wav2vec2-large-mms-1b-kazakh-speech2ner-ksc_synthetic-4b-10ep

This model is a fine-tuned version of facebook/mms-1b-all on the None dataset. It achieves the following results on the evaluation set:

• Loss: nan
• 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: 1e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0	0.15	8000	nan	1.0
0.0	0.3	16000	nan	1.0
0.0	0.44	24000	nan	1.0
0.0	0.59	32000	nan	1.0
0.0	0.74	40000	nan	1.0
0.0	0.89	48000	nan	1.0
0.0	1.03	56000	nan	1.0
0.0	1.18	64000	nan	1.0
0.0	1.33	72000	nan	1.0
0.0	1.48	80000	nan	1.0
0.0	1.62	88000	nan	1.0
0.0	1.77	96000	nan	1.0
0.0	1.92	104000	nan	1.0
0.0	2.07	112000	nan	1.0
0.0	2.21	120000	nan	1.0
0.0	2.36	128000	nan	1.0
0.0	2.51	136000	nan	1.0
0.0	2.66	144000	nan	1.0
0.0	2.8	152000	nan	1.0
0.0	2.95	160000	nan	1.0
0.0	3.1	168000	nan	1.0
0.0	3.25	176000	nan	1.0
0.0	3.39	184000	nan	1.0
0.0	3.54	192000	nan	1.0
0.0	3.69	200000	nan	1.0
0.0	3.84	208000	nan	1.0
0.0	3.98	216000	nan	1.0
0.0	4.13	224000	nan	1.0
0.0	4.28	232000	nan	1.0
0.0	4.43	240000	nan	1.0
0.0	4.57	248000	nan	1.0
0.0	4.72	256000	nan	1.0
0.0	4.87	264000	nan	1.0
0.0	5.02	272000	nan	1.0
0.0	5.16	280000	nan	1.0
0.0	5.31	288000	nan	1.0
0.0	5.46	296000	nan	1.0
0.0	5.61	304000	nan	1.0
0.0	5.75	312000	nan	1.0
0.0	5.9	320000	nan	1.0
0.0	6.05	328000	nan	1.0
0.0	6.2	336000	nan	1.0
0.0	6.34	344000	nan	1.0
0.0	6.49	352000	nan	1.0
0.0	6.64	360000	nan	1.0
0.0	6.79	368000	nan	1.0
0.0	6.93	376000	nan	1.0
0.0	7.08	384000	nan	1.0
0.0	7.23	392000	nan	1.0
0.0	7.38	400000	nan	1.0
0.0	7.52	408000	nan	1.0
0.0	7.67	416000	nan	1.0
0.0	7.82	424000	nan	1.0
0.0	7.97	432000	nan	1.0
0.0	8.11	440000	nan	1.0
0.0	8.26	448000	nan	1.0
0.0	8.41	456000	nan	1.0
0.0	8.56	464000	nan	1.0
0.0	8.7	472000	nan	1.0
0.0	8.85	480000	nan	1.0
0.0	9.0	488000	nan	1.0
0.0	9.15	496000	nan	1.0
0.0	9.29	504000	nan	1.0
0.0	9.44	512000	nan	1.0
0.0	9.59	520000	nan	1.0
0.0	9.74	528000	nan	1.0
0.0	9.88	536000	nan	1.0

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3