gitlab-mr-analysis-default-model

This model is a fine-tuned version of distilbert-base-uncased-distilled-squad on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3356
• Accuracy: 94.72759226713534%
• F1: 0.9612

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	284	0.4056	85.76449912126537%	0.7346
0.4789	2.0	568	0.2722	91.47627416520211%	0.9335
0.4789	3.0	852	0.2413	94.37609841827768%	0.9592
0.1185	4.0	1137	0.2776	94.37609841827768%	0.9574
0.1185	5.0	1421	0.3132	93.84885764499121%	0.9472
0.0378	6.0	1705	0.3323	94.28822495606327%	0.9582
0.0378	7.0	1989	0.3393	94.28822495606327%	0.9575
0.0123	8.0	2274	0.3363	94.5518453427065%	0.9607
0.0077	9.0	2558	0.3333	94.63971880492092%	0.9606
0.0077	9.99	2840	0.3356	94.72759226713534%	0.9612

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1
• Datasets 2.14.4
• Tokenizers 0.13.2