up

config.json

@@ -1,7 +1,6 @@
 {
   "_name_or_path": "facebook/wav2vec2-base-960h",
   "activation_dropout": 0.1,
-  "apply_spec_augment": true,
   "architectures": [
     "UniSpeechSatForPreTraining"
   ],
@@ -9,7 +8,6 @@
   "bos_token_id": 1,
   "classifier_proj_size": 256,
   "codevector_dim": 256,
-  "contrastive_logits_temperature": 0.1,
   "conv_bias": false,
   "conv_dim": [
     512,
@@ -38,9 +36,6 @@
     2,
     2
   ],
-  "ctc_loss_reduction": "sum",
-  "ctc_zero_infinity": false,
-  "diversity_loss_weight": 0.1,
   "do_stable_layer_norm": false,
   "eos_token_id": 2,
   "feat_extract_activation": "gelu",
@@ -49,7 +44,6 @@
   "feat_proj_dropout": 0.1,
   "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.1,
-  "gradient_checkpointing": false,
   "hidden_act": "gelu",
   "hidden_dropout": 0.1,
   "hidden_dropout_prob": 0.1,
@@ -74,9 +68,8 @@
   "num_negatives": 100,
   "pad_token_id": 0,
   "proj_codevector_dim": 256,
-  "replace_prob": 0.5,
   "torch_dtype": "float32",
   "transformers_version": "4.12.0.dev0",
   "use_weighted_layer_sum": false,
-  "vocab_size": 256
+  "tokenizer_class": "Wav2Vec2CTCTokenizer"
 }

preprocessor_config.json

@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "return_attention_mask": false,
+  "sampling_rate": 16000
+}