Upload Wav2Vec2BERTForPreTraining (#1)

- Upload Wav2Vec2BERTForPreTraining (b164dfe01704c51035c024730fcdf188e6323c48)

config.json

@@ -5,53 +5,24 @@
   "add_adapter": false,
   "apply_spec_augment": false,
   "architectures": [
-    "Wav2Vec2ConformerForPreTraining"
+    "Wav2Vec2BERTForPreTraining"
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 1,
   "classifier_proj_size": 768,
   "codevector_dim": 768,
-  "conformer_conv_dropout": 0.0,
+  "conformer_conv_dropout": 0.1,
   "contrastive_logits_temperature": 0.1,
-  "conv_bias": false,
   "conv_depthwise_kernel_size": 31,
-  "conv_dim": [
-    512,
-    512,
-    512,
-    512,
-    512,
-    512,
-    160
-  ],
-  "conv_kernel": [
-    10,
-    3,
-    3,
-    3,
-    3,
-    2,
-    2
-  ],
-  "conv_stride": [
-    5,
-    2,
-    2,
-    2,
-    2,
-    2,
-    2
-  ],
   "ctc_loss_reduction": "sum",
   "ctc_zero_infinity": false,
   "diversity_loss_weight": 0.1,
   "eos_token_id": 2,
-  "feat_extract_activation": "gelu",
-  "feat_extract_norm": "group",
   "feat_proj_dropout": 0.0,
   "feat_quantizer_dropout": 0.0,
+  "feature_projection_input_dim": 160,
   "final_dropout": 0.1,
-  "hidden_act": "swish",
+  "hidden_act": "gelu",
   "hidden_dropout": 0.0,
   "hidden_size": 1024,
   "initializer_range": 0.02,
@@ -64,17 +35,13 @@
   "mask_feature_prob": 0.0,
   "mask_time_length": 10,
   "mask_time_min_masks": 2,
-  "mask_time_prob": 0.00,
+  "mask_time_prob": 0.05,
   "max_source_positions": 5000,
-  "model_type": "wav2vec2-conformer",
-  "non_causal_depth_wise_conv": false,
-  "num_adapter_layers": 0,
+  "model_type": "wav2vec2-bert",
+  "num_adapter_layers": 1,
   "num_attention_heads": 16,
   "num_codevector_groups": 2,
   "num_codevectors_per_group": 320,
-  "num_conv_pos_embedding_groups": 16,
-  "num_conv_pos_embeddings": 128,
-  "num_feat_extract_layers": 7,
   "num_hidden_layers": 24,
   "num_negatives": 100,
   "output_hidden_size": 1024,
@@ -83,9 +50,6 @@
   "proj_codevector_dim": 768,
   "right_max_position_embeddings": 8,
   "rotary_embedding_base": 10000,
-  "skip_encoder_layer_norm": true,
-  "skip_feature_encoder": true,
-  "skip_pos_conv_embed": true,
   "tdnn_dilation": [
     1,
     2,
@@ -109,7 +73,7 @@
   ],
   "torch_dtype": "float32",
   "transformers_version": "4.37.0.dev0",
-  "use_adapter_with_attention": false,
+  "use_intermediate_ffn_before_adapter": false,
   "use_weighted_layer_sum": false,
   "vocab_size": null,
   "xvector_output_dim": 512

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f09e925ebc936861e5d1f7320aff6e7bdd7f059c3becc3cffe79e1163eb09d7
-size 2328985440
+oid sha256:4ce9f7bd65da973a161b8c1a276f8721eddd6c9add60970d4426065a44176e34
+size 2328981576