Hugging Face's logo

Merikatori
/
hate-speech-roberta

Text Classification
Safetensors
English
roberta
hate-speech
twitter
Model card Files
xet
 Community
Merikatori commited on
Commit
536f80b
·
verified ·
1 Parent(s): cae6e61

Upload folder using huggingface_hub

Browse files
Files changed (21) hide show
  1. README.md +35 -0
  2. checkpoint-465/config.json +39 -0
  3. checkpoint-465/model.safetensors +3 -0
  4. checkpoint-465/rng_state.pth +3 -0
  5. checkpoint-465/scaler.pt +3 -0
  6. checkpoint-465/tokenizer.json +0 -0
  7. checkpoint-465/tokenizer_config.json +16 -0
  8. checkpoint-465/trainer_state.json +154 -0
  9. checkpoint-465/training_args.bin +3 -0
  10. checkpoint-775/config.json +39 -0
  11. checkpoint-775/model.safetensors +3 -0
  12. checkpoint-775/rng_state.pth +3 -0
  13. checkpoint-775/scaler.pt +3 -0
  14. checkpoint-775/tokenizer.json +0 -0
  15. checkpoint-775/tokenizer_config.json +16 -0
  16. checkpoint-775/trainer_state.json +228 -0
  17. checkpoint-775/training_args.bin +3 -0
  18. config.json +39 -0
  19. model.safetensors +3 -0
  20. tokenizer.json +0 -0
  21. tokenizer_config.json +16 -0
README.md ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: en
3
+ tags:
4
+ - text-classification
5
+ - hate-speech
6
+ - twitter
7
+ - roberta
8
+ datasets:
9
+ - hate_speech_offensive
10
+ metrics:
11
+ - f1
12
+ ---
13
+
14
+ # Hate Speech Detector — fine-tuned RoBERTa
15
+
16
+ Fine-tuned từ `cardiffnlp/twitter-roberta-base-hate` trên dataset Davidson et al. (2017).
17
+
18
+ ## Labels
19
+ - **0 — Hate Speech**: ngôn ngữ thù ghét
20
+ - **1 — Offensive**: xúc phạm nhưng không phải hate speech
21
+ - **2 — Neither**: bình thường
22
+
23
+ ## Kết quả
24
+ | Metric | Score |
25
+ |--------|-------|
26
+ | Macro F1 | ~0.77 |
27
+ | Hate Speech F1 | ~0.48 |
28
+ | Accuracy | ~0.89 |
29
+
30
+ ## Cách dùng
31
+ ```python
32
+ from transformers import pipeline
33
+ clf = pipeline("text-classification", model="Merikatori/hate-speech-roberta")
34
+ clf("I hate all people like that")
35
+ ```
checkpoint-465/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "RobertaForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "dtype": "float32",
10
+ "eos_token_id": 2,
11
+ "gradient_checkpointing": false,
12
+ "hidden_act": "gelu",
13
+ "hidden_dropout_prob": 0.1,
14
+ "hidden_size": 768,
15
+ "id2label": {
16
+ "0": "LABEL_0",
17
+ "1": "LABEL_1",
18
+ "2": "LABEL_2"
19
+ },
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "is_decoder": false,
23
+ "label2id": {
24
+ "LABEL_0": 0,
25
+ "LABEL_1": 1,
26
+ "LABEL_2": 2
27
+ },
28
+ "layer_norm_eps": 1e-05,
29
+ "max_position_embeddings": 514,
30
+ "model_type": "roberta",
31
+ "num_attention_heads": 12,
32
+ "num_hidden_layers": 12,
33
+ "pad_token_id": 1,
34
+ "tie_word_embeddings": true,
35
+ "transformers_version": "5.2.0",
36
+ "type_vocab_size": 1,
37
+ "use_cache": false,
38
+ "vocab_size": 50265
39
+ }
checkpoint-465/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e33aa5ed6c97fd88a5380890559898a84fddcb456181718c26da4dccb056e01c
