Hugging Face's logo

CrisisNarratives
/
adapter-13classes-single_label

Text2Text Generation
Transformers
PyTorch
English
t5
text-generation-inference
Inference Endpoints
Model card Files Community
1
Faeze commited on
Commit
db7396c
·
verified ·
1 Parent(s): 6853175

Upload folder using huggingface_hub

Browse files
Files changed (16) hide show
  1. all_results.json +34 -0
  2. config.json +171 -0
  3. eval_instances_normalized_attn_scores.pt +3 -0
  4. eval_results.json +12 -0
  5. performance_results.json +4 -0
  6. predictions.csv +117 -0
  7. pytorch_model.bin +3 -0
  8. special_tokens_map.json +1 -0
  9. spiece.model +3 -0
  10. test_instances_normalized_attn_scores.pt +3 -0
  11. test_results.json +12 -0
  12. tokenizer.json +0 -0
  13. tokenizer_config.json +1 -0
  14. train_results.json +14 -0
  15. trainer_state.json +0 -0
  16. training_args.bin +3 -0
all_results.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 200.0,
3
+ "eval_accuracy": 76.72413793103449,
4
+ "eval_average_metrics": 70.1131066055635,
5
+ "eval_classification_report": "{\"accusation\": {\"precision\": 0.6666666666666666, \"recall\": 0.8333333333333334, \"f1-score\": 0.7407407407407408, \"support\": 12.0}, \"appreciation\": {\"precision\": 1.0, \"recall\": 0.75, \"f1-score\": 0.8571428571428571, \"support\": 8.0}, \"challenge\": {\"precision\": 0.7931034482758621, \"recall\": 0.7419354838709677, \"f1-score\": 0.7666666666666667, \"support\": 31.0}, \"evaluation\": {\"precision\": 1.0, \"recall\": 0.7777777777777778, \"f1-score\": 0.8750000000000001, \"support\": 9.0}, \"informing statement\": {\"precision\": 0.7297297297297297, \"recall\": 0.7297297297297297, \"f1-score\": 0.7297297297297297, \"support\": 37.0}, \"question\": {\"precision\": 1.0, \"recall\": 1.0, \"f1-score\": 1.0, \"support\": 1.0}, \"rejection\": {\"precision\": 0.7777777777777778, \"recall\": 0.7777777777777778, \"f1-score\": 0.7777777777777778, \"support\": 9.0}, \"request\": {\"precision\": 0.6666666666666666, \"recall\": 0.8888888888888888, \"f1-score\": 0.761904761904762, \"support\": 9.0}, \"accuracy\": 0.7672413793103449, \"macro avg\": {\"precision\": 0.8292430361395879, \"recall\": 0.8124303739223094, \"f1-score\": 0.8136203167453168, \"support\": 116.0}, \"weighted avg\": {\"precision\": 0.7809155766944115, \"recall\": 0.7672413793103449, \"f1-score\": 0.7693520799124248, \"support\": 116.0}}",
6
+ "eval_f1_macro": 50.06894256894257,
7
+ "eval_f1_micro": 76.72413793103448,
8
+ "eval_f1_weighted": 76.93520799124248,
9
+ "eval_loss": 0.9065346121788025,
10
+ "eval_runtime": 1.1668,
11
+ "eval_samples_per_second": 99.413,
12
+ "init_mem_cpu_alloc_delta": -542769152,
13
+ "init_mem_cpu_peaked_delta": 542777344,
14
+ "init_mem_gpu_alloc_delta": 891528192,
15
+ "init_mem_gpu_peaked_delta": 0,
16
+ "peak_memory": 5.305544921875,
17
+ "test_accuracy": 76.72413793103449,
18
+ "test_average_metrics": 70.1131066055635,
19
+ "test_classification_report": "{\"accusation\": {\"precision\": 0.6666666666666666, \"recall\": 0.8333333333333334, \"f1-score\": 0.7407407407407408, \"support\": 12.0}, \"appreciation\": {\"precision\": 1.0, \"recall\": 0.75, \"f1-score\": 0.8571428571428571, \"support\": 8.0}, \"challenge\": {\"precision\": 0.7931034482758621, \"recall\": 0.7419354838709677, \"f1-score\": 0.7666666666666667, \"support\": 31.0}, \"evaluation\": {\"precision\": 1.0, \"recall\": 0.7777777777777778, \"f1-score\": 0.8750000000000001, \"support\": 9.0}, \"informing statement\": {\"precision\": 0.7297297297297297, \"recall\": 0.7297297297297297, \"f1-score\": 0.7297297297297297, \"support\": 37.0}, \"question\": {\"precision\": 1.0, \"recall\": 1.0, \"f1-score\": 1.0, \"support\": 1.0}, \"rejection\": {\"precision\": 0.7777777777777778, \"recall\": 0.7777777777777778, \"f1-score\": 0.7777777777777778, \"support\": 9.0}, \"request\": {\"precision\": 0.6666666666666666, \"recall\": 0.8888888888888888, \"f1-score\": 0.761904761904762, \"support\": 9.0}, \"accuracy\": 0.7672413793103449, \"macro avg\": {\"precision\": 0.8292430361395879, \"recall\": 0.8124303739223094, \"f1-score\": 0.8136203167453168, \"support\": 116.0}, \"weighted avg\": {\"precision\": 0.7809155766944115, \"recall\": 0.7672413793103449, \"f1-score\": 0.7693520799124248, \"support\": 116.0}}",
20
+ "test_f1_macro": 50.06894256894257,
21
+ "test_f1_micro": 76.72413793103448,
22
+ "test_f1_weighted": 76.93520799124248,
23
+ "test_loss": 0.9065346121788025,
24
+ "test_runtime": 1.187,
25
+ "test_samples_per_second": 97.724,
26
+ "total_time in minutes ": 39.80907916666666,
27
+ "train_mem_cpu_alloc_delta": 937439232,
28
+ "train_mem_cpu_peaked_delta": 40960,
29
+ "train_mem_gpu_alloc_delta": 40215552,
30
+ "train_mem_gpu_peaked_delta": 4624273408,
31
+ "train_runtime": 2386.5323,
32
+ "train_samples": 488,
33
+ "train_samples_per_second": 1.341
34
+ }
config.json ADDED
@@ -0,0 +1,171 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "t5-base",
3
+ "architectures": [
4
+ "T5ForConditionalGeneration"
5
+ ],
6
+ "attn_method": "linear",
7
+ "attn_prefix_tuning": false,
8
+ "class_weights": [
9
+ 0.40939597315436244,
10
+ 0.46923076923076923,
11
+ 1.1296296296296295,
12
+ 1.525,
13
+ 1.525,
14
+ 1.6944444444444444,
15
+ 1.90625,
16
+ 8.714285714285714
17
+ ],
18
+ "d_ff": 3072,
19
+ "d_kv": 64,
20
+ "d_model": 768,
21
+ "decoder_start_token_id": 0,
22
+ "dropout_rate": 0.1,
23
+ "eos_token_id": 1,
24
+ "feed_forward_proj": "relu",
25
+ "fix_attention": false,
26
+ "gradient_checkpointing": false,
27
+ "id2label": {
28
+ "0": "LABEL_0",
29
+ "1": "LABEL_1",
30
+ "2": "LABEL_2",
31
+ "3": "LABEL_3",
32
+ "4": "LABEL_4",
33
+ "5": "LABEL_5",
34
+ "6": "LABEL_6",
35
+ "7": "LABEL_7",
36
+ "8": "LABEL_8",
37
+ "9": "LABEL_9",
38
+ "10": "LABEL_10",
39
+ "11": "LABEL_11",
40
+ "12": "LABEL_12"
41
+ },
42
+ "ignore_target": false,
43
+ "init_prefix_method": "random",
44
+ "initializer_factor": 1.0,
45
+ "is_contrastive": false,
46
+ "is_encoder_decoder": true,
47
+ "label2id": {
48
+ "LABEL_0": 0,
49
+ "LABEL_1": 1,
50
+ "LABEL_10": 10,
51
+ "LABEL_11": 11,
52
+ "LABEL_12": 12,
53
+ "LABEL_2": 2,
54
+ "LABEL_3": 3,
55
+ "LABEL_4": 4,
56
+ "LABEL_5": 5,
57
+ "LABEL_6": 6,
58
+ "LABEL_7": 7,
59
+ "LABEL_8": 8,
60
+ "LABEL_9": 9
61
+ },
62
+ "label_based": false,
63
+ "label_token_ids": [
64
+ [
65
+ 16,
66
+ 10454,
67
+ 2493,
68
+ 1
69
+ ],
70
+ [
71
+ 1921,
72
+ 1
73
+ ],
74
+ [
75
+ 17316,
76
+ 257,
77
+ 1
78
+ ],
79
+ [
80
+ 11746,
81
+ 1
82
+ ],
83
+ [
84
+ 1690,
85
+ 1
86
+ ],
87
+ [
88
+ 5002,
89
+ 1
90
+ ],
91
+ [
92
+ 21530,
93
+ 1
94
+ ],
95
+ [
96
+ 822,
97
+ 1
98
+ ],
99
+ [
100
+ 11122,
101
+ 1
102
+ ],
103
+ [
104
+ 6384,
105
+ 1
106
+ ],
107
+ [
108
+ 20,
109
+ 7419,
110
+ 1
111
+ ],
112
+ [
113
+ 7209,
114
+ 1
115
+ ],
116
+ [
117
+ 30430,
118
+ 1
119
+ ]
120
+ ],
121
+ "layer_norm_epsilon": 1e-06,
122
+ "learned_temperature": false,
123
+ "max_length": 4,
124
+ "model_type": "t5",
125
+ "n_positions": 512,
126
+ "normalize_prefixes": false,
127
+ "num_decoder_layers": 12,
128
+ "num_heads": 12,
129
+ "num_layers": 12,
130
+ "num_target": 1,
131
+ "output_past": true,
132
+ "pad_token_id": 0,
133
+ "prefix_num": 1,
134
+ "prefix_tuning": false,
135
+ "relative_attention_num_buckets": 32,
136
+ "shared_attn": false,
137
+ "task_specific_params": {
138
+ "summarization": {
139
+ "early_stopping": true,
140
+ "length_penalty": 2.0,
141
+ "max_length": 200,
142
+ "min_length": 30,
143
+ "no_repeat_ngram_size": 3,
144
+ "num_beams": 4,
145
+ "prefix": "summarize: "
146
+ },
147
+ "translation_en_to_de": {
148
+ "early_stopping": true,
149
+ "max_length": 300,
150
+ "num_beams": 4,
151
+ "prefix": "translate English to German: "
152
+ },
153
+ "translation_en_to_fr": {
154
+ "early_stopping": true,
155
+ "max_length": 300,
156
+ "num_beams": 4,
157
+ "prefix": "translate English to French: "
158
+ },
159
+ "translation_en_to_ro": {
160
+ "early_stopping": true,
161
+ "max_length": 300,
162
+ "num_beams": 4,
163
+ "prefix": "translate English to Romanian: "
164
+ }
165
+ },
166
+ "temperature": 2000,
167
+ "train_task_adapters": true,
168
+ "transformers_version": "4.6.0",
169
+ "use_cache": true,
170
+ "vocab_size": 32100
171
+ }
eval_instances_normalized_attn_scores.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50764592c09d3053c8ca12d3bd7108d14b2d3b3e93191cb4e40cfdac94418f81
3
+ size 1048
eval_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 200.0,
3
+ "eval_accuracy": 76.72413793103449,
4
+ "eval_average_metrics": 70.1131066055635,
5
+ "eval_classification_report": "{\"accusation\": {\"precision\": 0.6666666666666666, \"recall\": 0.8333333333333334, \"f1-score\": 0.7407407407407408, \"support\": 12.0}, \"appreciation\": {\"precision\": 1.0, \"recall\": 0.75, \"f1-score\": 0.8571428571428571, \"support\": 8.0}, \"challenge\": {\"precision\": 0.7931034482758621, \"recall\": 0.7419354838709677, \"f1-score\": 0.7666666666666667, \"support\": 31.0}, \"evaluation\": {\"precision\": 1.0, \"recall\": 0.7777777777777778, \"f1-score\": 0.8750000000000001, \"support\": 9.0}, \"informing statement\": {\"precision\": 0.7297297297297297, \"recall\": 0.7297297297297297, \"f1-score\": 0.7297297297297297, \"support\": 37.0}, \"question\": {\"precision\": 