dit-rvl_maveriq_tobacco3482_2023-07-05

This model is a fine-tuned version of microsoft/dit-base-finetuned-rvlcdip on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4530
• Accuracy: 0.94

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.96	3	2.2927	0.01
No log	1.96	6	2.2632	0.08
No log	2.96	9	2.2334	0.18
No log	3.96	12	2.2025	0.195
No log	4.96	15	2.1686	0.235
No log	5.96	18	2.1274	0.325
No log	6.96	21	2.0784	0.385
No log	7.96	24	2.0284	0.465
No log	8.96	27	1.9750	0.55
No log	9.96	30	1.9206	0.585
No log	10.96	33	1.8683	0.61
No log	11.96	36	1.8164	0.65
No log	12.96	39	1.7660	0.735
No log	13.96	42	1.7195	0.765
No log	14.96	45	1.6761	0.815
No log	15.96	48	1.6336	0.83
No log	16.96	51	1.5918	0.835
No log	17.96	54	1.5511	0.835
No log	18.96	57	1.5101	0.84
No log	19.96	60	1.4699	0.85
No log	20.96	63	1.4307	0.855
No log	21.96	66	1.3925	0.865
No log	22.96	69	1.3534	0.865
No log	23.96	72	1.3164	0.885
No log	24.96	75	1.2825	0.885
No log	25.96	78	1.2458	0.88
No log	26.96	81	1.2091	0.88
No log	27.96	84	1.1762	0.89
No log	28.96	87	1.1446	0.885
No log	29.96	90	1.1126	0.9
No log	30.96	93	1.0840	0.905
No log	31.96	96	1.0549	0.9
No log	32.96	99	1.0247	0.91
No log	33.96	102	0.9962	0.925
No log	34.96	105	0.9685	0.93
No log	35.96	108	0.9447	0.93
No log	36.96	111	0.9217	0.93
No log	37.96	114	0.9007	0.93
No log	38.96	117	0.8778	0.935
No log	39.96	120	0.8551	0.935
No log	40.96	123	0.8325	0.93
No log	41.96	126	0.8129	0.93
No log	42.96	129	0.7970	0.93
No log	43.96	132	0.7810	0.93
No log	44.96	135	0.7609	0.935
No log	45.96	138	0.7441	0.935
No log	46.96	141	0.7313	0.935
No log	47.96	144	0.7184	0.935
No log	48.96	147	0.7044	0.93
No log	49.96	150	0.6902	0.93
No log	50.96	153	0.6773	0.935
No log	51.96	156	0.6666	0.935
No log	52.96	159	0.6554	0.935
No log	53.96	162	0.6446	0.935
No log	54.96	165	0.6308	0.94
No log	55.96	168	0.6194	0.94
No log	56.96	171	0.6098	0.94
No log	57.96	174	0.6021	0.94
No log	58.96	177	0.5922	0.935
No log	59.96	180	0.5820	0.94
No log	60.96	183	0.5735	0.94
No log	61.96	186	0.5632	0.94
No log	62.96	189	0.5559	0.94
No log	63.96	192	0.5494	0.94
No log	64.96	195	0.5430	0.94
No log	65.96	198	0.5370	0.935
No log	66.96	201	0.5320	0.935
No log	67.96	204	0.5278	0.935
No log	68.96	207	0.5228	0.935
No log	69.96	210	0.5166	0.935
No log	70.96	213	0.5117	0.935
No log	71.96	216	0.5076	0.935
No log	72.96	219	0.5029	0.94
No log	73.96	222	0.4985	0.94
No log	74.96	225	0.4945	0.94
No log	75.96	228	0.4904	0.94
No log	76.96	231	0.4865	0.94
No log	77.96	234	0.4833	0.94
No log	78.96	237	0.4804	0.94
No log	79.96	240	0.4779	0.94
No log	80.96	243	0.4757	0.94
No log	81.96	246	0.4738	0.94
No log	82.96	249	0.4719	0.935
No log	83.96	252	0.4701	0.935
No log	84.96	255	0.4684	0.935
No log	85.96	258	0.4669	0.935
No log	86.96	261	0.4653	0.935
No log	87.96	264	0.4637	0.935
No log	88.96	267	0.4620	0.935
No log	89.96	270	0.4602	0.935
No log	90.96	273	0.4586	0.94
No log	91.96	276	0.4572	0.94
No log	92.96	279	0.4562	0.94
No log	93.96	282	0.4553	0.94
No log	94.96	285	0.4546	0.94
No log	95.96	288	0.4540	0.94
No log	96.96	291	0.4535	0.94
No log	97.96	294	0.4532	0.94
No log	98.96	297	0.4530	0.94
No log	99.96	300	0.4530	0.94

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1.post200
• Datasets 2.9.0
• Tokenizers 0.13.2