8f72f6797526115372be0870bef4a07d

This model is a fine-tuned version of google-bert/bert-base-chinese on the contemmcm/amazon_reviews_2013 [cell-phone] dataset. It achieves the following results on the evaluation set:

• Loss: 1.1055
• Data Size: 1.0
• Epoch Runtime: 107.9802
• Accuracy: 0.5518
• F1 Macro: 0.5005

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Accuracy	F1 Macro
No log	0	0	1.8557	0	7.5523	0.1495	0.0905
No log	1	1973	1.4686	0.0078	8.5170	0.3853	0.1112
0.0328	2	3946	1.4784	0.0156	8.9284	0.3853	0.1112
1.481	3	5919	1.4767	0.0312	10.7119	0.3907	0.1829
1.4287	4	7892	1.3822	0.0625	13.6778	0.3898	0.2126
1.2223	5	9865	1.1962	0.125	19.8552	0.5056	0.3502
1.1215	6	11838	1.0818	0.25	33.3395	0.5543	0.3941
1.0791	7	13811	1.0284	0.5	56.3698	0.5680	0.4771
1.0017	8.0	15784	0.9793	1.0	108.4160	0.6030	0.5123
0.9637	9.0	17757	0.9674	1.0	107.2610	0.6067	0.4771
0.8773	10.0	19730	0.9572	1.0	107.0164	0.6100	0.5261
0.8304	11.0	21703	0.9793	1.0	106.4088	0.6169	0.5265
0.8091	12.0	23676	1.0087	1.0	106.5335	0.6092	0.5176
0.8262	13.0	25649	1.0515	1.0	106.5963	0.5892	0.5301
0.7466	14.0	27622	1.1055	1.0	107.9802	0.5518	0.5005

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1