Hugging Face's logo

bigmorning
/
whisper_input_decoder_shift_r_labels_no_force__0040

Automatic Speech Recognition
Transformers
TensorFlow
whisper
generated_from_keras_callback
Inference Endpoints
Model card Files Community
bigmorning commited on
Commit
9e79181
•
1 Parent(s): b83a3a9

Upload TFWhisperForConditionalGeneration

Browse files
Files changed (4) hide show
  1. README.md +96 -0
  2. config.json +51 -0
  3. generation_config.json +117 -0
  4. tf_model.h5 +3 -0
README.md ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: openai/whisper-tiny
4
+ tags:
5
+ - generated_from_keras_callback
6
+ model-index:
7
+ - name: whisper_input_decoder_shift_r_labels_no_force__0040
8
+ results: []
9
+ ---
10
+
11
+ <!-- This model card has been generated automatically according to the information Keras had access to. You should
12
+ probably proofread and complete it, then remove this comment. -->
13
+
14
+ # whisper_input_decoder_shift_r_labels_no_force__0040
15
+
16
+ This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on an unknown dataset.
17
+ It achieves the following results on the evaluation set:
18
+ - Train Loss: 0.2291
19
+ - Train Accuracy: 0.0324
20
+ - Train Wermet: 0.0706
21
+ - Validation Loss: 0.7876
22
+ - Validation Accuracy: 0.0206
23
+ - Validation Wermet: 0.2755
24
+ - Epoch: 39
25
+
26
+ ## Model description
27
+
28
+ More information needed
29
+
30
+ ## Intended uses & limitations
31
+
32
+ More information needed
33
+
34
+ ## Training and evaluation data
35
+
36
+ More information needed
37
+
38
+ ## Training procedure
39
+
40
+ ### Training hyperparameters
41
+
42
+ The following hyperparameters were used during training:
43
+ - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': 1e-05, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False, 'weight_decay_rate': 0.01}
44
+ - training_precision: float32
45
+
46
+ ### Training results
47
+
48
+ | Train Loss | Train Accuracy | Train Wermet | Validation Loss | Validation Accuracy | Validation Wermet | Epoch |
49
+ |:----------:|:--------------:|:------------:|:---------------:|:-------------------:|:-----------------:|:-----:|
50
+ | 5.6348 | 0.0091 | 1.5865 | 4.2935 | 0.0093 | 0.9579 | 0 |
51
+ | 4.9212 | 0.0099 | 0.9054 | 4.1262 | 0.0097 | 0.9390 | 1 |
52
+ | 4.6819 | 0.0107 | 0.8319 | 3.9071 | 0.0103 | 0.8966 | 2 |
53
+ | 4.4443 | 0.0114 | 0.8310 | 3.7367 | 0.0106 | 0.8939 | 3 |
54
+ | 4.2479 | 0.0119 | 0.8226 | 3.6101 | 0.0109 | 0.8696 | 4 |
55
+ | 4.0911 | 0.0124 | 0.8103 | 3.5364 | 0.0110 | 0.8946 | 5 |
