+ "learning_rate": 0.00012931034482758618,
25
+ "loss": 1.2379,
26
+ "step": 500
27
+ },
28
+ {
29
+ "epoch": 0.06,
30
+ "learning_rate": 0.0001939655172413793,
31
+ "loss": 1.1796,
32
+ "step": 750
33
+ },
34
+ {
35
+ "epoch": 0.09,
36
+ "learning_rate": 0.00025862068965517237,
37
+ "loss": 1.1506,
38
+ "step": 1000
39
+ },
40
+ {
41
+ "epoch": 0.11,
42
+ "learning_rate": 0.00029999954465515996,
43
+ "loss": 1.1289,
44
+ "step": 1250
45
+ },
46
+ {
47
+ "epoch": 0.13,
48
+ "learning_rate": 0.0002999935015419672,
49
+ "loss": 1.1127,
50
+ "step": 1500
51
+ },
52
+ {
53
+ "epoch": 0.15,
54
+ "learning_rate": 0.0002999804318305346,
55
+ "loss": 1.0975,
56
+ "step": 1750
57
+ },
58
+ {
59
+ "epoch": 0.17,
60
+ "learning_rate": 0.00029996033613312583,
61
+ "loss": 1.0853,
62
+ "step": 2000
63
+ },
64
+ {
65
+ "epoch": 0.19,
66
+ "learning_rate": 0.00029993321539114454,
67
+ "loss": 1.0772,
68
+ "step": 2250
69
+ },
70
+ {
71
+ "epoch": 0.22,
72
+ "learning_rate": 0.00029989907087508934,
73
+ "loss": 1.0658,
74
+ "step": 2500
75
+ },
76
+ {
77
+ "epoch": 0.24,
78
+ "learning_rate": 0.00029985790418449485,
79
+ "loss": 1.0585,
80
+ "step": 2750
81
+ },
82
+ {
83
+ "epoch": 0.26,
84
+ "learning_rate": 0.00029980971724785665,
85
+ "loss": 1.0505,
86
+ "step": 3000
87
+ },
88
+ {
89
+ "epoch": 0.28,
90
+ "learning_rate": 0.00029975451232254073,
91
+ "loss": 1.0458,
92
+ "step": 3250
93
+ },
94
+ {
95
+ "epoch": 0.3,
96
+ "learning_rate": 0.0002996922919946781,
97
+ "loss": 1.0388,
98
+ "step": 3500
99
+ },
100
+ {
101
+ "epoch": 0.32,
102
+ "learning_rate": 0.0002996230591790433,
103
+ "loss": 1.0336,
104
+ "step": 3750
105
+ },
106
+ {
107
+ "epoch": 0.35,
108
+ "learning_rate": 0.0002995468171189181,
109
+ "loss": 1.0278,
110
+ "step": 4000
111
+ },
112
+ {
113
+ "epoch": 0.37,
114
+ "learning_rate": 0.0002994635693859393,
115
+ "loss": 1.0248,
116
+ "step": 4250
117
+ },
118
+ {
119
+ "epoch": 0.39,
120
+ "learning_rate": 0.00029937331987993176,
121
+ "loss": 1.0181,
122
+ "step": 4500
123
+ },
124
+ {
125
+ "epoch": 0.41,
126
+ "learning_rate": 0.00029927607282872543,
127
+ "loss": 1.0141,
128
+ "step": 4750
129
+ },
130
+ {
131
+ "epoch": 0.43,
132
+ "learning_rate": 0.00029917183278795734,
133
+ "loss": 1.0099,
134
+ "step": 5000
135
+ },
136
+ {
137
+ "epoch": 0.45,
138
+ "learning_rate": 0.0002990606046408584,
139
+ "loss": 1.0064,
140
+ "step": 5250
141
+ },
142
+ {
143
+ "epoch": 0.47,
144
+ "learning_rate": 0.0002989423935980243,
145
+ "loss": 1.0016,
146
+ "step": 5500
147
+ },
148
+ {
149
+ "epoch": 0.5,
150
+ "learning_rate": 0.0002988172051971717,
151
+ "loss": 0.9979,
152
+ "step": 5750
153
+ },
154
+ {
155
+ "epoch": 0.52,
156
+ "learning_rate": 0.00029868504530287856,
157
+ "loss": 0.9947,
158
+ "step": 6000
159
+ },
160
+ {
161
+ "epoch": 0.54,
162
+ "learning_rate": 0.00029854592010630966,
163
+ "loss": 0.9902,
164
+ "step": 6250
165
+ },
166
+ {
167
+ "epoch": 0.56,
168
+ "learning_rate": 0.0002983998361249265,
169
+ "loss": 0.9873,
170
+ "step": 6500
171
+ },
172
+ {
173
+ "epoch": 0.58,
174
+ "learning_rate": 0.0002982468002021818,
175
+ "loss": 0.9837,
176
+ "step": 6750
177
+ },
178
+ {
179
+ "epoch": 0.6,
180
+ "learning_rate": 0.00029808681950719917,
181
+ "loss": 0.9816,
182
+ "step": 7000
183
+ },
184
+ {
185
+ "epoch": 0.63,
186
+ "learning_rate": 0.0002979199015344371,
187
+ "loss": 0.9764,
188
+ "step": 7250
189
+ },
190
+ {
191
+ "epoch": 0.65,
192
+ "learning_rate": 0.0002977460541033379,
193
+ "loss": 0.9762,
194
+ "step": 7500
195
+ },
196
+ {
197
+ "epoch": 0.67,
198
+ "learning_rate": 0.0002975652853579614,
199
+ "loss": 0.9732,
200
+ "step": 7750
201
+ },
202
+ {
203
+ "epoch": 0.69,
204
+ "learning_rate": 0.0002973776037666034,
205
+ "loss": 0.968,
206
+ "step": 8000
207
+ },
208
+ {
209
+ "epoch": 0.71,
210
+ "learning_rate": 0.00029718301812139915,
211
+ "loss": 0.9665,
212
+ "step": 8250
213
+ },
214
+ {
215
+ "epoch": 0.73,
216
+ "learning_rate": 0.00029698153753791124,
217
+ "loss": 0.962,
218
+ "step": 8500
219
+ },
220
+ {
221
+ "epoch": 0.75,
222
+ "learning_rate": 0.00029677317145470266,
223
+ "loss": 0.9613,
224
+ "step": 8750
225
+ },
226
+ {
227
+ "epoch": 0.78,
228
+ "learning_rate": 0.0002965579296328946,
229
+ "loss": 0.9581,
230
+ "step": 9000
231
+ },
232
+ {
233
+ "epoch": 0.8,
234
+ "learning_rate": 0.00029633582215570934,
235
+ "loss": 0.955,
236
+ "step": 9250
237
+ },
238
+ {
239
+ "epoch": 0.82,
240
+ "learning_rate": 0.0002961068594279976,
241
+ "loss": 0.9544,
242
+ "step": 9500
243
+ },
244
+ {
245
+ "epoch": 0.84,
246
+ "learning_rate": 0.0002958710521757516,
247
+ "loss": 0.9511,
248
+ "step": 9750
249
+ },
250
+ {
251
+ "epoch": 0.86,
252
+ "learning_rate": 0.00029562841144560175,
253
+ "loss": 0.9486,
254
+ "step": 10000
255
+ },
256
+ {
257
+ "epoch": 0.88,
258
+ "learning_rate": 0.0002953789486043003,
259
+ "loss": 0.9458,
260
+ "step": 10250
261
+ },
262
+ {
263
+ "epoch": 0.91,
264
+ "learning_rate": 0.00029512267533818786,
265
+ "loss": 0.944,
266
+ "step": 10500
267
+ },
268
+ {
269
+ "epoch": 0.93,
270
+ "learning_rate": 0.0002948596036526464,
271
+ "loss": 0.9405,
272
+ "step": 10750
273
+ },
274
+ {
275
+ "epoch": 0.95,
276
+ "learning_rate": 0.0002945897458715367,
277
+ "loss": 0.9397,
278
+ "step": 11000
279
+ },
280
+ {
281
+ "epoch": 0.97,
282
+ "learning_rate": 0.0002943131146366213,
283
+ "loss": 0.9386,
284
+ "step": 11250
285
+ },
286
+ {
287
+ "epoch": 0.99,
288
+ "learning_rate": 0.0002940297229069719,
289
+ "loss": 0.9351,
290
+ "step": 11500
291
+ },
292
+ {
293
+ "epoch": 1.0,
294
+ "eval_alliteration_score": 0.4007220216606498,
295
+ "eval_harmonic_meter_score": 0.21012850306426992,
296
+ "eval_harmonic_rhyme_score": 0.2964593011664462,
297
+ "eval_meter_score": 0.4084058997732567,
298
+ "eval_rhyme_score": 0.7458416438967879,
299
+ "eval_runtime": 2271.5688,
300
+ "eval_samples_per_second": 1.189,
301
+ "eval_steps_per_second": 0.074,
302
+ "step": 11591
303
+ },
304
+ {
305
+ "epoch": 1.01,
306
+ "learning_rate": 0.00029373958395836246,
307
+ "loss": 0.9233,
308
+ "step": 11750
309
+ },
310
+ {
311
+ "epoch": 1.04,
312
+ "learning_rate": 0.0002934427113826474,
313
+ "loss": 0.9132,
314
+ "step": 12000
315
+ },
316
+ {
317
+ "epoch": 1.06,
318
+ "learning_rate": 0.00029313911908712464,
319
+ "loss": 0.9124,
320
+ "step": 12250
321
+ },
322
+ {
323
+ "epoch": 1.08,
324
+ "learning_rate": 0.00029282882129388416,
325
+ "loss": 0.9113,
326
+ "step": 12500
327
+ },
328
+ {
329
+ "epoch": 1.1,
330
+ "learning_rate": 0.0002925118325391421,
331
+ "loss": 0.9109,
332
+ "step": 12750
333
+ },
334
+ {
335
+ "epoch": 1.12,
336
+ "learning_rate": 0.0002921881676725592,
337
+ "loss": 0.9089,
338
+ "step": 13000
339
+ },
340
+ {
341
+ "epoch": 1.14,
342
+ "learning_rate": 0.00029185784185654554,
343
+ "loss": 0.9066,
344
+ "step": 13250
345
+ },
346
+ {
347
+ "epoch": 1.16,
348
+ "learning_rate": 0.0002915208705655501,
349
+ "loss": 0.9053,
350
+ "step": 13500
351
+ },
352
+ {
353
+ "epoch": 1.19,
354
+ "learning_rate": 0.0002911772695853361,
355
+ "loss": 0.9045,
356
+ "step": 13750
357
+ },
358
+ {
359
+ "epoch": 1.21,
360
+ "learning_rate": 0.00029082705501224114,
361
+ "loss": 0.9016,
362
+ "step": 14000
363
+ },
364
+ {
365
+ "epoch": 1.23,
366
+ "learning_rate": 0.00029047024325242333,
367
+ "loss": 0.9015,
368
+ "step": 14250
369
+ },
370
+ {
371
+ "epoch": 1.25,
372
+ "learning_rate": 0.00029010685102109267,
373
+ "loss": 0.8978,
374
+ "step": 14500
375
+ },
376
+ {
377
+ "epoch": 1.27,
378
+ "learning_rate": 0.0002897368953417281,
379
+ "loss": 0.8968,
380
+ "step": 14750
381
+ },
382
+ {
383
+ "epoch": 1.29,
384
+ "learning_rate": 0.00028936039354528003,
385
+ "loss": 0.8975,
386
+ "step": 15000
387
+ },
388
+ {
389
+ "epoch": 1.32,
390
+ "learning_rate": 0.0002889773632693582,
391
+ "loss": 0.8942,
392
+ "step": 15250
393
+ },
394
+ {
395
+ "epoch": 1.34,
396
+ "learning_rate": 0.0002885878224574058,
397
+ "loss": 0.8928,
398
+ "step": 15500
399
+ },
400
+ {
401
+ "epoch": 1.36,
402
+ "learning_rate": 0.00028819178935785857,
403
+ "loss": 0.89,
404
+ "step": 15750
405
+ },
406
+ {
407
+ "epoch": 1.38,
408
+ "learning_rate": 0.00028778928252329024,
409
+ "loss": 0.8891,
410
+ "step": 16000
411
+ },
412
+ {
413
+ "epoch": 1.4,
414
+ "learning_rate": 0.00028738032080954305,
415
+ "loss": 0.8869,
416
+ "step": 16250
417
+ },
418
+ {
419
+ "epoch": 1.42,
420
+ "learning_rate": 0.0002869649233748447,
421
+ "loss": 0.8853,
422
+ "step": 16500
423
+ },
424
+ {
425
+ "epoch": 1.45,
426
+ "learning_rate": 0.0002865431096789108,
427
+ "loss": 0.8846,
428
+ "step": 16750
429
+ },
430
+ {
431
+ "epoch": 1.47,
432
+ "learning_rate": 0.00028611489948203336,
433
+ "loss": 0.883,
434
+ "step": 17000
435
+ },
436
+ {
437
+ "epoch": 1.49,
438
+ "learning_rate": 0.0002856803128441548,
439
+ "loss": 0.8817,
440
+ "step": 17250
441
+ },
442
+ {
443
+ "epoch": 1.51,
444
+ "learning_rate": 0.0002852393701239284,
445
+ "loss": 0.8803,
446
+ "step": 17500
447
+ },
448
+ {
449
+ "epoch": 1.53,
450
+ "learning_rate": 0.0002847920919777649,
451
+ "loss": 0.8777,
452
+ "step": 17750
453
+ },
454
+ {
455
+ "epoch": 1.55,
456
+ "learning_rate": 0.000284338499358864,
457
+ "loss": 0.8774,
458
+ "step": 18000
459
+ },
460
+ {
461
+ "epoch": 1.57,
462
+ "learning_rate": 0.0002838786135162338,
463
+ "loss": 0.8758,
464
+ "step": 18250
465
+ },
466
+ {
467
+ "epoch": 1.6,
468
+ "learning_rate": 0.0002834124559936946,
469
+ "loss": 0.8737,
470
+ "step": 18500
471
+ },
472
+ {
473
+ "epoch": 1.62,
474
+ "learning_rate": 0.00028294004862887,
475
+ "loss": 0.8726,
476
+ "step": 18750
477
+ },
478
+ {
479
+ "epoch": 1.64,
480
+ "learning_rate": 0.0002824614135521637,
481
+ "loss": 0.8702,
482
+ "step": 19000
483
+ },
484
+ {
485
+ "epoch": 1.66,
486
+ "learning_rate": 0.0002819765731857232,
487
+ "loss": 0.8699,
488
+ "step": 19250
489
+ },
490
+ {
491
+ "epoch": 1.68,
492
+ "learning_rate": 0.0002814855502423888,
493
+ "loss": 0.8684,
494
+ "step": 19500
495
+ },
496
+ {
497
+ "epoch": 1.7,
498
+ "learning_rate": 0.00028098836772463007,
499
+ "loss": 0.8667,
500
+ "step": 19750
501
+ },
502
+ {
503
+ "epoch": 1.73,
504
+ "learning_rate": 0.00028048504892346806,
505
+ "loss": 0.8651,
506
+ "step": 20000
507
+ },
508
+ {
509
+ "epoch": 1.75,
510
+ "learning_rate": 0.0002799756174173843,
511
+ "loss": 0.8638,
512
+ "step": 20250
513
+ },
514
+ {
515
+ "epoch": 1.77,
516
+ "learning_rate": 0.0002794600970712162,
517
+ "loss": 0.8625,
518
+ "step": 20500
519
+ },
520
+ {
521
+ "epoch": 1.79,
522
+ "learning_rate": 0.000278938512035039,
523
+ "loss": 0.8591,
524
+ "step": 20750
525
+ },
526
+ {
527
+ "epoch": 1.81,
528
+ "learning_rate": 0.00027841088674303456,
529
+ "loss": 0.8587,
530
+ "step": 21000
531
+ },
532
+ {
533
+ "epoch": 1.83,
534
+ "learning_rate": 0.0002778772459123468,
535
+ "loss": 0.8579,
536
+ "step": 21250
537
+ },
538
+ {
539
+ "epoch": 1.85,
540
+ "learning_rate": 0.00027733761454192354,
541
+ "loss": 0.8572,
542
+ "step": 21500
543
+ },
544
+ {
545
+ "epoch": 1.88,
546
+ "learning_rate": 0.00027679201791134546,
547
+ "loss": 0.8538,
548
+ "step": 21750
549
+ },
550
+ {
551
+ "epoch": 1.9,
552
+ "learning_rate": 0.0002762404815796421,
553
+ "loss": 0.8535,
554
+ "step": 22000
555
+ },
556
+ {
557
+ "epoch": 1.92,
558
+ "learning_rate": 0.0002756830313840942,
559
+ "loss": 0.8511,
560
+ "step": 22250
561
+ },
562
+ {
563
+ "epoch": 1.94,
564
+ "learning_rate": 0.0002751196934390236,
565
+ "loss": 0.8516,
566
+ "step": 22500
567
+ },
568
+ {
569
+ "epoch": 1.96,
570
+ "learning_rate": 0.0002745504941345696,
571
+ "loss": 0.8492,
572
+ "step": 22750
573
+ },
574
+ {
575
+ "epoch": 1.98,
576
+ "learning_rate": 0.000273975460135453,
577
+ "loss": 0.8469,
578
+ "step": 23000
579
+ },
580
+ {
581
+ "epoch": 2.0,
582
+ "eval_alliteration_score": 0.43247815054976035,
583
+ "eval_harmonic_meter_score": 0.20177755686034296,
584
+ "eval_harmonic_rhyme_score": 0.4188003728656778,
585
+ "eval_meter_score": 0.4238294535271108,
586
+ "eval_rhyme_score": 0.8014050476978813,
587
+ "eval_runtime": 2299.0707,
588
+ "eval_samples_per_second": 1.174,
589
+ "eval_steps_per_second": 0.074,
590
+ "step": 23182
591
+ },
592
+ {
593
+ "epoch": 2.01,
594
+ "learning_rate": 0.0002733946183797268,
595
+ "loss": 0.8197,
596
+ "step": 23250
597
+ },
598
+ {
599
+ "epoch": 2.03,
600
+ "learning_rate": 0.0002728079960775141,
601
+ "loss": 0.8178,
602
+ "step": 23500
603
+ },
604
+ {
605
+ "epoch": 2.05,
606
+ "learning_rate": 0.0002722156207097338,
607
+ "loss": 0.8178,
608
+ "step": 23750
609
+ },
610
+ {
611
+ "epoch": 2.07,
612
+ "learning_rate": 0.00027161752002681283,
613
+ "loss": 0.8182,
614
+ "step": 24000
615
+ },
616
+ {
617
+ "epoch": 2.09,
618
+ "learning_rate": 0.0002710137220473864,
619
+ "loss": 0.8183,
620
+ "step": 24250
621
+ },
622
+ {
623
+ "epoch": 2.11,
624
+ "learning_rate": 0.00027040425505698554,
625
+ "loss": 0.8173,
626
+ "step": 24500
627
+ },
628
+ {
629
+ "epoch": 2.14,
630
+ "learning_rate": 0.00026978914760671156,
631
+ "loss": 0.8155,
632
+ "step": 24750
633
+ },
634
+ {
635
+ "epoch": 2.16,
636
+ "learning_rate": 0.00026916842851189906,
637
+ "loss": 0.8167,
638
+ "step": 25000
639
+ },
640
+ {
641
+ "epoch": 2.18,
642
+ "learning_rate": 0.0002685421268507658,
643
+ "loss": 0.8137,
644
+ "step": 25250
645
+ },
646
+ {
647
+ "epoch": 2.2,
648
+ "learning_rate": 0.00026791027196305055,
649
+ "loss": 0.8124,
650
+ "step": 25500
651
+ },
652
+ {
653
+ "epoch": 2.22,
654
+ "learning_rate": 0.00026727289344863857,
655
+ "loss": 0.8132,
656
+ "step": 25750
657
+ },
658
+ {
659
+ "epoch": 2.24,
660
+ "learning_rate": 0.0002666300211661751,
661
+ "loss": 0.8129,
662
+ "step": 26000
663
+ },
664
+ {
665
+ "epoch": 2.26,
666
+ "learning_rate": 0.00026598168523166653,
667
+ "loss": 0.8095,
668
+ "step": 26250
669
+ },
670
+ {
671
+ "epoch": 2.29,
672
+ "learning_rate": 0.0002653279160170696,
673
+ "loss": 0.8095,
674
+ "step": 26500
675
+ },
676
+ {
677
+ "epoch": 2.31,
678
+ "learning_rate": 0.00026466874414886847,
679
+ "loss": 0.8077,
680
+ "step": 26750
681
+ },
682
+ {
683
+ "epoch": 2.33,
684
+ "learning_rate": 0.00026400420050664024,
685
+ "loss": 0.8086,
686
+ "step": 27000
687
+ },
688
+ {
689
+ "epoch": 2.35,
690
+ "learning_rate": 0.0002633343162216082,
691
+ "loss": 0.8057,
692
+ "step": 27250
693
+ },
694
+ {
695
+ "epoch": 2.37,
696
+ "learning_rate": 0.0002626591226751835,
697
+ "loss": 0.8071,
698
+ "step": 27500
699
+ },
700
+ {
701
+ "epoch": 2.39,
702
+ "learning_rate": 0.00026197865149749495,
703
+ "loss": 0.8034,
704
+ "step": 27750
705
+ },
706
+ {
707
+ "epoch": 2.42,
708
+ "learning_rate": 0.00026129293456590756,
709
+ "loss": 0.8029,
710
+ "step": 28000
711
+ },
712
+ {
713
+ "epoch": 2.44,
714
+ "learning_rate": 0.00026060200400352876,
715
+ "loss": 0.8041,
716
+ "step": 28250
717
+ },
718
+ {
719
+ "epoch": 2.46,
720
+ "learning_rate": 0.00025990589217770406,
721
+ "loss": 0.8028,
722
+ "step": 28500
723
+ },
724
+ {
725
+ "epoch": 2.48,
726
+ "learning_rate": 0.00025920463169850033,
727
+ "loss": 0.8009,
728
+ "step": 28750
729
+ },
730
+ {
731
+ "epoch": 2.5,
732
+ "learning_rate": 0.00025849825541717853,
733
+ "loss": 0.798,
734
+ "step": 29000
735
+ },
736
+ {
737
+ "epoch": 2.52,
738
+ "learning_rate": 0.0002577867964246545,
739
+ "loss": 0.7971,
740
+ "step": 29250
741
+ },
742
+ {
743
+ "epoch": 2.55,
744
+ "learning_rate": 0.0002570702880499488,
745
+ "loss": 0.7977,
746
+ "step": 29500
747
+ },
748
+ {
749
+ "epoch": 2.57,
750
+ "learning_rate": 0.00025634876385862554,
751
+ "loss": 0.7952,
752
+ "step": 29750
753
+ },
754
+ {
755
+ "epoch": 2.59,
756
+ "learning_rate": 0.00025562225765121975,
757
+ "loss": 0.7944,
758
+ "step": 30000
759
+ },
760
+ {
761
+ "epoch": 2.61,
762
+ "learning_rate": 0.00025489080346165416,
763
+ "loss": 0.7941,
764
+ "step": 30250
765
+ },
766
+ {
767
+ "epoch": 2.63,
768
+ "learning_rate": 0.0002541544355556448,
769
+ "loss": 0.7902,
770
+ "step": 30500
771
+ },
772
+ {
773
+ "epoch": 2.65,
774
+ "learning_rate": 0.0002534131884290956,
775
+ "loss": 0.79,
776
+ "step": 30750
777
+ },
778
+ {
779
+ "epoch": 2.67,
780
+ "learning_rate": 0.0002526670968064827,
781
+ "loss": 0.7896,
782
+ "step": 31000
783
+ },
784
+ {
785
