distilbert-base-uncased-distilled-squad_07112024T120722

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

• Loss: 0.7930
• F1: 0.7643
• 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: 2e-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: 10
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Rate
No log	1.0	121	1.1749	0.5577	0.0000
No log	2.0	242	0.9801	0.6593	0.0000
No log	3.0	363	0.9184	0.6951	0.0000
No log	4.0	484	0.8064	0.7480	0.0000
0.9941	5.0	605	0.7930	0.7643	0.0000
0.9941	6.0	726	0.8313	0.7793	0.0000
0.9941	7.0	847	0.8955	0.7814	0.0000
0.9941	8.0	968	0.9010	0.7963	0.0000
0.3299	9.0	1089	0.9855	0.7945	0.0000
0.3299	10.0	1210	0.9753	0.8036	0.0000
0.3299	11.0	1331	1.0115	0.8095	0.0000
0.3299	12.0	1452	1.0508	0.8052	7e-06
0.0991	13.0	1573	1.1102	0.8125	0.0000
0.0991	14.0	1694	1.1321	0.8146	0.0000
0.0991	15.0	1815	1.2166	0.8085	3e-06
0.0991	16.0	1936	1.1935	0.8109	0.0000
0.0349	17.0	2057	1.2277	0.8074	0.0000
0.0349	18.0	2178	1.2145	0.8091	5e-07
0.0349	19.0	2299	1.2306	0.8082	1e-07
0.0349	20.0	2420	1.2334	0.8080	0.0

Framework versions

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