seamless-m4t_v2_iv_no_lang_id_indic_voices_49371_trial

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

• Loss: 0.7086
• Global Wer: 38.3315

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
8.5451	0.0122	50	2.5372	119.7383
6.5585	0.0244	100	2.2185	117.3937
3.6647	0.0365	150	2.0342	113.6314
1.7786	0.0487	200	1.0246	49.7819
1.0970	0.0609	250	0.8524	51.0360
1.0066	0.0731	300	0.8002	39.9128
0.8442	0.0853	350	0.7639	39.1494
0.8565	0.0974	400	0.7344	38.3315
0.8217	0.1096	450	0.7234	38.6587
0.8444	0.1218	500	0.7086	38.3315

Framework versions

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