Hugging Face's logo

arslanarjumand
/
wav2vec-read_aloud

Audio Classification
Transformers
Safetensors
wav2vec2-bert
Generated from Trainer
Inference Endpoints
Model card Files Community
arslanarjumand commited on
Commit
68b1505
•
1 Parent(s): e403151

arslanarjumand/wav2vec-read-aloud

Browse files
Files changed (4) hide show
  1. README.md +15 -15
  2. config.json +4 -4
  3. model.safetensors +2 -2
  4. training_args.bin +1 -1
README.md CHANGED
@@ -15,11 +15,11 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 973.4864
19
- - Pcc Accuracy: 0.7547
20
- - Pcc Fluency: 0.7664
21
- - Pcc Total Score: 0.8143
22
- - Pcc Content: nan
23
 
24
  ## Model description
25
 
@@ -38,7 +38,7 @@ More information needed
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
- - learning_rate: 5.5e-05
42
  - train_batch_size: 2
43
  - eval_batch_size: 6
44
  - seed: 42
@@ -46,20 +46,20 @@ The following hyperparameters were used during training:
46
  - total_train_batch_size: 24
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: cosine
49
- - lr_scheduler_warmup_ratio: 0.4
50
- - num_epochs: 15
51
 
52
  ### Training results
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Pcc Accuracy | Pcc Fluency | Pcc Total Score | Pcc Content |
55
  |:-------------:|:-----:|:----:|:---------------:|:------------:|:-----------:|:---------------:|:-----------:|
56
- | 2390.3109 | 1.95 | 500 | 2342.6951 | nan | 0.4815 | nan | nan |
57
- | 2164.6891 | 3.9 | 1000 | 2318.7217 | nan | 0.6461 | nan | nan |
58
- | 1078.8019 | 5.85 | 1500 | 1029.2085 | 0.6188 | 0.7014 | 0.6845 | nan |
59
- | 974.6556 | 7.8 | 2000 | 985.5543 | 0.7117 | 0.7355 | 0.7743 | nan |
60
- | 1002.623 | 9.75 | 2500 | 989.1628 | 0.7401 | 0.7533 | 0.7995 | nan |
61
- | 947.5643 | 11.7 | 3000 | 972.3806 | 0.7507 | 0.7628 | 0.8103 | nan |
62
- | 995.6286 | 13.65 | 3500 | 973.4864 | 0.7547 | 0.7664 | 0.8143 | nan |
63
 
64
 
65
  ### Framework versions
 
15
 
16
  This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Loss: 0.1115
19
+ - Pcc Accuracy: 0.7918
20
+ - Pcc Fluency: 0.7940
21
+ - Pcc Total Score: 0.8472
22
+ - Pcc Content: 0.8160
23
 
24
  ## Model description
25
 
 
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
+ - learning_rate: 0.00055
42
  - train_batch_size: 2
43
  - eval_batch_size: 6
44
  - seed: 42
 
46
  - total_train_batch_size: 24
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: cosine
49
+ - lr_scheduler_warmup_ratio: 0.25
50
+ - num_epochs: 14
51
 
52
  ### Training results
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Pcc Accuracy | Pcc Fluency | Pcc Total Score | Pcc Content |
55
  |:-------------:|:-----:|:----:|:---------------:|:------------:|:-----------:|:---------------:|:-----------:|
56
+ | 0.1483 | 1.94 | 500 | 0.1659 | 0.7256 | 0.6982 | 0.7616 | 0.7480 |
57
+ | 0.1338 | 3.89 | 1000 | 0.1369 | 0.7706 | 0.7680 | 0.8154 | 0.7835 |
58
+ | 0.124 | 5.83 | 1500 | 0.1754 | 0.6686 | 0.6459 | 0.7110 | 0.6823 |
59
+ | 0.1147 | 7.77 | 2000 | 0.1149 | 0.7838 | 0.7848 | 0.8368 | 0.8048 |
60
+ | 0.1024 | 9.72 | 2500 | 0.1135 | 0.7802 | 0.7819 | 0.8340 | 0.8048 |
61
+ | 0.0945 | 11.66 | 3000 | 0.1168 | 0.7891 | 0.7876 | 0.8418 | 0.8095 |
62
+ | 0.0945 | 13.61 | 3500 | 0.1115 | 0.7918 | 0.7940 | 0.8472 | 0.8160 |
63
 
64
 
65
  ### Framework versions
config.json CHANGED
@@ -11,7 +11,7 @@
11
  ],
12
  "attention_dropout": 0.0094,
13
  "bos_token_id": 1,
14
- "classifier_proj_size": 128,
15
  "codevector_dim": 768,
16
  "conformer_conv_dropout": 0.1,
17
  "contrastive_logits_temperature": 0.1,
@@ -46,17 +46,17 @@
46
  "left_max_position_embeddings": 64,
47
  "mask_feature_length": 5,
48
  "mask_feature_min_masks": 5,
49
- "mask_feature_prob": 0.0575,
50
  "mask_time_length": 5,
51
  "mask_time_min_masks": 5,
52
- "mask_time_prob": 0.0585,
53
  "max_source_positions": 5000,
54
  "model_type": "wav2vec2-bert",
55
  "num_adapter_layers": 1,
56
  "num_attention_heads": 16,
57
  "num_codevector_groups": 2,
58
  "num_codevectors_per_group": 320,
59
- "num_hidden_layers": 12,
60
  "num_negatives": 100,
61
  "output_hidden_size": 1024,
62
  "pad_token_id": 0,
 
11
  ],
12
  "attention_dropout": 0.0094,
13
  "bos_token_id": 1,
14
+ "classifier_proj_size": 64,
15
  "codevector_dim": 768,
16
  "conformer_conv_dropout": 0.1,
17
  "contrastive_logits_temperature": 0.1,
 
46
  "left_max_position_embeddings": 64,
47
  "mask_feature_length": 5,
48
  "mask_feature_min_masks": 5,
49
+ "mask_feature_prob": 0.00575,
50
  "mask_time_length": 5,
51
  "mask_time_min_masks": 5,
52
+ "mask_time_prob": 0.00585,
53
  "max_source_positions": 5000,
54
  "model_type": "wav2vec2-bert",
55
  "num_adapter_layers": 1,
56
  "num_attention_heads": 16,
57
  "num_codevector_groups": 2,
58
  "num_codevectors_per_group": 320,
59
+ "num_hidden_layers": 10,
60
  "num_negatives": 100,
61
  "output_hidden_size": 1024,
62
  "pad_token_id": 0,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d4660acd2a4de557833aa1ce4fbd7ab93333e7c01cf645ab396a4e3a220db11a
3
- size 1206994592
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51cea6e9fe9f1cd5e680f7f8cb4412cd742b203bb3c39686b62ab7ab1d3af4ca
3
+ size 1000576520
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:596191023888232fe0a588808a22c46817c7b68e607063d9fa88732a803e892a
3
  size 4728
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:602f8283ebf51735a04449e81f60a8b79bbd12594166992e39d583a5f9b0d583
3
  size 4728