seamless_m4t_v2_cv_no_lang_id_common_voices_49371_trial

This model is a fine-tuned version of facebook/seamless-m4t-v2-large on the common_voice_22_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4023
• Global Wer: 28.3223

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-06
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss	Global Wer
9.4728	0.0411	50	1.8962	112.7886
7.0816	0.0821	100	1.7341	111.2160
2.5123	0.1232	150	1.5378	109.3288
0.7678	0.1642	200	0.5763	32.2217
0.5966	0.2053	250	0.4997	30.9064
0.3945	0.2463	300	0.4659	29.7305
0.4612	0.2874	350	0.4367	29.1443
0.4049	0.3284	400	0.4207	28.6404
0.3421	0.3695	450	0.4130	28.5867
0.3216	0.4105	500	0.4023	28.3223

Framework versions

• Transformers 5.0.0.dev0
• Pytorch 2.9.0+cu126
• Datasets 3.6.0
• Tokenizers 0.22.2