56
+ | 3.9590 | 0.0127 | 0.7913 | 3.4556 | 0.0113 | 0.8388 | 6 |
57
+ | 3.8513 | 0.0130 | 0.7794 | 3.4106 | 0.0114 | 0.8515 | 7 |
58
+ | 3.7607 | 0.0133 | 0.7657 | 3.3507 | 0.0115 | 0.8261 | 8 |
59
+ | 3.6757 | 0.0136 | 0.7548 | 3.3141 | 0.0116 | 0.8400 | 9 |
60
+ | 3.6023 | 0.0138 | 0.7454 | 3.2711 | 0.0117 | 0.8006 | 10 |
61
+ | 3.5261 | 0.0140 | 0.7348 | 3.2391 | 0.0119 | 0.8101 | 11 |
62
+ | 3.4534 | 0.0143 | 0.7212 | 3.2070 | 0.0120 | 0.7870 | 12 |
63
+ | 3.3814 | 0.0146 | 0.7080 | 3.1505 | 0.0122 | 0.7826 | 13 |
64
+ | 3.3069 | 0.0148 | 0.6961 | 3.1102 | 0.0124 | 0.7609 | 14 |
65
+ | 3.2229 | 0.0152 | 0.6781 | 3.0542 | 0.0125 | 0.7532 | 15 |
66
+ | 3.1334 | 0.0156 | 0.6614 | 2.9840 | 0.0127 | 0.7448 | 16 |
67
+ | 3.0313 | 0.0160 | 0.6425 | 2.9032 | 0.0130 | 0.7123 | 17 |
68
+ | 2.9122 | 0.0166 | 0.6202 | 2.7986 | 0.0134 | 0.6930 | 18 |
69
+ | 2.7559 | 0.0173 | 0.5940 | 2.6337 | 0.0139 | 0.6673 | 19 |
70
+ | 2.5649 | 0.0182 | 0.5674 | 2.4490 | 0.0145 | 0.6383 | 20 |
71
+ | 2.3414 | 0.0193 | 0.5299 | 2.2785 | 0.0150 | 0.6183 | 21 |
72
+ | 2.0966 | 0.0206 | 0.4903 | 2.0460 | 0.0158 | 0.5649 | 22 |
73
+ | 1.8283 | 0.0220 | 0.4459 | 1.8369 | 0.0165 | 0.5306 | 23 |
74
+ | 1.5547 | 0.0235 | 0.3996 | 1.6356 | 0.0172 | 0.4848 | 24 |
75
+ | 1.3218 | 0.0249 | 0.3581 | 1.4682 | 0.0179 | 0.4510 | 25 |
76
+ | 1.1383 | 0.0260 | 0.3211 | 1.3465 | 0.0183 | 0.4226 | 26 |
77
+ | 0.9876 | 0.0270 | 0.2920 | 1.2323 | 0.0188 | 0.3966 | 27 |
78
+ | 0.8635 | 0.0278 | 0.2651 | 1.1482 | 0.0191 | 0.3749 | 28 |
79
+ | 0.7620 | 0.0284 | 0.2435 | 1.0816 | 0.0194 | 0.3565 | 29 |
80
+ | 0.6749 | 0.0290 | 0.2234 | 1.0187 | 0.0196 | 0.3433 | 30 |
81
+ | 0.5998 | 0.0295 | 0.2025 | 0.9761 | 0.0198 | 0.3319 | 31 |
82
+ | 0.5325 | 0.0300 | 0.1827 | 0.9326 | 0.0200 | 0.3213 | 32 |
83
+ | 0.4735 | 0.0305 | 0.1665 | 0.8942 | 0.0201 | 0.3110 | 33 |
84
+ | 0.4228 | 0.0308 | 0.1466 | 0.8735 | 0.0202 | 0.3026 | 34 |
85
+ | 0.3747 | 0.0312 | 0.1293 | 0.8408 | 0.0203 | 0.2931 | 35 |
86
+ | 0.3331 | 0.0316 | 0.1111 | 0.8253 | 0.0204 | 0.2891 | 36 |
87
+ | 0.2947 | 0.0319 | 0.0962 | 0.8084 | 0.0205 | 0.2849 | 37 |
88
+ | 0.2601 | 0.0322 | 0.0817 | 0.7906 | 0.0205 | 0.2783 | 38 |
89
+ | 0.2291 | 0.0324 | 0.0706 | 0.7876 | 0.0206 | 0.2755 | 39 |
90
+
91
+
92
+ ### Framework versions
93
+
94
+ - Transformers 4.34.0.dev0
95
+ - TensorFlow 2.13.0
96
+ - Tokenizers 0.13.3
config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-tiny",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 50257
12
+ ],
13
+ "bos_token_id": 50257,
14
