categorization-finetuned-20220721-164940-pruned-20220803-184651

This model is a fine-tuned version of carted-nlp/categorization-finetuned-20220721-164940 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4673
• Accuracy: 0.8760
• F1: 0.8751

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: 7e-06
• train_batch_size: 48
• eval_batch_size: 48
• seed: 314
• gradient_accumulation_steps: 6
• total_train_batch_size: 288
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 1000
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.3404	0.51	2000	0.4329	0.8872	0.8865
0.3433	1.01	4000	0.4280	0.8883	0.8876
0.3281	1.52	6000	0.4302	0.8890	0.8883
0.331	2.02	8000	0.4265	0.8891	0.8885
0.3224	2.53	10000	0.4300	0.8881	0.8874
0.3361	3.04	12000	0.4291	0.8889	0.8882
0.3323	3.54	14000	0.4337	0.8878	0.8871
0.3556	4.05	16000	0.4345	0.8857	0.8851
0.3663	4.56	18000	0.4417	0.8836	0.8828
0.3902	5.06	20000	0.4555	0.8789	0.8781
0.4036	5.57	22000	0.4556	0.8788	0.8779
0.4305	6.07	24000	0.4697	0.8751	0.8742
0.4501	6.58	26000	0.4763	0.8738	0.8725
0.4733	7.09	28000	0.4857	0.8710	0.8700
0.4851	7.59	30000	0.4863	0.8705	0.8695
0.4846	8.1	32000	0.4849	0.8708	0.8698
0.4856	8.61	34000	0.4835	0.8707	0.8695
0.4774	9.11	36000	0.4797	0.8719	0.8708
0.4635	9.62	38000	0.4776	0.8728	0.8717
0.4561	10.12	40000	0.4746	0.8739	0.8729
0.4475	10.63	42000	0.4705	0.8749	0.8740
0.4413	11.14	44000	0.4691	0.8754	0.8744
0.4389	11.64	46000	0.4679	0.8760	0.8750
0.4361	12.15	48000	0.4677	0.8759	0.8749
0.4362	12.65	50000	0.4672	0.8763	0.8753
0.4309	13.16	52000	0.4671	0.8761	0.8751
0.4316	13.67	54000	0.4670	0.8764	0.8754
0.4321	14.17	56000	0.4668	0.8764	0.8755
0.4311	14.68	58000	0.4668	0.8764	0.8754

Framework versions

• Transformers 4.18.0.dev0
• Pytorch 1.9.1+cu111
• Datasets 2.3.2
• Tokenizers 0.11.6