README.md

metadata

library_name: transformers
license: apache-2.0
base_model: facebook/wav2vec2-lv-60-espeak-cv-ft
tags:
  - generated_from_trainer
model-index:
  - name: wav2vec2-lv-60-espeak-cv-ft-custom_vocab-OtherDiacritics-ds-f4
    results: []

wav2vec2-lv-60-espeak-cv-ft-custom_vocab-OtherDiacritics-ds-f4

This model is a fine-tuned version of facebook/wav2vec2-lv-60-espeak-cv-ft on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2274
• Per: 0.2720

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 3e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Per
17.5919	0.7194	400	4.5477	0.9998
4.4013	1.4388	800	4.0848	0.9998
4.0797	2.1583	1200	3.6617	0.9998
3.2434	2.8777	1600	1.5953	0.4748
2.0985	3.5971	2000	0.9012	0.3548
1.6061	4.3165	2400	0.6137	0.3131
1.3735	5.0360	2800	0.4757	0.2970
1.2007	5.7554	3200	0.4016	0.2898
1.1208	6.4748	3600	0.3516	0.2822
1.0346	7.1942	4000	0.3281	0.2825
0.9946	7.9137	4400	0.3055	0.2807
0.9412	8.6331	4800	0.2925	0.2794
0.9012	9.3525	5200	0.2809	0.2776
0.8657	10.0719	5600	0.2775	0.2776
0.859	10.7914	6000	0.2680	0.2762
0.847	11.5108	6400	0.2662	0.2751
0.8133	12.2302	6800	0.2616	0.2753
0.78	12.9496	7200	0.2563	0.2744
0.7681	13.6691	7600	0.2550	0.2751
0.7648	14.3885	8000	0.2500	0.2743
0.7517	15.1079	8400	0.2485	0.2748
0.7606	15.8273	8800	0.2408	0.2731
0.7295	16.5468	9200	0.2407	0.2732
0.7193	17.2662	9600	0.2420	0.2718
0.7135	17.9856	10000	0.2376	0.2719
0.6955	18.7050	10400	0.2365	0.2726
0.6812	19.4245	10800	0.2368	0.2726
0.6962	20.1439	11200	0.2346	0.2727
0.6812	20.8633	11600	0.2360	0.2740
0.6825	21.5827	12000	0.2312	0.2729
0.6835	22.3022	12400	0.2307	0.2732
0.6704	23.0216	12800	0.2282	0.2732
0.6588	23.7410	13200	0.2299	0.2732
0.6738	24.4604	13600	0.2264	0.2719
0.6596	25.1799	14000	0.2259	0.2720
0.6559	25.8993	14400	0.2293	0.2719
0.6248	26.6187	14800	0.2266	0.2716
0.649	27.3381	15200	0.2283	0.2724
0.6443	28.0576	15600	0.2260	0.2725
0.6373	28.7770	16000	0.2273	0.2729
0.6233	29.4964	16400	0.2274	0.2720

Framework versions

• Transformers 4.57.6
• Pytorch 2.9.1+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2