mdd-arabic-e4-xlsr1b-l1-twostage

omarZACK/mdd-arabic-e4-xlsr1b-l1-twostage

mdd-arabic-e4-xlsr1b-l1-twostage is a machine learning model from omarZACK. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers.

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

Downloads · 30 days

22% of all-time downloads

All-time downloads

Public

Parameters

964M

3.9 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors3.9 GB · 100%

At a glance

Library: transformers
Model type: wav2vec2
Access: Public
Created: Aug 16, 2026
Updated: Aug 16, 2026
SHA: 6b39f5ab

Model card

From the Hugging Face model README

mdd-arabic-e4-xlsr1b-l1-twostage

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

Loss: 0.2217
Per: 0.1611
Accuracy: 0.8967
Precision: 0.4798
Recall: 0.5142
F1 Macro: 0.4809

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: 5e-05
train_batch_size: 16
eval_batch_size: 16
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: cosine
lr_scheduler_warmup_steps: 112
num_epochs: 20
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Per	Accuracy	Precision	Recall	F1 Macro
1.4239	1.0	113	0.4629	0.1889	0.8419	0.4169	0.4038	0.4092
1.9971	2.0	226	0.4099	0.1844	0.8461	0.4005	0.4333	0.4019
1.6982	3.0	339	0.3874	0.1888	0.8535	0.4170	0.4569	0.4253
1.5809	4.0	452	0.3696	0.1763	0.8586	0.4397	0.4347	0.4324
1.6826	5.0	565	0.3562	0.1794	0.8616	0.4503	0.4465	0.4442
1.3899	6.0	678	0.3482	0.1840	0.8623	0.4559	0.4506	0.4513
1.2339	7.0	791	0.3266	0.1719	0.8690	0.4634	0.4523	0.4519
1.4888	8.0	904	0.3042	0.1746	0.8718	0.4591	0.4672	0.4581
1.1724	9.0	1017	0.3066	0.1767	0.8734	0.4747	0.4610	0.4659
1.3263	10.0	1130	0.2804	0.1648	0.8817	0.4772	0.4816	0.4700
1.1646	11.0	1243	0.2656	0.1644	0.8845	0.4796	0.4856	0.4739
1.0695	12.0	1356	0.2573	0.1746	0.8842	0.4702	0.4999	0.4771
1.3837	13.0	1469	0.2479	0.1587	0.8898	0.4671	0.4978	0.4623
1.1185	14.0	1582	0.2388	0.1629	0.8916	0.4847	0.4996	0.4825
1.3106	15.0	1695	0.2340	0.1648	0.8936	0.4753	0.5204	0.4791
1.1250	16.0	1808	0.2275	0.1639	0.8956	0.4823	0.5124	0.4851
1.2460	17.0	1921	0.2233	0.1608	0.8978	0.4821	0.5159	0.4840
0.9342	18.0	2034	0.2217	0.1611	0.8967	0.4798	0.5142	0.4809

Framework versions

Transformers 5.15.0
Pytorch 2.13.0+cu130
Datasets 5.0.1
Tokenizers 0.22.2

Library: transformers
Type: wav2vec2
Created: Aug 16, 2026
Updated: Aug 16, 2026