+ "epoch": 2.7,
786
+ "learning_rate": 0.00025191619563922745,
787
+ "loss": 0.7877,
788
+ "step": 31250
789
+ },
790
+ {
791
+ "epoch": 2.72,
792
+ "learning_rate": 0.0002511605201040593,
793
+ "loss": 0.7864,
794
+ "step": 31500
795
+ },
796
+ {
797
+ "epoch": 2.74,
798
+ "learning_rate": 0.0002504001056013679,
799
+ "loss": 0.7874,
800
+ "step": 31750
801
+ },
802
+ {
803
+ "epoch": 2.76,
804
+ "learning_rate": 0.0002496349877535445,
805
+ "loss": 0.7848,
806
+ "step": 32000
807
+ },
808
+ {
809
+ "epoch": 2.78,
810
+ "learning_rate": 0.0002488652024033133,
811
+ "loss": 0.7828,
812
+ "step": 32250
813
+ },
814
+ {
815
+ "epoch": 2.8,
816
+ "learning_rate": 0.00024809078561205264,
817
+ "loss": 0.7825,
818
+ "step": 32500
819
+ },
820
+ {
821
+ "epoch": 2.83,
822
+ "learning_rate": 0.00024731177365810525,
823
+ "loss": 0.7805,
824
+ "step": 32750
825
+ },
826
+ {
827
+ "epoch": 2.85,
828
+ "learning_rate": 0.00024652820303507897,
829
+ "loss": 0.7791,
830
+ "step": 33000
831
+ },
832
+ {
833
+ "epoch": 2.87,
834
+ "learning_rate": 0.0002457401104501371,
835
+ "loss": 0.7783,
836
+ "step": 33250
837
+ },
838
+ {
839
+ "epoch": 2.89,
840
+ "learning_rate": 0.00024494753282227894,
841
+ "loss": 0.7774,
842
+ "step": 33500
843
+ },
844
+ {
845
+ "epoch": 2.91,
846
+ "learning_rate": 0.0002441505072806101,
847
+ "loss": 0.7764,
848
+ "step": 33750
849
+ },
850
+ {
851
+ "epoch": 2.93,
852
+ "learning_rate": 0.00024334907116260328,
853
+ "loss": 0.7753,
854
+ "step": 34000
855
+ },
856
+ {
857
+ "epoch": 2.95,
858
+ "learning_rate": 0.0002425432620123492,
859
+ "loss": 0.7734,
860
+ "step": 34250
861
+ },
862
+ {
863
+ "epoch": 2.98,
864
+ "learning_rate": 0.0002417331175787975,
865
+ "loss": 0.7729,
866
+ "step": 34500
867
+ },
868
+ {
869
+ "epoch": 3.0,
870
+ "learning_rate": 0.00024091867581398877,
871
+ "loss": 0.7719,
872
+ "step": 34750
873
+ },
874
+ {
875
+ "epoch": 3.0,
876
+ "eval_alliteration_score": 0.43172298240714885,
877
+ "eval_harmonic_meter_score": 0.208647467127026,
878
+ "eval_harmonic_rhyme_score": 0.4413259788497785,
879
+ "eval_meter_score": 0.457874360800918,
880
+ "eval_rhyme_score": 0.8102502944648488,
881
+ "eval_runtime": 2323.5118,
882
+ "eval_samples_per_second": 1.162,
883
+ "eval_steps_per_second": 0.145,
884
+ "step": 34773
885
+ },
886
+ {
887
+ "epoch": 3.02,
888
+ "learning_rate": 0.00024009997487127647,
889
+ "loss": 0.7391,
890
+ "step": 35000
891
+ },
892
+ {
893
+ "epoch": 3.04,
894
+ "learning_rate": 0.00023927705310353945,
895
+ "loss": 0.7344,
896
+ "step": 35250
897
+ },
898
+ {
899
+ "epoch": 3.06,
900
+ "learning_rate": 0.00023844994906138548,
901
+ "loss": 0.7352,
902
+ "step": 35500
903
+ },
904
+ {
905
+ "epoch": 3.08,
906
+ "learning_rate": 0.00023761870149134522,
907
+ "loss": 0.7349,
908
+ "step": 35750
909
+ },
910
+ {
911
+ "epoch": 3.11,
912
+ "learning_rate": 0.0002367833493340571,
913
+ "loss": 0.7381,
914
+ "step": 36000
915
+ },
916
+ {
917
+ "epoch": 3.13,
918
+ "learning_rate": 0.00023594393172244322,
919
+ "loss": 0.7372,
920
+ "step": 36250
921
+ },
922
+ {
923
+ "epoch": 3.15,
924
+ "learning_rate": 0.00023510048797987585,
925
+ "loss": 0.7338,
926
+ "step": 36500
927
+ },
928
+ {
929
+ "epoch": 3.17,
930
+ "learning_rate": 0.00023425305761833555,
931
+ "loss": 0.7338,
932
+ "step": 36750
933
+ },
934
+ {
935
+ "epoch": 3.19,
936
+ "learning_rate": 0.00023340168033656017,
937
+ "loss": 0.732,
938
+ "step": 37000
939
+ },
940
+ {
941
+ "epoch": 3.21,
942
+ "learning_rate": 0.0002325463960181849,
943
+ "loss": 0.7339,
944
+ "step": 37250
945
+ },
946
+ {
947
+ "epoch": 3.24,
948
+ "learning_rate": 0.00023168724472987423,
949
+ "loss": 0.7313,
950
+ "step": 37500
951
+ },
952
+ {
953
+ "epoch": 3.26,
954
+ "learning_rate": 0.00023082426671944457,
955
+ "loss": 0.7327,
956
+ "step": 37750
957
+ },
958
+ {
959
+ "epoch": 3.28,
960
+ "learning_rate": 0.00022995750241397924,
961
+ "loss": 0.7299,
962
+ "step": 38000
963
+ },
964
+ {
965
+ "epoch": 3.3,
966
+ "learning_rate": 0.0002290869924179342,
967
+ "loss": 0.7306,
968
+ "step": 38250
969
+ },
970
+ {
971
+ "epoch": 3.32,
972
+ "learning_rate": 0.00022821277751123635,
973
+ "loss": 0.7289,
974
+ "step": 38500
975
+ },
976
+ {
977
+ "epoch": 3.34,
978
+ "learning_rate": 0.0002273348986473726,
979
+ "loss": 0.7285,
980
+ "step": 38750
981
+ },
982
+ {
983
+ "epoch": 3.36,
984
+ "learning_rate": 0.00022645339695147193,
985
+ "loss": 0.7276,
986
+ "step": 39000
987
+ },
988
+ {
989
+ "epoch": 3.39,
990
+ "learning_rate": 0.0002255683137183783,
991
+ "loss": 0.7284,
992
+ "step": 39250
993
+ },
994
+ {
995
+ "epoch": 3.41,
996
+ "learning_rate": 0.0002246796904107168,
997
+ "loss": 0.7275,
998
+ "step": 39500
999
+ },
1000
+ {
1001
+ "epoch": 3.43,
1002
+ "learning_rate": 0.00022378756865695055,
1003
+ "loss": 0.7254,
1004
+ "step": 39750
1005
+ },
1006
+ {
1007
+ "epoch": 3.45,
1008
+ "learning_rate": 0.00022289199024943116,
1009
+ "loss": 0.7252,
1010
+ "step": 40000
1011
+ },
1012
+ {
1013
+ "epoch": 3.47,
1014
+ "learning_rate": 0.00022199299714244057,
1015
+ "loss": 0.7245,
1016
+ "step": 40250
1017
+ },
1018
+ {
1019
+ "epoch": 3.49,
1020
+ "learning_rate": 0.00022109063145022591,
1021
+ "loss": 0.7212,
1022
+ "step": 40500
1023
+ },
1024
+ {
1025
+ "epoch": 3.52,
1026
+ "learning_rate": 0.00022018493544502638,
1027
+ "loss": 0.7225,
1028
+ "step": 40750
1029
+ },
1030
+ {
1031
+ "epoch": 3.54,
1032
+ "learning_rate": 0.00021927595155509316,
1033
+ "loss": 0.7184,
1034
+ "step": 41000
1035
+ },
1036
+ {
1037
+ "epoch": 3.56,
1038
+ "learning_rate": 0.0002183637223627018,
1039
+ "loss": 0.72,
1040
+ "step": 41250
1041
+ },
1042
+ {
1043
+ "epoch": 3.58,
1044
+ "learning_rate": 0.00021744829060215716,
1045
+ "loss": 0.7186,
1046
+ "step": 41500
1047
+ },
1048
+ {
1049
+ "epoch": 3.6,
1050
+ "learning_rate": 0.00021652969915779183,
1051
+ "loss": 0.7181,
1052
+ "step": 41750
1053
+ },
1054
+ {
1055
+ "epoch": 3.62,
1056
+ "learning_rate": 0.0002156079910619569,
1057
+ "loss": 0.716,
1058
+ "step": 42000
1059
+ },
1060
+ {
1061
+ "epoch": 3.65,
1062
+ "learning_rate": 0.0002146832094930062,
1063
+ "loss": 0.7166,
1064
+ "step": 42250
1065
+ },
1066
+ {
1067
+ "epoch": 3.67,
1068
+ "learning_rate": 0.00021375539777327365,
1069
+ "loss": 0.7153,
1070
+ "step": 42500
1071
+ },
1072
+ {
1073
+ "epoch": 3.69,
1074
+ "learning_rate": 0.0002128245993670435,
1075
+ "loss": 0.7134,
1076
+ "step": 42750
1077
+ },
1078
+ {
1079
+ "epoch": 3.71,
1080
+ "learning_rate": 0.00021189085787851446,
1081
+ "loss": 0.713,
1082
+ "step": 43000
1083
+ },
1084
+ {
1085
+ "epoch": 3.73,
1086
+ "learning_rate": 0.00021095421704975687,
1087
+ "loss": 0.7101,
1088
+ "step": 43250
1089
+ },
1090
+ {
1091
+ "epoch": 3.75,
1092
+ "learning_rate": 0.00021001472075866362,
1093
+ "loss": 0.7097,
1094
+ "step": 43500
1095
+ },
1096
+ {
1097
+ "epoch": 3.77,
1098
+ "learning_rate": 0.00020907241301689473,
1099
+ "loss": 0.7093,
1100
+ "step": 43750
1101
+ },
1102
+ {
1103
+ "epoch": 3.8,
1104
+ "learning_rate": 0.00020812733796781542,
1105
+ "loss": 0.7072,
1106
+ "step": 44000
1107
+ },
1108
+ {
1109
+ "epoch": 3.82,
1110
+ "learning_rate": 0.00020717953988442822,
1111
+ "loss": 0.7073,
1112
+ "step": 44250
1113
+ },
1114
+ {
1115
+ "epoch": 3.84,
1116
+ "learning_rate": 0.000206229063167299,
1117
+ "loss": 0.7061,
1118
+ "step": 44500
1119
+ },
1120
+ {
1121
+ "epoch": 3.86,
1122
+ "learning_rate": 0.000205275952342477,
1123
+ "loss": 0.7021,
1124
+ "step": 44750
1125
+ },
1126
+ {
1127
+ "epoch": 3.88,
1128
+ "learning_rate": 0.00020432025205940874,
1129
+ "loss": 0.7021,
1130
+ "step": 45000
1131
+ },
1132
+ {
1133
+ "epoch": 3.9,
1134
+ "learning_rate": 0.00020336200708884674,
1135
+ "loss": 0.7005,
1136
+ "step": 45250
1137
+ },
1138
+ {
1139
+ "epoch": 3.93,
1140
+ "learning_rate": 0.00020240126232075196,
1141
+ "loss": 0.7011,
1142
+ "step": 45500
1143
+ },
1144
+ {
1145
+ "epoch": 3.95,
1146
+ "learning_rate": 0.0002014380627621909,
1147
+ "loss": 0.699,
1148
+ "step": 45750
1149
+ },
1150
+ {
1151
+ "epoch": 3.97,
1152
+ "learning_rate": 0.00020047245353522713,
1153
+ "loss": 0.6975,
1154
