twitter-roberta-base-sentiment-latest_12112024T172612

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

• Loss: 1.3788
• F1: 0.4290
• Learning Rate: 0.0

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 600
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Rate
No log	0.9942	86	1.8133	0.1353	0.0000
No log	2.0	173	1.8268	0.3232	0.0000
No log	2.9942	259	1.5445	0.4004	0.0000
No log	4.0	346	1.4478	0.4242	0.0000
No log	4.9942	432	1.4799	0.4227	0.0000
1.3607	6.0	519	1.3788	0.4290	0.0000
1.3607	6.9942	605	1.5873	0.4305	1e-05
1.3607	8.0	692	1.5440	0.4413	1e-05
1.3607	8.9942	778	1.5222	0.4604	0.0000
1.3607	10.0	865	1.5721	0.4613	0.0000
1.3607	10.9942	951	1.8378	0.4372	0.0000
0.6442	12.0	1038	1.8343	0.4509	0.0000
0.6442	12.9942	1124	1.8831	0.4564	0.0000
0.6442	14.0	1211	1.9827	0.4736	0.0000
0.6442	14.9942	1297	2.0701	0.4714	0.0000
0.6442	16.0	1384	2.0660	0.4637	0.0000
0.6442	16.9942	1470	2.0215	0.4878	6e-06
0.281	18.0	1557	2.3276	0.4730	0.0000
0.281	18.9942	1643	2.4235	0.4554	0.0000
0.281	20.0	1730	2.5274	0.4627	0.0000
0.281	20.9942	1816	2.7742	0.4463	0.0000
0.281	22.0	1903	2.6885	0.4622	0.0000
0.281	22.9942	1989	2.6585	0.4598	0.0000
0.1192	24.0	2076	2.7094	0.4588	0.0000
0.1192	24.9942	2162	2.7192	0.4667	0.0000
0.1192	26.0	2249	2.7276	0.4729	7e-07
0.1192	26.9942	2335	2.7288	0.4745	4e-07
0.1192	28.0	2422	2.7238	0.4774	2e-07
0.0744	28.9942	2508	2.7567	0.4760	0.0
0.0744	29.8266	2580	2.7577	0.4728	0.0

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.19.1