twitter_sentiment_small_4

This model is a fine-tuned version of microsoft/deberta-v3-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4844
• Accuracy: 0.824
• F1-score: 0.8057
• Precision: 0.8295
• Recall: 0.7833

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1-score	Precision	Recall
0.643	0.0889	100	0.5191	0.774	0.7466	0.7817	0.7146
0.4779	0.1778	200	0.4895	0.787	0.7677	0.7805	0.7554
0.4069	0.2667	300	0.4630	0.795	0.7745	0.7946	0.7554
0.4215	0.3556	400	0.4562	0.8	0.7669	0.8393	0.7060
0.4187	0.4444	500	0.4350	0.807	0.7992	0.7758	0.8240
0.4197	0.5333	600	0.4497	0.806	0.7785	0.8317	0.7318
0.4034	0.6222	700	0.4335	0.817	0.8111	0.7813	0.8433
0.4058	0.7111	800	0.4231	0.804	0.7996	0.7637	0.8391
0.4044	0.8	900	0.4404	0.805	0.8056	0.7523	0.8670
0.3678	0.8889	1000	0.4000	0.815	0.8095	0.7782	0.8433
0.3791	0.9778	1100	0.4451	0.814	0.814	0.7622	0.8734
0.3109	1.0667	1200	0.5034	0.817	0.8039	0.8030	0.8047
0.2999	1.1556	1300	0.4740	0.812	0.8105	0.7643	0.8627
0.2902	1.2444	1400	0.4517	0.825	0.8066	0.8314	0.7833
0.2664	1.3333	1500	0.4646	0.83	0.8225	0.8008	0.8455
0.2826	1.4222	1600	0.4844	0.824	0.8057	0.8295	0.7833

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.19.1