+ "step": 46000
1155
+ },
1156
+ {
1157
+ "epoch": 3.99,
1158
+ "learning_rate": 0.0001995044798748077,
1159
+ "loss": 0.6964,
1160
+ "step": 46250
1161
+ },
1162
+ {
1163
+ "epoch": 4.0,
1164
+ "eval_alliteration_score": 0.45886250690226393,
1165
+ "eval_harmonic_meter_score": 0.20919003027562127,
1166
+ "eval_harmonic_rhyme_score": 0.4443731034149568,
1167
+ "eval_meter_score": 0.4682188641643268,
1168
+ "eval_rhyme_score": 0.8157541677970918,
1169
+ "eval_runtime": 2407.9845,
1170
+ "eval_samples_per_second": 1.121,
1171
+ "eval_steps_per_second": 0.14,
1172
+ "step": 46364
1173
+ },
1174
+ {
1175
+ "epoch": 4.01,
1176
+ "learning_rate": 0.0001985341871266439,
1177
+ "loss": 0.6746,
1178
+ "step": 46500
1179
+ },
1180
+ {
1181
+ "epoch": 4.03,
1182
+ "learning_rate": 0.00019756162074508707,
1183
+ "loss": 0.6539,
1184
+ "step": 46750
1185
+ },
1186
+ {
1187
+ "epoch": 4.05,
1188
+ "learning_rate": 0.00019658682629099928,
1189
+ "loss": 0.6592,
1190
+ "step": 47000
1191
+ },
1192
+ {
1193
+ "epoch": 4.08,
1194
+ "learning_rate": 0.00019560984942961876,
1195
+ "loss": 0.6597,
1196
+ "step": 47250
1197
+ },
1198
+ {
1199
+ "epoch": 4.1,
1200
+ "learning_rate": 0.00019463073592842096,
1201
+ "loss": 0.6588,
1202
+ "step": 47500
1203
+ },
1204
+ {
1205
+ "epoch": 4.12,
1206
+ "learning_rate": 0.0001936495316549744,
1207
+ "loss": 0.6581,
1208
+ "step": 47750
1209
+ },
1210
+ {
1211
+ "epoch": 4.14,
1212
+ "learning_rate": 0.00019266628257479183,
1213
+ "loss": 0.6601,
1214
+ "step": 48000
1215
+ },
1216
+ {
1217
+ "epoch": 4.16,
1218
+ "learning_rate": 0.0001916810347491772,
1219
+ "loss": 0.6575,
1220
+ "step": 48250
1221
+ },
1222
+ {
1223
+ "epoch": 4.18,
1224
+ "learning_rate": 0.0001906938343330677,
1225
+ "loss": 0.6583,
1226
+ "step": 48500
1227
+ },
1228
+ {
1229
+ "epoch": 4.21,
1230
+ "learning_rate": 0.0001897047275728715,
1231
+ "loss": 0.6563,
1232
+ "step": 48750
1233
+ },
1234
+ {
1235
+ "epoch": 4.23,
1236
+ "learning_rate": 0.00018871376080430155,
1237
+ "loss": 0.6566,
1238
+ "step": 49000
1239
+ },
1240
+ {
1241
+ "epoch": 4.25,
1242
+ "learning_rate": 0.00018772098045020464,
1243
+ "loss": 0.6549,
1244
+ "step": 49250
1245
+ },
1246
+ {
1247
+ "epoch": 4.27,
1248
+ "learning_rate": 0.00018672643301838695,
1249
+ "loss": 0.6557,
1250
+ "step": 49500
1251
+ },
1252
+ {
1253
+ "epoch": 4.29,
1254
+ "learning_rate": 0.0001857301650994351,
1255
+ "loss": 0.656,
1256
+ "step": 49750
1257
+ },
1258
+ {
1259
+ "epoch": 4.31,
1260
+ "learning_rate": 0.00018473222336453377,
1261
+ "loss": 0.6535,
1262
+ "step": 50000
1263
+ },
1264
+ {
1265
+ "epoch": 4.34,
1266
+ "learning_rate": 0.00018373265456327918,
1267
+ "loss": 0.6536,
1268
+ "step": 50250
1269
+ },
1270
+ {
1271
+ "epoch": 4.36,
1272
+ "learning_rate": 0.00018273150552148925,
1273
+ "loss": 0.6525,
1274
+ "step": 50500
1275
+ },
1276
+ {
1277
+ "epoch": 4.38,
1278
+ "learning_rate": 0.00018172882313900965,
1279
+ "loss": 0.6541,
1280
+ "step": 50750
1281
+ },
1282
+ {
1283
+ "epoch": 4.4,
1284
+ "learning_rate": 0.0001807246543875172,
1285
+ "loss": 0.6521,
1286
+ "step": 51000
1287
+ },
1288
+ {
1289
+ "epoch": 4.42,
1290
+ "learning_rate": 0.00017971904630831908,
1291
+ "loss": 0.651,
1292
+ "step": 51250
1293
+ },
1294
+ {
1295
+ "epoch": 4.44,
1296
+ "learning_rate": 0.0001787120460101492,
1297
+ "loss": 0.6511,
1298
+ "step": 51500
1299
+ },
1300
+ {
1301
+ "epoch": 4.46,
1302
+ "learning_rate": 0.0001777037006669615,
1303
+ "loss": 0.6472,
1304
+ "step": 51750
1305
+ },
1306
+ {
1307
+ "epoch": 4.49,
1308
+ "learning_rate": 0.00017669405751571975,
1309
+ "loss": 0.646,
1310
+ "step": 52000
1311
+ },
1312
+ {
1313
+ "epoch": 4.51,
1314
+ "learning_rate": 0.00017568316385418507,
1315
+ "loss": 0.6476,
1316
+ "step": 52250
1317
+ },
1318
+ {
1319
+ "epoch": 4.53,
1320
+ "learning_rate": 0.0001746710670386998,
1321
+ "loss": 0.6465,
1322
+ "step": 52500
1323
+ },
1324
+ {
1325
+ "epoch": 4.55,
1326
+ "learning_rate": 0.0001736578144819695,
1327
+ "loss": 0.6447,
1328
+ "step": 52750
1329
+ },
1330
+ {
1331
+ "epoch": 4.57,
1332
+ "learning_rate": 0.00017264345365084128,
1333
+ "loss": 0.6471,
1334
+ "step": 53000
1335
+ },
1336
+ {
1337
+ "epoch": 4.59,
1338
+ "learning_rate": 0.0001716280320640808,
1339
+ "loss": 0.6429,
1340
+ "step": 53250
1341
+ },
1342
+ {
1343
+ "epoch": 4.62,
1344
+ "learning_rate": 0.00017061159729014568,
1345
+ "loss": 0.6396,
1346
+ "step": 53500
1347
+ },
1348
+ {
1349
+ "epoch": 4.64,
1350
+ "learning_rate": 0.00016959419694495754,
1351
+ "loss": 0.6434,
1352
+ "step": 53750
1353
+ },
1354
+ {
1355
+ "epoch": 4.66,
1356
+ "learning_rate": 0.00016857587868967087,
1357
+ "loss": 0.6406,
1358
+ "step": 54000
1359
+ },
1360
+ {
1361
+ "epoch": 4.68,
1362
+ "learning_rate": 0.00016755669022844086,
1363
+ "loss": 0.6382,
1364
+ "step": 54250
1365
+ },
1366
+ {
1367
+ "epoch": 4.7,
1368
+ "learning_rate": 0.00016653667930618827,
1369
+ "loss": 0.639,
1370
+ "step": 54500
1371
+ },
1372
+ {
1373
+ "epoch": 4.72,
1374
+ "learning_rate": 0.00016551589370636293,
1375
+ "loss": 0.6376,
1376
+ "step": 54750
1377
+ },
1378
+ {
1379
+ "epoch": 4.75,
1380
+ "learning_rate": 0.00016449438124870516,
1381
+ "loss": 0.637,
1382
+ "step": 55000
1383
+ },
1384
+ {
1385
+ "epoch": 4.77,
1386
+ "learning_rate": 0.0001634721897870057,
1387
+ "loss": 0.6345,
1388
+ "step": 55250
1389
+ },
1390
+ {
1391
+ "epoch": 4.79,
1392
+ "learning_rate": 0.00016244936720686393,
1393
+ "loss": 0.6334,
1394
+ "step": 55500
1395
+ },
1396
+ {
1397
+ "epoch": 4.81,
1398
+ "learning_rate": 0.00016142596142344466,
1399
+ "loss": 0.6325,
1400
+ "step": 55750
1401
+ },
1402
+ {
1403
+ "epoch": 4.83,
1404
+ "learning_rate": 0.00016040202037923343,
1405
+ "loss": 0.6303,
1406
+ "step": 56000
1407
+ },
1408
+ {
1409
+ "epoch": 4.85,
1410
+ "learning_rate": 0.00015937759204179055,
1411
+ "loss": 0.6301,
1412
+ "step": 56250
1413
+ },
1414
+ {
1415
+ "epoch": 4.87,
1416
+ "learning_rate": 0.0001583527244015042,
1417
+ "loss": 0.6291,
1418
+ "step": 56500
1419
+ },
1420
+ {
1421
+ "epoch": 4.9,
1422
+ "learning_rate": 0.000157327465469342,
1423
+ "loss": 0.6279,
1424
+ "step": 56750
1425
+ },
1426
+ {
1427
+ "epoch": 4.92,
1428
+ "learning_rate": 0.00015630186327460213,
1429
+ "loss": 0.626,
1430
+ "step": 57000
1431
+ },
1432
+ {
1433
+ "epoch": 4.94,
1434
+ "learning_rate": 0.00015527596586266328,
1435
+ "loss": 0.6246,
1436
+ "step": 57250
1437
+ },
1438
+ {
1439
+ "epoch": 4.96,
1440
+ "learning_rate": 0.00015424982129273382,
1441
+ "loss": 0.6256,
1442
+ "step": 57500
1443
+ },
1444
+ {
1445
+ "epoch": 4.98,
1446
+ "learning_rate": 0.00015322347763560055,
1447
+ "loss": 0.6247,
1448
+ "step": 57750
1449
+ },
1450
+ {
1451
+ "epoch": 5.0,
1452
+ "eval_alliteration_score": 0.4378994165045846,
1453
+ "eval_harmonic_meter_score": 0.20452236335398252,
1454
+ "eval_harmonic_rhyme_score": 0.4254891583797342,
1455
+ "eval_meter_score": 0.46303572443803803,
1456
+ "eval_rhyme_score": 0.8150200876525034,
1457
+ "eval_runtime": 2319.352,
1458
+ "eval_samples_per_second": 1.164,
1459
+ "eval_steps_per_second": 0.146,
1460
+ "step": 57955
1461
+ },
1462
+ {
1463
+ "epoch": 5.0,
1464
+ "learning_rate": 0.00015219698297137664,
1465
+ "loss": 0.6162,
1466
+ "step": 58000
1467
+ },
1468
+ {
1469
+ "epoch": 5.03,
1470
+ "learning_rate": 0.00015117038538724938,
1471
+ "loss": 0.5811,
1472
+ "step": 58250
1473
+ },
1474
+ {
1475
+ "epoch": 5.05,
1476
+ "learning_rate": 0.00015014373297522746,
1477
+ "loss": 0.5826,
1478
+ "step": 58500
1479
+ },
1480
+ {
1481
+ "epoch": 5.07,
1482
+ "learning_rate": 0.00014911707382988802,
1483
+ "loss": 0.5835,
1484
+ "step": 58750
1485
+ },
1486
+ {
1487
+ "epoch": 5.09,
1488
+ "learning_rate": 0.00014809045604612363,
1489
+ "loss": 0.585,
1490
+ "step": 59000
1491
+ },
1492
+ {
1493
+ "epoch": 5.11,
1494
+ "learning_rate": 0.00014706392771688935,
1495
+ "loss": 0.5839,
1496
+ "step": 59250
1497
+ },
1498
+ {
1499
+ "epoch": 5.13,
1500
+ "learning_rate": 0.00014603753693094937,
1501
+ "loss": 0.5838,
1502
+ "step": 59500
1503
+ },
1504
+ {
1505
+ "epoch": 5.15,
1506
+ "learning_rate": 0.00014501133177062474,
