Emotion-Image-Classification-V2 open-source emotional image classification model, accurately identifying emotions in images

Emotion Image Classification V2

Developed by jhoppanne

A fine-tuned emotion image classification model based on Google's ViT model, achieving an accuracy of 59.38% on the validation set.

Image Classification

Transformers

Open Source License:Apache-2.0 #ViT fine-tuned model #Image emotion recognition #Moderate accuracy

Downloads 2,176

Release Time : 5/30/2024

Model Overview

This model is a fine-tuned emotion image classification model based on Google's ViT architecture, primarily used for classifying and recognizing emotions in images.

Model Features

Based on ViT Architecture

Uses Google's Vision Transformer (ViT) as the base model, offering excellent image feature extraction capabilities.

Emotion Classification

Fine-tuned specifically for image emotion recognition tasks, suitable for analyzing emotional content in images.

Moderate Accuracy

Achieves an accuracy of 59.38% on the validation set, suitable for general emotion recognition applications.

Model Capabilities

Image classification

Emotion recognition

Visual feature extraction

Use Cases

Social Media Analysis

User-uploaded Image Emotion Analysis

Analyzes the emotional tendencies in images uploaded by social media users

Can recognize 59.38% of emotion categories

Market Research

Ad Image Emotional Impact Assessment

Evaluates the emotional reactions elicited by advertisement images

🚀 Emotion-Image-Classification-V2

This model is a fine - tuned version of [google/vit - base - patch16 - 224 - in21k](https://huggingface.co/google/vit - base - patch16 - 224 - in21k) for image emotion classification, achieving certain accuracy on the imagefolder dataset.

🚀 Quick Start

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 imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 1.2748
Accuracy: 0.5938

✨ Features

Based on the pre - trained model [google/vit - base - patch16 - 224 - in21k](https://huggingface.co/google/vit - base - patch16 - 224 - in21k), it can perform emotion classification on images.
Achieved an accuracy of 0.5938 on the evaluation set.

📚 Documentation

Model Information

Property	Details
Model Type	Fine - tuned version of [google/vit - base - patch16 - 224 - in21k](https://huggingface.co/google/vit - base - patch16 - 224 - in21k)
Training Data	imagefolder dataset
Metrics	Accuracy

Training Procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 7e - 07
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
num_epochs: 1750

