Training in progress, step 1000

config.json

@@ -1,20 +1,17 @@
 {
-  "_name_or_path": "facebook/wav2vec2-xls-r-300m",
-  "activation_dropout": 0.0,
-  "adapter_attn_dim": null,
-  "adapter_kernel_size": 3,
-  "adapter_stride": 2,
-  "add_adapter": false,
+  "_name_or_path": "team-lucid/hubert-base-korean",
+  "activation_dropout": 0.1,
   "apply_spec_augment": true,
   "architectures": [
-    "Wav2Vec2ForCTC"
+    "HubertForCTC"
   ],
-  "attention_dropout": 0.0,
+  "attention_dropout": 0.1,
+  "auto_map": {
+    "FlaxAutoModel": "team-lucid/hubert-base-korean--modeling_flax_hubert.FlaxHubertModel"
+  },
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "codevector_dim": 768,
-  "contrastive_logits_temperature": 0.1,
-  "conv_bias": true,
+  "conv_bias": false,
   "conv_dim": [
     512,
     512,
@@ -42,68 +39,42 @@
     2,
     2
   ],
-  "ctc_loss_reduction": "mean",
+  "ctc_loss_reduction": "sum",
   "ctc_zero_infinity": false,
-  "diversity_loss_weight": 0.1,
-  "do_stable_layer_norm": true,
+  "do_stable_layer_norm": false,
   "eos_token_id": 2,
   "feat_extract_activation": "gelu",
   "feat_extract_dropout": 0.0,
-  "feat_extract_norm": "layer",
-  "feat_proj_dropout": 0.0,
-  "feat_quantizer_dropout": 0.0,
-  "final_dropout": 0.0,
+  "feat_extract_norm": "group",
+  "feat_proj_dropout": 0.1,
+  "feat_proj_layer_norm": true,
+  "final_dropout": 0.1,
+  "final_proj_size": 256,
   "gradient_checkpointing": false,
   "hidden_act": "gelu",
-  "hidden_dropout": 0.0,
-  "hidden_size": 1024,
+  "hidden_dropout": 0.1,
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
   "initializer_range": 0.02,
-  "intermediate_size": 4096,
+  "intermediate_size": 3072,
   "layer_norm_eps": 1e-05,
-  "layerdrop": 0.0,
+  "layerdrop": 0.1,
   "mask_feature_length": 10,
   "mask_feature_min_masks": 0,
   "mask_feature_prob": 0.0,
   "mask_time_length": 10,
   "mask_time_min_masks": 2,
   "mask_time_prob": 0.05,
-  "model_type": "wav2vec2",
-  "num_adapter_layers": 3,
-  "num_attention_heads": 16,
-  "num_codevector_groups": 2,
-  "num_codevectors_per_group": 320,
+  "model_type": "hubert",
+  "num_attention_heads": 12,
   "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,
-  "pad_token_id": 49,
-  "proj_codevector_dim": 768,
-  "tdnn_dilation": [
-    1,
-    2,
-    3,
-    1,
-    1
-  ],
-  "tdnn_dim": [
-    512,
-    512,
-    512,
-    512,
-    1500
-  ],
-  "tdnn_kernel": [
-    5,
-    3,
-    3,
-    1,
-    1
-  ],
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "tokenizer_class": "Wav2Vec2CTCTokenizer",
   "torch_dtype": "float32",
   "transformers_version": "4.44.2",
   "use_weighted_layer_sum": false,
-  "vocab_size": 52,
-  "xvector_output_dim": 512
+  "vocab_size": 53
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2d5739e8c9fe1e10513d682c0251e2944049230dfc48789219f20d003a268e77
-size 1262020680
+oid sha256:d812d5b3d52cf248d9d642410a22b87d113979ddd815eef4805863e137e711fa
+size 377675308

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5e55eeba14618f8026e7588c3b3c64acd8e267cb78986913c8b9be28bda60807
+oid sha256:4d7ebf4fee5cac7a1229d195422f37577810acf20f9cc6397d9217415776b86a
 size 5112