1507
+ "loss": 0.5836,
1508
+ "step": 59750
1509
+ },
1510
+ {
1511
+ "epoch": 5.18,
1512
+ "learning_rate": 0.00014398536030954076,
1513
+ "loss": 0.5828,
1514
+ "step": 60000
1515
+ },
1516
+ {
1517
+ "epoch": 5.2,
1518
+ "learning_rate": 0.0001429596706103746,
1519
+ "loss": 0.5842,
1520
+ "step": 60250
1521
+ },
1522
+ {
1523
+ "epoch": 5.22,
1524
+ "learning_rate": 0.0001419343107226042,
1525
+ "loss": 0.5839,
1526
+ "step": 60500
1527
+ },
1528
+ {
1529
+ "epoch": 5.24,
1530
+ "learning_rate": 0.00014090932868025707,
1531
+ "loss": 0.5819,
1532
+ "step": 60750
1533
+ },
1534
+ {
1535
+ "epoch": 5.26,
1536
+ "learning_rate": 0.00013988477249966025,
1537
+ "loss": 0.5839,
1538
+ "step": 61000
1539
+ },
1540
+ {
1541
+ "epoch": 5.28,
1542
+ "learning_rate": 0.0001388606901771907,
1543
+ "loss": 0.5818,
1544
+ "step": 61250
1545
+ },
1546
+ {
1547
+ "epoch": 5.31,
1548
+ "learning_rate": 0.00013783712968702718,
1549
+ "loss": 0.5816,
1550
+ "step": 61500
1551
+ },
1552
+ {
1553
+ "epoch": 5.33,
1554
+ "learning_rate": 0.00013681413897890257,
1555
+ "loss": 0.5811,
1556
+ "step": 61750
1557
+ },
1558
+ {
1559
+ "epoch": 5.35,
1560
+ "learning_rate": 0.00013579176597585784,
1561
+ "loss": 0.5789,
1562
+ "step": 62000
1563
+ },
1564
+ {
1565
+ "epoch": 5.37,
1566
+ "learning_rate": 0.0001347700585719968,
1567
+ "loss": 0.5798,
1568
+ "step": 62250
1569
+ },
1570
+ {
1571
+ "epoch": 5.39,
1572
+ "learning_rate": 0.0001337490646302426,
1573
+ "loss": 0.577,
1574
+ "step": 62500
1575
+ },
1576
+ {
1577
+ "epoch": 5.41,
1578
+ "learning_rate": 0.00013272883198009572,
1579
+ "loss": 0.5776,
1580
+ "step": 62750
1581
+ },
1582
+ {
1583
+ "epoch": 5.44,
1584
+ "learning_rate": 0.00013170940841539296,
1585
+ "loss": 0.5753,
1586
+ "step": 63000
1587
+ },
1588
+ {
1589
+ "epoch": 5.46,
1590
+ "learning_rate": 0.00013069084169206873,
1591
+ "loss": 0.5761,
1592
+ "step": 63250
1593
+ },
1594
+ {
1595
+ "epoch": 5.48,
1596
+ "learning_rate": 0.00012967317952591788,
1597
+ "loss": 0.577,
1598
+ "step": 63500
1599
+ },
1600
+ {
1601
+ "epoch": 5.5,
1602
+ "learning_rate": 0.00012865646959036038,
1603
+ "loss": 0.5739,
1604
+ "step": 63750
1605
+ },
1606
+ {
1607
+ "epoch": 5.52,
1608
+ "learning_rate": 0.00012764075951420792,
1609
+ "loss": 0.5737,
1610
+ "step": 64000
1611
+ },
1612
+ {
1613
+ "epoch": 5.54,
1614
+ "learning_rate": 0.00012662609687943277,
1615
+ "loss": 0.572,
1616
+ "step": 64250
1617
+ },
1618
+ {
1619
+ "epoch": 5.56,
1620
+ "learning_rate": 0.00012561252921893876,
1621
+ "loss": 0.5706,
1622
+ "step": 64500
1623
+ },
1624
+ {
1625
+ "epoch": 5.59,
1626
+ "learning_rate": 0.00012460010401433463,
1627
+ "loss": 0.5703,
1628
+ "step": 64750
1629
+ },
1630
+ {
1631
+ "epoch": 5.61,
1632
+ "learning_rate": 0.00012358886869370933,
1633
+ "loss": 0.5699,
1634
+ "step": 65000
1635
+ },
1636
+ {
1637
+ "epoch": 5.63,
1638
+ "learning_rate": 0.0001225788706294108,
1639
+ "loss": 0.5689,
1640
+ "step": 65250
1641
+ },
1642
+ {
1643
+ "epoch": 5.65,
1644
+ "learning_rate": 0.00012157015713582626,
1645
+ "loss": 0.568,
1646
+ "step": 65500
1647
+ },
1648
+ {
1649
+ "epoch": 5.67,
1650
+ "learning_rate": 0.00012056277546716603,
1651
+ "loss": 0.5662,
1652
+ "step": 65750
1653
+ },
1654
+ {
1655
+ "epoch": 5.69,
1656
+ "learning_rate": 0.0001195567728152496,
1657
+ "loss": 0.5642,
1658
+ "step": 66000
1659
+ },
1660
+ {
1661
+ "epoch": 5.72,
1662
+ "learning_rate": 0.0001185521963072951,
1663
+ "loss": 0.5638,
1664
+ "step": 66250
1665
+ },
1666
+ {
1667
+ "epoch": 5.74,
1668
+ "learning_rate": 0.00011754909300371153,
1669
+ "loss": 0.5639,
1670
+ "step": 66500
1671
+ },
1672
+ {
1673
+ "epoch": 5.76,
1674
+ "learning_rate": 0.0001165475098958941,
1675
+ "loss": 0.5617,
1676
+ "step": 66750
1677
+ },
1678
+ {
1679
+ "epoch": 5.78,
1680
+ "learning_rate": 0.00011554749390402273,
1681
+ "loss": 0.5592,
1682
+ "step": 67000
1683
+ },
1684
+ {
1685
+ "epoch": 5.8,
1686
+ "learning_rate": 0.0001145490918748645,
1687
+ "loss": 0.5616,
1688
+ "step": 67250
1689
+ },
1690
+ {
1691
+ "epoch": 5.82,
1692
+ "learning_rate": 0.00011355235057957859,
1693
+ "loss": 0.5614,
1694
+ "step": 67500
1695
+ },
1696
+ {
1697
+ "epoch": 5.85,
1698
+ "learning_rate": 0.00011255731671152548,
1699
+ "loss": 0.5582,
1700
+ "step": 67750
1701
+ },
1702
+ {
1703
+ "epoch": 5.87,
1704
+ "learning_rate": 0.00011156403688407935,
1705
+ "loss": 0.5568,
1706
+ "step": 68000
1707
+ },
1708
+ {
1709
+ "epoch": 5.89,
1710
+ "learning_rate": 0.0001105725576284447,
1711
+ "loss": 0.556,
1712
+ "step": 68250
1713
+ },
1714
+ {
1715
+ "epoch": 5.91,
1716
+ "learning_rate": 0.00010958292539147633,
1717
+ "loss": 0.5538,
1718
+ "step": 68500
1719
+ },
1720
+ {
1721
+ "epoch": 5.93,
1722
+ "learning_rate": 0.00010859518653350373,
1723
+ "loss": 0.5535,
1724
+ "step": 68750
1725
+ },
1726
+ {
1727
+ "epoch": 5.95,
1728
+ "learning_rate": 0.00010760938732615886,
1729
+ "loss": 0.5536,
1730
+ "step": 69000
1731
+ },
1732
+ {
1733
+ "epoch": 5.97,
1734
+ "learning_rate": 0.00010662557395020896,
1735
+ "loss": 0.5536,
1736
+ "step": 69250
1737
+ },
1738
+ {
1739
+ "epoch": 6.0,
1740
+ "learning_rate": 0.00010564379249339304,
1741
+ "loss": 0.551,
1742
+ "step": 69500
1743
+ },
1744
+ {
1745
+ "epoch": 6.0,
1746
+ "eval_alliteration_score": 0.4355458036204059,
1747
+ "eval_harmonic_meter_score": 0.184149411235037,
1748
+ "eval_harmonic_rhyme_score": 0.44674866827401805,
1749
+ "eval_meter_score": 0.44256677255617083,
1750
+ "eval_rhyme_score": 0.8221214981988054,
1751
+ "eval_runtime": 2288.5532,
1752
+ "eval_samples_per_second": 1.18,
1753
+ "eval_steps_per_second": 0.148,
1754
+ "step": 69546
1755
+ },
1756
+ {
1757
+ "epoch": 6.02,
1758
+ "learning_rate": 0.00010466408894826249,
1759
+ "loss": 0.5183,
1760
+ "step": 69750
1761
+ },
1762
+ {
1763
+ "epoch": 6.04,
1764
+ "learning_rate": 0.0001036865092100271,
1765
+ "loss": 0.5118,
1766
+ "step": 70000
1767
+ },
1768
+ {
1769
+ "epoch": 6.06,
1770
+ "learning_rate": 0.00010271109907440465,
1771
+ "loss": 0.5148,
1772
+ "step": 70250
1773
+ },
1774
+ {
1775
+ "epoch": 6.08,
1776
+ "learning_rate": 0.00010173790423547578,
1777
+ "loss": 0.512,
1778
+ "step": 70500
1779
+ },
1780
+ {
1781
+ "epoch": 6.1,
1782
+ "learning_rate": 0.00010076697028354309,
1783
+ "loss": 0.5148,
1784
+ "step": 70750
1785
+ },
1786
+ {
1787
+ "epoch": 6.13,
1788
+ "learning_rate": 9.97983427029958e-05,
1789
+ "loss": 0.5154,
1790
+ "step": 71000
1791
+ },
1792
+ {
1793
+ "epoch": 6.15,
1794
+ "learning_rate": 9.88320668701788e-05,
1795
+ "loss": 0.5114,
1796
+ "step": 71250
1797
+ },
1798
+ {
1799
+ "epoch": 6.17,
1800
+ "learning_rate": 9.786818805126695e-05,
1801
+ "loss": 0.5137,
1802
+ "step": 71500
1803
+ },
1804
+ {
1805
+ "epoch": 6.19,
1806
+ "learning_rate": 9.690675140014445e-05,
1807
+ "loss": 0.5132,
1808
+ "step": 71750
1809
+ },
1810
+ {
1811
+ "epoch": 6.21,
1812
+ "learning_rate": 9.594780195628981e-05,
1813
+ "loss": 0.5121,
1814
+ "step": 72000
1815
+ },
1816
+ {
1817
+ "epoch": 6.23,
1818
+ "learning_rate": 9.499138464266571e-05,
1819
+ "loss": 0.5125,
1820
+ "step": 72250
1821
+ },
1822
+ {
1823
+ "epoch": 6.25,
1824
+ "learning_rate": 9.403754426361472e-05,
1825
+ "loss": 0.5128,
1826
+ "step": 72500
1827
+ },
1828
+ {
1829
+ "epoch": 6.28,
1830
+ "learning_rate": 9.308632550276002e-05,
1831
+ "loss": 0.5132,
1832
+ "step": 72750
1833
+ },
1834
+ {
1835
+ "epoch": 6.3,
1836
+ "learning_rate": 9.213777292091273e-05,
1837
+ "loss": 0.513,
1838
+ "step": 73000
1839
+ },
1840
+ {
1841
+ "epoch": 6.32,
1842
+ "learning_rate": 9.119193095398385e-05,
1843
+ "loss": 0.5114,
1844
+ "step": 73250
1845
+ },
1846
+ {
1847
+ "epoch": 6.34,
1848
+ "learning_rate": 9.024884391090313e-05,
1849
+ "loss": 0.5095,
1850
+ "step": 73500
1851
+ },
1852
+ {
1853
+ "epoch": 6.36,
1854
+ "learning_rate": 8.930855597154284e-05,
1855
+ "loss": 0.5069,
1856
+ "step": 73750
1857
+ },
1858
+ {
1859
+ "epoch": 6.38,
1860
+ "learning_rate": 8.837111118464855e-05,
1861
+ "loss": 0.5066,
1862
+ "step": 74000
1863
+ },
1864
+ {
1865
+ "epoch": 6.41,
1866
+ "learning_rate": 8.743655346577544e-05,
1867
