Hugging Face's logo

navidved
/
ExHubert-fine-tuned-persian_v2

Audio Classification
Transformers
TensorBoard
Safetensors
hubert
Generated from Trainer
custom_code
Model card Files Metrics Community
navidved commited on
Commit
dfc2951
·
verified ·
1 Parent(s): add9d17

navidved/ExHubert-fine-tuned-persian_v2

Browse files
Files changed (6) hide show
  1. README.md +83 -0
  2. config.json +95 -0
  3. model.safetensors +3 -0
  4. preprocessor_config.json +9 -0
  5. runs/Jan04_10-59-20_9e601cdb5e9b/events.out.tfevents.1735988362.9e601cdb5e9b.499.0 +3 -0
  6. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: cc-by-nc-sa-4.0
4
+ base_model: amiriparian/ExHuBERT
5
+ tags:
6
+ - generated_from_trainer
7
+ metrics:
8
+ - accuracy
9
+ - precision
10
+ - recall
11
+ - f1
12
+ model-index:
13
+ - name: ExHubert-fine-tuned-persian_v2
14
+ results: []
15
+ ---
16
+
17
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
18
+ should probably proofread and complete it, then remove this comment. -->
19
+
20
+ # ExHubert-fine-tuned-persian_v2
21
+
22
+ This model is a fine-tuned version of [amiriparian/ExHuBERT](https://huggingface.co/amiriparian/ExHuBERT) on an unknown dataset.
23
+ It achieves the following results on the evaluation set:
24
+ - Loss: 0.6595
25
+ - Accuracy: 0.8160
26
+ - Precision: 0.8627
27
+ - Recall: 0.6471
28
+ - F1: 0.7395
29
+ - Precision Neutral: 0.7957
30
+ - Recall Neutral: 0.9303
31
+ - F1 Neutral: 0.8578
32
+ - Precision Anger: 0.8627
33
+ - Recall Anger: 0.6471
34
+ - F1 Anger: 0.7395
35
+
36
+ ## Model description
37
+
38
+ More information needed
39
+
40
+ ## Intended uses & limitations
41
+
42
+ More information needed
43
+
44
+ ## Training and evaluation data
45
+
46
+ More information needed
47
+
48
+ ## Training procedure
49
+
50
+ ### Training hyperparameters
51
+
52
+ The following hyperparameters were used during training:
53
+ - learning_rate: 2e-05
54
+ - train_batch_size: 4
55
+ - eval_batch_size: 4
56
+ - seed: 42
57
+ - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
58
+ - lr_scheduler_type: cosine
59
+ - num_epochs: 10
60
+ - mixed_precision_training: Native AMP
61
+
62
+ ### Training results
63
+
64
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Precision Neutral | Recall Neutral | F1 Neutral | Precision Anger | Recall Anger | F1 Anger |
65
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:-----------------:|:--------------:|:----------:|:---------------:|:------------:|:--------:|
66
+ | 0.6249 | 1.0 | 337 | 0.5633 | 0.7953 | 0.8602 | 0.5882 | 0.6987 | 0.7705 | 0.9353 | 0.8449 | 0.8602 | 0.5882 | 0.6987 |
67
+ | 0.4361 | 2.0 | 674 | 0.6575 | 0.7923 | 0.8511 | 0.5882 | 0.6957 | 0.7695 | 0.9303 | 0.8423 | 0.8511 | 0.5882 | 0.6957 |
68
+ | 0.3807 | 3.0 | 1011 | 0.6459 | 0.7923 | 0.6737 | 0.9412 | 0.7853 | 0.9456 | 0.6915 | 0.7989 | 0.6737 | 0.9412 | 0.7853 |
69
+ | 0.3778 | 4.0 | 1348 | 0.5804 | 0.8042 | 0.9070 | 0.5735 | 0.7027 | 0.7689 | 0.9602 | 0.8540 | 0.9070 | 0.5735 | 0.7027 |
70
+ | 0.3353 | 5.0 | 1685 | 0.6511 | 0.7745 | 0.8846 | 0.5074 | 0.6449 | 0.7413 | 0.9552 | 0.8348 | 0.8846 | 0.5074 | 0.6449 |
71
+ | 0.3446 | 6.0 | 2022 | 0.5828 | 0.8042 | 0.7273 | 0.8235 | 0.7724 | 0.8689 | 0.7910 | 0.8281 | 0.7273 | 0.8235 | 0.7724 |
72
+ | 0.2655 | 7.0 | 2359 | 0.6738 | 0.7953 | 0.8681 | 0.5809 | 0.6960 | 0.7683 | 0.9403 | 0.8456 | 0.8681 | 0.5809 | 0.6960 |
73
+ | 0.2426 | 8.0 | 2696 | 0.6541 | 0.8101 | 0.8158 | 0.6838 | 0.744 | 0.8072 | 0.8955 | 0.8491 | 0.8158 | 0.6838 | 0.744 |
74
+ | 0.1998 | 9.0 | 3033 | 0.6532 | 0.8190 | 0.8641 | 0.6544 | 0.7448 | 0.7991 | 0.9303 | 0.8598 | 0.8641 | 0.6544 | 0.7448 |
75
+ | 0.3289 | 10.0 | 3370 | 0.6595 | 0.8160 | 0.8627 | 0.6471 | 0.7395 | 0.7957 | 0.9303 | 0.8578 | 0.8627 | 0.6471 | 0.7395 |
76
+
77
+
78
+ ### Framework versions
79
+
80
+ - Transformers 4.47.1
81
+ - Pytorch 2.5.1+cu121
82
+ - Datasets 3.2.0
83
+ - Tokenizers 0.21.0
config.json ADDED
@@ -0,0 +1,95 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "amiriparian/ExHuBERT",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "ExHuBERT"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "auto_map": {
10
+ "AutoModelForAudioClassification": "amiriparian/ExHuBERT--ExHuBERT_model.ExHuBERT"
11
+ },
12
+ "bos_token_id": 1,
13
+ "classifier_proj_size": 256,
14
+ "conv_bias": true,
15
+ "conv_dim": [
16
+ 512,
17
+ 512,
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512
23
+ ],
24
+ "conv_kernel": [
25
+ 10,
26
+ 3,
27
+ 3,
28
+ 3,
29
+ 3,
30
+ 2,
31
+ 2
32
+ ],
33
+ "conv_stride": [
34
+ 5,
35
+ 2,
36
+ 2,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2
41
+ ],
42
+ "ctc_loss_reduction": "sum",
43
+ "ctc_zero_infinity": false,
44
+ "diversity_loss_weight": 0.1,
45
+ "do_stable_layer_norm": true,
46
+ "eos_token_id": 2,
47
+ "feat_extract_activation": "gelu",
48
+ "feat_extract_dropout": 0.0,
49
+ "feat_extract_norm": "layer",
50
+ "feat_proj_dropout": 0.1,
51
+ "feat_proj_layer_norm": true,
52
+ "final_dropout": 0.1,
53
+ "gradient_checkpointing": false,
54
+ "hidden_act": "gelu",
55
+ "hidden_dropout": 0.1,
56
+ "hidden_dropout_prob": 0.1,
57
+ "hidden_size": 1024,
58
+ "id2label": {
59
+ "0": "LABEL_0",
60
+ "1": "LABEL_1",
61
+ "2": "LABEL_2",
62
+ "3": "LABEL_3",
63
+ "4": "LABEL_4",
64
+ "5": "LABEL_5"
65
+ },
66
+ "initializer_range": 0.02,
67
+ "intermediate_size": 4096,
68
+ "label2id": {
69
+ "LABEL_0": 0,
70
+ "LABEL_1": 1,
71
+ "LABEL_2": 2,
72
+ "LABEL_3": 3,
73
+ "LABEL_4": 4,
74
+ "LABEL_5": 5
75
+ },
76
+ "layer_norm_eps": 1e-05,
77
+ "layerdrop": 0.1,
78
+ "mask_feature_length": 10,
79
+ "mask_feature_min_masks": 0,
80
+ "mask_feature_prob": 0.0,
81
+ "mask_time_length": 10,
82
+ "mask_time_min_masks": 2,
83
+ "mask_time_prob": 0.05,
84
+ "model_type": "hubert",
85
+ "num_attention_heads": 16,
86
+ "num_conv_pos_embedding_groups": 16,
87
+ "num_conv_pos_embeddings": 128,
88
+ "num_feat_extract_layers": 7,
89
+ "num_hidden_layers": 48,
90
+ "pad_token_id": 0,
91
+ "torch_dtype": "float32",
92
+ "transformers_version": "4.47.1",
93
+ "use_weighted_layer_sum": false,
94
+ "vocab_size": 32
95
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d62830173116e0c9773ebc762f9c1422e668880fe3ab84b10c47140adf9f3e0
3
+ size 2673678432
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/Jan04_10-59-20_9e601cdb5e9b/events.out.tfevents.1735988362.9e601cdb5e9b.499.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b68d8fa8564cc2e4b44fc99af0535aef647e3ae9557d17497410f13c61983dd
3
+ size 85912
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:678309c84afd05d761e8c12ab20ced207eaa6195326e89d2e1b75ac3796f6fce
3
+ size 5432