Lakoc commited on
Commit
a087e10
1 Parent(s): 8f5bb44

End of training

Browse files
Files changed (33) hide show
  1. .gitattributes +1 -0
  2. README.md +88 -0
  3. config.json +275 -0
  4. generation_config.json +15 -0
  5. model.safetensors +3 -0
  6. predictions_common_voice_13_en_common_voice_13_en_test_wer100.00.csv +5 -0
  7. predictions_common_voice_13_en_common_voice_13_en_test_wer100.00_hyp.trn +4 -0
  8. predictions_common_voice_13_en_common_voice_13_en_test_wer100.00_ref.trn +4 -0
  9. predictions_common_voice_13_en_common_voice_13_en_test_wer108.70.csv +5 -0
  10. predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn +4 -0
  11. predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn.dtl +129 -0
  12. predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn.snt.utt +85 -0
  13. predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn.sys +18 -0
  14. predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_ref.trn +4 -0
  15. predictions_common_voice_13_en_common_voice_13_en_test_wer156.52.csv +5 -0
  16. predictions_common_voice_13_en_common_voice_13_en_test_wer156.52_hyp.trn +4 -0
  17. predictions_common_voice_13_en_common_voice_13_en_test_wer156.52_ref.trn +4 -0
  18. predictions_common_voice_13_en_common_voice_13_en_test_wer21.29.csv +27 -0
  19. predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn +26 -0
  20. predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn.dtl +86 -0
  21. predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn.snt.utt +503 -0
  22. predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn.sys +18 -0
  23. predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_ref.trn +26 -0
  24. predictions_common_voice_13_en_common_voice_13_en_test_wer26.31.csv +0 -0
  25. predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn +0 -0
  26. predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn.dtl +0 -0
  27. predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn.snt.utt +3 -0
  28. predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn.sys +18 -0
  29. predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_ref.trn +0 -0
  30. predictions_common_voice_13_en_common_voice_13_en_test_wer99.13.csv +0 -0
  31. predictions_common_voice_13_en_common_voice_13_en_test_wer99.13_hyp.trn +0 -0
  32. predictions_common_voice_13_en_common_voice_13_en_test_wer99.13_ref.trn +0 -0
  33. training_args.bin +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn.snt.utt filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ datasets:
5
+ - common_voice_13_0
6
+ metrics:
7
+ - wer
8
+ model-index:
9
+ - name: DeCRED_small_cv
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # DeCRED_small_cv
17
+
18
+ This model is a fine-tuned version of [](https://huggingface.co/) on the common_voice_13_0 dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Cer: 0.0855
21
+ - Deletions: 4133
22
+ - Hits: 123172
23
+ - Insertions: 3671
24
+ - Loss: 1.1174
25
+ - Mer: 0.1889
26
+ - Substitutions: 20876
27
+ - Wer: 0.1935
28
+ - Wil: 0.3069
29
+ - Wip: 0.6931
30
+
31
+ ## Model description
32
+
33
+ More information needed
34
+
35
+ ## Intended uses & limitations
36
+
37
+ More information needed
38
+
39
+ ## Training and evaluation data
40
+
41
+ More information needed
42
+
43
+ ## Training procedure
44
+
45
+ ### Training hyperparameters
46
+
47
+ The following hyperparameters were used during training:
48
+ - learning_rate: 0.002
49
+ - train_batch_size: 256
50
+ - eval_batch_size: 128
51
+ - seed: 42
52
+ - distributed_type: multi-GPU
53
+ - num_devices: 2
54
+ - total_train_batch_size: 512
55
+ - total_eval_batch_size: 256
56
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
57
+ - lr_scheduler_type: linear
58
+ - lr_scheduler_warmup_steps: 10000
59
+ - num_epochs: 20.0
60
+
61
+ ### Training results
62
+
63
+ | Training Loss | Epoch | Step | Cer | Deletions | Hits | Insertions | Validation Loss | Mer | Substitutions | Wer | Wil | Wip |
64
+ |:-------------:|:-----:|:-----:|:------:|:---------:|:------:|:----------:|:---------------:|:------:|:-------------:|:------:|:------:|:------:|
65
+ | 1.3459 | 5.0 | 7445 | 0.1676 | 7277 | 102705 | 6830 | 1.5193 | 0.3374 | 38199 | 0.3530 | 0.5182 | 0.4818 |
66
+ | 1.3006 | 6.0 | 8934 | 0.1584 | 6381 | 105485 | 7074 | 1.4779 | 0.3206 | 36315 | 0.3359 | 0.4956 | 0.5044 |
67
+ | 1.7549 | 7.0 | 10423 | 0.1620 | 6967 | 104307 | 6557 | 1.4874 | 0.3259 | 36907 | 0.3403 | 0.5031 | 0.4969 |
68
+ | 1.5621 | 8.0 | 11912 | 0.1361 | 5203 | 111002 | 6489 | 1.3687 | 0.2823 | 31976 | 0.2947 | 0.4437 | 0.5563 |
69
+ | 1.4988 | 9.0 | 13401 | 0.1298 | 5025 | 112303 | 6136 | 1.3308 | 0.2723 | 30853 | 0.2835 | 0.4299 | 0.5701 |
70
+ | 1.4495 | 10.0 | 14890 | 0.1212 | 5505 | 113957 | 5063 | 1.2901 | 0.2564 | 28719 | 0.2651 | 0.4068 | 0.5932 |
71
+ | 1.4092 | 11.0 | 16379 | 0.1158 | 5162 | 115248 | 4910 | 1.2625 | 0.2472 | 27771 | 0.2554 | 0.3941 | 0.6059 |
72
+ | 1.3638 | 12.0 | 17868 | 0.1088 | 4408 | 117314 | 5173 | 1.2336 | 0.2350 | 26459 | 0.2432 | 0.3764 | 0.6236 |
73
+ | 1.335 | 13.0 | 19357 | 0.1060 | 4868 | 117783 | 4467 | 1.2177 | 0.2284 | 25530 | 0.2353 | 0.3665 | 0.6335 |
74
+ | 1.2966 | 14.0 | 20846 | 0.1006 | 4561 | 119385 | 4599 | 1.1945 | 0.2186 | 24235 | 0.2254 | 0.3511 | 0.6489 |
75
+ | 1.2826 | 15.0 | 22335 | 0.0969 | 4455 | 119983 | 4111 | 1.1767 | 0.2122 | 23743 | 0.2180 | 0.3429 | 0.6571 |
76
+ | 1.2529 | 16.0 | 23824 | 0.0954 | 4229 | 120627 | 4312 | 1.1664 | 0.2090 | 23325 | 0.2150 | 0.3377 | 0.6623 |
77
+ | 1.2103 | 17.0 | 25313 | 0.0907 | 4218 | 121711 | 3998 | 1.1458 | 0.2002 | 22252 | 0.2056 | 0.3244 | 0.6756 |
78
+ | 1.18 | 18.0 | 26802 | 0.0891 | 4153 | 122139 | 3883 | 1.1355 | 0.1968 | 21889 | 0.2019 | 0.3194 | 0.6806 |
79
+ | 1.1529 | 19.0 | 28291 | 0.0864 | 4132 | 122860 | 3681 | 1.1236 | 0.1910 | 21189 | 0.1957 | 0.3105 | 0.6895 |
80
+ | 1.1342 | 20.0 | 29780 | 0.0855 | 4133 | 123172 | 3671 | 1.1174 | 0.1889 | 20876 | 0.1935 | 0.3069 | 0.6931 |
81
+
82
+
83
+ ### Framework versions
84
+
85
+ - Transformers 4.40.0.dev0
86
+ - Pytorch 2.2.0+rocm5.6
87
+ - Datasets 2.18.0
88
+ - Tokenizers 0.15.2
config.json ADDED
@@ -0,0 +1,275 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "JointCTCAttentionEncoderDecoder"
4
+ ],
5
+ "bos_token_id": 0,
6
+ "ctc_weight": 0.3,
7
+ "decoder": {
8
+ "_name_or_path": "Lakoc/gpt2_256h_6l_add_head3_04",
9
+ "activation_function": "gelu_new",
10
+ "add_cross_attention": true,
11
+ "architectures": null,
12
+ "attn_pdrop": 0.1,
13
+ "average_logits": false,
14
+ "bad_words_ids": null,
15
+ "begin_suppress_tokens": null,
16
+ "bos_token_id": 0,
17
+ "chunk_size_feed_forward": 0,
18
+ "cross_attention_hidden_size": null,
19
+ "decoder_start_token_id": null,
20
+ "diversity_penalty": 0.0,
21
+ "do_sample": false,
22
+ "early_stopping": false,
23
+ "embd_pdrop": 0.1,
24
+ "encoder_no_repeat_ngram_size": 0,
25
+ "eos_token_id": 1,
26
+ "exponential_decay_length_penalty": null,
27
+ "finetuning_task": null,
28
+ "forced_bos_token_id": null,
29
+ "forced_eos_token_id": null,
30
+ "head_locations": [
31
+ 3
32
+ ],
33
+ "head_weights": [
34
+ 0.6,
35
+ 0.4
36
+ ],
37
+ "id2label": {
38
+ "0": "LABEL_0",
39
+ "1": "LABEL_1"
40
+ },
41
+ "initializer_range": 0.02,
42
+ "is_decoder": true,
43
+ "is_encoder_decoder": false,
44
+ "label2id": {
45
+ "LABEL_0": 0,
46
+ "LABEL_1": 1
47
+ },
48
+ "layer_norm_epsilon": 1e-05,
49
+ "length_penalty": 1.0,
50
+ "max_length": 20,
51
+ "min_length": 0,
52
+ "model_type": "gpt2-multi-head",
53
+ "n_embd": 256,
54
+ "n_head": 4,
55
+ "n_inner": 2048,
56
+ "n_layer": 6,
57
+ "n_positions": 1024,
58
+ "no_repeat_ngram_size": 0,
59
+ "num_beam_groups": 1,
60
+ "num_beams": 1,
61
+ "num_return_sequences": 1,
62
+ "output_attentions": false,
63
+ "output_hidden_states": false,
64
+ "output_scores": false,
65
+ "pad_token_id": null,
66
+ "pos_emb_fixed": true,
67
+ "prefix": null,
68
+ "problem_type": null,
69
+ "pruned_heads": {},
70
+ "remove_invalid_values": false,
71
+ "reorder_and_upcast_attn": false,
72
+ "repetition_penalty": 1.0,
73
+ "resid_pdrop": 0.1,
74
+ "return_dict": true,
75
+ "return_dict_in_generate": false,
76
+ "scale_attn_by_inverse_layer_idx": false,
77
+ "scale_attn_weights": true,
78
+ "sep_token_id": null,
79
+ "summary_activation": null,
80
+ "summary_first_dropout": 0.1,
81
+ "summary_proj_to_labels": true,
82
+ "summary_type": "cls_index",
83
+ "summary_use_proj": true,
84
+ "suppress_tokens": null,
85
+ "task_specific_params": null,
86
+ "temperature": 1.0,
87
+ "tf_legacy_loss": false,
88
+ "tie_additional_weights": false,
89
+ "tie_encoder_decoder": false,
90
+ "tie_word_embeddings": false,
91
+ "tokenizer_class": null,
92
+ "top_k": 50,
93
+ "top_p": 1.0,
94
+ "torch_dtype": null,
95
+ "torchscript": false,
96
+ "typical_p": 1.0,
97
+ "use_bfloat16": false,
98
+ "use_cache": true,
99
+ "vocab_size": 1000
100
+ },
101
+ "decoder_start_token_id": 0,
102
+ "encoder": {
103
+ "_name_or_path": "Lakoc/fisher_ebranchformer_enc_12_layers_fixed",
104
+ "activation_dropout": 0.1,
105
+ "adapter_attn_dim": null,
106
+ "adapter_kernel_size": 3,
107
+ "adapter_stride": 2,
108
+ "add_adapter": false,
109
+ "add_cross_attention": false,
110
+ "apply_spec_augment": false,
111
+ "architectures": null,
112
+ "attention_dropout": 0.1,
113
+ "bad_words_ids": null,
114
+ "begin_suppress_tokens": null,
115
+ "bos_token_id": 1,
116
+ "chunk_size_feed_forward": 0,
117
+ "classifier_proj_size": 256,
118
+ "codevector_dim": 256,
119
+ "conformer_conv_dropout": 0.1,
120
+ "contrastive_logits_temperature": 0.1,
121
+ "conv_bias": false,
122
+ "conv_depthwise_kernel_size": 31,
123
+ "conv_dim": [
124
+ 256,
125
+ 256
126
+ ],
127
+ "conv_kernel": [
128
+ 3,
129
+ 3
130
+ ],
131
+ "conv_padding": [
132
+ 1,
133
+ 1
134
+ ],
135
+ "conv_stride": [
136
+ 2,
137
+ 2
138
+ ],
139
+ "cross_attention_hidden_size": null,
140
+ "csgu_activation": "identity",
141
+ "csgu_conv_dropout": 0.1,
142
+ "csgu_kernel_size": 31,
143
+ "csgu_use_linear_after_conv": false,
144
+ "ctc_loss_reduction": "mean",
145
+ "ctc_zero_infinity": true,
146
+ "decoder_start_token_id": null,
147
+ "diversity_loss_weight": 0.1,
148
+ "diversity_penalty": 0.0,
149
+ "do_sample": false,
150
+ "do_stable_layer_norm": false,
151
+ "early_stopping": false,
152
+ "encoder_no_repeat_ngram_size": 0,
153
+ "eos_token_id": 2,
154
+ "exponential_decay_length_penalty": null,
155
+ "fe_position_embeddings": false,
156
+ "feat_extract_activation": "gelu",
157
+ "feat_extract_norm": "group",
158
+ "feat_proj_dropout": 0.0,
159
+ "feat_quantizer_dropout": 0.0,
160
+ "final_dropout": 0.1,
161
+ "finetuning_task": null,
162
+ "forced_bos_token_id": null,
163
+ "forced_eos_token_id": null,
164
+ "hidden_act": "gelu",
165
+ "hidden_dropout": 0.1,
166
+ "hidden_size": 256,
167
+ "id2label": {
168
+ "0": "LABEL_0",
169
+ "1": "LABEL_1"
170
+ },
171
+ "initializer_range": 0.02,
172
+ "intermediate_size": 1024,
173
+ "is_causal": false,
174
+ "is_decoder": false,
175
+ "is_encoder_decoder": false,
176
+ "label2id": {
177
+ "LABEL_0": 0,
178
+ "LABEL_1": 1
179
+ },
180
+ "layer_norm_eps": 1e-05,
181
+ "layerdrop": 0.0,
182
+ "length_penalty": 1.0,
183
+ "mask_feature_length": 10,
184
+ "mask_feature_min_masks": 0,
185
+ "mask_feature_prob": 0.0,
186
+ "mask_time_length": 10,
187
+ "mask_time_min_masks": 2,
188
+ "mask_time_prob": 0.05,
189
+ "max_length": 20,
190
+ "max_source_positions": 1024,
191
+ "merge_conv_kernel": 31,
192
+ "min_length": 0,
193
+ "model_type": "wav2vec2-ebranchformer",
194
+ "no_repeat_ngram_size": 0,
195
+ "num_adapter_layers": 3,
196
+ "num_attention_heads": 4,
197
+ "num_beam_groups": 1,
198
+ "num_beams": 1,
199
+ "num_codevector_groups": 2,
200
+ "num_codevectors_per_group": 320,
201
+ "num_conv_pos_embedding_groups": 16,
202
+ "num_conv_pos_embeddings": 128,
203
+ "num_fbanks": 80,
204
+ "num_feat_extract_layers": 2,
205
+ "num_hidden_layers": 12,
206
+ "num_mel_bins": 80,
207
+ "num_negatives": 100,
208
+ "num_return_sequences": 1,
209
+ "output_attentions": false,
210
+ "output_hidden_size": 256,
211
+ "output_hidden_states": false,
212
+ "output_scores": false,
213
+ "pad_token_id": 3,
214
+ "position_embeddings_type": "relative",
215
+ "prefix": null,
216
+ "problem_type": null,
217
+ "proj_codevector_dim": 256,
218
+ "pruned_heads": {},
219
+ "remove_invalid_values": false,
220
+ "repetition_penalty": 1.0,
221
+ "return_dict": true,
222
+ "return_dict_in_generate": false,
223
+ "rotary_embedding_base": 10000,
224
+ "sep_token_id": null,
225
+ "suppress_tokens": null,
226
+ "task_specific_params": null,
227
+ "tdnn_dilation": [
228
+ 1,
229
+ 2,
230
+ 3,
231
+ 1,
232
+ 1
233
+ ],
234
+ "tdnn_dim": [
235
+ 512,
236
+ 512,
237
+ 512,
238
+ 512,
239
+ 1500
240
+ ],
241
+ "tdnn_kernel": [
242
+ 5,
243
+ 3,
244
+ 3,
245
+ 1,
246
+ 1
247
+ ],
248
+ "temperature": 1.0,
249
+ "tf_legacy_loss": false,
250
+ "tie_encoder_decoder": false,
251
+ "tie_word_embeddings": true,
252
+ "tokenizer_class": null,
253
+ "top_k": 50,
254
+ "top_p": 1.0,
255
+ "torch_dtype": null,
256
+ "torchscript": false,
257
+ "typical_p": 1.0,
258
+ "use_bfloat16": false,
259
+ "use_fbanks": true,
260
+ "use_macaron_ff": true,
261
+ "use_weighted_layer_sum": false,
262
+ "vocab_size": 1000,
263
+ "xvector_output_dim": 512
264
+ },
265
+ "eos_token_id": 1,
266
+ "is_encoder_decoder": true,
267
+ "lsm_factor": 0.1,
268
+ "mask_token_id": 4,
269
+ "model_type": "joint_aed_ctc_speech-encoder-decoder",
270
+ "pad_token_id": 3,
271
+ "shared_lm_head": false,
272
+ "tie_word_embeddings": false,
273
+ "torch_dtype": "float32",
274
+ "transformers_version": "4.40.0.dev0"
275
+ }
generation_config.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "apply_eos_space_trick": false,
3
+ "bos_token_id": 0,
4
+ "ctc_margin": 0,
5
+ "ctc_weight": 0.0,
6
+ "decoder_start_token_id": 0,
7
+ "eos_space_trick_weight": 0.0,
8
+ "eos_token_id": 1,
9
+ "lm_model": null,
10
+ "lm_weight": 0.0,
11
+ "max_length": 512,
12
+ "pad_token_id": 3,
13
+ "space_token_id": -1,
14
+ "transformers_version": "4.40.0.dev0"
15
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:497190bb024e75b6d0e560b39282d01acefc76355f3940c66b2ac852a3d3a9c5
3
+ size 144243216
predictions_common_voice_13_en_common_voice_13_en_test_wer100.00.csv ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ label,prediction
2
+ joe keaton disapproved of films and buster also had reservations about the medium,
3
+ shell be all right,
4
+ six,
5
+ alls well that ends well,
predictions_common_voice_13_en_common_voice_13_en_test_wer100.00_hyp.trn ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ (utterance_0)
2
+ (utterance_1)
3
+ (utterance_2)
4
+ (utterance_3)
predictions_common_voice_13_en_common_voice_13_en_test_wer100.00_ref.trn ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_0)
2
+ shell be all right (utterance_1)
3
+ six (utterance_2)
4
+ alls well that ends well (utterance_3)
predictions_common_voice_13_en_common_voice_13_en_test_wer108.70.csv ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ label,prediction
2
+ joe keaton disapproved of films and buster also had reservations about the medium,a was available moa moas thea ends theas thels thes
3
+ shell be all right,hus
4
+ six,asasas theaslsas thes
5
+ alls well that ends well,a wasasasasasaslasasasas californias alsasls alsas
predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ a was available moa moas thea ends theas thels thes (utterance_0)
2
+ hus (utterance_1)
3
+ asasas theaslsas thes (utterance_2)
4
+ a wasasasasasaslasasasas californias alsasls alsas (utterance_3)
predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn.dtl ADDED
@@ -0,0 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ DETAILED OVERALL REPORT FOR THE SYSTEM: /scratch/project_465000836/ipoloka/huggingface_asr/experiments/decred/commonvoice/DeCRED_small_cv/predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn
2
+
3
+ SENTENCE RECOGNITION PERFORMANCE
4
+
5
+ sentences 4
6
+ with errors 100.0% ( 4)
7
+
8
+ with substitutions 100.0% ( 4)
9
+ with deletions 50.0% ( 2)
10
+ with insertions 25.0% ( 1)
11
+
12
+
13
+ WORD RECOGNITION PERFORMANCE
14
+
15
+ Percent Total Error = 108.7% ( 25)
16
+
17
+ Percent Correct = 0.0% ( 0)
18
+
19
+ Percent Substitution = 73.9% ( 17)
20
+ Percent Deletions = 26.1% ( 6)
21
+ Percent Insertions = 8.7% ( 2)
22
+ Percent Word Accuracy = -8.7%
23
+
24
+
25
+ Ref. words = ( 23)
26
+ Hyp. words = ( 19)
27
+ Aligned words = ( 25)
28
+
29
+ CONFUSION PAIRS Total (17)
30
+ With >= 1 occurrences (17)
31
+
32
+ 1: 1 -> about ==> theas
33
+ 2: 1 -> alls ==> a
34
+ 3: 1 -> also ==> moas
35
+ 4: 1 -> and ==> available
36
+ 5: 1 -> buster ==> moa
37
+ 6: 1 -> ends ==> alsasls
38
+ 7: 1 -> films ==> was
39
+ 8: 1 -> had ==> thea
40
+ 9: 1 -> medium ==> thes
41
+ 10: 1 -> of ==> a
42
+ 11: 1 -> reservations ==> ends
43
+ 12: 1 -> right ==> hus
44
+ 13: 1 -> six ==> thes
45
+ 14: 1 -> that ==> californias
46
+ 15: 1 -> the ==> thels
47
+ 16: 1 -> well ==> alsas
48
+ 17: 1 -> well ==> wasasasasasaslasasasas
49
+ -------
50
+ 17
51
+
52
+
53
+
54
+ INSERTIONS Total (2)
55
+ With >= 1 occurrences (2)
56
+
57
+ 1: 1 -> asasas
58
+ 2: 1 -> theaslsas
59
+ -------
60
+ 2
61
+
62
+
63
+
64
+ DELETIONS Total (6)
65
+ With >= 1 occurrences (6)
66
+
67
+ 1: 1 -> all
68
+ 2: 1 -> be
69
+ 3: 1 -> disapproved
70
+ 4: 1 -> joe
71
+ 5: 1 -> keaton
72
+ 6: 1 -> shell
73
+ -------
74
+ 6
75
+
76
+
77
+
78
+ SUBSTITUTIONS Total (16)
79
+ With >= 1 occurrences (16)
80
+
81
+ 1: 2 -> well
82
+ 2: 1 -> about
83
+ 3: 1 -> alls
84
+ 4: 1 -> also
85
+ 5: 1 -> and
86
+ 6: 1 -> buster
87
+ 7: 1 -> ends
88
+ 8: 1 -> films
89
+ 9: 1 -> had
90
+ 10: 1 -> medium
91
+ 11: 1 -> of
92
+ 12: 1 -> reservations
93
+ 13: 1 -> right
94
+ 14: 1 -> six
95
+ 15: 1 -> that
96
+ 16: 1 -> the
97
+ -------
98
+ 17
99
+
100
+
101
+ * NOTE: The 'Substitution' words are those reference words
102
+ for which the recognizer supplied an incorrect word.
103
+
104
+
105
+ FALSELY RECOGNIZED Total (15)
106
+ With >= 1 occurrences (15)
107
+
108
+ 1: 2 -> a
109
+ 2: 2 -> thes
110
+ 3: 1 -> alsas
111
+ 4: 1 -> alsasls
112
+ 5: 1 -> available
113
+ 6: 1 -> californias
114
+ 7: 1 -> ends
115
+ 8: 1 -> hus
116
+ 9: 1 -> moa
117
+ 10: 1 -> moas
118
+ 11: 1 -> thea
119
+ 12: 1 -> theas
120
+ 13: 1 -> thels
121
+ 14: 1 -> was
122
+ 15: 1 -> wasasasasasaslasasasas
123
+ -------
124
+ 17
125
+
126
+
127
+ * NOTE: The 'Falsely Recognized' words are those hypothesis words
128
+ which the recognizer incorrectly substituted for a reference word.
129
+
predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn.snt.utt ADDED
@@ -0,0 +1,85 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ===============================================================================
2
+
3
+ SENTENCE LEVEL REPORT FOR THE SYSTEM:
4
+ Name: /scratch/project_465000836/ipoloka/huggingface_asr/experiments/decred/commonvoice/DeCRED_small_cv/predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn
5
+
6
+ ===============================================================================
7
+
8
+
9
+ SPEAKER utt
10
+ id: (utterance_0)
11
+ Scores: (#C #S #D #I) 0 10 3 0
12
+ REF: JOE KEATON DISAPPROVED OF FILMS AND BUSTER ALSO HAD RESERVATIONS ABOUT THE MEDIUM
13
+ HYP: *** ****** *********** A WAS AVAILABLE MOA MOAS THEA ENDS THEAS THELS THES
14
+ Eval: D D D S S S S S S S S S S
15
+
16
+ Correct = 0.0% 0 ( 0)
17
+ Substitutions = 76.9% 10 ( 10)
18
+ Deletions = 23.1% 3 ( 3)
19
+ Insertions = 0.0% 0 ( 0)
20
+
21
+ Errors = 100.0% 13 ( 13)
22
+
23
+ Ref. words = 13 ( 13)
24
+ Hyp. words = 10 ( 10)
25
+ Aligned words = 13 ( 13)
26
+
27
+ -------------------------------------------------------------------------------
28
+
29
+ id: (utterance_1)
30
+ Scores: (#C #S #D #I) 0 1 3 0
31
+ REF: SHELL BE ALL RIGHT
32
+ HYP: ***** ** *** HUS
33
+ Eval: D D D S
34
+
35
+ Correct = 0.0% 0 ( 0)
36
+ Substitutions = 25.0% 1 ( 11)
37
+ Deletions = 75.0% 3 ( 6)
38
+ Insertions = 0.0% 0 ( 0)
39
+
40
+ Errors = 100.0% 4 ( 17)
41
+
42
+ Ref. words = 4 ( 17)
43
+ Hyp. words = 1 ( 11)
44
+ Aligned words = 4 ( 17)
45
+
46
+ -------------------------------------------------------------------------------
47
+
48
+ id: (utterance_2)
49
+ Scores: (#C #S #D #I) 0 1 0 2
50
+ REF: ****** ********* SIX
51
+ HYP: ASASAS THEASLSAS THES
52
+ Eval: I I S
53
+
54
+ Correct = 0.0% 0 ( 0)
55
+ Substitutions = 100.0% 1 ( 12)
56
+ Deletions = 0.0% 0 ( 6)
57
+ Insertions = 200.0% 2 ( 2)
58
+
59
+ Errors = 300.0% 3 ( 20)
60
+
61
+ Ref. words = 1 ( 18)
62
+ Hyp. words = 3 ( 14)
63
+ Aligned words = 3 ( 20)
64
+
65
+ -------------------------------------------------------------------------------
66
+
67
+ id: (utterance_3)
68
+ Scores: (#C #S #D #I) 0 5 0 0
69
+ REF: ALLS WELL THAT ENDS WELL
70
+ HYP: A WASASASASASASLASASASAS CALIFORNIAS ALSASLS ALSAS
71
+ Eval: S S S S S
72
+
73
+ Correct = 0.0% 0 ( 0)
74
+ Substitutions = 100.0% 5 ( 17)
75
+ Deletions = 0.0% 0 ( 6)
76
+ Insertions = 0.0% 0 ( 2)
77
+
78
+ Errors = 100.0% 5 ( 25)
79
+
80
+ Ref. words = 5 ( 23)
81
+ Hyp. words = 5 ( 19)
82
+ Aligned words = 5 ( 25)
83
+
84
+ -------------------------------------------------------------------------------
85
+
predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn.sys ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+
3
+
4
+ SYSTEM SUMMARY PERCENTAGES by SPEAKER
5
+
6
+ ,--------------------------------------------------------------------------------------------------------------------------------------------------------------------------.
