Upload Wav2Vec2ForCTC

README.md

@@ -1,13 +1,13 @@
 ---
+base_model: facebook/wav2vec2-base-960h
+datasets:
+- FYP/LJ-Speech1111LJ
 language:
 - eng
 license: apache-2.0
-base_model: facebook/wav2vec2-base-960h
 tags:
 - '[finetuned_model, lj_speech11]'
 - generated_from_trainer
-datasets:
-- FYP/LJ-Speech1111LJ
 model-index:
 - name: SpeechT5 STT Wav2Vec2
   results: []

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "facebook/wav2vec2-base-960h",
+  "_name_or_path": "facebook/wav2vec2-large-960h",
   "activation_dropout": 0.1,
   "adapter_attn_dim": null,
   "adapter_kernel_size": 3,
@@ -50,16 +50,15 @@
   "feat_extract_activation": "gelu",
   "feat_extract_dropout": 0.0,
   "feat_extract_norm": "group",
-  "feat_proj_dropout": 0.1,
+  "feat_proj_dropout": 0.0,
   "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.1,
-  "gradient_checkpointing": false,
   "hidden_act": "gelu",
   "hidden_dropout": 0.1,
   "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "hidden_size": 1024,
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
   "layerdrop": 0.1,
   "mask_feature_length": 10,
@@ -70,15 +69,15 @@
   "mask_time_prob": 0.05,
   "model_type": "wav2vec2",
   "num_adapter_layers": 3,
-  "num_attention_heads": 12,
+  "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": 12,
+  "num_hidden_layers": 24,
   "num_negatives": 100,
-  "output_hidden_size": 768,
+  "output_hidden_size": 1024,
   "pad_token_id": 0,
   "proj_codevector_dim": 256,
   "tdnn_dilation": [

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d84fee48c0b18002daa7ca6fe3f6300d2743e5717de5801a2d6039385ffd591d
-size 377611120
+oid sha256:29efe5b81b4080d96694ddee2154e817419674ce8286162e8b712130d20cf306
+size 1261897368