Hugging Face's logo

Hatman
/
audio-emotion-detection

Audio Classification
Transformers
Safetensors
wav2vec2
Inference Endpoints
Model card Files Community
hatmanstack commited on
Commit
95f7c25
•
1 Parent(s): 489aaad

Initial

Browse files
Files changed (4) hide show
  1. README.md +48 -0
  2. config.json +134 -0
  3. model.safetensors +3 -0
  4. preprocessor_config.json +10 -0
README.md CHANGED
@@ -1,3 +1,51 @@
1
  ---
2
  license: apache-2.0
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: apache-2.0
3
+ base_model: facebook/wav2vec2-large-xlsr-53
4
+ metrics:
5
+ - accuracy
6
+ model-index:
7
+ - name: audio-emotion-detection
8
+ results: []
9
  ---
10
+
11
+ # Audio Emotion Detection
12
+
13
+ This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53).
14
+
15
+ It achieves the following results on the evaluation set:
16
+ - Loss: 0.9555
17
+ - Accuracy: 0.6262
18
+
19
+ ## Model description
20
+
21
+ A model that returns Labels for Angry, Disgusted, Fearful, Happy, Neutral, Sad, Suprised. All aduio was trained at a sampling rate of 16000 and all inputs should be transformed to work properly.
22
+
23
+ ## Training and evaluation data
24
+
25
+ - **mozilla-foundation/common_voice_6_0**
26
+ - **speech-recognition-community-v2/dev_data**
27
+
28
+ ### Training hyperparameters
29
+
30
+ The following hyperparameters were used during training:
31
+ - learning_rate: 0.0005
32
+ - train_batch_size: 32
33
+ - eval_batch_size: 32
34
+ - seed: 42
35
+ - gradient_accumulation_steps: 8
36
+ - total_train_batch_size: 256
37
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
38
+ - lr_scheduler_type: linear
39
+ - lr_scheduler_warmup_ratio: 0.01
40
+ - num_epochs: 4
41
+
42
+ ### Training results
43
+
44
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
45
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
46
+ | 1.5875 | 1.0 | 40 | 1.2574 | 0.5133 |
47
+ | 1.1637 | 2.0 | 80 | 1.0852 | 0.5590 |
48
+ | 0.9827 | 3.0 | 120 | 1.0048 | 0.6090 |
49
+ | 0.8683 | 4.0 | 160 | 0.9555 | 0.6262 |
50
+
51
+
config.json ADDED
@@ -0,0 +1,134 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "jonatasgrosman/wav2vec2-large-xlsr-53-english",
3
+ "activation_dropout": 0.05,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForSequenceClassification"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 256,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "mean",
46
+ "ctc_zero_infinity": true,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.05,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.05,
58
+ "hidden_size": 1024,
59
+ "id2label": {
60
+ "0": "Angry",
61
+ "1": "Disgusted",
62
+ "2": "Fearful",
63
+ "3": "Happy",
64
+ "4": "Neutral",
65
+ "5": "Sad",
66
+ "6": "Suprised"
67
+ },
68
+ "initializer_range": 0.02,
69
+ "intermediate_size": 4096,
70
+ "label2id": {
71
+ "Angry": "0",
72
+ "Disgusted": "1",
73
+ "Fearful": "2",
74
+ "Happy": "3",
75
+ "Neutral": "4",
76
+ "Sad": "5",
77
+ "Suprised": "6"
78
+ },
79
+ "layer_norm_eps": 1e-05,
80
+ "layerdrop": 0.05,
81
+ "mask_channel_length": 10,
82
+ "mask_channel_min_space": 1,
83
+ "mask_channel_other": 0.0,
84
+ "mask_channel_prob": 0.0,
85
+ "mask_channel_selection": "static",
86
+ "mask_feature_length": 10,
87
+ "mask_feature_min_masks": 0,
88
+ "mask_feature_prob": 0.0,
89
+ "mask_time_length": 10,
90
+ "mask_time_min_masks": 2,
91
+ "mask_time_min_space": 1,
92
+ "mask_time_other": 0.0,
93
+ "mask_time_prob": 0.05,
94
+ "mask_time_selection": "static",
95
+ "model_type": "wav2vec2",
96
+ "num_adapter_layers": 3,
97
+ "num_attention_heads": 16,
98
+ "num_codevector_groups": 2,
99
+ "num_codevectors_per_group": 320,
100
+ "num_conv_pos_embedding_groups": 16,
101
+ "num_conv_pos_embeddings": 128,
102
+ "num_feat_extract_layers": 7,
103
+ "num_hidden_layers": 24,
104
+ "num_negatives": 100,
105
+ "output_hidden_size": 1024,
106
+ "pad_token_id": 0,
107
+ "proj_codevector_dim": 256,
108
+ "tdnn_dilation": [
109
+ 1,
110
+ 2,
111
+ 3,
112
+ 1,
113
+ 1
114
+ ],
115
+ "tdnn_dim": [
116
+ 512,
117
+ 512,
118
+ 512,
119
+ 512,
120
+ 1500
121
+ ],
122
+ "tdnn_kernel": [
123
+ 5,
124
+ 3,
125
+ 3,
126
+ 1,
127
+ 1
128
+ ],
129
+ "torch_dtype": "float32",
130
+ "transformers_version": "4.39.3",
131
+ "use_weighted_layer_sum": false,
132
+ "vocab_size": 33,
133
+ "xvector_output_dim": 512
134
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f4768b4e6d1036cbf2aab3a212749a6e04031ed39adb6c1d4656c29e14f6fec
3
+ size 1262864444
preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "processor_class": "Wav2Vec2ProcessorWithLM",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }