add new models with high performance

README.md

@@ -23,6 +23,9 @@ model-index:
          value:  20.89%
 ---
 ## Evaluation on Common Voice FR Test
+The script used for training and evaluation can be found here: https://github.com/irebai/wav2vec2
+
+
 ```python
 import torch
 import torchaudio
@@ -73,6 +76,20 @@ print(wer.compute(predictions=result["predicted"], references=result["target"]))
 
 ## Results
 
+# v0.1
+
 WER=18.29%
 
 SER=71.44%
+
+# v1.0
+
+WER=15.97%
+
+CER=5.51%
+
+# v2.0
+
+WER=14.71%
+
+CER=5.06%

v1.0/config.json

@@ -0,0 +1,78 @@
+{
+  "_name_or_path": "/workspace/output_models/wav2vec2-large-xlsr-53/checkpoint-71900",
+  "activation_dropout": 0.055,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.094,
+  "bos_token_id": 1,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": false,
+  "do_stable_layer_norm": true,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.04,
+  "final_dropout": 0.0,
+  "gradient_checkpointing": true,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.047,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.041,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "mask_feature_length": 10,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.4,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_attention_heads": 16,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "pad_token_id": 46,
+  "pooling_type": "max",
+  "time_pooling_size": 1,
+  "transformers_version": "4.5.0.dev0",
+  "vocab_size": 47
+}

v1.0/preprocessor_config.json

@@ -0,0 +1,8 @@
+{
+  "do_normalize": true,
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

v1.0/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0922012f44439677c60c929ef48938ec2c3ae6ec717523cae13988a2d22bfd99
+size 1262144708

v1.0/results.txt

@@ -0,0 +1,2 @@
+WER=15.97816949508748
+CER=5.518223907734944

v1.0/special_tokens_map.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "pad_token": "<pad>"}

v1.0/tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": null, "eos_token": null, "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|"}

v1.0/vocab.json

@@ -0,0 +1 @@
+{"a": 0, "e": 1, "i": 2, "o": 3, "u": 4, "y": 5, "b": 6, "c": 7, "d": 8, "f": 9, "g": 10, "h": 11, "j": 12, "k": 13, "l": 14, "m": 15, "n": 16, "p": 17, "q": 18, "r": 19, "s": 20, "t": 21, "v": 22, "w": 23, "x": 24, "z": 25, "à": 26, "â": 27, "æ": 28, "ç": 29, "è": 30, "é": 31, "ê": 32, "ë": 33, "î": 34, "ï": 35, "ô": 36, "œ": 37, "ù": 38, "û": 39, "ü": 40, "ÿ": 41, "'": 42, "-": 43, "|": 44, "<unk>": 45, "<pad>": 46}

v2.0/config.json

@@ -0,0 +1,78 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
+  "activation_dropout": 0.055,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.094,
+  "bos_token_id": 1,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": false,
+  "do_stable_layer_norm": true,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.04,
+  "final_dropout": 0.0,
+  "gradient_checkpointing": true,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.047,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.041,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "mask_feature_length": 10,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.2,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_attention_heads": 16,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "pad_token_id": 1,
+  "pooling_type": "max",
+  "time_pooling_size": 1,
+  "transformers_version": "4.5.0.dev0",
+  "vocab_size": 47
+}

v2.0/preprocessor_config.json

@@ -0,0 +1,8 @@
+{
+  "do_normalize": true,
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

v2.0/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:536491b9fe83421402af342b175daba8ea082099c742492c4e5000403833b7bf
+size 1262144708

v2.0/results.txt

@@ -0,0 +1,2 @@
+WER=14.711290915777859
+CER=5.068944791354929

v2.0/special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<bos>", "eos_token": "<eos>", "unk_token": "<unk>", "pad_token": "<pad>"}

v2.0/tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<bos>", "eos_token": "<eos>", "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|"}

v2.0/vocab.json

@@ -0,0 +1 @@
+{"<unk>": 0, "<pad>": 1, "a": 2, "e": 3, "i": 4, "o": 5, "u": 6, "y": 7, "b": 8, "c": 9, "d": 10, "f": 11, "g": 12, "h": 13, "j": 14, "k": 15, "l": 16, "m": 17, "n": 18, "p": 19, "q": 20, "r": 21, "s": 22, "t": 23, "v": 24, "w": 25, "x": 26, "z": 27, "à": 28, "â": 29, "æ": 30, "ç": 31, "è": 32, "é": 33, "ê": 34, "ë": 35, "î": 36, "ï": 37, "ô": 38, "œ": 39, "ù": 40, "û": 41, "ü": 42, "ÿ": 43, "'": 44, "-": 45, "|": 46}