scenario-KD-SCR-CDF-CL-D2_data-cl-cardiff_cl_only66

This model is a fine-tuned version of haryoaw/scenario-MDBT-TCR_data-cl-cardiff_cl_only on the None dataset. It achieves the following results on the evaluation set:

• Loss: 342.1574
• Accuracy: 0.3588
• F1: 0.3024

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 32
• seed: 66
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0870	250	586.5452	0.3349	0.1940
590.5338	2.1739	500	550.6793	0.3387	0.3077
590.5338	3.2609	750	525.7099	0.3356	0.2186
484.7646	4.3478	1000	504.6014	0.3318	0.1716
484.7646	5.4348	1250	486.0431	0.3326	0.3053
428.6661	6.5217	1500	471.5125	0.3333	0.1791
428.6661	7.6087	1750	457.1420	0.3380	0.1987
388.6495	8.6957	2000	443.9854	0.3356	0.2083
388.6495	9.7826	2250	432.0132	0.3410	0.2399
357.031	10.8696	2500	420.9312	0.3356	0.2598
357.031	11.9565	2750	411.4677	0.3619	0.3083
331.4649	13.0435	3000	402.7120	0.3465	0.2608
331.4649	14.1304	3250	394.5621	0.3480	0.2748
310.7491	15.2174	3500	387.3688	0.3565	0.2908
310.7491	16.3043	3750	379.8443	0.3580	0.3025
294.0918	17.3913	4000	373.5259	0.3326	0.1972
294.0918	18.4783	4250	368.8514	0.3403	0.2242
280.7149	19.5652	4500	363.9998	0.3465	0.2808
280.7149	20.6522	4750	360.2543	0.3650	0.3035
270.0493	21.7391	5000	355.6851	0.3503	0.2586
270.0493	22.8261	5250	352.3521	0.3588	0.2877
261.8392	23.9130	5500	349.3174	0.3688	0.2932
261.8392	25.0	5750	347.2019	0.3356	0.2419
255.6845	26.0870	6000	345.0442	0.3580	0.3081
255.6845	27.1739	6250	343.5341	0.3549	0.3119
251.6646	28.2609	6500	342.7285	0.3596	0.2783
251.6646	29.3478	6750	342.1574	0.3588	0.3024

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.19.1