README.md

metadata

license: mit
base_model: microsoft/deberta-v3-base
tags:
  - generated_from_trainer
metrics:
  - f1
  - precision
  - recall
  - accuracy
model-index:
  - name: debert-imeocap
    results: []

debert-imeocap

This model is a fine-tuned version of microsoft/deberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.3914
• F1: 0.6372
• Precision: 0.6448
• Recall: 0.6365
• Accuracy: 0.6365

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Precision	Recall	Accuracy
1.5405	1.0	74	1.4488	0.3206	0.2386	0.4885	0.4885
1.3156	2.0	148	1.1964	0.5541	0.5627	0.575	0.575
1.0728	3.0	222	1.1077	0.6001	0.6189	0.5981	0.5981
0.9239	4.0	296	1.0742	0.6324	0.6361	0.6365	0.6365
0.7802	5.0	370	1.0834	0.6073	0.6333	0.6058	0.6058
0.661	6.0	444	1.1733	0.5984	0.6166	0.5962	0.5962
0.602	7.0	518	1.1786	0.5911	0.6193	0.5885	0.5885
0.5391	8.0	592	1.2171	0.6156	0.6251	0.6154	0.6154
0.4815	9.0	666	1.2566	0.6259	0.6399	0.625	0.625
0.4548	10.0	740	1.2927	0.6233	0.6417	0.6212	0.6212
0.4538	11.0	814	1.2969	0.6385	0.6461	0.6385	0.6385
0.4119	12.0	888	1.3455	0.6376	0.6464	0.6365	0.6365
0.3968	13.0	962	1.3709	0.6304	0.6413	0.6288	0.6288
0.352	14.0	1036	1.3823	0.6246	0.6360	0.6231	0.6231
0.3551	15.0	1110	1.3914	0.6372	0.6448	0.6365	0.6365

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2