End of training

README.md

@@ -7,7 +7,9 @@ tags:
 - '[finetuned_model, lj_speech11]'
 - generated_from_trainer
 datasets:
-- FYP/LJ-Speech1111
+- FYP/LJ-SpeechLJ
+metrics:
+- wer
 model-index:
 - name: SpeechT5 STT Wav2Vec2
   results: []
@@ -20,7 +22,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [facebook/wav2vec2-base-960h](https://huggingface.co/facebook/wav2vec2-base-960h) on the Lj-Speech dataset.
 It achieves the following results on the evaluation set:
-- Loss: nan
+- Loss: 252.7729
+- Wer: 1.0
 
 ## Model description
 
@@ -40,28 +43,35 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 0.0001
-- train_batch_size: 16
+- train_batch_size: 4
 - eval_batch_size: 8
 - seed: 42
-- gradient_accumulation_steps: 2
-- total_train_batch_size: 32
+- gradient_accumulation_steps: 4
+- 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: 100
-- num_epochs: 3
+- num_epochs: 5
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss |
-|:-------------:|:------:|:----:|:---------------:|
-| 0.0           | 1.1277 | 53   | nan             |
-| 0.0           | 2.2553 | 106  | nan             |
+| Training Loss | Epoch  | Step | Validation Loss | Wer    |
+|:-------------:|:------:|:----:|:---------------:|:------:|
+| 130.9264      | 0.5319 | 50   | 256.5228        | 0.9827 |
+| 134.1007      | 1.0638 | 100  | 256.2832        | 0.9827 |
+| 131.0841      | 1.5957 | 150  | 253.9561        | 0.9827 |
+| 132.4283      | 2.1277 | 200  | 254.4677        | 0.9827 |
+| 137.3693      | 2.6596 | 250  | 254.6855        | 1.0    |
+| 128.1369      | 3.1915 | 300  | 252.8348        | 1.0    |
+| 132.3826      | 3.7234 | 350  | 254.7122        | 1.0    |
+| 130.9401      | 4.2553 | 400  | 254.6629        | 1.0    |
+| 129.5693      | 4.7872 | 450  | 252.7729        | 1.0    |
 
 
 ### Framework versions
 
-- Transformers 4.41.2
+- Transformers 4.42.3
 - Pytorch 2.3.0+cu121
-- Datasets 2.19.2
+- Datasets 2.20.0
 - Tokenizers 0.19.1

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "facebook/wav2vec2-base-960h",
+  "_name_or_path": "Asim037/wav222vec222v2-stt",
   "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": [
@@ -103,7 +102,7 @@
     1
   ],
   "torch_dtype": "float32",
-  "transformers_version": "4.41.2",
+  "transformers_version": "4.42.3",
   "use_weighted_layer_sum": false,
   "vocab_size": 32,
   "xvector_output_dim": 512

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5db199d38dd677248551a11622537efee70e5c2f384ce3df3b342d75eba77c06
-size 377611120
+oid sha256:23187bc5c6795853de6f1b62ef6069d376bd0ba90923af2cb1475ee137ac80d3
+size 1261897368

preprocessor_config.json

@@ -4,6 +4,7 @@
   "feature_size": 1,
   "padding_side": "right",
   "padding_value": 0.0,
+  "processor_class": "Wav2Vec2Processor",
   "return_attention_mask": false,
   "sampling_rate": 16000
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76f27f16df85037b23d8b7afc5f0ecd8816397670aeed96a8b00316bdc351167
+oid sha256:3d7424636e70377767859147c81158f1e94c4224475503466285ac1090eb7530
 size 5048