scenario-kd-po-ner-half-xlmr_data-univner_full66

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

• Loss: 119.1294
• Precision: 0.4508
• Recall: 0.4147
• F1: 0.4320
• Accuracy: 0.9488

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
284.146	0.29	500	212.3445	0.0	0.0	0.0	0.9241
195.0927	0.58	1000	183.6382	0.1678	0.0104	0.0196	0.9248
175.6188	0.87	1500	169.2953	0.2481	0.0462	0.0778	0.9266
164.5343	1.16	2000	160.0481	0.25	0.0763	0.1169	0.9295
155.7977	1.46	2500	153.3116	0.2572	0.1597	0.1971	0.9334
150.8056	1.75	3000	147.9380	0.2919	0.1860	0.2272	0.9354
145.886	2.04	3500	144.0042	0.3123	0.2203	0.2584	0.9363
141.6606	2.33	4000	140.6623	0.3362	0.2148	0.2621	0.9378
138.9315	2.62	4500	137.7953	0.3395	0.2557	0.2917	0.9390
136.3904	2.91	5000	135.3107	0.3432	0.2940	0.3167	0.9407
133.4631	3.2	5500	133.3555	0.3602	0.3239	0.3411	0.9419
131.8994	3.49	6000	131.5568	0.3743	0.3372	0.3548	0.9421
130.0259	3.78	6500	130.1601	0.3590	0.3480	0.3534	0.9425
128.8084	4.08	7000	128.5461	0.3878	0.3197	0.3505	0.9431
126.9491	4.37	7500	127.3171	0.4126	0.3611	0.3852	0.9444
125.7889	4.66	8000	126.2976	0.3964	0.3598	0.3773	0.9449
125.1757	4.95	8500	125.3498	0.3985	0.3718	0.3847	0.9448
124.1594	5.24	9000	124.6697	0.4158	0.3617	0.3869	0.9457
122.9069	5.53	9500	123.8015	0.4246	0.3735	0.3975	0.9457
121.9428	5.82	10000	123.1179	0.4038	0.3846	0.3940	0.9461
121.4957	6.11	10500	122.5535	0.4256	0.3835	0.4035	0.9468
120.8021	6.4	11000	121.9703	0.4315	0.3887	0.4090	0.9467
120.6592	6.69	11500	121.5849	0.4290	0.4178	0.4233	0.9478
119.7714	6.99	12000	121.0852	0.4380	0.4115	0.4243	0.9479
119.6204	7.28	12500	120.7726	0.4477	0.3894	0.4165	0.9478
118.8015	7.57	13000	120.4320	0.4449	0.4093	0.4264	0.9480
118.8596	7.86	13500	120.1600	0.4454	0.4053	0.4244	0.9481
118.376	8.15	14000	119.9419	0.4406	0.4047	0.4219	0.9484
118.1562	8.44	14500	119.6902	0.4427	0.4099	0.4257	0.9483
117.8396	8.73	15000	119.5728	0.4410	0.4155	0.4279	0.9484
117.496	9.02	15500	119.4226	0.4450	0.4188	0.4315	0.9486
117.5852	9.31	16000	119.3681	0.4498	0.4096	0.4288	0.9486
117.6006	9.61	16500	119.2253	0.4467	0.4048	0.4247	0.9487
117.3602	9.9	17000	119.1294	0.4508	0.4147	0.4320	0.9488

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.13.3