7
+ |/scratch/project_465000836/ipoloka/huggingface_asr/experiments/decred/commonvoice/DeCRED_small_cv/predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_hyp.trn|
8
+ |--------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
9
+ | SPKR | # Snt # Wrd | Corr Sub Del Ins Err S.Err |
10
+ |--------------------+------------------------------------+----------------------------------------------------------------------------------------------------------------|
11
+ | utt | 4 23 | 0.0 73.9 26.1 8.7 108.7 100.0 |
12
+ |==========================================================================================================================================================================|
13
+ | Sum/Avg | 4 23 | 0.0 73.9 26.1 8.7 108.7 100.0 |
14
+ |==========================================================================================================================================================================|
15
+ | Mean | 4.0 23.0 | 0.0 73.9 26.1 8.7 108.7 100.0 |
16
+ | S.D. | 0.0 0.0 | 0.0 0.0 0.0 0.0 0.0 0.0 |
17
+ | Median | 4.0 23.0 | 0.0 73.9 26.1 8.7 108.7 100.0 |
18
+ `--------------------------------------------------------------------------------------------------------------------------------------------------------------------------'
predictions_common_voice_13_en_common_voice_13_en_test_wer108.70_ref.trn ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_0)
2
+ shell be all right (utterance_1)
3
+ six (utterance_2)
4
+ alls well that ends well (utterance_3)
predictions_common_voice_13_en_common_voice_13_en_test_wer156.52.csv ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ label,prediction
2
+ joe keaton disapproved of films and buster also had reservations about the medium,s ofs thes ofs year
3
+ shell be all right,thes thes withs ofs year
4
+ six,thes withsss ofss
5
+ alls well that ends well,s ofs thes withs ofs ofs her withs thess thesss thes thess thesss thes thes
predictions_common_voice_13_en_common_voice_13_en_test_wer156.52_hyp.trn ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ s ofs thes ofs year (utterance_0)
2
+ thes thes withs ofs year (utterance_1)
3
+ thes withsss ofss (utterance_2)
4
+ s ofs thes withs ofs ofs her withs thess thesss thes thess thesss thes thes (utterance_3)
predictions_common_voice_13_en_common_voice_13_en_test_wer156.52_ref.trn ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_0)
2
+ shell be all right (utterance_1)
3
+ six (utterance_2)
4
+ alls well that ends well (utterance_3)
predictions_common_voice_13_en_common_voice_13_en_test_wer21.29.csv ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ label,prediction
2
+ joe keaton disapproved of films and buster also had reservations about the medium,joel keaton disapproved of films and buster also had reservations about the medium
3
+ shell be all right,j be a right
4
+ six,six
5
+ alls well that ends well,oswell that ends well
6
+ joe keaton disapproved of films and buster also had reservations about the medium,joel keaton disapproved of films and buster also had reservations about the medium
7
+ shell be all right,j be a right
8
+ six,six
9
+ alls well that ends well,oswell that ends well
10
+ joe keaton disapproved of films and buster also had reservations about the medium,joel keaton disapproved of films and buster also had reservations about the medium
11
+ shell be all right,j be a right
12
+ six,six
13
+ alls well that ends well,oswell that ends well
14
+ joe keaton disapproved of films and buster also had reservations about the medium,joel keaton disapproved of films and buster also had reservations about the medium
15
+ shell be all right,j be a right
16
+ six,six
17
+ alls well that ends well,oswell that ends well
18
+ joe keaton disapproved of films and buster also had reservations about the medium,joel keaton disapproved of films and buster also had reservations about the medium
19
+ shell be all right,j be a right
20
+ six,six
21
+ alls well that ends well,oswell that ends well
22
+ joe keaton disapproved of films and buster also had reservations about the medium,joel keaton disapproved of films and buster also had reservations about the medium
23
+ shell be all right,j be a right
24
+ six,six
25
+ alls well that ends well,oswell that ends well
26
+ joe keaton disapproved of films and buster also had reservations about the medium,joel keaton disapproved of films and buster also had reservations about the medium
27
+ shell be all right,j be a right
predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ joel keaton disapproved of films and buster also had reservations about the medium (utterance_0)
2
+ j be a right (utterance_1)
3
+ six (utterance_2)
4
+ oswell that ends well (utterance_3)
5
+ joel keaton disapproved of films and buster also had reservations about the medium (utterance_4)
6
+ j be a right (utterance_5)
7
+ six (utterance_6)
8
+ oswell that ends well (utterance_7)
9
+ joel keaton disapproved of films and buster also had reservations about the medium (utterance_8)
10
+ j be a right (utterance_9)
11
+ six (utterance_10)
12
+ oswell that ends well (utterance_11)
13
+ joel keaton disapproved of films and buster also had reservations about the medium (utterance_12)
14
+ j be a right (utterance_13)
15
+ six (utterance_14)
16
+ oswell that ends well (utterance_15)
17
+ joel keaton disapproved of films and buster also had reservations about the medium (utterance_16)
18
+ j be a right (utterance_17)
19
+ six (utterance_18)
20
+ oswell that ends well (utterance_19)
21
+ joel keaton disapproved of films and buster also had reservations about the medium (utterance_20)
22
+ j be a right (utterance_21)
23
+ six (utterance_22)
24
+ oswell that ends well (utterance_23)
25
+ joel keaton disapproved of films and buster also had reservations about the medium (utterance_24)
26
+ j be a right (utterance_25)
predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn.dtl ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ DETAILED OVERALL REPORT FOR THE SYSTEM: /scratch/project_465000836/ipoloka/huggingface_asr/experiments/decred/commonvoice/DeCRED_small_cv/predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn
2
+
3
+ SENTENCE RECOGNITION PERFORMANCE
4
+
5
+ sentences 26
6
+ with errors 76.9% ( 20)
7
+
8
+ with substitutions 76.9% ( 20)
9
+ with deletions 23.1% ( 6)
10
+ with insertions 0.0% ( 0)
11
+
12
+
13
+ WORD RECOGNITION PERFORMANCE
14
+
15
+ Percent Total Error = 21.3% ( 33)
16
+
17
+ Percent Correct = 78.7% ( 122)
18
+
19
+ Percent Substitution = 17.4% ( 27)
20
+ Percent Deletions = 3.9% ( 6)
21
+ Percent Insertions = 0.0% ( 0)
22
+ Percent Word Accuracy = 78.7%
23
+
24
+
25
+ Ref. words = ( 155)
26
+ Hyp. words = ( 149)
27
+ Aligned words = ( 155)
28
+
29
+ CONFUSION PAIRS Total (4)
30
+ With >= 1 occurrences (4)
31
+
32
+ 1: 7 -> all ==> a
33
+ 2: 7 -> joe ==> joel
34
+ 3: 7 -> shell ==> j
35
+ 4: 6 -> well ==> oswell
36
+ -------
37
+ 27
38
+
39
+
40
+
41
+ INSERTIONS Total (0)
42
+ With >= 1 occurrences (0)
43
+
44
+ -------
45
+ 0
46
+
47
+
48
+
49
+ DELETIONS Total (1)
50
+ With >= 1 occurrences (1)
51
+
52
+ 1: 6 -> alls
53
+ -------
54
+ 6
55
+
56
+
57
+
58
+ SUBSTITUTIONS Total (4)
59
+ With >= 1 occurrences (4)
60
+
61
+ 1: 7 -> all
62
+ 2: 7 -> joe
63
+ 3: 7 -> shell
64
+ 4: 6 -> well
65
+ -------
66
+ 27
67
+
68
+
69
+ * NOTE: The 'Substitution' words are those reference words
70
+ for which the recognizer supplied an incorrect word.
71
+
72
+
73
+ FALSELY RECOGNIZED Total (4)
74
+ With >= 1 occurrences (4)
75
+
76
+ 1: 7 -> a
77
+ 2: 7 -> j
78
+ 3: 7 -> joel
79
+ 4: 6 -> oswell
80
+ -------
81
+ 27
82
+
83
+
84
+ * NOTE: The 'Falsely Recognized' words are those hypothesis words
85
+ which the recognizer incorrectly substituted for a reference word.
86
+
predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn.snt.utt ADDED
@@ -0,0 +1,503 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ===============================================================================
2
+
3
+ SENTENCE LEVEL REPORT FOR THE SYSTEM:
4
+ Name: /scratch/project_465000836/ipoloka/huggingface_asr/experiments/decred/commonvoice/DeCRED_small_cv/predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn
5
+
6
+ ===============================================================================
7
+
8
+
9
+ SPEAKER utt
10
+ id: (utterance_0)
11
+ Scores: (#C #S #D #I) 12 1 0 0
12
+ REF: JOE keaton disapproved of films and buster also had reservations about the medium
13
+ HYP: JOEL keaton disapproved of films and buster also had reservations about the medium
14
+ Eval: S
15
+
16
+ Correct = 92.3% 12 ( 12)
17
+ Substitutions = 7.7% 1 ( 1)
18
+ Deletions = 0.0% 0 ( 0)
19
+ Insertions = 0.0% 0 ( 0)
20
+
21
+ Errors = 7.7% 1 ( 1)
22
+
23
+ Ref. words = 13 ( 13)
24
+ Hyp. words = 13 ( 13)
25
+ Aligned words = 13 ( 13)
26
+
27
+ -------------------------------------------------------------------------------
28
+
29
+ id: (utterance_1)
30
+ Scores: (#C #S #D #I) 2 2 0 0
31
+ REF: SHELL be ALL right
32
+ HYP: J be A right
33
+ Eval: S S
34
+
35
+ Correct = 50.0% 2 ( 14)
36
+ Substitutions = 50.0% 2 ( 3)
37
+ Deletions = 0.0% 0 ( 0)
38
+ Insertions = 0.0% 0 ( 0)
39
+
40
+ Errors = 50.0% 2 ( 3)
41
+
42
+ Ref. words = 4 ( 17)
43
+ Hyp. words = 4 ( 17)
44
+ Aligned words = 4 ( 17)
45
+
46
+ -------------------------------------------------------------------------------
47
+
48
+ id: (utterance_2)
49
+ Scores: (#C #S #D #I) 1 0 0 0
50
+ REF: six
51
+ HYP: six
52
+ Eval:
53
+
54
+ Correct = 100.0% 1 ( 15)
55
+ Substitutions = 0.0% 0 ( 3)
56
+ Deletions = 0.0% 0 ( 0)
57
+ Insertions = 0.0% 0 ( 0)
58
+
59
+ Errors = 0.0% 0 ( 3)
60
+
61
+ Ref. words = 1 ( 18)
62
+ Hyp. words = 1 ( 18)
63
+ Aligned words = 1 ( 18)
64
+
65
+ -------------------------------------------------------------------------------
66
+
67
+ id: (utterance_3)
68
+ Scores: (#C #S #D #I) 3 1 1 0
69
+ REF: ALLS WELL that ends well
70
+ HYP: **** OSWELL that ends well
71
+ Eval: D S
72
+
73
+ Correct = 60.0% 3 ( 18)
74
+ Substitutions = 20.0% 1 ( 4)
75
+ Deletions = 20.0% 1 ( 1)
76
+ Insertions = 0.0% 0 ( 0)
77
+
78
+ Errors = 40.0% 2 ( 5)
79
+
80
+ Ref. words = 5 ( 23)
81
+ Hyp. words = 4 ( 22)
82
+ Aligned words = 5 ( 23)
83
+
84
+ -------------------------------------------------------------------------------
85
+
86
+ id: (utterance_4)
87
+ Scores: (#C #S #D #I) 12 1 0 0
88
+ REF: JOE keaton disapproved of films and buster also had reservations about the medium
89
+ HYP: JOEL keaton disapproved of films and buster also had reservations about the medium
90
+ Eval: S
91
+
92
+ Correct = 92.3% 12 ( 30)
93
+ Substitutions = 7.7% 1 ( 5)
94
+ Deletions = 0.0% 0 ( 1)
95
+ Insertions = 0.0% 0 ( 0)
96
+
97
+ Errors = 7.7% 1 ( 6)
98
+
99
+ Ref. words = 13 ( 36)
100
+ Hyp. words = 13 ( 35)
101
+ Aligned words = 13 ( 36)
102
+
103
+ -------------------------------------------------------------------------------
104
+
105
+ id: (utterance_5)
106
+ Scores: (#C #S #D #I) 2 2 0 0
107
+ REF: SHELL be ALL right
108
+ HYP: J be A right
109
+ Eval: S S
110
+
111
+ Correct = 50.0% 2 ( 32)
112
+ Substitutions = 50.0% 2 ( 7)
113
+ Deletions = 0.0% 0 ( 1)
114
+ Insertions = 0.0% 0 ( 0)
115
+
116
+ Errors = 50.0% 2 ( 8)
117
+
118
+ Ref. words = 4 ( 40)
119
+ Hyp. words = 4 ( 39)
120
+ Aligned words = 4 ( 40)
121
+
122
+ -------------------------------------------------------------------------------
123
+
124
+ id: (utterance_6)
125
+ Scores: (#C #S #D #I) 1 0 0 0
126
+ REF: six
127
+ HYP: six
128
+ Eval:
129
+
130
+ Correct = 100.0% 1 ( 33)
131
+ Substitutions = 0.0% 0 ( 7)
132
+ Deletions = 0.0% 0 ( 1)
133
+ Insertions = 0.0% 0 ( 0)
134
+
135
+ Errors = 0.0% 0 ( 8)
136
+
137
+ Ref. words = 1 ( 41)
138
+ Hyp. words = 1 ( 40)
139
+ Aligned words = 1 ( 41)
140
+
141
+ -------------------------------------------------------------------------------
142
+
143
+ id: (utterance_7)
144
+ Scores: (#C #S #D #I) 3 1 1 0
145
+ REF: ALLS WELL that ends well
146
+ HYP: **** OSWELL that ends well
147
+ Eval: D S
148
+
149
+ Correct = 60.0% 3 ( 36)
150
+ Substitutions = 20.0% 1 ( 8)
151
+ Deletions = 20.0% 1 ( 2)
152
+ Insertions = 0.0% 0 ( 0)
153
+
154
+ Errors = 40.0% 2 ( 10)
155
+
156
+ Ref. words = 5 ( 46)
157
+ Hyp. words = 4 ( 44)
158
+ Aligned words = 5 ( 46)
159
+
160
+ -------------------------------------------------------------------------------
161
+
162
+ id: (utterance_8)
163
+ Scores: (#C #S #D #I) 12 1 0 0
164
+ REF: JOE keaton disapproved of films and buster also had reservations about the medium
165
+ HYP: JOEL keaton disapproved of films and buster also had reservations about the medium
166
+ Eval: S
167
+
168
+ Correct = 92.3% 12 ( 48)
169
+ Substitutions = 7.7% 1 ( 9)
170
+ Deletions = 0.0% 0 ( 2)
171
+ Insertions = 0.0% 0 ( 0)
172
+
173
+ Errors = 7.7% 1 ( 11)
174
+
175
+ Ref. words = 13 ( 59)
176
+ Hyp. words = 13 ( 57)
177
+ Aligned words = 13 ( 59)
178
+
179
+ -------------------------------------------------------------------------------
180
+
181
+ id: (utterance_9)
182
+ Scores: (#C #S #D #I) 2 2 0 0
183
+ REF: SHELL be ALL right
184
+ HYP: J be A right
185
+ Eval: S S
186
+
187
+ Correct = 50.0% 2 ( 50)
188
+ Substitutions = 50.0% 2 ( 11)
189
+ Deletions = 0.0% 0 ( 2)
190
+ Insertions = 0.0% 0 ( 0)
191
+
192
+ Errors = 50.0% 2 ( 13)
193
+
194
+ Ref. words = 4 ( 63)
195
+ Hyp. words = 4 ( 61)
196
+ Aligned words = 4 ( 63)
197
+
198
+ -------------------------------------------------------------------------------
199
+
200
+ id: (utterance_10)
201
+ Scores: (#C #S #D #I) 1 0 0 0
202
+ REF: six
203
+ HYP: six
204
+ Eval:
205
+
206
+ Correct = 100.0% 1 ( 51)
207
+ Substitutions = 0.0% 0 ( 11)
208
+ Deletions = 0.0% 0 ( 2)
209
+ Insertions = 0.0% 0 ( 0)
210
+
211
+ Errors = 0.0% 0 ( 13)
212
+
213
+ Ref. words = 1 ( 64)
214
+ Hyp. words = 1 ( 62)
215
+ Aligned words = 1 ( 64)
216
+
217
+ -------------------------------------------------------------------------------
218
+
219
+ id: (utterance_11)
220
+ Scores: (#C #S #D #I) 3 1 1 0
221
+ REF: ALLS WELL that ends well
222
+ HYP: **** OSWELL that ends well
223
+ Eval: D S
224
+
225
+ Correct = 60.0% 3 ( 54)
226
+ Substitutions = 20.0% 1 ( 12)
227
+ Deletions = 20.0% 1 ( 3)
228
+ Insertions = 0.0% 0 ( 0)
229
+
230
+ Errors = 40.0% 2 ( 15)
231
+
232
+ Ref. words = 5 ( 69)
233
+ Hyp. words = 4 ( 66)
234
+ Aligned words = 5 ( 69)
235
+
236
+ -------------------------------------------------------------------------------
237
+
238
+ id: (utterance_12)
239
+ Scores: (#C #S #D #I) 12 1 0 0
240
+ REF: JOE keaton disapproved of films and buster also had reservations about the medium
241
+ HYP: JOEL keaton disapproved of films and buster also had reservations about the medium
242
+ Eval: S
243
+
244
+ Correct = 92.3% 12 ( 66)
245
+ Substitutions = 7.7% 1 ( 13)
246
+ Deletions = 0.0% 0 ( 3)
247
+ Insertions = 0.0% 0 ( 0)
248
+
249
+ Errors = 7.7% 1 ( 16)
250
+
251
+ Ref. words = 13 ( 82)
252
+ Hyp. words = 13 ( 79)
253
+ Aligned words = 13 ( 82)
254
+
255
+ -------------------------------------------------------------------------------
256
+
257
+ id: (utterance_13)
258
+ Scores: (#C #S #D #I) 2 2 0 0
259
+ REF: SHELL be ALL right
260
+ HYP: J be A right
261
+ Eval: S S
262
+
263
+ Correct = 50.0% 2 ( 68)
264
+ Substitutions = 50.0% 2 ( 15)
265
+ Deletions = 0.0% 0 ( 3)
266
+ Insertions = 0.0% 0 ( 0)
267
+
268
+ Errors = 50.0% 2 ( 18)
269
+
270
+ Ref. words = 4 ( 86)
271
+ Hyp. words = 4 ( 83)
272
+ Aligned words = 4 ( 86)
273
+
274
+ -------------------------------------------------------------------------------
275
+
276
+ id: (utterance_14)
277
+ Scores: (#C #S #D #I) 1 0 0 0
278
+ REF: six
279
+ HYP: six
280
+ Eval:
281
+
282
+ Correct = 100.0% 1 ( 69)
283
+ Substitutions = 0.0% 0 ( 15)
284
+ Deletions = 0.0% 0 ( 3)
285
+ Insertions = 0.0% 0 ( 0)
286
+
287
+ Errors = 0.0% 0 ( 18)
288
+
289
+ Ref. words = 1 ( 87)
290
+ Hyp. words = 1 ( 84)
291
+ Aligned words = 1 ( 87)
292
+
293
+ -------------------------------------------------------------------------------
294
+
295
+ id: (utterance_15)
296
+ Scores: (#C #S #D #I) 3 1 1 0
297
+ REF: ALLS WELL that ends well
298
+ HYP: **** OSWELL that ends well
299
+ Eval: D S
300
+
301
+ Correct = 60.0% 3 ( 72)
302
+ Substitutions = 20.0% 1 ( 16)
303
+ Deletions = 20.0% 1 ( 4)
304
+ Insertions = 0.0% 0 ( 0)
305
+
306
+ Errors = 40.0% 2 ( 20)
307
+
308
+ Ref. words = 5 ( 92)
309
+ Hyp. words = 4 ( 88)
310
+ Aligned words = 5 ( 92)
311
+
312
+ -------------------------------------------------------------------------------
313
+
314
+ id: (utterance_16)
315
+ Scores: (#C #S #D #I) 12 1 0 0
316
+ REF: JOE keaton disapproved of films and buster also had reservations about the medium
317
+ HYP: JOEL keaton disapproved of films and buster also had reservations about the medium
318
+ Eval: S
319
+
320
+ Correct = 92.3% 12 ( 84)
321
+ Substitutions = 7.7% 1 ( 17)
322
+ Deletions = 0.0% 0 ( 4)
323
+ Insertions = 0.0% 0 ( 0)
324
+
325
+ Errors = 7.7% 1 ( 21)
326
+
327
+ Ref. words = 13 (105)
328
+ Hyp. words = 13 (101)
329
+ Aligned words = 13 (105)
330
+
331
+ -------------------------------------------------------------------------------
332
+
333
+ id: (utterance_17)
334
+ Scores: (#C #S #D #I) 2 2 0 0
335
+ REF: SHELL be ALL right
336
+ HYP: J be A right
337
+ Eval: S S
338
+
339
+ Correct = 50.0% 2 ( 86)
340
+ Substitutions = 50.0% 2 ( 19)
341
+ Deletions = 0.0% 0 ( 4)
342
+ Insertions = 0.0% 0 ( 0)
343
+
344
+ Errors = 50.0% 2 ( 23)
345
+
346
+ Ref. words = 4 (109)
347
+ Hyp. words = 4 (105)
348
+ Aligned words = 4 (109)
349
+
350
+ -------------------------------------------------------------------------------
351
+
352
+ id: (utterance_18)
353
+ Scores: (#C #S #D #I) 1 0 0 0
354
+ REF: six
355
+ HYP: six
356
+ Eval:
357
+
358
+ Correct = 100.0% 1 ( 87)
359
+ Substitutions = 0.0% 0 ( 19)
360
+ Deletions = 0.0% 0 ( 4)
361
+ Insertions = 0.0% 0 ( 0)
362
+
363
+ Errors = 0.0% 0 ( 23)
364
+
365
+ Ref. words = 1 (110)
366
+ Hyp. words = 1 (106)
367
+ Aligned words = 1 (110)
368
+
369
+ -------------------------------------------------------------------------------
370
+
371
+ id: (utterance_19)
372
+ Scores: (#C #S #D #I) 3 1 1 0
373
+ REF: ALLS WELL that ends well
374
+ HYP: **** OSWELL that ends well
375
+ Eval: D S
376
+
377
+ Correct = 60.0% 3 ( 90)
378
+ Substitutions = 20.0% 1 ( 20)
379
+ Deletions = 20.0% 1 ( 5)
380
+ Insertions = 0.0% 0 ( 0)
381
+
382
+ Errors = 40.0% 2 ( 25)
383
+
384
+ Ref. words = 5 (115)
385
+ Hyp. words = 4 (110)
386
+ Aligned words = 5 (115)
387
+
388
+ -------------------------------------------------------------------------------
389
+
390
+ id: (utterance_20)
391
+ Scores: (#C #S #D #I) 12 1 0 0
392
+ REF: JOE keaton disapproved of films and buster also had reservations about the medium
393
+ HYP: JOEL keaton disapproved of films and buster also had reservations about the medium
394
+ Eval: S
395
+
396
+ Correct = 92.3% 12 (102)
397
+ Substitutions = 7.7% 1 ( 21)
398
+ Deletions = 0.0% 0 ( 5)
399
+ Insertions = 0.0% 0 ( 0)
400
+
401
+ Errors = 7.7% 1 ( 26)
402
+
403
+ Ref. words = 13 (128)
404
+ Hyp. words = 13 (123)
405
+ Aligned words = 13 (128)
406
+
407
+ -------------------------------------------------------------------------------
408
+
409
+ id: (utterance_21)
410
+ Scores: (#C #S #D #I) 2 2 0 0
411
+ REF: SHELL be ALL right
412
+ HYP: J be A right
413
+ Eval: S S
414
+
415
+ Correct = 50.0% 2 (104)
416
+ Substitutions = 50.0% 2 ( 23)
417
+ Deletions = 0.0% 0 ( 5)
418
+ Insertions = 0.0% 0 ( 0)
419
+
420
+ Errors = 50.0% 2 ( 28)
421
+
422
+ Ref. words = 4 (132)
423
+ Hyp. words = 4 (127)
424
+ Aligned words = 4 (132)
425
+
426
+ -------------------------------------------------------------------------------
427
+
428
+ id: (utterance_22)
429
+ Scores: (#C #S #D #I) 1 0 0 0
430
+ REF: six
431
+ HYP: six
432
+ Eval:
433
+
434
+ Correct = 100.0% 1 (105)
435
+ Substitutions = 0.0% 0 ( 23)
436
+ Deletions = 0.0% 0 ( 5)
437
+ Insertions = 0.0% 0 ( 0)
438
+
439
+ Errors = 0.0% 0 ( 28)
440
+
441
+ Ref. words = 1 (133)
442
+ Hyp. words = 1 (128)
443
+ Aligned words = 1 (133)
444
+
445
+ -------------------------------------------------------------------------------
446
+
447
+ id: (utterance_23)
448
+ Scores: (#C #S #D #I) 3 1 1 0
449
+ REF: ALLS WELL that ends well
450
+ HYP: **** OSWELL that ends well
451
+ Eval: D S
452
+
453
+ Correct = 60.0% 3 (108)
454
+ Substitutions = 20.0% 1 ( 24)
455
+ Deletions = 20.0% 1 ( 6)
456
+ Insertions = 0.0% 0 ( 0)
457
+
458
+ Errors = 40.0% 2 ( 30)
459
+
460
+ Ref. words = 5 (138)
461
+ Hyp. words = 4 (132)
462
+ Aligned words = 5 (138)
463
+
464
+ -------------------------------------------------------------------------------
465
+
466
+ id: (utterance_24)
467
+ Scores: (#C #S #D #I) 12 1 0 0
468
+ REF: JOE keaton disapproved of films and buster also had reservations about the medium
469
+ HYP: JOEL keaton disapproved of films and buster also had reservations about the medium
