FT_3

This model is a fine-tuned version of sheepy928/default on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7726
• Accuracy: 0.7393
• F1: 0.6292
• Recall: 0.7393
• Precision: 0.7147
• Combined Score: 0.7056

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: 8
• eval_batch_size: 32
• seed: 42
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10000
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Recall	Precision	Combined Score
0.6785	1.6	300	0.7930	0.7387	0.6276	0.7387	0.5456	0.6627
0.5583	3.19	600	0.6910	0.7613	0.7316	0.7613	0.7072	0.7404
0.7857	4.79	900	0.6515	0.7387	0.6276	0.7387	0.5456	0.6627
0.6309	6.38	1200	0.5592	0.848	0.8270	0.848	0.8382	0.8403
0.2216	7.98	1500	0.5708	0.8773	0.8432	0.8773	0.8496	0.8619
0.3214	9.57	1800	0.4550	0.896	0.8584	0.896	0.8776	0.8820
0.7521	11.17	2100	0.3819	0.884	0.8423	0.884	0.8059	0.8541
0.5048	12.77	2400	0.6582	0.7387	0.6276	0.7387	0.5456	0.6627
0.6435	14.36	2700	0.5365	0.8467	0.8092	0.8467	0.7798	0.8206
0.9304	15.96	3000	0.7577	0.7387	0.6289	0.7387	0.6302	0.6841
0.7902	17.55	3300	0.7684	0.7387	0.6289	0.7387	0.6302	0.6841
0.6364	19.15	3600	0.7638	0.7387	0.6289	0.7387	0.6302	0.6841
0.6738	20.74	3900	0.7769	0.7393	0.6292	0.7393	0.7147	0.7056
0.8142	22.34	4200	0.7443	0.7393	0.6292	0.7393	0.7147	0.7056
0.8184	23.94	4500	0.7635	0.7393	0.6292	0.7393	0.7147	0.7056
0.7562	25.53	4800	0.7467	0.7393	0.6292	0.7393	0.7147	0.7056
0.5699	27.13	5100	0.7867	0.7393	0.6292	0.7393	0.7147	0.7056
0.761	28.72	5400	0.7570	0.7393	0.6292	0.7393	0.7147	0.7056

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0
• Datasets 2.14.5
• Tokenizers 0.14.1