learning_rate: 5e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
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: 150
training_steps: 15000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.9401	1.3528	250	1.6130	0.7243	0.2977
0.9312	2.7056	500	1.4246	0.7076	0.2833
0.8869	4.0543	750	1.3945	0.6974	0.2798
0.8336	5.4071	1000	1.3563	0.6888	0.2753
0.8381	6.7598	1250	1.3552	0.6835	0.2730
0.8299	8.1085	1500	1.3419	0.6820	0.2725
0.8211	9.4613	1750	1.3233	0.6783	0.2703
0.8279	10.8141	2000	1.3269	0.6768	0.2708
0.8293	12.1628	2250	1.3268	0.6750	0.2699
0.8453	13.5156	2500	1.3184	0.6768	0.2702
0.7781	14.8684	2750	1.3205	0.6768	0.2695
0.8421	16.2171	3000	1.3023	0.6744	0.2687
0.8303	17.5699	3250	1.3095	0.6755	0.2689
0.8275	18.9227	3500	1.3081	0.6749	0.2690
0.826	20.2714	3750	1.3081	0.6747	0.2684
0.7939	21.6242	4000	1.3161	0.6753	0.2683
0.812	22.9769	4250	1.3081	0.6746	0.2683
0.7883	24.3256	4500	1.3087	0.6736	0.2681
0.8132	25.6784	4750	1.3077	0.6731	0.2680
0.8081	27.0271	5000	1.3100	0.6721	0.2676
0.8255	28.3799	5250	1.3150	0.6724	0.2681
0.7873	29.7327	5500	1.3116	0.6737	0.2682
0.7799	31.0814	5750	1.3161	0.6731	0.2678
0.8041	32.4342	6000	1.3123	0.6747	0.2680
0.8082	33.7870	6250	1.3108	0.6734	0.2682
0.8392	35.1357	6500	1.3091	0.6749	0.2677
0.7876	36.4885	6750	1.3133	0.6743	0.2681
0.7956	37.8412	7000	1.3102	0.6739	0.2680
0.8283	39.1900	7250	1.3107	0.6733	0.2676
0.7912	40.5427	7500	1.3103	0.6724	0.2677
0.8154	41.8955	7750	1.3107	0.6732	0.2677
0.8217	43.2442	8000	1.3071	0.6729	0.2675
0.8535	44.5970	8250	1.3120	0.6739	0.2677
0.8453	45.9498	8500	1.3116	0.6735	0.2675
0.8383	47.2985	8750	1.3095	0.6735	0.2681
0.8032	48.6513	9000	1.3111	0.6740	0.2676
0.7492	50.0	9250	1.3111	0.6716	0.2677
0.7978	51.3528	9500	1.3135	0.6734	0.2681
0.8083	52.7056	9750	1.3118	0.6733	0.2676
0.7924	54.0543	10000	1.3119	0.6732	0.2675
0.8447	55.4071	10250	1.3115	0.6727	0.2676
0.8506	56.7598	10500	1.3110	0.6733	0.2678
0.781	58.1085	10750	1.3111	0.6736	0.2680
0.8132	59.4613	11000	1.3114	0.6732	0.2674
0.8167	60.8141	11250	1.3099	0.6732	0.2678
0.8181	62.1628	11500	1.3099	0.6740	0.2677
0.7717	63.5156	11750	1.3114	0.6738	0.2676
0.793	64.8684	12000	1.3112	0.6740	0.2676
0.8516	66.2171	12250	1.3111	0.6735	0.2677
0.7858	67.5699	12500	1.3114	0.6739	0.2676
0.8055	68.9227	12750	1.3115	0.6735	0.2674
0.7659	70.2714	13000	1.3120	0.6737	0.2675
0.7435	71.6242	13250	1.3124	0.6743	0.2675
0.8563	72.9769	13500	1.3112	0.6733	0.2677
0.8208	74.3256	13750	1.3117	0.6735	0.2678
0.7837	75.6784	14000	1.3115	0.6737	0.2674
0.7335	77.0271	14250	1.3116	0.6735	0.2677
0.7955	78.3799	14500	1.3122	0.6737	0.2679
0.7779	79.7327	14750	1.3113	0.6746	0.2680
0.767	81.0814	15000	1.3116	0.6743	0.2680

Framework versions

Transformers 4.52.0
Pytorch 2.7.1+cu118
Datasets 3.6.0
Tokenizers 0.21.4

KasuleTrevor/cdli-xlsr-lg-xlsr-full5e5-a40

cdli-xlsr-lg-xlsr-full5e5-a40

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

KasuleTrevor/
cdli-xlsr-lg-xlsr-full5e5-a40