1.0, \"recall\": 1.0, \"f1-score\": 1.0, \"support\": 1.0}, \"rejection\": {\"precision\": 0.7777777777777778, \"recall\": 0.7777777777777778, \"f1-score\": 0.7777777777777778, \"support\": 9.0}, \"request\": {\"precision\": 0.6666666666666666, \"recall\": 0.8888888888888888, \"f1-score\": 0.761904761904762, \"support\": 9.0}, \"accuracy\": 0.7672413793103449, \"macro avg\": {\"precision\": 0.8292430361395879, \"recall\": 0.8124303739223094, \"f1-score\": 0.8136203167453168, \"support\": 116.0}, \"weighted avg\": {\"precision\": 0.7809155766944115, \"recall\": 0.7672413793103449, \"f1-score\": 0.7693520799124248, \"support\": 116.0}}",
6
+ "eval_f1_macro": 50.06894256894257,
7
+ "eval_f1_micro": 76.72413793103448,
8
+ "eval_f1_weighted": 76.93520799124248,
9
+ "eval_loss": 0.9065346121788025,
10
+ "eval_runtime": 1.1668,
11
+ "eval_samples_per_second": 99.413
12
+ }
performance_results.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "peak_memory": 5.305544921875,
3
+ "total_time in minutes ": 39.80907916666666
4
+ }
predictions.csv ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ,id,pred
2
+ 0,91,<
3
+ 1,24,<
4
+ 2,124,<
5
+ 3,289,<
6
+ 4,152,<
7
+ 5,178,<
8
+ 6,213,<
9
+ 7,160,<
10
+ 8,45,<
11
+ 9,651,<
12
+ 10,270,<
13
+ 11,72,<
14
+ 12,611,<
15
+ 13,449,<
16
+ 14,501,<
17
+ 15,209,<
18
+ 16,677,<
19
+ 17,144,<
20
+ 18,331,<
21
+ 19,62,<
22
+ 20,436,<
23
+ 21,554,<
24
+ 22,108,<
25
+ 23,479,<
26
+ 24,641,<
27
+ 25,466,<
28
+ 26,589,<
29
+ 27,643,<
30
+ 28,246,<
31
+ 29,287,<
32
+ 30,462,<
33
+ 31,5,<
34
+ 32,379,<
35
+ 33,560,<
36
+ 34,305,<
37
+ 35,616,<
38
+ 36,47,<
39
+ 37,100,<
40
+ 38,358,<
41
+ 39,645,<
42
+ 40,2,<
43
+ 41,619,<
44
+ 42,329,<
45
+ 43,682,<
46
+ 44,550,<
47
+ 45,444,<
48
+ 46,16,<
49
+ 47,655,<
50
+ 48,21,<
51
+ 49,342,<
52
+ 50,671,<
53
+ 51,448,<
54
+ 52,548,<
55
+ 53,109,<
56
+ 54,404,<
57
+ 55,129,<
58
+ 56,348,<
59
+ 57,565,<
60
+ 58,268,<
61
+ 59,673,<
62
+ 60,627,<
63
+ 61,255,<
64
+ 62,360,<
65
+ 63,95,<
66
+ 64,130,<
67
+ 65,572,<
68
+ 66,512,<
69
+ 67,199,<
70
+ 68,288,<
71
+ 69,633,<
72
+ 70,159,<
73
+ 71,639,<
74
+ 72,174,<
75
+ 73,567,<
76
+ 74,106,<
77
+ 75,263,<
78
+ 76,343,<
79
+ 77,194,<
80
+ 78,587,<
81
+ 79,612,<
82
+ 80,226,<
83
+ 81,653,<
84
+ 82,186,<
85
+ 83,482,<
86
+ 84,652,<
87
+ 85,333,<
88
+ 86,530,<
89
+ 87,592,<
90
+ 88,494,<
91
+ 89,549,<
92
+ 90,415,<
93
+ 91,9,<
94
+ 92,139,<
95
+ 93,617,<
96
+ 94,326,<
97
+ 95,59,<
98
+ 96,614,<
99
+ 97,64,<
100
+ 98,327,<
101
+ 99,189,<
102
+ 100,214,<
103
+ 101,258,<
104
+ 102,230,<
105
+ 103,216,<
106
+ 104,57,<
107
+ 105,523,<
108
+ 106,313,<
109
+ 107,498,<
110
+ 108,336,<
111
+ 109,623,<
112
+ 110,632,<
113
+ 111,115,<
114
+ 112,600,<
115
+ 113,259,<
116
+ 114,193,<
117
+ 115,257,<
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:058007ccf1dff0f709016fb19dc3d857b4508b7913b6bfa41fa4e03d51d4f40f
3
+ size 898972686
