qwen2_5_omni_text_image_half

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7678
• Token Acc: 0.7693

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: 0.0001
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• total_eval_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Token Acc
0.9669	0.4211	50	0.9984	0.7193
0.9312	0.8421	100	0.9554	0.7274
0.7948	1.2611	150	0.9151	0.7375
0.7381	1.6821	200	0.8547	0.7479
0.6022	2.1011	250	0.8082	0.7591
0.6003	2.5221	300	0.7770	0.7660
0.5722	2.9432	350	0.7678	0.7692
0.5801	3.0	357	0.7678	0.7693

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.22.1