470
+ Eval: S
471
+
472
+ Correct = 92.3% 12 (120)
473
+ Substitutions = 7.7% 1 ( 25)
474
+ Deletions = 0.0% 0 ( 6)
475
+ Insertions = 0.0% 0 ( 0)
476
+
477
+ Errors = 7.7% 1 ( 31)
478
+
479
+ Ref. words = 13 (151)
480
+ Hyp. words = 13 (145)
481
+ Aligned words = 13 (151)
482
+
483
+ -------------------------------------------------------------------------------
484
+
485
+ id: (utterance_25)
486
+ Scores: (#C #S #D #I) 2 2 0 0
487
+ REF: SHELL be ALL right
488
+ HYP: J be A right
489
+ Eval: S S
490
+
491
+ Correct = 50.0% 2 (122)
492
+ Substitutions = 50.0% 2 ( 27)
493
+ Deletions = 0.0% 0 ( 6)
494
+ Insertions = 0.0% 0 ( 0)
495
+
496
+ Errors = 50.0% 2 ( 33)
497
+
498
+ Ref. words = 4 (155)
499
+ Hyp. words = 4 (149)
500
+ Aligned words = 4 (155)
501
+
502
+ -------------------------------------------------------------------------------
503
+
predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn.sys ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+
3
+
4
+ SYSTEM SUMMARY PERCENTAGES by SPEAKER
5
+
6
+ ,-------------------------------------------------------------------------------------------------------------------------------------------------------------------------.
7
+ |/scratch/project_465000836/ipoloka/huggingface_asr/experiments/decred/commonvoice/DeCRED_small_cv/predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_hyp.trn|
8
+ |-------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
9
+ | SPKR | # Snt # Wrd | Corr Sub Del Ins Err S.Err |
10
+ |--------------------+------------------------------------+---------------------------------------------------------------------------------------------------------------|
11
+ | utt | 26 155 | 78.7 17.4 3.9 0.0 21.3 76.9 |
12
+ |=========================================================================================================================================================================|
13
+ | Sum/Avg | 26 155 | 78.7 17.4 3.9 0.0 21.3 76.9 |
14
+ |=========================================================================================================================================================================|
15
+ | Mean | 26.0 155.0 | 78.7 17.4 3.9 0.0 21.3 76.9 |
16
+ | S.D. | 0.0 0.0 | 0.0 0.0 0.0 0.0 0.0 0.0 |
17
+ | Median | 26.0 155.0 | 78.7 17.4 3.9 0.0 21.3 76.9 |
18
+ `-------------------------------------------------------------------------------------------------------------------------------------------------------------------------'
predictions_common_voice_13_en_common_voice_13_en_test_wer21.29_ref.trn ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_0)
2
+ shell be all right (utterance_1)
3
+ six (utterance_2)
4
+ alls well that ends well (utterance_3)
5
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_4)
6
+ shell be all right (utterance_5)
7
+ six (utterance_6)
8
+ alls well that ends well (utterance_7)
9
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_8)
10
+ shell be all right (utterance_9)
11
+ six (utterance_10)
12
+ alls well that ends well (utterance_11)
13
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_12)
14
+ shell be all right (utterance_13)
15
+ six (utterance_14)
16
+ alls well that ends well (utterance_15)
17
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_16)
18
+ shell be all right (utterance_17)
19
+ six (utterance_18)
20
+ alls well that ends well (utterance_19)
21
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_20)
22
+ shell be all right (utterance_21)
23
+ six (utterance_22)
24
+ alls well that ends well (utterance_23)
25
+ joe keaton disapproved of films and buster also had reservations about the medium (utterance_24)
26
+ shell be all right (utterance_25)
predictions_common_voice_13_en_common_voice_13_en_test_wer26.31.csv ADDED
The diff for this file is too large to render. See raw diff
 
predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn ADDED
The diff for this file is too large to render. See raw diff
 
predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn.dtl ADDED
The diff for this file is too large to render. See raw diff
 
predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn.snt.utt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2a2e626069436b4700a4c81206d42dd9fac798290a90be354f999f37e9be969
3
+ size 11492598
predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn.sys ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+
3
+
4
+ SYSTEM SUMMARY PERCENTAGES by SPEAKER
5
+
6
+ ,-------------------------------------------------------------------------------------------------------------------------------------------------------------------------.
7
+ |/scratch/project_465000836/ipoloka/huggingface_asr/experiments/decred/commonvoice/DeCRED_small_cv/predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_hyp.trn|
8
+ |-------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
9
+ | SPKR | # Snt # Wrd | Corr Sub Del Ins Err S.Err |
10
+ |-------------------+----------------------------------------+------------------------------------------------------------------------------------------------------------|
11
+ | utt | 16384 144172 | 77.2 17.8 5.1 3.5 26.3 69.5 |
12
+ |=========================================================================================================================================================================|
13
+ | Sum/Avg | 16384 144172 | 77.2 17.8 5.1 3.5 26.3 69.5 |
14
+ |=========================================================================================================================================================================|
15
+ | Mean | 16384.0 144172.0 | 77.2 17.8 5.1 3.5 26.3 69.5 |
16
+ | S.D. | 0.0 0.0 | 0.0 0.0 0.0 0.0 0.0 0.0 |
17
+ | Median | 16384.0 144172.0 | 77.2 17.8 5.1 3.5 26.3 69.5 |
18
+ `-------------------------------------------------------------------------------------------------------------------------------------------------------------------------'
predictions_common_voice_13_en_common_voice_13_en_test_wer26.31_ref.trn ADDED
The diff for this file is too large to render. See raw diff
 
predictions_common_voice_13_en_common_voice_13_en_test_wer99.13.csv ADDED
The diff for this file is too large to render. See raw diff
 
predictions_common_voice_13_en_common_voice_13_en_test_wer99.13_hyp.trn ADDED
The diff for this file is too large to render. See raw diff
 
predictions_common_voice_13_en_common_voice_13_en_test_wer99.13_ref.trn ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:610498329fd4c2cbe33c90d47b3294863967e0d466b0680feb038c21e1d4e281
3
+ size 5816