xuliu15/English_32r_LoRA_10mins_new

LoRA-English-10m

This model is a fine-tuned version of openai/whisper-small on the librispeech dataset. It achieves the following results on the evaluation set:

• Loss: 0.2052

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: 0.001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	6	2.4022
No log	2.0	12	1.7738
No log	3.0	18	1.0358
No log	4.0	24	0.7506
0.9696	5.0	30	0.6126
0.9696	6.0	36	0.5719
0.9696	7.0	42	0.5161
0.9696	8.0	48	0.3983
0.2101	9.0	54	0.2375
0.2101	10.0	60	0.2141
0.2101	11.0	66	0.1953
0.2101	12.0	72	0.1938
0.0125	13.0	78	0.1997
0.0125	14.0	84	0.2035
0.0125	15.0	90	0.2038
0.0125	16.0	96	0.2036
0.0015	17.0	102	0.2023
0.0015	18.0	108	0.2017
0.0015	19.0	114	0.2016
0.0015	20.0	120	0.2015
0.0009	21.0	126	0.2017
0.0009	22.0	132	0.2022
0.0009	23.0	138	0.2025
0.0009	24.0	144	0.2029
0.0007	25.0	150	0.2033
0.0007	26.0	156	0.2034
0.0007	27.0	162	0.2036
0.0007	28.0	168	0.2039
0.0007	29.0	174	0.2042
0.0007	30.0	180	0.2045
0.0007	31.0	186	0.2047
0.0007	32.0	192	0.2048
0.0007	33.0	198	0.2048
0.0006	34.0	204	0.2049
0.0006	35.0	210	0.2049
0.0006	36.0	216	0.2050
0.0006	37.0	222	0.2051
0.0006	38.0	228	0.2051
0.0006	39.0	234	0.2052
0.0006	40.0	240	0.2052

Framework versions

• PEFT 0.10.0
• Transformers 4.40.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1