+ "loss": 0.5082,
1868
+ "step": 74250
1869
+ },
1870
+ {
1871
+ "epoch": 6.43,
1872
+ "learning_rate": 8.650492659523093e-05,
1873
+ "loss": 0.5065,
1874
+ "step": 74500
1875
+ },
1876
+ {
1877
+ "epoch": 6.45,
1878
+ "learning_rate": 8.557627421602401e-05,
1879
+ "loss": 0.5043,
1880
+ "step": 74750
1881
+ },
1882
+ {
1883
+ "epoch": 6.47,
1884
+ "learning_rate": 8.465063983182055e-05,
1885
+ "loss": 0.5071,
1886
+ "step": 75000
1887
+ },
1888
+ {
1889
+ "epoch": 6.49,
1890
+ "learning_rate": 8.37280668049054e-05,
1891
+ "loss": 0.5069,
1892
+ "step": 75250
1893
+ },
1894
+ {
1895
+ "epoch": 6.51,
1896
+ "learning_rate": 8.280859835415095e-05,
1897
+ "loss": 0.5025,
1898
+ "step": 75500
1899
+ },
1900
+ {
1901
+ "epoch": 6.54,
1902
+ "learning_rate": 8.189227755299259e-05,
1903
+ "loss": 0.5021,
1904
+ "step": 75750
1905
+ },
1906
+ {
1907
+ "epoch": 6.56,
1908
+ "learning_rate": 8.097914732741086e-05,
1909
+ "loss": 0.5031,
1910
+ "step": 76000
1911
+ },
1912
+ {
1913
+ "epoch": 6.58,
1914
+ "learning_rate": 8.006925045392063e-05,
1915
+ "loss": 0.5002,
1916
+ "step": 76250
1917
+ },
1918
+ {
1919
+ "epoch": 6.6,
1920
+ "learning_rate": 7.916262955756686e-05,
1921
+ "loss": 0.5015,
1922
+ "step": 76500
1923
+ },
1924
+ {
1925
+ "epoch": 6.62,
1926
+ "learning_rate": 7.825932710992824e-05,
1927
+ "loss": 0.4984,
1928
+ "step": 76750
1929
+ },
1930
+ {
1931
+ "epoch": 6.64,
1932
+ "learning_rate": 7.735938542712732e-05,
1933
+ "loss": 0.5004,
1934
+ "step": 77000
1935
+ },
1936
+ {
1937
+ "epoch": 6.66,
1938
+ "learning_rate": 7.64628466678482e-05,
1939
+ "loss": 0.4982,
1940
+ "step": 77250
1941
+ },
1942
+ {
1943
+ "epoch": 6.69,
1944
+ "learning_rate": 7.556975283136148e-05,
1945
+ "loss": 0.4965,
1946
+ "step": 77500
1947
+ },
1948
+ {
1949
+ "epoch": 6.71,
1950
+ "learning_rate": 7.468014575555687e-05,
1951
+ "loss": 0.4987,
1952
+ "step": 77750
1953
+ },
1954
+ {
1955
+ "epoch": 6.73,
1956
+ "learning_rate": 7.379406711498332e-05,
1957
+ "loss": 0.4947,
1958
+ "step": 78000
1959
+ },
1960
+ {
1961
+ "epoch": 6.75,
1962
+ "learning_rate": 7.291155841889657e-05,
1963
+ "loss": 0.4918,
1964
+ "step": 78250
1965
+ },
1966
+ {
1967
+ "epoch": 6.77,
1968
+ "learning_rate": 7.203266100931468e-05,
1969
+ "loss": 0.4957,
1970
+ "step": 78500
1971
+ },
1972
+ {
1973
+ "epoch": 6.79,
1974
+ "learning_rate": 7.115741605908132e-05,
1975
+ "loss": 0.4927,
1976
+ "step": 78750
1977
+ },
1978
+ {
1979
+ "epoch": 6.82,
1980
+ "learning_rate": 7.028586456993707e-05,
1981
+ "loss": 0.4912,
1982
+ "step": 79000
1983
+ },
1984
+ {
1985
+ "epoch": 6.84,
1986
+ "learning_rate": 6.941804737059833e-05,
1987
+ "loss": 0.4907,
1988
+ "step": 79250
1989
+ },
1990
+ {
1991
+ "epoch": 6.86,
1992
+ "learning_rate": 6.855400511484516e-05,
1993
+ "loss": 0.4908,
1994
+ "step": 79500
1995
+ },
1996
+ {
1997
+ "epoch": 6.88,
1998
+ "learning_rate": 6.769377827961639e-05,
1999
+ "loss": 0.4881,
2000
+ "step": 79750
2001
+ },
2002
+ {
2003
+ "epoch": 6.9,
2004
+ "learning_rate": 6.683740716311371e-05,
2005
+ "loss": 0.4878,
2006
+ "step": 80000
2007
+ },
2008
+ {
2009
+ "epoch": 6.92,
2010
+ "learning_rate": 6.598493188291358e-05,
2011
+ "loss": 0.4861,
2012
+ "step": 80250
2013
+ },
2014
+ {
2015
+ "epoch": 6.95,
2016
+ "learning_rate": 6.513639237408814e-05,
2017
+ "loss": 0.4843,
2018
+ "step": 80500
2019
+ },
2020
+ {
2021
+ "epoch": 6.97,
2022
+ "learning_rate": 6.429182838733434e-05,
2023
+ "loss": 0.484,
2024
+ "step": 80750
2025
+ },
2026
+ {
2027
+ "epoch": 6.99,
2028
+ "learning_rate": 6.345127948711172e-05,
2029
+ "loss": 0.4821,
2030
+ "step": 81000
2031
+ },
2032
+ {
2033
+ "epoch": 7.0,
2034
+ "eval_alliteration_score": 0.4303659202621518,
2035
+ "eval_harmonic_meter_score": 0.1753914687780149,
2036
+ "eval_harmonic_rhyme_score": 0.43386940648753536,
2037
+ "eval_meter_score": 0.4408589702287603,
2038
+ "eval_rhyme_score": 0.815514426880565,
2039
+ "eval_runtime": 2426.169,
2040
+ "eval_samples_per_second": 1.113,
2041
+ "eval_steps_per_second": 0.139,
2042
+ "step": 81137
2043
+ },
2044
+ {
2045
+ "epoch": 7.01,
2046
+ "learning_rate": 6.261478504978898e-05,
2047
+ "loss": 0.4684,
2048
+ "step": 81250
2049
+ },
2050
+ {
2051
+ "epoch": 7.03,
2052
+ "learning_rate": 6.178238426179942e-05,
2053
+ "loss": 0.4529,
2054
+ "step": 81500
2055
+ },
2056
+ {
2057
+ "epoch": 7.05,
2058
+ "learning_rate": 6.0954116117805176e-05,
2059
+ "loss": 0.4513,
2060
+ "step": 81750
2061
+ },
2062
+ {
2063
+ "epoch": 7.07,
2064
+ "learning_rate": 6.0130019418870444e-05,
2065
+ "loss": 0.452,
2066
+ "step": 82000
2067
+ },
2068
+ {
2069
+ "epoch": 7.1,
2070
+ "learning_rate": 5.931013277064377e-05,
2071
+ "loss": 0.4517,
2072
+ "step": 82250
2073
+ },
2074
+ {
2075
+ "epoch": 7.12,
2076
+ "learning_rate": 5.8494494581549705e-05,
2077
+ "loss": 0.4501,
2078
+ "step": 82500
2079
+ },
2080
+ {
2081
+ "epoch": 7.14,
2082
+ "learning_rate": 5.7683143060989395e-05,
2083
+ "loss": 0.4495,
2084
+ "step": 82750
2085
+ },
2086
+ {
2087
+ "epoch": 7.16,
2088
+ "learning_rate": 5.6876116217550485e-05,
2089
+ "loss": 0.451,
2090
+ "step": 83000
2091
+ },
2092
+ {
2093
+ "epoch": 7.18,
2094
+ "learning_rate": 5.6073451857226904e-05,
2095
+ "loss": 0.4504,
2096
+ "step": 83250
2097
+ },
2098
+ {
2099
+ "epoch": 7.2,
2100
+ "learning_rate": 5.527518758164754e-05,
2101
+ "loss": 0.4509,
2102
+ "step": 83500
2103
+ },
2104
+ {
2105
+ "epoch": 7.23,
2106
+ "learning_rate": 5.4481360786314834e-05,
2107
+ "loss": 0.4507,
2108
+ "step": 83750
2109
+ },
2110
+ {
2111
+ "epoch": 7.25,
2112
+ "learning_rate": 5.369200865885296e-05,
2113
+ "loss": 0.4511,
2114
+ "step": 84000
2115
+ },
2116
+ {
2117
+ "epoch": 7.27,
2118
+ "learning_rate": 5.2907168177265706e-05,
2119
+ "loss": 0.4489,
2120
+ "step": 84250
2121
+ },
2122
+ {
2123
+ "epoch": 7.29,
2124
+ "learning_rate": 5.212687610820424e-05,
2125
+ "loss": 0.448,
2126
+ "step": 84500
2127
+ },
2128
+ {
2129
+ "epoch": 7.31,
2130
+ "learning_rate": 5.135116900524474e-05,
2131
+ "loss": 0.4468,
2132
+ "step": 84750
2133
+ },
2134
+ {
2135
+ "epoch": 7.33,
2136
+ "learning_rate": 5.058008320717585e-05,
2137
+ "loss": 0.4472,
2138
+ "step": 85000
2139
+ },
2140
+ {
2141
+ "epoch": 7.35,
2142
+ "learning_rate": 4.981365483629661e-05,
2143
+ "loss": 0.4503,
2144
+ "step": 85250
2145
+ },
2146
+ {
2147
+ "epoch": 7.38,
2148
+ "learning_rate": 4.905191979672413e-05,
2149
+ "loss": 0.4455,
2150
+ "step": 85500
2151
+ },
2152
+ {
2153
+ "epoch": 7.4,
2154
+ "learning_rate": 4.829491377271165e-05,
2155
+ "loss": 0.447,
2156
+ "step": 85750
2157
+ },
2158
+ {
2159
+ "epoch": 7.42,
2160
+ "learning_rate": 4.7542672226976754e-05,
2161
+ "loss": 0.4447,
2162
+ "step": 86000
2163
+ },
2164
+ {
2165
+ "epoch": 7.44,
2166
+ "learning_rate": 4.679523039904028e-05,
2167
+ "loss": 0.4465,
2168
+ "step": 86250
2169
+ },
2170
+ {
2171
+ "epoch": 7.46,
2172
+ "learning_rate": 4.605262330357542e-05,
2173
+ "loss": 0.4443,
2174
+ "step": 86500
2175
+ },
2176
+ {
2177
+ "epoch": 7.48,
2178
+ "learning_rate": 4.531488572876734e-05,
2179
+ "loss": 0.4434,
2180
+ "step": 86750
2181
+ },
2182
+ {
2183
+ "epoch": 7.51,
2184
+ "learning_rate": 4.458205223468363e-05,
2185
+ "loss": 0.4461,
2186
+ "step": 87000
2187
+ },
2188
+ {
2189
+ "epoch": 7.53,
2190
+ "learning_rate": 4.3854157151655144e-05,
2191
+ "loss": 0.4438,
2192
+ "step": 87250
2193
+ },
2194
+ {
2195
+ "epoch": 7.55,
2196
+ "learning_rate": 4.313123457866796e-05,
2197
+ "loss": 0.4409,
2198
+ "step": 87500
2199
+ },
2200
+ {
2201
+ "epoch": 7.57,
2202
+ "learning_rate": 4.241331838176564e-05,
2203
+ "loss": 0.4428,
2204
+ "step": 87750
2205
+ },
2206
+ {
2207
+ "epoch": 7.59,
2208
+ "learning_rate": 4.170044219246323e-05,
2209
+ "loss": 0.441,
2210
+ "step": 88000
2211
+ },
2212
+ {
2213
+ "epoch": 7.61,
2214
+ "learning_rate": 4.0992639406171303e-05,
2215
+ "loss": 0.4405,
2216
+ "step": 88250
2217
+ },
2218
+ {
2219
+ "epoch": 7.64,
2220
+ "learning_rate": 4.028994318063184e-05,
2221
+ "loss": 0.4412,
2222
+ "step": 88500
2223
+ },
2224
+ {
2225
+ "epoch": 7.66,
2226
+ "learning_rate": 3.9592386434364634e-05,
2227
+ "loss": 0.4378,
2228
