cls-comment-phobert-base-v2-v1.0

This model is a fine-tuned version of vinai/phobert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3315
• Accuracy: 0.9418
• F1 Score: 0.8961
• Recall: 0.9194
• Precision: 0.8741

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Score	Recall	Precision
0.3255	1.0	234	0.3272	0.8649	0.7583	0.7766	0.7408
0.2667	2.0	469	0.2380	0.9037	0.8231	0.8211	0.8251
0.2047	3.0	703	0.2150	0.9216	0.8550	0.8475	0.8627
0.1666	4.0	938	0.1988	0.9320	0.8724	0.8524	0.8934
0.0982	5.0	1172	0.2151	0.9306	0.8769	0.9052	0.8503
0.0869	6.0	1407	0.2543	0.9201	0.8640	0.9296	0.8070
0.0663	7.0	1641	0.2238	0.9417	0.8935	0.8959	0.8911
0.0724	8.0	1876	0.2402	0.9345	0.8844	0.9184	0.8529
0.0558	9.0	2110	0.2477	0.9309	0.8800	0.9286	0.8363
0.0317	10.0	2345	0.2638	0.9381	0.8901	0.9184	0.8635
0.0261	11.0	2579	0.2889	0.9334	0.8837	0.9267	0.8445
0.0703	12.0	2814	0.2500	0.9406	0.8935	0.9120	0.8756
0.0323	13.0	3048	0.2570	0.9334	0.8840	0.9291	0.8430
0.0286	14.0	3283	0.3078	0.9316	0.8817	0.9345	0.8346
0.0213	15.0	3517	0.2986	0.9392	0.8925	0.9252	0.8620
0.0279	16.0	3752	0.2928	0.9400	0.8927	0.9150	0.8715
0.016	17.0	3986	0.3136	0.9365	0.8888	0.9296	0.8514
0.0257	18.0	4221	0.3127	0.9410	0.8944	0.9150	0.8748
0.0126	19.0	4455	0.3221	0.9421	0.8961	0.9145	0.8784
0.0205	19.96	4680	0.3315	0.9418	0.8961	0.9194	0.8741

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2