wav2vec2-large-xls-r-300m-hungarian-colab

README.md

@@ -1,6 +1,6 @@
 ---
 license: apache-2.0
-base_model: facebook/wav2vec2-xls-r-300m
+base_model: facebook/wav2vec2-lv-60-espeak-cv-ft
 tags:
 - generated_from_trainer
 datasets:
@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # wav2vec2-large-xls-r-300m-hungarian-colab
 
-This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the voxpopuli dataset.
+This model is a fine-tuned version of [facebook/wav2vec2-lv-60-espeak-cv-ft](https://huggingface.co/facebook/wav2vec2-lv-60-espeak-cv-ft) on the voxpopuli dataset.
 
 ## Model description
 
@@ -35,15 +35,15 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 0.0003
-- train_batch_size: 16
+- train_batch_size: 8
 - eval_batch_size: 8
 - seed: 42
 - gradient_accumulation_steps: 2
-- total_train_batch_size: 32
+- total_train_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 500
-- num_epochs: 30
+- num_epochs: 50
 - mixed_precision_training: Native AMP
 
 ### Training results

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "facebook/wav2vec2-xls-r-300m",
+  "_name_or_path": "facebook/wav2vec2-lv-60-espeak-cv-ft",
   "activation_dropout": 0.0,
   "adapter_attn_dim": null,
   "adapter_kernel_size": 3,
@@ -9,7 +9,7 @@
   "architectures": [
     "Wav2Vec2ForCTC"
   ],
-  "attention_dropout": 0.0,
+  "attention_dropout": 0.1,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
   "codevector_dim": 768,
@@ -42,7 +42,7 @@
     2,
     2
   ],
-  "ctc_loss_reduction": "mean",
+  "ctc_loss_reduction": "sum",
   "ctc_zero_infinity": false,
   "diversity_loss_weight": 0.1,
   "do_stable_layer_norm": true,
@@ -50,23 +50,31 @@
   "feat_extract_activation": "gelu",
   "feat_extract_dropout": 0.0,
   "feat_extract_norm": "layer",
-  "feat_proj_dropout": 0.0,
+  "feat_proj_dropout": 0.1,
   "feat_quantizer_dropout": 0.0,
   "final_dropout": 0.0,
   "gradient_checkpointing": false,
   "hidden_act": "gelu",
-  "hidden_dropout": 0.0,
+  "hidden_dropout": 0.1,
   "hidden_size": 1024,
   "initializer_range": 0.02,
   "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
-  "layerdrop": 0.0,
+  "layerdrop": 0.1,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
   "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,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.075,
+  "mask_time_selection": "static",
   "model_type": "wav2vec2",
   "num_adapter_layers": 3,
   "num_attention_heads": 16,
@@ -78,7 +86,7 @@
   "num_hidden_layers": 24,
   "num_negatives": 100,
   "output_hidden_size": 1024,
-  "pad_token_id": 41,
+  "pad_token_id": 0,
   "proj_codevector_dim": 768,
   "tdnn_dilation": [
     1,
@@ -104,6 +112,6 @@
   "torch_dtype": "float32",
   "transformers_version": "4.35.2",
   "use_weighted_layer_sum": false,
-  "vocab_size": 44,
+  "vocab_size": 392,
   "xvector_output_dim": 512
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb7a49a25129d13c6765a153cbf37d66478ab25315057c487ab07ed2767bb256
-size 1261987880
+oid sha256:6f3dfac983bf67770f10ee2dd4b43d55b5037feb8e1166577b7fcb8e0d71efac
+size 1263414696

runs/Nov24_19-43-06_ac4520cd39c0/events.out.tfevents.1700855129.ac4520cd39c0.422.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1aab0ed53ed3d4271510336bdff1a6aa2bff93791652651d928bafe52d56c3de
+size 6585

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8e40e29aa39353bb449257b458c8bd60036226d296e2815f19711043b3e666db
+oid sha256:ea3358a91e8443cd7cde2a0a90a81fdd3ec4b8ccfc82b36fdd6d7282efe28a9f
 size 4600