quote-type-sentence-model

This model is a fine-tuned version of roberta-base to perform sentence level classification on a news dataset. The following sentence level tags are identified:

  'No Quote'
  'Direct Quote'
  'Published Work/Press Report'
  'Indirect Quote'
  'Statement/Public Speech'
  'Background/Narrative'
  'Other'
  'Proposal/Order/Law'
  'Email/Social Media Post'
  'Court Proceeding'
  'Direct Observation'

It achieves the following results on the evaluation set:

• Loss: 0.8795
• F1: 0.5407

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: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	0.19	100	1.1174	0.2831
No log	0.38	200	1.1066	0.3356
No log	0.57	300	1.0490	0.4126
No log	0.76	400	1.0280	0.3778
1.0973	0.95	500	0.9378	0.4492
1.0973	1.14	600	1.0546	0.4650
1.0973	1.33	700	0.9806	0.4619
1.0973	1.52	800	0.8989	0.5176
1.0973	1.7	900	0.9531	0.5078
0.8155	1.89	1000	0.9482	0.4781
0.8155	2.08	1100	0.8935	0.5084
0.8155	2.27	1200	0.9059	0.5236
0.8155	2.46	1300	0.9483	0.5127
0.8155	2.65	1400	0.8961	0.5355
0.6225	2.84	1500	0.8795	0.5407

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3