bart-large-cnn-prompt_generation-2.0

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

• Loss: 2.6403
• Actual score: 0.8766
• Predction score: 0.5039
• Score difference: 0.3727

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-07
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 75
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Actual score	Predction score	Score difference
No log	1.0	8	3.6549	0.8766	-0.2093	1.0859
No log	2.0	16	3.6012	0.8766	-0.1961	1.0728
No log	3.0	24	3.5331	0.8766	-0.1613	1.0379
No log	4.0	32	3.4417	0.8766	-0.1132	0.9899
No log	5.0	40	3.3501	0.8766	-0.1821	1.0587
No log	6.0	48	3.2904	0.8766	-0.1653	1.0419
No log	7.0	56	3.2418	0.8766	-0.4566	1.3332
No log	8.0	64	3.1620	0.8766	-0.2897	1.1663
No log	9.0	72	3.0925	0.8766	-0.5185	1.3951
No log	10.0	80	3.0442	0.8766	-0.7127	1.5893
No log	11.0	88	3.0064	0.8766	-0.4893	1.3659
No log	12.0	96	2.9742	0.8766	-0.6391	1.5157
No log	13.0	104	2.9475	0.8766	-0.4873	1.3640
No log	14.0	112	2.9254	0.8766	-0.2786	1.1552
No log	15.0	120	2.9061	0.8766	-0.1893	1.0660
No log	16.0	128	2.8887	0.8766	-0.2202	1.0968
No log	17.0	136	2.8730	0.8766	-0.2009	1.0775
No log	18.0	144	2.8588	0.8766	-0.2101	1.0867
No log	19.0	152	2.8461	0.8766	-0.3374	1.2140
No log	20.0	160	2.8337	0.8766	-0.2005	1.0772
No log	21.0	168	2.8216	0.8766	-0.2570	1.1336
No log	22.0	176	2.8104	0.8766	-0.3601	1.2367
No log	23.0	184	2.7996	0.8766	-0.4823	1.3589
No log	24.0	192	2.7895	0.8766	-0.4451	1.3217
No log	25.0	200	2.7798	0.8766	-0.3621	1.2388
No log	26.0	208	2.7706	0.8766	-0.4108	1.2874
No log	27.0	216	2.7625	0.8766	-0.4750	1.3517
No log	28.0	224	2.7547	0.8766	-0.4004	1.2771
No log	29.0	232	2.7471	0.8766	-0.4535	1.3301
No log	30.0	240	2.7393	0.8766	-0.5414	1.4180
No log	31.0	248	2.7328	0.8766	-0.5666	1.4433
No log	32.0	256	2.7268	0.8766	-0.6630	1.5396
No log	33.0	264	2.7211	0.8766	-0.4073	1.2839
No log	34.0	272	2.7160	0.8766	-0.5464	1.4230
No log	35.0	280	2.7113	0.8766	-0.3629	1.2396
No log	36.0	288	2.7065	0.8766	-0.2926	1.1692
No log	37.0	296	2.7025	0.8766	-0.2596	1.1362
No log	38.0	304	2.6981	0.8766	-0.1478	1.0244
No log	39.0	312	2.6939	0.8766	-0.2252	1.1018
No log	40.0	320	2.6901	0.8766	-0.2750	1.1516
No log	41.0	328	2.6867	0.8766	-0.0900	0.9667
No log	42.0	336	2.6836	0.8766	-0.2377	1.1144
No log	43.0	344	2.6804	0.8766	-0.3135	1.1901
No log	44.0	352	2.6774	0.8766	-0.1023	0.9789
No log	45.0	360	2.6745	0.8766	-0.0386	0.9152
No log	46.0	368	2.6714	0.8766	0.1602	0.7164
No log	47.0	376	2.6689	0.8766	0.2508	0.6258
No log	48.0	384	2.6668	0.8766	0.1577	0.7190
No log	49.0	392	2.6648	0.8766	0.0565	0.8201
No log	50.0	400	2.6627	0.8766	0.2379	0.6387
No log	51.0	408	2.6607	0.8766	0.2343	0.6423
No log	52.0	416	2.6588	0.8766	0.2719	0.6048
No log	53.0	424	2.6570	0.8766	0.2214	0.6552
No log	54.0	432	2.6555	0.8766	0.2729	0.6037
No log	55.0	440	2.6541	0.8766	0.2798	0.5968
No log	56.0	448	2.6528	0.8766	0.0662	0.8104
No log	57.0	456	2.6514	0.8766	0.0377	0.8390
No log	58.0	464	2.6502	0.8766	0.2886	0.5880
No log	59.0	472	2.6491	0.8766	0.2257	0.6509
No log	60.0	480	2.6481	0.8766	0.2561	0.6206
No log	61.0	488	2.6471	0.8766	0.2683	0.6083
No log	62.0	496	2.6461	0.8766	0.2897	0.5869
2.5848	63.0	504	2.6453	0.8766	0.2974	0.5793
2.5848	64.0	512	2.6445	0.8766	0.2946	0.5820
2.5848	65.0	520	2.6438	0.8766	0.3021	0.5745
2.5848	66.0	528	2.6433	0.8766	0.2679	0.6087
2.5848	67.0	536	2.6428	0.8766	0.3133	0.5633
2.5848	68.0	544	2.6423	0.8766	0.3398	0.5368
2.5848	69.0	552	2.6418	0.8766	0.4149	0.4617
2.5848	70.0	560	2.6413	0.8766	0.4674	0.4092
2.5848	71.0	568	2.6410	0.8766	0.4929	0.3838
2.5848	72.0	576	2.6407	0.8766	0.4974	0.3793
2.5848	73.0	584	2.6406	0.8766	0.4948	0.3818
2.5848	74.0	592	2.6404	0.8766	0.4623	0.4143
2.5848	75.0	600	2.6403	0.8766	0.5039	0.3727

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1