20-finetuned-spiderTraining50-200

This model is a fine-tuned version of facebook/convnextv2-tiny-22k-384 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4409
• Accuracy: 0.8909
• Precision: 0.8899
• Recall: 0.8922
• F1: 0.8881

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: 0.0005
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.4597	1.0	125	1.1828	0.6617	0.7065	0.6632	0.6495
1.2938	2.0	250	1.0667	0.6827	0.7244	0.6845	0.6701
1.1316	3.0	375	0.9465	0.7257	0.7828	0.7291	0.7280
0.7827	4.0	500	0.8576	0.7397	0.7701	0.7394	0.7372
0.7407	5.0	625	0.8084	0.7728	0.7876	0.7728	0.7636
0.6481	6.0	750	0.7537	0.7798	0.7999	0.7783	0.7765
0.5868	7.0	875	0.6406	0.8258	0.8341	0.8266	0.8224
0.4461	8.0	1000	0.7555	0.7768	0.7953	0.7736	0.7679
0.4984	9.0	1125	0.6601	0.8128	0.8260	0.8120	0.8059
0.3898	10.0	1250	0.7017	0.8108	0.8296	0.8079	0.8059
0.3262	11.0	1375	0.6199	0.8258	0.8341	0.8268	0.8212
0.3243	12.0	1500	0.6561	0.8188	0.8316	0.8256	0.8191
0.2914	13.0	1625	0.6037	0.8368	0.8504	0.8429	0.8351
0.2627	14.0	1750	0.5609	0.8529	0.8588	0.8557	0.8501
0.2457	15.0	1875	0.5266	0.8639	0.8674	0.8666	0.8613
0.2294	16.0	2000	0.5475	0.8589	0.8658	0.8608	0.8560
0.2088	17.0	2125	0.4929	0.8699	0.8726	0.8678	0.8672
0.2101	18.0	2250	0.4488	0.8799	0.8782	0.8775	0.8752
0.1767	19.0	2375	0.4543	0.8829	0.8824	0.8808	0.8782
0.1445	20.0	2500	0.4409	0.8909	0.8899	0.8922	0.8881

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3