3
+ size 498615876
checkpoint-465/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22ed563928bb8049d07cfb6385208376710406a5d82aec95f63136e694d96a48
3
+ size 14645
checkpoint-465/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2aaaf715d1dc34d09e6d12501bac02a106a5b96a2df14d7c4704b5e0c632d8e
3
+ size 1383
checkpoint-465/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-465/tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<s>",
5
+ "cls_token": "<s>",
6
+ "eos_token": "</s>",
7
+ "errors": "replace",
8
+ "is_local": false,
9
+ "mask_token": "<mask>",
10
+ "model_max_length": 1000000000000000019884624838656,
11
+ "pad_token": "<pad>",
12
+ "sep_token": "</s>",
13
+ "tokenizer_class": "RobertaTokenizer",
14
+ "trim_offsets": true,
15
+ "unk_token": "<unk>"
16
+ }
checkpoint-465/trainer_state.json ADDED
@@ -0,0 +1,154 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 465,
3
+ "best_metric": 0.7698383516941082,
4
+ "best_model_checkpoint": "/kaggle/working/transformer_output/checkpoint-465",
5
+ "epoch": 3.0,
6
+ "eval_steps": 500,
7
+ "global_step": 465,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.3225806451612903,
14
+ "grad_norm": 3.0797810554504395,
15
+ "learning_rate": 1.2727272727272728e-05,
16
+ "loss": 1.1631031799316407,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.6451612903225806,
21
+ "grad_norm": 3.368655204772949,
22
+ "learning_rate": 1.9951016547387286e-05,
23
+ "loss": 0.6313877105712891,
24
+ "step": 100
25
+ },
26
+ {
27
+ "epoch": 0.967741935483871,
28
+ "grad_norm": 5.035543918609619,
29
+ "learning_rate": 1.947950097086825e-05,
30
+ "loss": 0.5252903747558594,
31
+ "step": 150
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.8886423239862821,
36
+ "eval_f1": 0.8959192934594974,
37
+ "eval_hate_f1": 0.47107438016528924,
38
+ "eval_hate_prec": 0.3886363636363636,
39
+ "eval_hate_rec": 0.5979020979020979,
40
+ "eval_loss": 0.26823195815086365,
41
+ "eval_macro_f1": 0.7588550972474853,
42
+ "eval_precision": 0.9065499982342927,
43
+ "eval_recall": 0.8886423239862821,
44
+ "eval_runtime": 8.5862,
45
+ "eval_samples_per_second": 577.325,
46
+ "eval_steps_per_second": 9.084,
47
+ "step": 155
48
+ },
49
+ {
50
+ "epoch": 1.2903225806451613,
51
+ "grad_norm": 2.4748122692108154,
52
+ "learning_rate": 1.8529927230890757e-05,
53
+ "loss": 0.501370506286621,
54
+ "step": 200
55
+ },
56
+ {
57
+ "epoch": 1.6129032258064515,
58
+ "grad_norm": 2.85589599609375,
59
+ "learning_rate": 1.715018302529852e-05,
60
+ "loss": 0.4371611404418945,
61
+ "step": 250
62
+ },
63
+ {
64
+ "epoch": 1.935483870967742,
65
+ "grad_norm": 2.5343403816223145,
66
+ "learning_rate": 1.5409849862969994e-05,
67
+ "loss": 0.44288436889648436,
68
+ "step": 300
69
+ },
70
+ {
71
+ "epoch": 2.0,
72
+ "eval_accuracy": 0.8912648779503732,
73
+ "eval_f1": 0.8998036744615823,
74
+ "eval_hate_f1": 0.4732620320855615,
75
+ "eval_hate_prec": 0.38311688311688313,
76
+ "eval_hate_rec": 0.6188811188811189,
77
+ "eval_loss": 0.2427479773759842,
78
+ "eval_macro_f1": 0.7651698753431324,
79
+ "eval_precision": 0.913809087207257,
80
+ "eval_recall": 0.8912648779503732,
81
+ "eval_runtime": 8.6774,
82
+ "eval_samples_per_second": 571.253,
83
+ "eval_steps_per_second": 8.989,
84
+ "step": 310
85
+ },
86
+ {
87
+ "epoch": 2.258064516129032,
88
+ "grad_norm": 6.783071994781494,
89
+ "learning_rate": 1.3396694017461708e-05,
90
+ "loss": 0.4083076477050781,
91
+ "step": 350
92
+ },
93
+ {
94
+ "epoch": 2.5806451612903225,
95
+ "grad_norm": 4.064752578735352,
96
+ "learning_rate": 1.1212240409675825e-05,
97
+ "loss": 0.36163642883300784,
98
+ "step": 400
99
+ },
100
+ {
101
+ "epoch": 2.903225806451613,
102
+ "grad_norm": 3.120911121368408,
103
+ "learning_rate": 8.966652631884506e-06,
104
+ "loss": 0.3937630081176758,
105
+ "step": 450
106
+ },
107
+ {
108
+ "epoch": 3.0,
109
+ "eval_accuracy": 0.8912648779503732,
110
+ "eval_f1": 0.8998510802952099,
111
+ "eval_hate_f1": 0.48548812664907653,
112
+ "eval_hate_prec": 0.3898305084745763,
113
+ "eval_hate_rec": 0.6433566433566433,
114
+ "eval_loss": 0.24792158603668213,
115
+ "eval_macro_f1": 0.7698383516941082,
116
+ "eval_precision": 0.9136908950590453,
117
+ "eval_recall": 0.8912648779503732,
118
+ "eval_runtime": 8.7831,
119
+ "eval_samples_per_second": 564.382,
120
+ "eval_steps_per_second": 8.881,
121
+ "step": 465
122
+ }
123
+ ],
124
+ "logging_steps": 50,
125
+ "max_steps": 775,
126
+ "num_input_tokens_seen": 0,
127
+ "num_train_epochs": 5,
128
+ "save_steps": 500,
129
+ "stateful_callbacks": {
130
+ "EarlyStoppingCallback": {
131
+ "args": {
132
+ "early_stopping_patience": 2,
133
+ "early_stopping_threshold": 0.0
134
+ },
135
+ "attributes": {
136
+ "early_stopping_patience_counter": 0
137
+ }
138
+ },
139
+ "TrainerControl": {
140
+ "args": {
141
+ "should_epoch_stop": false,
142
+ "should_evaluate": false,
143
+ "should_log": false,
144
+ "should_save": true,
145
+ "should_training_stop": false
146
+ },
147
+ "attributes": {}
148
+ }
149
+ },
150
+ "total_flos": 1467136861839936.0,
151
+ "train_batch_size": 64,
152
+ "trial_name": null,
153
+ "trial_params": null
154
+ }
checkpoint-465/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b0eee9e6e1a3712626f958410bea2b5a2002e579aab47745ad3abf8793b4938b
3
+ size 5201
checkpoint-775/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "RobertaForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "dtype": "float32",
10
+ "eos_token_id": 2,
11
+ "gradient_checkpointing": false,
12
+ "hidden_act": "gelu",
13
+ "hidden_dropout_prob": 0.1,
14
+ "hidden_size": 768,
15
+ "id2label": {
16
+ "0": "LABEL_0",
17
+ "1": "LABEL_1",
18
+ "2": "LABEL_2"
19
+ },
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "is_decoder": false,
23
+ "label2id": {
24
+ "LABEL_0": 0,
25
+ "LABEL_1": 1,
26
+ "LABEL_2": 2
27
+ },
28
+ "layer_norm_eps": 1e-05,
29
+ "max_position_embeddings": 514,
30
+ "model_type": "roberta",
31
+ "num_attention_heads": 12,
32
+ "num_hidden_layers": 12,
33
+ "pad_token_id": 1,
34
+ "tie_word_embeddings": true,
35
+ "transformers_version": "5.2.0",
36
+ "type_vocab_size": 1,
37
+ "use_cache": false,
38
+ "vocab_size": 50265
39
+ }
checkpoint-775/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:978e7ae0fadc8be849b6c3e538ece208a59c0f28adc3155b34ef030d5d4083dd
3
+ size 498615876
checkpoint-775/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a783403dc4579b8b2f91342a021235335fbb4a7ec7c10f645201a3369f6fe603
3
+ size 14645
checkpoint-775/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f3d80ae95be3cb164bc91930a1290a7a924974d7ed911d61faba43e067d13c8
3
+ size 1383
checkpoint-775/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-775/tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<s>",
5
+ "cls_token": "<s>",
6
+ "eos_token": "</s>",
7
+ "errors": "replace",
8
+ "is_local": false,
9
+ "mask_token": "<mask>",
10
+ "model_max_length": 1000000000000000019884624838656,
11
+ "pad_token": "<pad>",
12
+ "sep_token": "</s>",
13
+ "tokenizer_class": "RobertaTokenizer",
14
+ "trim_offsets": true,
15
+ "unk_token": "<unk>"
16
+ }
checkpoint-775/trainer_state.json ADDED
@@ -0,0 +1,228 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 465,
3
+ "best_metric": 0.7698383516941082,
4
+ "best_model_checkpoint": "/kaggle/working/transformer_output/checkpoint-465",
5
+ "epoch": 5.0,
6
+ "eval_steps": 500,
7
+ "global_step": 775,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.3225806451612903,
14
+ "grad_norm": 3.0797810554504395,
15
+ "learning_rate": 1.2727272727272728e-05,
16
+ "loss": 1.1631031799316407,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.6451612903225806,
21
+ "grad_norm": 3.368655204772949,
22
+ "learning_rate": 1.9951016547387286e-05,
23
+ "loss": 0.6313877105712891,
24
+ "step": 100
25
+ },
26
+ {
27
+ "epoch": 0.967741935483871,
28
+ "grad_norm": 5.035543918609619,
29
+ "learning_rate": 1.947950097086825e-05,
30
+ "loss": 0.5252903747558594,
31
+ "step": 150
32
+ },
33
+ {
34
+ "epoch": 1.0,
35
+ "eval_accuracy": 0.8886423239862821,
36
+ "eval_f1": 0.8959192934594974,
37
+ "eval_hate_f1": 0.47107438016528924,
38
+ "eval_hate_prec": 0.3886363636363636,
39
+ "eval_hate_rec": 0.5979020979020979,
40
+ "eval_loss": 0.26823195815086365,
41
+ "eval_macro_f1": 0.7588550972474853,
42
+ "eval_precision": 0.9065499982342927,
43
+ "eval_recall": 0.8886423239862821,
44
+ "eval_runtime": 8.5862,
45
+ "eval_samples_per_second": 577.325,
46
+ "eval_steps_per_second": 9.084,
47
+ "step": 155
48
+ },
49
+ {
50
+ "epoch": 1.2903225806451613,
51
+ "grad_norm": 2.4748122692108154,
52
+ "learning_rate": 1.8529927230890757e-05,
53
+ "loss": 0.501370506286621,
54
+ "step": 200
55
+ },
56
+ {
57
+ "epoch": 1.6129032258064515,
58
+ "grad_norm": 2.85589599609375,
59
+ "learning_rate": 1.715018302529852e-05,
60
+ "loss": 0.4371611404418945,
61
+ "step": 250
62
+ },
63
+ {
64
+ "epoch": 1.935483870967742,
65
+ "grad_norm": 2.5343403816223145,
66
+ "learning_rate": 1.5409849862969994e-05,
67
+ "loss": 0.44288436889648436,
68
+ "step": 300
69
+ },
70
+ {
71
+ "epoch": 2.0,
72
+ "eval_accuracy": 0.8912648779503732,
73
+ "eval_f1": 0.8998036744615823,
74
+ "eval_hate_f1": 0.4732620320855615,
75
+ "eval_hate_prec": 0.38311688311688313,
76
+ "eval_hate_rec": 0.6188811188811189,
77
+ "eval_loss": 0.2427479773759842,
78
+ "eval_macro_f1": 0.7651698753431324,
79
+ "eval_precision": 0.913809087207257,
80
+ "eval_recall": 0.8912648779503732,
81
+ "eval_runtime": 8.6774,
82
+ "eval_samples_per_second": 571.253,
83
+ "eval_steps_per_second": 8.989,
84
+ "step": 310
85
+ },
86
+ {
87
+ "epoch": 2.258064516129032,
88
+ "grad_norm": 6.783071994781494,
89
+ "learning_rate": 1.3396694017461708e-05,
90
+ "loss": 0.4083076477050781,
91
+ "step": 350
92
+ },
93
+ {
94
+ "epoch": 2.5806451612903225,
95
+ "grad_norm": 4.064752578735352,
96
+ "learning_rate": 1.1212240409675825e-05,
97
+ "loss": 0.36163642883300784,
98
+ "step": 400
99
+ },
100
+ {
101
+ "epoch": 2.903225806451613,
102
+ "grad_norm": 3.120911121368408,
103
+ "learning_rate": 8.966652631884506e-06,
104
+ "loss": 0.3937630081176758,
105
+ "step": 450
106
+ },
107
+ {
108
+ "epoch": 3.0,
109
+ "eval_accuracy": 0.8912648779503732,
110
+ "eval_f1": 0.8998510802952099,
111
+ "eval_hate_f1": 0.48548812664907653,
112
+ "eval_hate_prec": 0.3898305084745763,
113
+ "eval_hate_rec": 0.6433566433566433,
114
+ "eval_loss": 0.24792158603668213,
115
+ "eval_macro_f1": 0.7698383516941082,
116
+ "eval_precision": 0.9136908950590453,
117
+ "eval_recall": 0.8912648779503732,
118
+ "eval_runtime": 8.7831,
119
+ "eval_samples_per_second": 564.382,
120
+ "eval_steps_per_second": 8.881,
121
+ "step": 465
122
+ },
123
+ {
124
+ "epoch": 3.225806451612903,
125
+ "grad_norm": 2.958275556564331,
126
+ "learning_rate": 6.773177317242257e-06,
127
+ "loss": 0.3346284103393555,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 3.5483870967741935,
132
+ "grad_norm": 2.155022144317627,
133
+ "learning_rate": 4.742433029292856e-06,
134
+ "loss": 0.3073293113708496,
135
+ "step": 550
136
+ },
137
+ {
138
+ "epoch": 3.870967741935484,
139
+ "grad_norm": 3.33176589012146,
140
+ "learning_rate": 2.976831686947884e-06,
141
+ "loss": 0.31076955795288086,
142
+ "step": 600
143
+ },
144
+ {
145
+ "epoch": 4.0,
146
+ "eval_accuracy": 0.8809763970143232,
147
+ "eval_f1": 0.8935393778095193,
148
+ "eval_hate_f1": 0.464839094159714,
149
+ "eval_hate_prec": 0.352622061482821,
150
+ "eval_hate_rec": 0.6818181818181818,
151
+ "eval_loss": 0.2574857175350189,
152
+ "eval_macro_f1": 0.760244573017305,
153
+ "eval_precision": 0.9147072848857565,
154
+ "eval_recall": 0.8809763970143232,
155
+ "eval_runtime": 8.7972,
156
+ "eval_samples_per_second": 563.474,
157
+ "eval_steps_per_second": 8.866,
158
+ "step": 620
159
+ },
160
+ {
161
+ "epoch": 4.193548387096774,
162
+ "grad_norm": 3.568920612335205,
163
+ "learning_rate": 1.565413856553095e-06,
164
+ "loss": 0.3023691749572754,
165
+ "step": 650
166
+ },
167
+ {
168
+ "epoch": 4.516129032258064,
169
+ "grad_norm": 3.1700286865234375,
170
+ "learning_rate": 5.79358371039106e-07,
171
+ "loss": 0.2629020118713379,
172
+ "step": 700
173
+ },
174
+ {
175