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "additional_special_tokens": ["<extra_id_0>", "<extra_id_1>", "<extra_id_2>", "<extra_id_3>", "<extra_id_4>", "<extra_id_5>", "<extra_id_6>", "<extra_id_7>", "<extra_id_8>", "<extra_id_9>", "<extra_id_10>", "<extra_id_11>", "<extra_id_12>", "<extra_id_13>", "<extra_id_14>", "<extra_id_15>", "<extra_id_16>", "<extra_id_17>", "<extra_id_18>", "<extra_id_19>", "<extra_id_20>", "<extra_id_21>", "<extra_id_22>", "<extra_id_23>", "<extra_id_24>", "<extra_id_25>", "<extra_id_26>", "<extra_id_27>", "<extra_id_28>", "<extra_id_29>", "<extra_id_30>", "<extra_id_31>", "<extra_id_32>", "<extra_id_33>", "<extra_id_34>", "<extra_id_35>", "<extra_id_36>", "<extra_id_37>", "<extra_id_38>", "<extra_id_39>", "<extra_id_40>", "<extra_id_41>", "<extra_id_42>", "<extra_id_43>", "<extra_id_44>", "<extra_id_45>", "<extra_id_46>", "<extra_id_47>", "<extra_id_48>", "<extra_id_49>", "<extra_id_50>", "<extra_id_51>", "<extra_id_52>", "<extra_id_53>", "<extra_id_54>", "<extra_id_55>", "<extra_id_56>", "<extra_id_57>", "<extra_id_58>", "<extra_id_59>", "<extra_id_60>", "<extra_id_61>", "<extra_id_62>", "<extra_id_63>", "<extra_id_64>", "<extra_id_65>", "<extra_id_66>", "<extra_id_67>", "<extra_id_68>", "<extra_id_69>", "<extra_id_70>", "<extra_id_71>", "<extra_id_72>", "<extra_id_73>", "<extra_id_74>", "<extra_id_75>", "<extra_id_76>", "<extra_id_77>", "<extra_id_78>", "<extra_id_79>", "<extra_id_80>", "<extra_id_81>", "<extra_id_82>", "<extra_id_83>", "<extra_id_84>", "<extra_id_85>", "<extra_id_86>", "<extra_id_87>", "<extra_id_88>", "<extra_id_89>", "<extra_id_90>", "<extra_id_91>", "<extra_id_92>", "<extra_id_93>", "<extra_id_94>", "<extra_id_95>", "<extra_id_96>", "<extra_id_97>", "<extra_id_98>", "<extra_id_99>"]}
spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d60acb128cf7b7f2536e8f38a5b18a05535c9e14c7a355904270e15b0945ea86
3
+ size 791656
test_instances_normalized_attn_scores.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0cf18366eb389698056f39a6158bd36abae1b39c77767057ccaca0a3c413cd26
3
+ size 1048
test_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 200.0,
3
+ "test_accuracy": 76.72413793103449,
4
+ "test_average_metrics": 70.1131066055635,
5
+ "test_classification_report": "{\"accusation\": {\"precision\": 0.6666666666666666, \"recall\": 0.8333333333333334, \"f1-score\": 0.7407407407407408, \"support\": 12.0}, \"appreciation\": {\"precision\": 1.0, \"recall\": 0.75, \"f1-score\": 0.8571428571428571, \"support\": 8.0}, \"challenge\": {\"precision\": 0.7931034482758621, \"recall\": 0.7419354838709677, \"f1-score\": 0.7666666666666667, \"support\": 31.0}, \"evaluation\": {\"precision\": 1.0, \"recall\": 0.7777777777777778, \"f1-score\": 0.8750000000000001, \"support\": 9.0}, \"informing statement\": {\"precision\": 0.7297297297297297, \"recall\": 0.7297297297297297, \"f1-score\": 0.7297297297297297, \"support\": 37.0}, \"question\": {\"precision\": 1.0, \"recall\": 1.0, \"f1-score\": 1.0, \"support\": 1.0}, \"rejection\": {\"precision\": 0.7777777777777778, \"recall\": 0.7777777777777778, \"f1-score\": 0.7777777777777778, \"support\": 9.0}, \"request\": {\"precision\": 0.6666666666666666, \"recall\": 0.8888888888888888, \"f1-score\": 0.761904761904762, \"support\": 9.0}, \"accuracy\": 0.7672413793103449, \"macro avg\": {\"precision\": 0.8292430361395879, \"recall\": 0.8124303739223094, \"f1-score\": 0.8136203167453168, \"support\": 116.0}, \"weighted avg\": {\"precision\": 0.7809155766944115, \"recall\": 0.7672413793103449, \"f1-score\": 0.7693520799124248, \"support\": 116.0}}",
