abhishtagatya commited on
Commit
0f16e00
1 Parent(s): ffad0c6

End of training

Browse files
README.md ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: facebook/hubert-base-ls960
4
+ tags:
5
+ - audio-classification
6
+ - deepfake
7
+ - audio-spoof
8
+ - generated_from_trainer
9
+ metrics:
10
+ - accuracy
11
+ model-index:
12
+ - name: hubert-base-960h-itw-deepfake
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # hubert-base-960h-itw-deepfake
20
+
21
+ This model is a fine-tuned version of [facebook/hubert-base-ls960](https://huggingface.co/facebook/hubert-base-ls960) on an unknown dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 0.0756
24
+ - Accuracy: 0.9873
25
+ - Far: 0.0083
26
+ - Frr: 0.0203
27
+ - Eer: 0.0143
28
+
29
+ ## Model description
30
+
31
+ More information needed
32
+
33
+ ## Intended uses & limitations
34
+
35
+ More information needed
36
+
37
+ ## Training and evaluation data
38
+
39
+ More information needed
40
+
41
+ ## Training procedure
42
+
43
+ ### Training hyperparameters
44
+
45
+ The following hyperparameters were used during training:
46
+ - learning_rate: 1e-06
47
+ - train_batch_size: 2
48
+ - eval_batch_size: 2
49
+ - seed: 42
50
+ - gradient_accumulation_steps: 2
51
+ - total_train_batch_size: 4
52
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
53
+ - lr_scheduler_type: linear
54
+ - num_epochs: 2.0
55
+
56
+ ### Training results
57
+
58
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy | Far | Frr | Eer |
59
+ |:-------------:|:-----:|:-----:|:---------------:|:--------:|:------:|:------:|:------:|
60
+ | 0.4081 | 0.39 | 2500 | 0.1152 | 0.9722 | 0.0285 | 0.0267 | 0.0276 |
61
+ | 0.1168 | 0.79 | 5000 | 0.0822 | 0.9844 | 0.0120 | 0.0216 | 0.0168 |
62
+ | 0.0979 | 1.18 | 7500 | 0.0896 | 0.9846 | 0.0130 | 0.0195 | 0.0162 |
63
+ | 0.0983 | 1.57 | 10000 | 0.1007 | 0.9833 | 0.0155 | 0.0186 | 0.0171 |
64
+ | 0.0901 | 1.97 | 12500 | 0.0756 | 0.9873 | 0.0083 | 0.0203 | 0.0143 |
65
+
66
+
67
+ ### Framework versions
68
+
69
+ - Transformers 4.38.0.dev0
70
+ - Pytorch 2.1.2+cu121
71
+ - Datasets 2.16.2.dev0
72
+ - Tokenizers 0.15.1
config.json ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/hubert-base-ls960",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "HubertForSequenceClassification"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 1,
10
+ "classifier_proj_size": 256,
11
+ "conv_bias": false,
12
+ "conv_dim": [
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512,
19
+ 512
20
+ ],
21
+ "conv_kernel": [
22
+ 10,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 3,
27
+ 2,
28
+ 2
29
+ ],
30
+ "conv_stride": [
31
+ 5,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2,
37
+ 2
38
+ ],
39
+ "ctc_loss_reduction": "sum",
40
+ "ctc_zero_infinity": false,
41
+ "do_stable_layer_norm": false,
42
+ "eos_token_id": 2,
43
+ "feat_extract_activation": "gelu",
44
+ "feat_extract_dropout": 0.0,
45
+ "feat_extract_norm": "group",
46
+ "feat_proj_dropout": 0.1,
47
+ "feat_proj_layer_norm": true,
48
+ "final_dropout": 0.1,
49
+ "finetuning_task": "hubert_clf",
50
+ "gradient_checkpointing": false,
51
+ "hidden_act": "gelu",
52
+ "hidden_dropout": 0.1,
53
+ "hidden_dropout_prob": 0.1,
54
+ "hidden_size": 768,
55
+ "id2label": {
56
+ "0": "bona-fide",
57
+ "1": "spoof"
58
+ },
59
+ "initializer_range": 0.02,
60
+ "intermediate_size": 3072,
61
+ "label2id": {
62
+ "bona-fide": 0,
63
+ "spoof": 1
64
+ },
65
+ "layer_norm_eps": 1e-05,
66
+ "layerdrop": 0.1,
67
+ "mask_feature_length": 10,
68
+ "mask_feature_min_masks": 0,
69
+ "mask_feature_prob": 0.0,
70
+ "mask_time_length": 10,
71
+ "mask_time_min_masks": 2,
72
+ "mask_time_prob": 0.05,
73
+ "model_type": "hubert",
74
+ "num_attention_heads": 12,
75
+ "num_conv_pos_embedding_groups": 16,
76
+ "num_conv_pos_embeddings": 128,
77
+ "num_feat_extract_layers": 7,
78
+ "num_hidden_layers": 12,
79
+ "pad_token_id": 0,
80
+ "pooling_mode": "mean",
81
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
82
+ "torch_dtype": "float32",
83
+ "transformers_version": "4.38.0.dev0",
84
+ "use_weighted_layer_sum": false,
85
+ "vocab_size": 32
86
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:16fb8cf389110d097a934a5c50da468fce96e7bbd3e2ffb0a08aaa5d0f02548c
3
+ size 378301944
preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "return_attention_mask": false,
8
+ "sampling_rate": 16000
9
+ }
runs/Jan25_18-36-59_38073bf969c7/events.out.tfevents.1706207819.38073bf969c7.34438.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:566b52ed6d4f86673b09cd4a12f2d77628d10506c240a9d5975d292da9d8a006
3
+ size 8983
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e16a8a37aa2a1ee4c34691d53591cbeac09b5f42d2ddd6108722183ed14db0f4
3
+ size 4792