med_v1_M01

mprzibilla/med_v1_M01

med_v1_M01 is a automatic speech recognition model from mprzibilla. Use it when you need speech turned into text. It is set up for transformers.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

2% of all-time downloads

All-time downloads

524

Public

Repo size

114 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin1.5 GB · 68%

At a glance

Task: Automatic Speech Recognition
Library: transformers
Model type: wav2vec2
Access: Public
Created: Oct 20, 2022
Updated: Dec 3, 2022
SHA: 80b2b7e0

Model card

From the Hugging Face model README

med_v1_M01

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

Cer: 0.8926
Loss: 3.2885
Wer: 1.0

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: 0.0001
train_batch_size: 20
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: 13500
num_epochs: 2000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Cer	Validation Loss	Wer
16.9472	200.0	27000	0.9041	3.2161	1.0
2.6444	400.0	54000	0.8959	3.2246	1.0
2.5434	600.0	81000	0.8942	3.2352	1.0
2.4925	800.0	108000	0.8959	3.2468	1.0
2.4631	1000.0	135000	0.8942	3.2656	1.0
2.4434	1200.0	162000	0.8942	3.2616	1.0
2.4292	1400.0	189000	0.8942	3.2840	1.0
2.4203	1600.0	216000	0.8926	3.2791	1.0
2.4148	1800.0	243000	0.8926	3.2866	1.0
2.4112	2000.0	270000	0.8926	3.2885	1.0

Framework versions

Transformers 4.18.0
Pytorch 1.10.2+cu102
Datasets 2.3.2
Tokenizers 0.12.1

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
Created: Oct 20, 2022
Updated: Dec 3, 2022