Icelandic-noAug-to-withAug

Chenchuhui/Icelandic-noAug-to-withAug

Icelandic-noAug-to-withAug is a automatic speech recognition model from Chenchuhui. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as cc-by-4.0.

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

Downloads · 30 days

20% of all-time downloads

All-time downloads

Public

Parameters

315M

31.5 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors1.3 GB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: cc-by-4.0
Model type: wav2vec2
Access: Public
Created: Dec 19, 2023
Updated: Dec 19, 2023
SHA: 36a6678b

Base models

Chenchuhui/Icelandic-finetuned-no-data-augmentation

Model card

From the Hugging Face model README

Icelandic-noAug-to-withAug

This model is a fine-tuned version of Chenchuhui/Icelandic-finetuned-no-data-augmentation on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.2435
Wer: 0.2215

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: 0.0003
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
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: 500
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.2	10	0.6673	0.4206
No log	0.4	20	0.5317	0.3602
No log	0.6	30	0.4229	0.3121
No log	0.8	40	0.3765	0.2819
0.7888	1.0	50	0.3486	0.2606
0.7888	1.2	60	0.3058	0.2394
0.7888	1.4	70	0.2923	0.2371
0.7888	1.6	80	0.2790	0.2271
0.7888	1.8	90	0.2507	0.2215
0.3794	2.0	100	0.2319	0.2114
0.3794	2.2	110	0.2187	0.2002
0.3794	2.4	120	0.2383	0.1991
0.3794	2.6	130	0.2320	0.1980
0.3794	2.8	140	0.2346	0.1980
0.2625	3.0	150	0.2286	0.1935
0.2625	3.2	160	0.2379	0.1924
0.2625	3.4	170	0.2134	0.1857
0.2625	3.6	180	0.2144	0.1890
0.2625	3.8	190	0.2218	0.1890
0.2221	4.0	200	0.2112	0.1935
0.2221	4.2	210	0.2041	0.1890
0.2221	4.4	220	0.1923	0.1745
0.2221	4.6	230	0.2031	0.1913
0.2221	4.8	240	0.2254	0.2036
0.1735	5.0	250	0.2232	0.2025
0.1735	5.2	260	0.1915	0.1812
0.1735	5.4	270	0.2135	0.1946
0.1735	5.6	280	0.2215	0.1969
0.1735	5.8	290	0.2204	0.1991
0.1813	6.0	300	0.2215	0.2047
0.1813	6.2	310	0.2201	0.2002
0.1813	6.4	320	0.2308	0.1991
0.1813	6.6	330	0.2474	0.2069
0.1813	6.8	340	0.2355	0.2081
0.2036	7.0	350	0.2274	0.2047
0.2036	7.2	360	0.2398	0.2081
0.2036	7.4	370	0.2162	0.1957
0.2036	7.6	380	0.2284	0.2047
0.2036	7.8	390	0.2402	0.1980
0.1953	8.0	400	0.1947	0.1946
0.1953	8.2	410	0.2114	0.2069
0.1953	8.4	420	0.2306	0.2237
0.1953	8.6	430	0.2182	0.2170
0.1953	8.8	440	0.2094	0.2170
0.1905	9.0	450	0.2671	0.2315
0.1905	9.2	460	0.3168	0.2136
0.1905	9.4	470	0.3005	0.2248
0.1905	9.6	480	0.2486	0.2260
0.1905	9.8	490	0.2349	0.2047
0.201	10.0	500	0.2435	0.2215

Framework versions

Transformers 4.35.2
Pytorch 2.1.0+cu121
Datasets 2.15.0
Tokenizers 0.15.0

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: cc-by-4.0
Created: Dec 19, 2023
Updated: Dec 19, 2023