Training in progress, epoch 1

config.json

@@ -1,6 +1,6 @@
 {
-  "_name_or_path": "facebook/wav2vec2-base-el-voxpopuli-v2",
-  "activation_dropout": 0.0,
+  "_name_or_path": "jonatasgrosman/wav2vec2-large-xlsr-53-english",
+  "activation_dropout": 0.05,
   "adapter_attn_dim": null,
   "adapter_kernel_size": 3,
   "adapter_stride": 2,
@@ -14,7 +14,7 @@
   "classifier_proj_size": 256,
   "codevector_dim": 256,
   "contrastive_logits_temperature": 0.1,
-  "conv_bias": false,
+  "conv_bias": true,
   "conv_dim": [
     512,
     512,
@@ -42,20 +42,20 @@
     2,
     2
   ],
-  "ctc_loss_reduction": "sum",
-  "ctc_zero_infinity": false,
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": true,
   "diversity_loss_weight": 0.1,
-  "do_stable_layer_norm": false,
+  "do_stable_layer_norm": true,
   "eos_token_id": 2,
   "feat_extract_activation": "gelu",
-  "feat_extract_norm": "group",
-  "feat_proj_dropout": 0.1,
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.05,
   "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.0,
-  "freeze_feat_extract_train": true,
   "hidden_act": "gelu",
-  "hidden_dropout": 0.1,
-  "hidden_size": 768,
+  "hidden_dropout": 0.05,
+  "hidden_size": 1024,
   "id2label": {
     "0": "angry",
     "1": "sad",
@@ -65,7 +65,7 @@
     "5": "happy"
   },
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "label2id": {
     "angry": 0,
     "disgust": 2,
@@ -75,7 +75,7 @@
     "sad": 1
   },
   "layer_norm_eps": 1e-05,
-  "layerdrop": 0.0,
+  "layerdrop": 0.05,
   "mask_channel_length": 10,
   "mask_channel_min_space": 1,
   "mask_channel_other": 0.0,
@@ -91,18 +91,16 @@
   "mask_time_prob": 0.05,
   "mask_time_selection": "static",
   "model_type": "wav2vec2",
-  "no_mask_channel_overlap": false,
-  "no_mask_time_overlap": false,
   "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": [
@@ -129,6 +127,6 @@
   "torch_dtype": "float32",
   "transformers_version": "4.39.3",
   "use_weighted_layer_sum": false,
-  "vocab_size": 32,
+  "vocab_size": 33,
   "xvector_output_dim": 512
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f7b958f17b76c1787a07e0258b5c29ce20d066d74553d5686ea0517e816c8b3
-size 378306480
+oid sha256:6a5b62438e9e4ebb0170990d16e272d57a770a6c2aa5808954c088d1e3d9f2c0
+size 1262863416

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1025033b3c8580dad7e47c601c7fc591c12d6d799c08770a0faa6889b8c82d2f
+oid sha256:decb07fd61901eb7248ec185cc960ae2d2777a8530880439ac805c206e177d0d
 size 4920