+ "epoch": 4.838709677419355,
176
+ "grad_norm": 3.6812124252319336,
177
+ "learning_rate": 6.839272934511143e-08,
178
+ "loss": 0.27848766326904295,
179
+ "step": 750
180
+ },
181
+ {
182
+ "epoch": 5.0,
183
+ "eval_accuracy": 0.878353843050232,
184
+ "eval_f1": 0.8912320589311761,
185
+ "eval_hate_f1": 0.4591715976331361,
186
+ "eval_hate_prec": 0.3470483005366726,
187
+ "eval_hate_rec": 0.6783216783216783,
188
+ "eval_loss": 0.2623799443244934,
189
+ "eval_macro_f1": 0.7572637429327783,
190
+ "eval_precision": 0.9126253015348699,
191
+ "eval_recall": 0.878353843050232,
192
+ "eval_runtime": 8.7768,
193
+ "eval_samples_per_second": 564.787,
194
+ "eval_steps_per_second": 8.887,
195
+ "step": 775
196
+ }
197
+ ],
198
+ "logging_steps": 50,
199
+ "max_steps": 775,
200
+ "num_input_tokens_seen": 0,
201
+ "num_train_epochs": 5,
202
+ "save_steps": 500,
203
+ "stateful_callbacks": {
204
+ "EarlyStoppingCallback": {
205
+ "args": {
206
+ "early_stopping_patience": 2,
207
+ "early_stopping_threshold": 0.0
208
+ },
209
+ "attributes": {
210
+ "early_stopping_patience_counter": 2
211
+ }
212
+ },
213
+ "TrainerControl": {
214
+ "args": {
215
+ "should_epoch_stop": false,
216
+ "should_evaluate": false,
217
+ "should_log": false,
218
+ "should_save": true,
219
+ "should_training_stop": true
220
+ },
221
+ "attributes": {}
222
+ }
223
+ },
224
+ "total_flos": 2445228103066560.0,
225
+ "train_batch_size": 64,
226
+ "trial_name": null,
227
+ "trial_params": null
228
+ }
checkpoint-775/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b0eee9e6e1a3712626f958410bea2b5a2002e579aab47745ad3abf8793b4938b
3
+ size 5201
config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "RobertaForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": null,
9
+ "dtype": "float32",
10
+ "eos_token_id": 2,
11
+ "gradient_checkpointing": false,
12
+ "hidden_act": "gelu",
13
+ "hidden_dropout_prob": 0.1,
14
+ "hidden_size": 768,
15
+ "id2label": {
16
+ "0": "LABEL_0",
17
+ "1": "LABEL_1",
18
+ "2": "LABEL_2"
19
+ },
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "is_decoder": false,
23
+ "label2id": {
24
+ "LABEL_0": 0,
25
+ "LABEL_1": 1,
26
+ "LABEL_2": 2
27
+ },
28
+ "layer_norm_eps": 1e-05,
29
+ "max_position_embeddings": 514,
30
+ "model_type": "roberta",
31
+ "num_attention_heads": 12,
32
+ "num_hidden_layers": 12,
33
+ "pad_token_id": 1,
34
+ "tie_word_embeddings": true,
35
+ "transformers_version": "5.2.0",
36
+ "type_vocab_size": 1,
37
+ "use_cache": false,
38
+ "vocab_size": 50265
39
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a51961eb1fe4d73716ad778fef4daf24ba1cd8312926aba06364162a90c7b512
3
+ size 498615876
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<s>",
5
+ "cls_token": "<s>",
6
+ "eos_token": "</s>",
7
+ "errors": "replace",
8
+ "is_local": false,
9
+ "mask_token": "<mask>",
10
+ "model_max_length": 1000000000000000019884624838656,
11
+ "pad_token": "<pad>",
12
+ "sep_token": "</s>",
13
+ "tokenizer_class": "RobertaTokenizer",
14
+ "trim_offsets": true,
15
+ "unk_token": "<unk>"
16
+ }