Open-source image classification model of hf_train_output - Accurately identify rock glaciers with an accuracy rate of up to 92.58%

Hf Train Output

Developed by venetis

This model is a fine-tuned image classification model based on Google's ViT-base-patch16-224-in21k on a rock glacier dataset, achieving an accuracy of 92.58%.

Image Classification

Transformers

Open Source License:Apache-2.0 #Rock Glacier Classification #High-precision ViT #Remote Sensing Image Analysis

Downloads 28

Release Time : 11/19/2022

Model Overview

This is a Vision Transformer model specifically designed for rock glacier image classification, suitable for geological remote sensing image analysis tasks.

Model Features

High-precision Classification

Achieves 92.58% classification accuracy on the rock glacier dataset.

Based on ViT Architecture

Utilizes the Vision Transformer base architecture with powerful feature extraction capabilities.

Fine-tuned Optimization

Obtains optimal performance through 50 training epochs and hyperparameter tuning.

Model Capabilities

Rock Glacier Image Classification

Geological Remote Sensing Image Analysis

High-resolution Image Processing

Use Cases

Geological Research

Glacier Change Monitoring

Used to identify and classify different types of rock glaciers.

Accuracy: 92.58%

Environmental Monitoring

Glacier Degradation Assessment

Assists researchers in evaluating glacier change trends.

🚀 hf_train_output

This model is a fine - tuned version of [google/vit - base - patch16 - 224 - in21k](https://huggingface.co/google/vit - base - patch16 - 224 - in21k) on the rock - glacier - dataset. It offers high - accuracy image classification capabilities, achieving excellent results on the evaluation set.

🚀 Quick Start

This model is already fine - tuned on the rock - glacier - dataset. You can directly use it for image classification tasks.

✨ Features

High Accuracy: Achieves an accuracy of 0.9258 on the evaluation set.
Fine - Tuned: Based on the pre - trained model [google/vit - base - patch16 - 224 - in21k], fine - tuned on the rock - glacier - dataset.

📚 Documentation

Model Performance

This model achieves the following results on the evaluation set:

Loss: 0.3894
Accuracy: 0.9258

Model Index

Property	Details
Model Name	hf_train_output
Task	Image Classification
Dataset	rock - glacier - dataset
Metrics	Accuracy: 0.9258241758241759

Training and Evaluation Data

The model is trained and evaluated on the rock - glacier - dataset.

Training Procedure

Training Hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e - 05
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
num_epochs: 50
mixed_precision_training: Native AMP

Training Results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5619	0.55	50	0.5432	0.7692
0.4582	1.1	100	0.4435	0.8352
0.3548	1.65	150	0.3739	0.8599
0.217	2.2	200	0.2913	0.9093
0.1709	2.75	250	0.2619	0.9148
0.0919	3.3	300	0.2475	0.9148
0.0652	3.85	350	0.3275	0.8901
0.0495	4.4	400	0.2515	0.9093
0.0321	4.95	450	0.2878	0.9066
0.0247	5.49	500	0.2612	0.9148
0.017	6.04	550	0.2687	0.9176
0.0131	6.59	600	0.3062	0.9093
0.0113	7.14	650	0.2587	0.9231
0.0099	7.69	700	0.2815	0.9203
0.009	8.24	750	0.2675	0.9286
0.0084	8.79	800	0.2711	0.9286
0.0077	9.34	850	0.2663	0.9313
0.0073	9.89	900	0.3003	0.9258
0.0069	10.44	950	0.2758	0.9313
0.0064	10.99	1000	0.2999	0.9258
0.0061	11.54	1050	0.2931	0.9313
0.0057	12.09	1100	0.2989	0.9313
0.0056	12.64	1150	0.2974	0.9313
0.0053	13.19	1200	0.3099	0.9258
0.005	13.74	1250	0.3131	0.9313
0.0049	14.29	1300	0.3201	0.9258
0.0046	14.84	1350	0.3109	0.9313
0.0045	15.38	1400	0.3168	0.9313
0.0043	15.93	1450	0.3226	0.9231
0.0042	16.48	1500	0.3234	0.9231
0.0041	17.03	1550	0.3283	0.9258
0.0039	17.58	1600	0.3304	0.9258
0.0038	18.13	1650	0.3321	0.9231
0.0037	18.68	1700	0.3362	0.9231
0.0036	19.23	1750	0.3307	0.9286
0.0035	19.78	1800	0.3357	0.9231
0.0034	20.33	1850	0.3244	0.9313
0.0033	20.88	1900	0.3497	0.9231
0.0032	21.43	1950	0.3443	0.9231
0.0031	21.98	2000	0.3398	0.9286
0.003	22.53	2050	0.3388	0.9286
0.003	23.08	2100	0.3399	0.9286
0.0029	23.63	2150	0.3548	0.9231
0.0028	24.18	2200	0.3475	0.9286
0.0028	24.73	2250	0.3480	0.9286
0.0027	25.27	2300	0.3542	0.9231
0.0026	25.82	2350	0.3589	0.9231
0.0026	26.37	2400	0.3449	0.9286
0.0025	26.92	2450	0.3604	0.9231
0.0025	27.47	2500	0.3493	0.9286
0.0024	28.02	2550	0.3631	0.9258
0.0024	28.57	2600	0.3590	0.9258
0.0023	29.12	2650	0.3604	0.9258
0.0023	29.67	2700	0.3667	0.9258
0.0022	30.22	2750	0.3571	0.9286
0.0022	30.77	2800	0.3660	0.9258
0.0021	31.32	2850	0.3638	0.9286
0.0021	31.87	2900	0.3729	0.9258
0.0021	32.42	2950	0.3706	0.9258
0.002	32.97	3000	0.3669	0.9286
0.002	33.52	3050	0.3740	0.9258
0.002	34.07	3100	0.3693	0.9286
0.002	34.62	3150	0.3700	0.9286
0.0019	35.16	3200	0.3752	0.9258
0.0019	35.71	3250	0.3753	0.9258
0.0019	36.26	3300	0.3721	0.9286
0.0018	36.81	3350	0.3764	0.9258
0.0018	37.36	3400	0.3758	0.9258
0.0018	37.91	3450	0.3775	0.9258
0.0018	38.46	3500	0.3812	0.9258
0.0018	39.01	3550	0.3817	0.9258
0.0017	39.56	3600	0.3815	0.9258
0.0017	40.11	3650	0.3825	0.9258
0.0017	40.66	3700	0.3852	0.9258
0.0017	41.21	3750	0.3854	0.9258
0.0017	41.76	3800	0.3823	0.9258
0.0016	42.31	3850	0.3829	0.9258
0.0016	42.86	3900	0.3873	0.9258
0.0016	43.41	3950	0.3842	0.9258
0.0016	43.96	4000	0.3857	0.9258
0.0016	44.51	4050	0.3873	0.9258
0.0016	45.05	4100	0.3878	0.9258
0.0016	45.6	4150	0.3881	0.9258
0.0016	46.15	4200	0.3888	0.9258
0.0016	46.7	4250	0.3891	0.9258
0.0016	47.25	4300	0.3878	0.9258
0.0016	47.8	4350	0.3890	0.9258
0.0016	48.35	4400	0.3890	0.9258
0.0015	48.9	4450	0.3895	0.9258
0.0015	49.45	4500	0.3896	0.9258
0.0015	50.0	4550	0.3894	0.9258

Framework Versions

Transformers 4.24.0
Pytorch 1.12.1+cu113
Datasets 2.7.0
Tokenizers 0.13.2

📄 License

This project is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご