README.md

metadata

library_name: transformers
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: MSP-Fusion
    results: []

MSP-Fusion

This model is a fine-tuned version of on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4092
• Wer: 0.2350
• Cer: 0.1167

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: 0.0002
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• 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: 500
• training_steps: 10000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.9948	0.05	500	0.4387	0.2571	0.1277
1.1610	0.1	1000	0.4477	0.2550	0.1237
0.9327	0.15	1500	0.4221	0.2414	0.1188
1.0052	0.2	2000	0.4231	0.2476	0.1231
0.9512	0.25	2500	0.4339	0.2556	0.1288
0.9423	0.3	3000	0.4389	0.2580	0.1278
0.8131	0.35	3500	0.4282	0.2489	0.1247
0.9237	0.4	4000	0.4336	0.2464	0.1219
0.9842	0.45	4500	0.4305	0.2559	0.1290
1.0400	0.5	5000	0.4241	0.2426	0.1199
0.9045	0.55	5500	0.4265	0.2449	0.1215
0.9655	0.6	6000	0.4266	0.2430	0.1204
0.8683	0.65	6500	0.4139	0.2379	0.1178
0.8457	0.7	7000	0.4124	0.2372	0.1178
0.9675	0.75	7500	0.4133	0.2369	0.1189
0.9740	0.8	8000	0.4178	0.2406	0.1208
0.9536	0.85	8500	0.4092	0.2350	0.1167
0.9858	0.9	9000	0.4144	0.2378	0.1184
0.9754	0.95	9500	0.4113	0.2366	0.1174
0.9603	1.0	10000	0.4126	0.2373	0.1177

Framework versions

• Transformers 5.10.2
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2