stereo_detect_rm

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

• Loss: 0.0122
• Mse: 0.0122

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: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Mse
0.0427	1.0	21	0.0278	0.0278
0.0271	2.0	42	0.0204	0.0204
0.0222	3.0	63	0.0180	0.0180
0.0209	4.0	84	0.0133	0.0133
0.0184	5.0	105	0.0114	0.0114
0.015	6.0	126	0.0145	0.0145
0.0119	7.0	147	0.0132	0.0132
0.0109	8.0	168	0.0158	0.0158
0.0075	9.0	189	0.0145	0.0145
0.007	10.0	210	0.0129	0.0129
0.007	11.0	231	0.0125	0.0125
0.0061	12.0	252	0.0133	0.0133
0.0056	13.0	273	0.0150	0.0150
0.005	14.0	294	0.0126	0.0126
0.0042	15.0	315	0.0126	0.0126
0.0043	16.0	336	0.0142	0.0142
0.0036	17.0	357	0.0155	0.0155
0.0039	18.0	378	0.0136	0.0136
0.0039	19.0	399	0.0127	0.0127
0.004	20.0	420	0.0132	0.0132
0.0036	21.0	441	0.0126	0.0126
0.0034	22.0	462	0.0121	0.0121
0.0032	23.0	483	0.0128	0.0128
0.003	24.0	504	0.0114	0.0114
0.003	25.0	525	0.0129	0.0129
0.0029	26.0	546	0.0114	0.0114
0.0027	27.0	567	0.0121	0.0121
0.0029	28.0	588	0.0120	0.0120
0.0026	29.0	609	0.0118	0.0118
0.0025	30.0	630	0.0122	0.0122

Framework versions

• Transformers 4.57.1
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1