roberta-large-lora-token-classification

This model is a fine-tuned version of roberta-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2546
• Precision: 0.9604
• Recall: 0.9600
• F1-score: 0.9601
• Accuracy: 0.9598

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1-score	Accuracy
No log	0.11	25	0.9686	0.6063	0.5640	0.5392	0.5617
No log	0.23	50	0.9781	0.5396	0.5383	0.4969	0.5362
No log	0.34	75	0.9558	0.6942	0.4627	0.3789	0.4651
No log	0.46	100	0.8439	0.6372	0.5855	0.5856	0.5871
No log	0.57	125	0.6714	0.7624	0.7533	0.7539	0.7520
No log	0.69	150	0.6038	0.7961	0.7635	0.7693	0.7641
No log	0.8	175	0.3989	0.8356	0.8362	0.8358	0.8351
No log	0.92	200	0.3601	0.8801	0.8600	0.8523	0.8579
No log	1.03	225	0.3550	0.8900	0.8797	0.8752	0.8780
No log	1.15	250	0.3047	0.8921	0.8894	0.8897	0.8887
No log	1.26	275	0.3499	0.9155	0.9162	0.9150	0.9155
No log	1.38	300	0.3978	0.8969	0.8890	0.8885	0.8887
No log	1.49	325	0.2146	0.9293	0.9220	0.9206	0.9209
No log	1.61	350	0.2954	0.9274	0.9283	0.9273	0.9276
No log	1.72	375	0.4710	0.9100	0.9005	0.9022	0.9008
No log	1.83	400	0.2987	0.9193	0.9174	0.9160	0.9169
No log	1.95	425	0.2549	0.9442	0.9403	0.9401	0.9397
No log	2.06	450	0.3173	0.9396	0.9401	0.9397	0.9397
No log	2.18	475	0.3349	0.9491	0.9484	0.9477	0.9477
0.5562	2.29	500	0.3155	0.9440	0.9442	0.9438	0.9437
0.5562	2.41	525	0.2871	0.9440	0.9442	0.9438	0.9437
0.5562	2.52	550	0.2983	0.9499	0.9496	0.9491	0.9491
0.5562	2.64	575	0.2736	0.9504	0.9482	0.9481	0.9477
0.5562	2.75	600	0.3900	0.9397	0.9386	0.9390	0.9383
0.5562	2.87	625	0.3460	0.9511	0.9510	0.9503	0.9504
0.5562	2.98	650	0.3509	0.9477	0.9445	0.9437	0.9437
0.5562	3.1	675	0.2368	0.9558	0.9551	0.9545	0.9544
0.5562	3.21	700	0.2279	0.9601	0.9591	0.9585	0.9584
0.5562	3.33	725	0.2538	0.9539	0.9537	0.9531	0.9531
0.5562	3.44	750	0.2577	0.9504	0.9509	0.9506	0.9504
0.5562	3.56	775	0.1994	0.9548	0.9548	0.9546	0.9544
0.5562	3.67	800	0.2247	0.9625	0.9628	0.9626	0.9625
0.5562	3.78	825	0.2145	0.9577	0.9576	0.9572	0.9571
0.5562	3.9	850	0.2466	0.9603	0.9591	0.9584	0.9584
0.5562	4.01	875	0.2373	0.9647	0.9643	0.9639	0.9638
0.5562	4.13	900	0.2058	0.9653	0.9655	0.9653	0.9651
0.5562	4.24	925	0.2640	0.9641	0.9642	0.9640	0.9638
0.5562	4.36	950	0.3120	0.9496	0.9459	0.9447	0.9450
0.5562	4.47	975	0.1946	0.9587	0.9590	0.9585	0.9584
0.2573	4.59	1000	0.2094	0.9652	0.9655	0.9653	0.9651
0.2573	4.7	1025	0.2405	0.9594	0.9577	0.9572	0.9571
0.2573	4.82	1050	0.2205	0.9628	0.9627	0.9627	0.9625
0.2573	4.93	1075	0.2546	0.9604	0.9600	0.9601	0.9598

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1