+ "classifier_proj_size": 256,
15
+ "d_model": 384,
16
+ "decoder_attention_heads": 6,
17
+ "decoder_ffn_dim": 1536,
18
+ "decoder_layerdrop": 0.0,
19
+ "decoder_layers": 4,
20
+ "decoder_start_token_id": 50258,
21
+ "dropout": 0.0,
22
+ "encoder_attention_heads": 6,
23
+ "encoder_ffn_dim": 1536,
24
+ "encoder_layerdrop": 0.0,
25
+ "encoder_layers": 4,
26
+ "eos_token_id": 50257,
27
+ "forced_decoder_ids": [],
28
+ "init_std": 0.02,
29
+ "is_encoder_decoder": true,
30
+ "mask_feature_length": 10,
31
+ "mask_feature_min_masks": 0,
32
+ "mask_feature_prob": 0.0,
33
+ "mask_time_length": 10,
34
+ "mask_time_min_masks": 2,
35
+ "mask_time_prob": 0.05,
36
+ "max_length": 448,
37
+ "max_source_positions": 1500,
38
+ "max_target_positions": 448,
39
+ "median_filter_width": 7,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 4,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.34.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
generation_config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "begin_suppress_tokens": [
4
+ 220,
5
+ 50257
6
+ ],
7
+ "bos_token_id": 50257,
8
+ "decoder_start_token_id": 50258,
9
+ "eos_token_id": 50257,
10
+ "forced_decoder_ids": [
11
+ [
12
+ 1,
13
+ 50259
14
+ ],
15
+ [
16
+ 2,
17
+ 50359
18
+ ],
19
+ [
20
+ 3,
21
+ 50363
22
+ ]
23
+ ],
24
+ "max_length": 448,
25
+ "pad_token_id": 50257,
26
+ "suppress_tokens": [
27
+ 1,
28
+ 2,
29
+ 7,
30
+ 8,
31
+ 9,
32
+ 10,
33
+ 14,
34
+ 25,
35
+ 26,
36
+ 27,
37
+ 28,
38
+ 29,
39
+ 31,
40
+ 58,
41
+ 59,
42
+ 60,
43
+ 61,
44
+ 62,
45
+ 63,
46
+ 90,
47
+ 91,
48
+ 92,
49
+ 93,
50
+ 359,
51
+ 503,
52
+ 522,
53
+ 542,
54
+ 873,
55
+ 893,
56
+ 902,
57
+ 918,
58
+ 922,
59
+ 931,
60
+ 1350,
61
+ 1853,
62
+ 1982,
63
+ 2460,
64
+ 2627,
65
+ 3246,
66
+ 3253,
67
+ 3268,
68
+ 3536,
69
+ 3846,
70
+ 3961,
71
+ 4183,
72
+ 4667,
73
+ 6585,
74
+ 6647,
75
+ 7273,
76
+ 9061,
77
+ 9383,
78
+ 10428,
79
+ 10929,
80
+ 11938,
81
+ 12033,
82
+ 12331,
83
+ 12562,
84
+ 13793,
85
+ 14157,
86
+ 14635,
87
+ 15265,
88
+ 15618,
89
+ 16553,
90
+ 16604,
91
+ 18362,
92
+ 18956,
93
+ 20075,
94
+ 21675,
95
+ 22520,
96
+ 26130,
97
+ 26161,
98
+ 26435,
99
+ 28279,
100
+ 29464,
101
+ 31650,
102
+ 32302,
103
+ 32470,
104
+ 36865,
105
+ 42863,
106
+ 47425,
107
+ 49870,
108
+ 50254,
109
+ 50258,
110
+ 50358,
111
+ 50359,
112
+ 50360,
113
+ 50361,
114
+ 50362
115
+ ],
116
+ "transformers_version": "4.34.0.dev0"
117
+ }
tf_model.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5515fff55df7c568f53773426b72598af49d455cf242ce557ca9bb85a2750c7
3
+ size 151253960