ROE_Patent_Breeze7B_Type2_V2

This model is a fine-tuned version of MediaTek-Research/Breeze-7B-Instruct-v1_0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9433

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-05
• train_batch_size: 1
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Use paged_adamw_32bit with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
10.716	1.0	13	1.2232
9.2635	2.0	26	1.1253
8.187	3.0	39	1.0637
8.0029	4.0	52	1.0235
7.4224	5.0	65	0.9961
7.2189	6.0	78	0.9753
7.1858	7.0	91	0.9579
6.2129	8.0	104	0.9465
6.2479	9.0	117	0.9392
6.2023	9.2424	120	0.9433

Framework versions

• PEFT 0.14.1.dev0
• Transformers 4.48.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0