Training in progress, epoch 1

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "facebook/hubert-large-ls960-ft",
+  "_name_or_path": "facebook/hubert-base-ls960",
   "activation_dropout": 0.1,
   "apply_spec_augment": true,
   "architectures": [
@@ -8,7 +8,7 @@
   "attention_dropout": 0.1,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "conv_bias": true,
+  "conv_bias": false,
   "conv_dim": [
     512,
     512,
@@ -38,12 +38,11 @@
   ],
   "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_extract_norm": "group",
   "feat_proj_dropout": 0.1,
   "feat_proj_layer_norm": true,
   "final_dropout": 0.1,
@@ -51,9 +50,9 @@
   "hidden_act": "gelu",
   "hidden_dropout": 0.1,
   "hidden_dropout_prob": 0.1,
-  "hidden_size": 1024,
+  "hidden_size": 768,
   "initializer_range": 0.02,
-  "intermediate_size": 4096,
+  "intermediate_size": 3072,
   "layer_norm_eps": 1e-05,
   "layerdrop": 0.1,
   "mask_feature_length": 10,
@@ -63,12 +62,13 @@
   "mask_time_min_masks": 2,
   "mask_time_prob": 0.05,
   "model_type": "hubert",
-  "num_attention_heads": 16,
+  "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_hidden_layers": 12,
   "pad_token_id": 0,
+  "tokenizer_class": "Wav2Vec2CTCTokenizer",
   "torch_dtype": "float32",
   "transformers_version": "4.40.2",
   "use_weighted_layer_sum": false,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:910cd1821ea03c46663da9af41f44b322f6cac6a82612adb37d6695ee1f1e3e2
-size 1262858408
+oid sha256:11709c2a0520ca65f2aa17c352d034ba21ebcb59993972436ab1e30b91661c5c
+size 378301896

preprocessor_config.json

@@ -4,6 +4,6 @@
   "feature_size": 1,
   "padding_side": "right",
   "padding_value": 0,
-  "return_attention_mask": true,
+  "return_attention_mask": false,
   "sampling_rate": 16000
 }

runs/Jun16_15-01-28_WIN-CEJF2NB9KFH/events.out.tfevents.1718521288.WIN-CEJF2NB9KFH.86804.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91b59b189d8b799e962c96879c214ecf4ce16f121302074d400cea289d63f19c
+size 26324

runs/Jun17_19-15-38_WIN-CEJF2NB9KFH/events.out.tfevents.1718622939.WIN-CEJF2NB9KFH.94304.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2f1a0e76028e047d0ecc70ce12382072e4a262eb5694ab5003a9f799e5486c8
+size 5608

runs/Jun17_19-19-29_WIN-CEJF2NB9KFH/events.out.tfevents.1718623169.WIN-CEJF2NB9KFH.52692.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e4501e13a8d54a0aba9b114bf2d9d9246f304356da87117828f2ce71ecbe0e9
+size 5608

runs/Jun17_19-23-58_WIN-CEJF2NB9KFH/events.out.tfevents.1718623438.WIN-CEJF2NB9KFH.57652.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01fdd37b0f2c8b821fd9a02ce23671f3a6051d83b3508c3fad0ab0e7118f0c7a
+size 6141

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a036e8ae43a4f80d452b222297f95d085931cbdaa111203794f6c2278197abe
+oid sha256:82fda100521421396e64d2151a6e860ea8cd8a754f7d202b6f22021dc2feb7bc
 size 4539