+ "step": 88750
2229
+ },
2230
+ {
2231
+ "epoch": 7.68,
2232
+ "learning_rate": 3.890000184512546e-05,
2233
+ "loss": 0.4388,
2234
+ "step": 89000
2235
+ },
2236
+ {
2237
+ "epoch": 7.7,
2238
+ "learning_rate": 3.8212821848375086e-05,
2239
+ "loss": 0.4364,
2240
+ "step": 89250
2241
+ },
2242
+ {
2243
+ "epoch": 7.72,
2244
+ "learning_rate": 3.753087863575993e-05,
2245
+ "loss": 0.4395,
2246
+ "step": 89500
2247
+ },
2248
+ {
2249
+ "epoch": 7.74,
2250
+ "learning_rate": 3.6854204153603714e-05,
2251
+ "loss": 0.4353,
2252
+ "step": 89750
2253
+ },
2254
+ {
2255
+ "epoch": 7.76,
2256
+ "learning_rate": 3.6182830101411315e-05,
2257
+ "loss": 0.435,
2258
+ "step": 90000
2259
+ },
2260
+ {
2261
+ "epoch": 7.79,
2262
+ "learning_rate": 3.551678793038358e-05,
2263
+ "loss": 0.435,
2264
+ "step": 90250
2265
+ },
2266
+ {
2267
+ "epoch": 7.81,
2268
+ "learning_rate": 3.485610884194389e-05,
2269
+ "loss": 0.4359,
2270
+ "step": 90500
2271
+ },
2272
+ {
2273
+ "epoch": 7.83,
2274
+ "learning_rate": 3.420082378627649e-05,
2275
+ "loss": 0.4338,
2276
+ "step": 90750
2277
+ },
2278
+ {
2279
+ "epoch": 7.85,
2280
+ "learning_rate": 3.3550963460876776e-05,
2281
+ "loss": 0.4353,
2282
+ "step": 91000
2283
+ },
2284
+ {
2285
+ "epoch": 7.87,
2286
+ "learning_rate": 3.2906558309113134e-05,
2287
+ "loss": 0.4316,
2288
+ "step": 91250
2289
+ },
2290
+ {
2291
+ "epoch": 7.89,
2292
+ "learning_rate": 3.226763851880079e-05,
2293
+ "loss": 0.4325,
2294
+ "step": 91500
2295
+ },
2296
+ {
2297
+ "epoch": 7.92,
2298
+ "learning_rate": 3.163423402078752e-05,
2299
+ "loss": 0.4318,
2300
+ "step": 91750
2301
+ },
2302
+ {
2303
+ "epoch": 7.94,
2304
+ "learning_rate": 3.100637448755172e-05,
2305
+ "loss": 0.4313,
2306
+ "step": 92000
2307
+ },
2308
+ {
2309
+ "epoch": 7.96,
2310
+ "learning_rate": 3.0384089331812305e-05,
2311
+ "loss": 0.4301,
2312
+ "step": 92250
2313
+ },
2314
+ {
2315
+ "epoch": 7.98,
2316
+ "learning_rate": 2.976740770515067e-05,
2317
+ "loss": 0.4322,
2318
+ "step": 92500
2319
+ },
2320
+ {
2321
+ "epoch": 8.0,
2322
+ "eval_alliteration_score": 0.42516268980477223,
2323
+ "eval_harmonic_meter_score": 0.19115865795897424,
2324
+ "eval_harmonic_rhyme_score": 0.4073988829309175,
2325
+ "eval_meter_score": 0.4526465021452277,
2326
+ "eval_rhyme_score": 0.8084745294411801,
2327
+ "eval_runtime": 2422.0648,
2328
+ "eval_samples_per_second": 1.115,
2329
+ "eval_steps_per_second": 0.14,
2330
+ "step": 92728
2331
+ },
2332
+ {
2333
+ "epoch": 8.0,
2334
+ "learning_rate": 2.9156358496645276e-05,
2335
+ "loss": 0.4282,
2336
+ "step": 92750
2337
+ },
2338
+ {
2339
+ "epoch": 8.02,
2340
+ "learning_rate": 2.8550970331518202e-05,
2341
+ "loss": 0.4058,
2342
+ "step": 93000
2343
+ },
2344
+ {
2345
+ "epoch": 8.05,
2346
+ "learning_rate": 2.795127156979429e-05,
2347
+ "loss": 0.4054,
2348
+ "step": 93250
2349
+ },
2350
+ {
2351
+ "epoch": 8.07,
2352
+ "learning_rate": 2.735729030497229e-05,
2353
+ "loss": 0.406,
2354
+ "step": 93500
2355
+ },
2356
+ {
2357
+ "epoch": 8.09,
2358
+ "learning_rate": 2.676905436270911e-05,
2359
+ "loss": 0.4057,
2360
+ "step": 93750
2361
+ },
2362
+ {
2363
+ "epoch": 8.11,
2364
+ "learning_rate": 2.6186591299516186e-05,
2365
+ "loss": 0.4054,
2366
+ "step": 94000
2367
+ },
2368
+ {
2369
+ "epoch": 8.13,
2370
+ "learning_rate": 2.5609928401468555e-05,
2371
+ "loss": 0.4061,
2372
+ "step": 94250
2373
+ },
2374
+ {
2375
+ "epoch": 8.15,
2376
+ "learning_rate": 2.503909268292653e-05,
2377
+ "loss": 0.4045,
2378
+ "step": 94500
2379
+ },
2380
+ {
2381
+ "epoch": 8.17,
2382
+ "learning_rate": 2.4474110885270343e-05,
2383
+ "loss": 0.403,
2384
+ "step": 94750
2385
+ },
2386
+ {
2387
+ "epoch": 8.2,
2388
+ "learning_rate": 2.3915009475647362e-05,
2389
+ "loss": 0.4041,
2390
+ "step": 95000
2391
+ },
2392
+ {
2393
+ "epoch": 8.22,
2394
+ "learning_rate": 2.3361814645732168e-05,
2395
+ "loss": 0.4047,
2396
+ "step": 95250
2397
+ },
2398
+ {
2399
+ "epoch": 8.24,
2400
+ "learning_rate": 2.2814552310499563e-05,
2401
+ "loss": 0.4044,
2402
+ "step": 95500
2403
+ },
2404
+ {
2405
+ "epoch": 8.26,
2406
+ "learning_rate": 2.2273248107010633e-05,
2407
+ "loss": 0.4035,
2408
+ "step": 95750
2409
+ },
2410
+ {
2411
+ "epoch": 8.28,
2412
+ "learning_rate": 2.173792739321174e-05,
2413
+ "loss": 0.4022,
2414
+ "step": 96000
2415
+ },
2416
+ {
2417
+ "epoch": 8.3,
2418
+ "learning_rate": 2.1208615246746545e-05,
2419
+ "loss": 0.4034,
2420
+ "step": 96250
2421
+ },
2422
+ {
2423
+ "epoch": 8.33,
2424
+ "learning_rate": 2.0685336463781288e-05,
2425
+ "loss": 0.4029,
2426
+ "step": 96500
2427
+ },
2428
+ {
2429
+ "epoch": 8.35,
2430
+ "learning_rate": 2.016811555784314e-05,
2431
+ "loss": 0.4009,
2432
+ "step": 96750
2433
+ },
2434
+ {
2435
+ "epoch": 8.37,
2436
+ "learning_rate": 1.9656976758671865e-05,
2437
+ "loss": 0.402,
2438
+ "step": 97000
2439
+ },
2440
+ {
2441
+ "epoch": 8.39,
2442
+ "learning_rate": 1.9151944011084696e-05,
2443
+ "loss": 0.4025,
2444
+ "step": 97250
2445
+ },
2446
+ {
2447
+ "epoch": 8.41,
2448
+ "learning_rate": 1.865304097385472e-05,
2449
+ "loss": 0.4032,
2450
+ "step": 97500
2451
+ },
2452
+ {
2453
+ "epoch": 8.43,
2454
+ "learning_rate": 1.8160291018602512e-05,
2455
+ "loss": 0.4006,
2456
+ "step": 97750
2457
+ },
2458
+ {
2459
+ "epoch": 8.45,
2460
+ "learning_rate": 1.767371722870125e-05,
2461
+ "loss": 0.4006,
2462
+ "step": 98000
2463
+ },
2464
+ {
2465
+ "epoch": 8.48,
2466
+ "learning_rate": 1.7193342398195286e-05,
2467
+ "loss": 0.4017,
2468
+ "step": 98250
2469
+ },
2470
+ {
2471
+ "epoch": 8.5,
2472
+ "learning_rate": 1.67191890307325e-05,
2473
+ "loss": 0.4001,
2474
+ "step": 98500
2475
+ },
2476
+ {
2477
+ "epoch": 8.52,
2478
+ "learning_rate": 1.625127933850997e-05,
2479
+ "loss": 0.3977,
2480
+ "step": 98750
2481
+ },
2482
+ {
2483
+ "epoch": 8.54,
2484
+ "learning_rate": 1.5789635241233484e-05,
2485
+ "loss": 0.4004,
2486
+ "step": 99000
2487
+ },
2488
+ {
2489
+ "epoch": 8.56,
2490
+ "learning_rate": 1.5334278365090597e-05,
2491
+ "loss": 0.4011,
2492
+ "step": 99250
2493
+ },
2494
+ {
2495
+ "epoch": 8.58,
2496
+ "learning_rate": 1.488523004173765e-05,
2497
+ "loss": 0.3995,
2498
+ "step": 99500
2499
+ },
2500
+ {
2501
+ "epoch": 8.61,
2502
+ "learning_rate": 1.4442511307300353e-05,
2503
+ "loss": 0.4012,
2504
+ "step": 99750
2505
+ },
2506
+ {
2507
+ "epoch": 8.63,
2508
+ "learning_rate": 1.4006142901388467e-05,
2509
+ "loss": 0.3993,
2510
+ "step": 100000
2511
+ },
2512
+ {
2513
+ "epoch": 8.65,
2514
+ "learning_rate": 1.3576145266124039e-05,
2515
+ "loss": 0.3996,
2516
+ "step": 100250
2517
+ },
2518
+ {
2519
+ "epoch": 8.67,
2520
+ "learning_rate": 1.3152538545183928e-05,
2521
+ "loss": 0.3973,
2522
+ "step": 100500
2523
+ },
2524
+ {
2525
+ "epoch": 8.69,
2526
+ "learning_rate": 1.2735342582856161e-05,
2527
+ "loss": 0.397,
2528
+ "step": 100750
2529
+ },
2530
+ {
2531
+ "epoch": 8.71,
2532
+ "learning_rate": 1.2324576923110202e-05,
2533
+ "loss": 0.3989,
2534
+ "step": 101000
2535
+ },
2536
+ {
2537
+ "epoch": 8.74,
2538
+ "learning_rate": 1.1920260808681403e-05,
2539
+ "loss": 0.3979,
2540
+ "step": 101250
2541
+ },
2542
+ {
2543
+ "epoch": 8.76,
2544
+ "learning_rate": 1.152241318016967e-05,
2545
+ "loss": 0.3945,
2546
+ "step": 101500
2547
+ },
2548
+ {
2549
+ "epoch": 8.78,
2550
+ "learning_rate": 1.1131052675152103e-05,
2551
+ "loss": 0.3966,
2552
+ "step": 101750
2553
+ },
2554
+ {
2555
+ "epoch": 8.8,
2556
+ "learning_rate": 1.0746197627309855e-05,
2557
+ "loss": 0.3985,
2558
+ "step": 102000
2559
+ },
2560
+ {
2561
+ "epoch": 8.82,
2562
+ "learning_rate": 1.0367866065569347e-05,
2563
+ "loss": 0.3964,
2564
+ "step": 102250
2565
+ },
2566
+ {
2567
+ "epoch": 8.84,
2568
+ "learning_rate": 9.996075713257652e-06,
2569
+ "loss": 0.3948,
2570
+ "step": 102500
2571
+ },
2572
+ {
2573
+ "epoch": 8.86,
2574
+ "learning_rate": 9.630843987272257e-06,
2575
+ "loss": 0.3952,
2576
+ "step": 102750
2577
+ },
2578
+ {
2579
+ "epoch": 8.89,
2580
+ "learning_rate": 9.272187997265013e-06,
2581
+ "loss": 0.3948,
2582
+ "step": 103000
2583
+ },
2584
+ {
2585
+ "epoch": 8.91,
2586
+ "learning_rate": 8.920124544840857e-06,
2587
+ "loss": 0.3965,
2588
