scenario-kd-po-ner-full-mdeberta_data-univner_half66

This model is a fine-tuned version of haryoaw/scenario-TCR-NER_data-univner_half on the None dataset. It achieves the following results on the evaluation set:

• Loss: 61.4326
• Precision: 0.7814
• Recall: 0.7843
• F1: 0.7828
• Accuracy: 0.9782

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: 3e-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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
134.73	0.5828	500	105.2633	0.6123	0.4383	0.5109	0.9484
96.7289	1.1655	1000	90.9748	0.6784	0.7067	0.6922	0.9701
84.9234	1.7483	1500	84.4287	0.7249	0.7498	0.7372	0.9740
78.4497	2.3310	2000	79.9555	0.7503	0.7396	0.7449	0.9748
73.644	2.9138	2500	76.3575	0.7372	0.7624	0.7496	0.9760
69.6629	3.4965	3000	73.5820	0.7391	0.7533	0.7461	0.9755
66.745	4.0793	3500	70.9258	0.7720	0.7482	0.7599	0.9767
63.9726	4.6620	4000	68.9640	0.7699	0.7423	0.7558	0.9763
61.778	5.2448	4500	67.0742	0.7621	0.7782	0.7701	0.9769
60.0151	5.8275	5000	65.6493	0.7804	0.7687	0.7745	0.9769
58.5554	6.4103	5500	64.3968	0.7767	0.7814	0.7791	0.9779
57.4554	6.9930	6000	63.7199	0.7844	0.7664	0.7753	0.9775
56.436	7.5758	6500	62.7811	0.7705	0.7847	0.7776	0.9777
55.6938	8.1585	7000	62.2272	0.7806	0.7791	0.7798	0.9781
55.105	8.7413	7500	61.7833	0.7791	0.7839	0.7815	0.9780
54.7106	9.3240	8000	61.5408	0.7820	0.7895	0.7858	0.9782
54.5482	9.9068	8500	61.4326	0.7814	0.7843	0.7828	0.9782

Framework versions

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