6
+ "test_f1_macro": 50.06894256894257,
7
+ "test_f1_micro": 76.72413793103448,
8
+ "test_f1_weighted": 76.93520799124248,
9
+ "test_loss": 0.9065346121788025,
10
+ "test_runtime": 1.187,
11
+ "test_samples_per_second": 97.724
12
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>", "extra_ids": 100, "additional_special_tokens": ["<extra_id_0>", "<extra_id_1>", "<extra_id_2>", "<extra_id_3>", "<extra_id_4>", "<extra_id_5>", "<extra_id_6>", "<extra_id_7>", "<extra_id_8>", "<extra_id_9>", "<extra_id_10>", "<extra_id_11>", "<extra_id_12>", "<extra_id_13>", "<extra_id_14>", "<extra_id_15>", "<extra_id_16>", "<extra_id_17>", "<extra_id_18>", "<extra_id_19>", "<extra_id_20>", "<extra_id_21>", "<extra_id_22>", "<extra_id_23>", "<extra_id_24>", "<extra_id_25>", "<extra_id_26>", "<extra_id_27>", "<extra_id_28>", "<extra_id_29>", "<extra_id_30>", "<extra_id_31>", "<extra_id_32>", "<extra_id_33>", "<extra_id_34>", "<extra_id_35>", "<extra_id_36>", "<extra_id_37>", "<extra_id_38>", "<extra_id_39>", "<extra_id_40>", "<extra_id_41>", "<extra_id_42>", "<extra_id_43>", "<extra_id_44>", "<extra_id_45>", "<extra_id_46>", "<extra_id_47>", "<extra_id_48>", "<extra_id_49>", "<extra_id_50>", "<extra_id_51>", "<extra_id_52>", "<extra_id_53>", "<extra_id_54>", "<extra_id_55>", "<extra_id_56>", "<extra_id_57>", "<extra_id_58>", "<extra_id_59>", "<extra_id_60>", "<extra_id_61>", "<extra_id_62>", "<extra_id_63>", "<extra_id_64>", "<extra_id_65>", "<extra_id_66>", "<extra_id_67>", "<extra_id_68>", "<extra_id_69>", "<extra_id_70>", "<extra_id_71>", "<extra_id_72>", "<extra_id_73>", "<extra_id_74>", "<extra_id_75>", "<extra_id_76>", "<extra_id_77>", "<extra_id_78>", "<extra_id_79>", "<extra_id_80>", "<extra_id_81>", "<extra_id_82>", "<extra_id_83>", "<extra_id_84>", "<extra_id_85>", "<extra_id_86>", "<extra_id_87>", "<extra_id_88>", "<extra_id_89>", "<extra_id_90>", "<extra_id_91>", "<extra_id_92>", "<extra_id_93>", "<extra_id_94>", "<extra_id_95>", "<extra_id_96>", "<extra_id_97>", "<extra_id_98>", "<extra_id_99>"], "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "t5-base"}
train_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 200.0,
3
+ "init_mem_cpu_alloc_delta": -542769152,
4
+ "init_mem_cpu_peaked_delta": 542777344,
5
+ "init_mem_gpu_alloc_delta": 891528192,
6
+ "init_mem_gpu_peaked_delta": 0,
7
+ "train_mem_cpu_alloc_delta": 937439232,
8
+ "train_mem_cpu_peaked_delta": 40960,
9
+ "train_mem_gpu_alloc_delta": 40215552,
10
+ "train_mem_gpu_peaked_delta": 4624273408,
11
+ "train_runtime": 2386.5323,
12
+ "train_samples": 488,
13
+ "train_samples_per_second": 1.341
14
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3b9951f71ca48d7687634e7a779a4fd4658bef911aa0337c63653c755a61d9f
3
+ size 3448