+ "step": 103250
2589
+ },
2590
+ {
2591
+ "epoch": 8.93,
2592
+ "learning_rate": 8.574670122770521e-06,
2593
+ "loss": 0.3968,
2594
+ "step": 103500
2595
+ },
2596
+ {
2597
+ "epoch": 8.95,
2598
+ "learning_rate": 8.23584091421805e-06,
2599
+ "loss": 0.3945,
2600
+ "step": 103750
2601
+ },
2602
+ {
2603
+ "epoch": 8.97,
2604
+ "learning_rate": 7.903652791982501e-06,
2605
+ "loss": 0.3941,
2606
+ "step": 104000
2607
+ },
2608
+ {
2609
+ "epoch": 8.99,
2610
+ "learning_rate": 7.578121317754593e-06,
2611
+ "loss": 0.3928,
2612
+ "step": 104250
2613
+ },
2614
+ {
2615
+ "epoch": 9.0,
2616
+ "eval_alliteration_score": 0.4286099865047234,
2617
+ "eval_harmonic_meter_score": 0.1830439988800133,
2618
+ "eval_harmonic_rhyme_score": 0.3932058401427829,
2619
+ "eval_meter_score": 0.4515833277501449,
2620
+ "eval_rhyme_score": 0.8040101942334149,
2621
+ "eval_runtime": 2368.5368,
2622
+ "eval_samples_per_second": 1.14,
2623
+ "eval_steps_per_second": 0.143,
2624
+ "step": 104319
2625
+ },
2626
+ {
2627
+ "epoch": 9.02,
2628
+ "learning_rate": 7.25926174138754e-06,
2629
+ "loss": 0.3855,
2630
+ "step": 104500
2631
+ },
2632
+ {
2633
+ "epoch": 9.04,
2634
+ "learning_rate": 6.947089000182754e-06,
2635
+ "loss": 0.3792,
2636
+ "step": 104750
2637
+ },
2638
+ {
2639
+ "epoch": 9.06,
2640
+ "learning_rate": 6.641617718189974e-06,
2641
+ "loss": 0.3828,
2642
+ "step": 105000
2643
+ },
2644
+ {
2645
+ "epoch": 9.08,
2646
+ "learning_rate": 6.342862205522359e-06,
2647
+ "loss": 0.3822,
2648
+ "step": 105250
2649
+ },
2650
+ {
2651
+ "epoch": 9.1,
2652
+ "learning_rate": 6.050836457685948e-06,
2653
+ "loss": 0.383,
2654
+ "step": 105500
2655
+ },
2656
+ {
2657
+ "epoch": 9.12,
2658
+ "learning_rate": 5.765554154924096e-06,
2659
+ "loss": 0.3809,
2660
+ "step": 105750
2661
+ },
2662
+ {
2663
+ "epoch": 9.15,
2664
+ "learning_rate": 5.487028661576637e-06,
2665
+ "loss": 0.3791,
2666
+ "step": 106000
2667
+ },
2668
+ {
2669
+ "epoch": 9.17,
2670
+ "learning_rate": 5.2152730254538036e-06,
2671
+ "loss": 0.382,
2672
+ "step": 106250
2673
+ },
2674
+ {
2675
+ "epoch": 9.19,
2676
+ "learning_rate": 4.9502999772249475e-06,
2677
+ "loss": 0.3824,
2678
+ "step": 106500
2679
+ },
2680
+ {
2681
+ "epoch": 9.21,
2682
+ "learning_rate": 4.692121929822185e-06,
2683
+ "loss": 0.3797,
2684
+ "step": 106750
2685
+ },
2686
+ {
2687
+ "epoch": 9.23,
2688
+ "learning_rate": 4.4407509778589256e-06,
2689
+ "loss": 0.3819,
2690
+ "step": 107000
2691
+ },
2692
+ {
2693
+ "epoch": 9.25,
2694
+ "learning_rate": 4.1961988970632465e-06,
2695
+ "loss": 0.3797,
2696
+ "step": 107250
2697
+ },
2698
+ {
2699
+ "epoch": 9.27,
2700
+ "learning_rate": 3.95847714372628e-06,
2701
+ "loss": 0.3805,
2702
+ "step": 107500
2703
+ },
2704
+ {
2705
+ "epoch": 9.3,
2706
+ "learning_rate": 3.7275968541655097e-06,
2707
+ "loss": 0.3804,
2708
+ "step": 107750
2709
+ },
2710
+ {
2711
+ "epoch": 9.32,
2712
+ "learning_rate": 3.5035688442030917e-06,
2713
+ "loss": 0.3802,
2714
+ "step": 108000
2715
+ },
2716
+ {
2717
+ "epoch": 9.34,
2718
+ "learning_rate": 3.286403608659155e-06,
2719
+ "loss": 0.3803,
2720
+ "step": 108250
2721
+ },
2722
+ {
2723
+ "epoch": 9.36,
2724
+ "learning_rate": 3.0761113208601807e-06,
2725
+ "loss": 0.3783,
2726
+ "step": 108500
2727
+ },
2728
+ {
2729
+ "epoch": 9.38,
2730
+ "learning_rate": 2.872701832162433e-06,
2731
+ "loss": 0.3813,
2732
+ "step": 108750
2733
+ },
2734
+ {
2735
+ "epoch": 9.4,
2736
+ "learning_rate": 2.6761846714904277e-06,
2737
+ "loss": 0.3806,
2738
+ "step": 109000
2739
+ },
2740
+ {
2741
+ "epoch": 9.43,
2742
+ "learning_rate": 2.4865690448905573e-06,
2743
+ "loss": 0.3812,
2744
+ "step": 109250
2745
+ },
2746
+ {
2747
+ "epoch": 9.45,
2748
+ "learning_rate": 2.303863835099867e-06,
2749
+ "loss": 0.381,
2750
+ "step": 109500
2751
+ },
2752
+ {
2753
+ "epoch": 9.47,
2754
+ "learning_rate": 2.128077601129841e-06,
2755
+ "loss": 0.3803,
2756
+ "step": 109750
2757
+ },
2758
+ {
2759
+ "epoch": 9.49,
2760
+ "learning_rate": 1.959218577865518e-06,
2761
+ "loss": 0.3793,
2762
+ "step": 110000
2763
+ },
2764
+ {
2765
+ "epoch": 9.51,
2766
+ "learning_rate": 1.7972946756797058e-06,
2767
+ "loss": 0.3799,
2768
+ "step": 110250
2769
+ },
2770
+ {
2771
+ "epoch": 9.53,
2772
+ "learning_rate": 1.6423134800624237e-06,
2773
+ "loss": 0.3811,
2774
+ "step": 110500
2775
+ },
2776
+ {
2777
+ "epoch": 9.55,
2778
+ "learning_rate": 1.4942822512654884e-06,
2779
+ "loss": 0.3794,
2780
+ "step": 110750
2781
+ },
2782
+ {
2783
+ "epoch": 9.58,
2784
+ "learning_rate": 1.3532079239625016e-06,
2785
+ "loss": 0.3798,
2786
+ "step": 111000
2787
+ },
2788
+ {
2789
+ "epoch": 9.6,
2790
+ "learning_rate": 1.2190971069238942e-06,
2791
+ "loss": 0.38,
2792
+ "step": 111250
2793
+ },
2794
+ {
2795
+ "epoch": 9.62,
2796
+ "learning_rate": 1.0919560827073903e-06,
2797
+ "loss": 0.3784,
2798
+ "step": 111500
2799
+ },
2800
+ {
2801
+ "epoch": 9.64,
2802
+ "learning_rate": 9.717908073636594e-07,
2803
+ "loss": 0.3785,
2804
+ "step": 111750
2805
+ },
2806
+ {
2807
+ "epoch": 9.66,
2808
+ "learning_rate": 8.586069101573556e-07,
2809
+ "loss": 0.3785,
2810
+ "step": 112000
2811
+ },
2812
+ {
2813
+ "epoch": 9.68,
2814
+ "learning_rate": 7.524096933033131e-07,
2815
+ "loss": 0.3785,
2816
+ "step": 112250
2817
+ },
2818
+ {
2819
+ "epoch": 9.71,
2820
+ "learning_rate": 6.532041317182435e-07,
2821
+ "loss": 0.3811,
2822
+ "step": 112500
2823
+ },
2824
+ {
2825
+ "epoch": 9.73,
2826
+ "learning_rate": 5.609948727876567e-07,
2827
+ "loss": 0.3803,
2828
+ "step": 112750
2829
+ },
2830
+ {
2831
+ "epoch": 9.75,
2832
+ "learning_rate": 4.757862361481346e-07,
2833
+ "loss": 0.3803,
2834
+ "step": 113000
2835
+ },
2836
+ {
2837
+ "epoch": 9.77,
2838
+ "learning_rate": 3.97582213484976e-07,
2839
+ "loss": 0.3797,
2840
+ "step": 113250
2841
+ },
2842
+ {
2843
+ "epoch": 9.79,
2844
+ "learning_rate": 3.2638646834523e-07,
2845
+ "loss": 0.3795,
2846
+ "step": 113500
2847
+ },
2848
+ {
2849
+ "epoch": 9.81,
2850
+ "learning_rate": 2.6220233596603303e-07,
2851
+ "loss": 0.3783,
2852
+ "step": 113750
2853
+ },
2854
+ {
2855
+ "epoch": 9.84,
2856
+ "learning_rate": 2.0503282311838397e-07,
2857
+ "loss": 0.3783,
2858
+ "step": 114000
2859
+ },
2860
+ {
2861
+ "epoch": 9.86,
2862
+ "learning_rate": 1.5488060796630675e-07,
2863
+ "loss": 0.3764,
2864
+ "step": 114250
2865
+ },
2866
+ {
2867
+ "epoch": 9.88,
2868
+ "learning_rate": 1.1174803994136727e-07,
2869
+ "loss": 0.3796,
2870
+ "step": 114500
2871
+ },
2872
+ {
2873
+ "epoch": 9.9,
2874
+ "learning_rate": 7.563713963261164e-08,
2875
+ "loss": 0.3794,
2876
+ "step": 114750
2877
+ },
2878
+ {
2879
+ "epoch": 9.92,
2880
+ "learning_rate": 4.654959869190844e-08,
2881
+ "loss": 0.3776,
2882
+ "step": 115000
2883
+ },
2884
+ {
2885
+ "epoch": 9.94,
2886
+ "learning_rate": 2.448677975472879e-08,
2887
+ "loss": 0.3787,
2888
+ "step": 115250
2889
+ },
2890
+ {
2891
+ "epoch": 9.96,
2892
+ "learning_rate": 9.449716376264128e-09,
2893
+ "loss": 0.3794,
2894
+ "step": 115500
2895
+ },
2896
+ {
2897
+ "epoch": 9.99,
2898
+ "learning_rate": 1.4391129830815251e-09,
2899
+ "loss": 0.3801,
2900
+ "step": 115750
2901
+ },
2902
+ {
2903
+ "epoch": 10.0,
2904
+ "eval_alliteration_score": 0.42713157177880684,
2905
+ "eval_harmonic_meter_score": 0.18179456186285856,
2906
+ "eval_harmonic_rhyme_score": 0.3997335158862321,
2907
+ "eval_meter_score": 0.4478478237036721,
2908
+ "eval_rhyme_score": 0.8034619629050801,
2909
+ "eval_runtime": 2381.8504,
2910
+ "eval_samples_per_second": 1.134,
2911
+ "eval_steps_per_second": 0.142,
2912
+ "step": 115910
2913
+ },
2914
+ {
2915
+ "epoch": 10.0,
2916
+ "step": 115910,
2917
+ "total_flos": 5.348698838905389e+18,
2918
+ "train_loss": 0.4458793824766618,
2919
+ "train_runtime": 86475.3469,
2920
+ "train_samples_per_second": 171.576,
2921
+ "train_steps_per_second": 1.34
2922
+ }
2923
+ ],
2924
+ "max_steps": 115910,
2925
+ "num_train_epochs": 10,
2926
+ "total_flos": 5.348698838905389e+18,
2927
+ "trial_name": null,
2928
+ "trial_params": null
2929
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cdc190cc82e16ca55f910f071560f0b92129dd8e5683b5c118da263abbcbc20e
3
+ size 3439