scenario-non-kd-scr-ner-full-xlmr_data-univner_half55

This model is a fine-tuned version of FacebookAI/xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6057
• Precision: 0.3164
• Recall: 0.4190
• F1: 0.3605
• Accuracy: 0.9241

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: 32
• eval_batch_size: 32
• seed: 55
• 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	Precision	Recall	F1	Accuracy
0.3684	0.5828	500	0.3201	0.3037	0.1297	0.1818	0.9250
0.2862	1.1655	1000	0.4033	0.1414	0.2340	0.1763	0.8857
0.2494	1.7483	1500	0.3582	0.1714	0.2489	0.2030	0.8971
0.2134	2.3310	2000	0.3811	0.1504	0.2642	0.1917	0.8876
0.1988	2.9138	2500	0.3442	0.1903	0.2610	0.2201	0.9067
0.1746	3.4965	3000	0.3546	0.1975	0.3053	0.2399	0.8995
0.1638	4.0793	3500	0.3659	0.2012	0.2925	0.2384	0.9024
0.1346	4.6620	4000	0.3343	0.2510	0.3200	0.2813	0.9165
0.1103	5.2448	4500	0.3847	0.2141	0.3679	0.2707	0.9001
0.0941	5.8275	5000	0.4124	0.2094	0.3936	0.2733	0.8914
0.0726	6.4103	5500	0.3687	0.2647	0.3927	0.3162	0.9141
0.0685	6.9930	6000	0.4199	0.2254	0.4031	0.2891	0.8982
0.0502	7.5758	6500	0.3970	0.2593	0.4017	0.3152	0.9123
0.0473	8.1585	7000	0.4260	0.2633	0.3972	0.3167	0.9144
0.0373	8.7413	7500	0.4215	0.2643	0.4113	0.3218	0.9123
0.0326	9.3240	8000	0.4395	0.2800	0.3894	0.3258	0.9180
0.0302	9.9068	8500	0.4467	0.2813	0.4014	0.3308	0.9166
0.025	10.4895	9000	0.4828	0.2676	0.4246	0.3283	0.9138
0.0225	11.0723	9500	0.4949	0.2713	0.4099	0.3265	0.9159
0.0184	11.6550	10000	0.4994	0.2641	0.4243	0.3255	0.9108
0.0172	12.2378	10500	0.4814	0.2957	0.4154	0.3455	0.9196
0.0148	12.8205	11000	0.5016	0.2707	0.4138	0.3273	0.9125
0.012	13.4033	11500	0.4947	0.2987	0.4108	0.3459	0.9213
0.0119	13.9860	12000	0.5388	0.2649	0.4363	0.3297	0.9101
0.0097	14.5688	12500	0.5064	0.3258	0.3862	0.3535	0.9279
0.009	15.1515	13000	0.5358	0.2916	0.4054	0.3392	0.9192
0.0083	15.7343	13500	0.5042	0.3260	0.3998	0.3592	0.9258
0.0075	16.3170	14000	0.5587	0.2790	0.4212	0.3356	0.9147
0.0072	16.8998	14500	0.5629	0.2705	0.4313	0.3325	0.9114
0.0052	17.4825	15000	0.5383	0.3214	0.4014	0.3570	0.9257
0.0057	18.0653	15500	0.5904	0.2784	0.4271	0.3371	0.9146
0.0049	18.6480	16000	0.6059	0.2663	0.4370	0.3310	0.9111
0.0046	19.2308	16500	0.5595	0.3023	0.4109	0.3484	0.9213
0.0041	19.8135	17000	0.5653	0.3122	0.4106	0.3547	0.9235
0.0033	20.3963	17500	0.5755	0.3079	0.4064	0.3504	0.9230
0.0035	20.9790	18000	0.5869	0.2899	0.4222	0.3438	0.9179
0.0029	21.5618	18500	0.5756	0.3057	0.4181	0.3532	0.9217
0.0025	22.1445	19000	0.5888	0.3189	0.4125	0.3597	0.9249
0.0024	22.7273	19500	0.5776	0.3043	0.4256	0.3549	0.9216
0.0023	23.3100	20000	0.5625	0.3258	0.4157	0.3653	0.9270
0.0019	23.8928	20500	0.5871	0.3203	0.4158	0.3619	0.9253
0.0018	24.4755	21000	0.5951	0.3152	0.4157	0.3585	0.9239
0.0017	25.0583	21500	0.5927	0.3169	0.4204	0.3614	0.9240
0.0015	25.6410	22000	0.5854	0.3164	0.4271	0.3635	0.9239
0.0015	26.2238	22500	0.5955	0.3077	0.4223	0.3560	0.9223
0.0011	26.8065	23000	0.5695	0.3575	0.4007	0.3779	0.9323
0.0011	27.3893	23500	0.5847	0.3338	0.4121	0.3688	0.9280
0.001	27.9720	24000	0.6023	0.3146	0.4154	0.3581	0.9245
0.0008	28.5548	24500	0.6122	0.3059	0.4232	0.3551	0.9217
0.0008	29.1375	25000	0.6104	0.3083	0.4186	0.3551	0.9226
0.0008	29.7203	25500	0.6057	0.3164	0.4190	0.3605	0.9241

Framework versions

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