wav2vec2-efeat-300m-norwegian-colab

README.md

@@ -4,7 +4,7 @@ base_model: facebook/wav2vec2-lv-60-espeak-cv-ft
 tags:
 - generated_from_trainer
 datasets:
-- nb_samtale
+- voxpopuli
 metrics:
 - wer
 model-index:
@@ -14,15 +14,15 @@ model-index:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
     dataset:
-      name: nb_samtale
-      type: nb_samtale
-      config: annotations
+      name: voxpopuli
+      type: voxpopuli
+      config: fi
       split: test
-      args: annotations
+      args: fi
     metrics:
     - name: Wer
       type: wer
-      value: 1.0
+      value: 0.5082158970645575
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -30,10 +30,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # wav2vec2-efeat-300m-norwegian-colab
 
-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 nb_samtale 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.
 It achieves the following results on the evaluation set:
-- Loss: nan
-- Wer: 1.0
+- Loss: 2050.9434
+- Wer: 0.5082
 
 ## Model description
 
@@ -60,15 +60,15 @@ The following hyperparameters were used during training:
 - total_train_batch_size: 32
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
+- lr_scheduler_warmup_steps: 1000
 - num_epochs: 5
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss | Wer |
-|:-------------:|:-----:|:----:|:---------------:|:---:|
-| 0.0           | 2.57  | 400  | nan             | 1.0 |
+| Training Loss | Epoch | Step | Validation Loss | Wer    |
+|:-------------:|:-----:|:----:|:---------------:|:------:|
+| 1368.1253     | 2.57  | 400  | 2050.9434       | 0.5082 |
 
 
 ### Framework versions

config.json

@@ -1,6 +1,6 @@
 {
-  "_name_or_path": "facebook/wav2vec2-base-960h",
-  "activation_dropout": 0.1,
+  "_name_or_path": "facebook/wav2vec2-lv-60-espeak-cv-ft",
+  "activation_dropout": 0.0,
   "adapter_attn_dim": null,
   "adapter_kernel_size": 3,
   "adapter_stride": 2,
@@ -12,9 +12,9 @@
   "attention_dropout": 0.1,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "codevector_dim": 256,
+  "codevector_dim": 768,
   "contrastive_logits_temperature": 0.1,
-  "conv_bias": false,
+  "conv_bias": true,
   "conv_dim": [
     512,
     512,
@@ -45,42 +45,49 @@
   "ctc_loss_reduction": "sum",
   "ctc_zero_infinity": false,
   "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_dropout": 0.0,
-  "feat_extract_norm": "group",
+  "feat_extract_norm": "layer",
   "feat_proj_dropout": 0.1,
   "feat_quantizer_dropout": 0.0,
-  "final_dropout": 0.1,
+  "final_dropout": 0.0,
   "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_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": 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,
+  "proj_codevector_dim": 768,
   "tdnn_dilation": [
     1,
     2,
@@ -105,6 +112,6 @@
   "torch_dtype": "float32",
   "transformers_version": "4.35.2",
   "use_weighted_layer_sum": false,
-  "vocab_size": 32,
+  "vocab_size": 392,
   "xvector_output_dim": 512
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4a5e99a4ef90862044b90732650a471780d80a92a4a0d9280653a143ea208416
-size 377611120
+oid sha256:c954e97b96779ab1aaead05771a5e8a285650eeaa1f2436668083242695346f6
+size 1263414696

preprocessor_config.json

@@ -3,7 +3,7 @@
   "feature_extractor_type": "Wav2Vec2FeatureExtractor",
   "feature_size": 1,
   "padding_side": "right",
-  "padding_value": 0.0,
-  "return_attention_mask": false,
+  "padding_value": 0,
+  "return_attention_mask": true,
   "sampling_rate": 16000
 }

runs/Dec04_16-58-12_9bf356752520/events.out.tfevents.1701709165.9bf356752520.297.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b2816aaf5421fb42064cdd208c1859d03ca875cc041f8083563f964a4173535
+size 6821

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:45f76a17c798876d8e9909d4bdfcd3e7d4c35e0d0169ba97552e63bf35fda7a0
+oid sha256:a7ab742eb299af50a3a024b9a8865d8a1af061f29941fc31af85b97ae5cd7a4c
 size 4600