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	20	2.0722	0.1875
No log	2.0	40	2.0667	0.2062
No log	3.0	60	2.0633	0.1938
No log	4.0	80	2.0682	0.1938
No log	5.0	100	2.0627	0.2
No log	6.0	120	2.0618	0.2062
No log	7.0	140	2.0573	0.2062
No log	8.0	160	2.0589	0.2062
No log	9.0	180	2.0585	0.1938
No log	10.0	200	2.0576	0.1875
No log	11.0	220	2.0578	0.2125
No log	12.0	240	2.0498	0.2625
No log	13.0	260	2.0542	0.2062
No log	14.0	280	2.0534	0.2
No log	15.0	300	2.0474	0.2562
No log	16.0	320	2.0513	0.225
No log	17.0	340	2.0472	0.2062
No log	18.0	360	2.0465	0.2562
No log	19.0	380	2.0379	0.275
No log	20.0	400	2.0416	0.2375
No log	21.0	420	2.0442	0.2687
No log	22.0	440	2.0399	0.2313
No log	23.0	460	2.0358	0.2625
No log	24.0	480	2.0316	0.2437
2.0458	25.0	500	2.0314	0.2687
2.0458	26.0	520	2.0382	0.2437
2.0458	27.0	540	2.0246	0.275
2.0458	28.0	560	2.0211	0.3187
2.0458	29.0	580	2.0240	0.2625
2.0458	30.0	600	2.0198	0.2875
2.0458	31.0	620	2.0204	0.2625
2.0458	32.0	640	2.0210	0.275
2.0458	33.0	660	2.0046	0.3125
2.0458	34.0	680	2.0153	0.2625
2.0458	35.0	700	2.0098	0.2625
2.0458	36.0	720	2.0085	0.2812
2.0458	37.0	740	1.9978	0.3187
2.0458	38.0	760	1.9962	0.2938
2.0458	39.0	780	1.9955	0.3
2.0458	40.0	800	1.9923	0.3
2.0458	41.0	820	1.9913	0.3063
2.0458	42.0	840	1.9844	0.325
2.0458	43.0	860	1.9922	0.275
2.0458	44.0	880	1.9731	0.3187
2.0458	45.0	900	1.9833	0.3
2.0458	46.0	920	1.9763	0.3187
2.0458	47.0	940	1.9686	0.35
2.0458	48.0	960	1.9707	0.3187
2.0458	49.0	980	1.9631	0.3312
1.9677	50.0	1000	1.9561	0.2812
1.9677	51.0	1020	1.9527	0.3312
1.9677	52.0	1040	1.9447	0.3563
1.9677	53.0	1060	1.9542	0.35
1.9677	54.0	1080	1.9509	0.375
1.9677	55.0	1100	1.9378	0.3937
1.9677	56.0	1120	1.9431	0.3563
1.9677	57.0	1140	1.9397	0.3312
1.9677	58.0	1160	1.9322	0.4
1.9677	59.0	1180	1.9252	0.3688
1.9677	60.0	1200	1.9209	0.3063
1.9677	61.0	1220	1.9210	0.3563
1.9677	62.0	1240	1.9226	0.3688
1.9677	63.0	1260	1.9054	0.3688
1.9677	64.0	1280	1.9153	0.35
1.9677	65.0	1300	1.8993	0.3563
1.9677	66.0	1320	1.9008	0.375
1.9677	67.0	1340	1.9008	0.35
1.9677	68.0	1360	1.8866	0.3625
1.9677	69.0	1380	1.8771	0.3937
1.9677	70.0	1400	1.8758	0.4
1.9677	71.0	1420	1.8682	0.3937
1.9677	72.0	1440	1.8604	0.3688
1.9677	73.0	1460	1.8660	0.3937
1.9677	74.0	1480	1.8702	0.3375
1.8562	75.0	1500	1.8483	0.4313
1.8562	76.0	1520	1.8525	0.3875
1.8562	77.0	1540	1.8467	0.3937
1.8562	78.0	1560	1.8481	0.3812
1.8562	79.0	1580	1.8282	0.3812
1.8562	80.0	1600	1.8395	0.3875
1.8562	81.0	1620	1.8251	0.375
1.8562	82.0	1640	1.8215	0.4125
1.8562	83.0	1660	1.8179	0.3625
1.8562	84.0	1680	1.8130	0.4188
1.8562	85.0	1700	1.8066	0.4
1.8562	86.0	1720	1.7993	0.4062
1.8562	87.0	1740	1.7954	0.4188
1.8562	88.0	1760	1.7936	0.3937
1.8562	89.0	1780	1.7972	0.4188
1.8562	90.0	1800	1.7876	0.3937
1.8562	91.0	1820	1.7810	0.3937
1.8562	92.0	1840	1.7838	0.3937
1.8562	93.0	1860	1.7711	0.3812
1.8562	94.0	1880	1.7780	0.3688
1.8562	95.0	1900	1.7426	0.4062
1.8562	96.0	1920	1.7399	0.425
1.8562	97.0	1940	1.7561	0.375
1.8562	98.0	1960	1.7287	0.3937
1.8562	99.0	1980	1.7425	0.4437
1.7294	100.0	2000	1.7308	0.4188
1.7294	101.0	2020	1.7389	0.375
1.7294	102.0	2040	1.7249	0.4375
1.7294	103.0	2060	1.7297	0.4188
1.7294	104.0	2080	1.7361	0.3875
1.7294	105.0	2100	1.7188	0.4188
1.7294	106.0	2120	1.7181	0.4562
1.7294	107.0	2140	1.7044	0.425
1.7294	108.0	2160	1.7030	0.4188
1.7294	109.0	2180	1.7070	0.425
1.7294	110.0	2200	1.7006	0.4437
1.7294	111.0	2220	1.6862	0.4688
1.7294	112.0	2240	1.6881	0.4437
1.7294	113.0	2260	1.6798	0.45
1.7294	114.0	2280	1.6982	0.3937
1.7294	115.0	2300	1.6812	0.4375
1.7294	116.0	2320	1.6751	0.45
1.7294	117.0	2340	1.6849	0.45
1.7294	118.0	2360	1.6690	0.4375
1.7294	119.0	2380	1.6583	0.4688
1.7294	120.0	2400	1.6666	0.4375
1.7294	121.0	2420	1.6651	0.45
1.7294	122.0	2440	1.6477	0.4688
1.7294	123.0	2460	1.6520	0.475
1.7294	124.0	2480	1.6642	0.4125
1.6198	125.0	2500	1.6633	0.475
1.6198	126.0	2520	1.6443	0.4375
1.6198	127.0	2540	1.6398	0.4813
1.6198	128.0	2560	1.6516	0.4437
1.6198	129.0	2580	1.6496	0.4562
1.6198	130.0	2600	1.6342	0.4625
1.6198	131.0	2620	1.6330	0.4437
1.6198	132.0	2640	1.6341	0.4625
1.6198	133.0	2660	1.6167	0.475
1.6198	134.0	2680	1.6270	0.4562
1.6198	135.0	2700	1.6322	0.4562
1.6198	136.0	2720	1.6252	0.4188
1.6198	137.0	2740	1.6045	0.4562
1.6198	138.0	2760	1.6107	0.5
1.6198	139.0	2780	1.6170	0.45
1.6198	140.0	2800	1.6067	0.4813
1.6198	141.0	2820	1.6116	0.4437
1.6198	142.0	2840	1.5962	0.4813
1.6198	143.0	2860	1.6096	0.425
1.6198	144.0	2880	1.6051	0.4562
1.6198	145.0	2900	1.5939	0.525
1.6198	146.0	2920	1.5871	0.5062
1.6198	147.0	2940	1.5917	0.4688
1.6198	148.0	2960	1.5911	0.4938
1.6198	149.0	2980	1.5950	0.4437
1.5343	150.0	3000	1.5974	0.4625
1.5343	151.0	3020	1.5713	0.4813
1.5343	152.0	3040	1.5637	0.5437
1.5343	153.0	3060	1.5830	0.4688
1.5343	154.0	3080	1.5770	0.4938
1.5343	155.0	3100	1.5800	0.5188
1.5343	156.0	3120	1.5782	0.4625
1.5343	157.0	3140	1.5769	0.4562
1.5343	158.0	3160	1.5751	0.5
1.5343	159.0	3180	1.5739	0.4938
1.5343	160.0	3200	1.5555	0.5
1.5343	161.0	3220	1.5561	0.475
1.5343	162.0	3240	1.5521	0.4938
1.5343	163.0	3260	1.5670	0.4688
1.5343	164.0	3280	1.5621	0.475
1.5343	165.0	3300	1.5622	0.4813
1.5343	166.0	3320	1.5497	0.4562
1.5343	167.0	3340	1.5535	0.4813
1.5343	168.0	3360	1.5526	0.4875
1.5343	169.0	3380	1.5301	0.5
1.5343	170.0	3400	1.5487	0.4813
1.5343	171.0	3420	1.5489	0.4625
1.5343	172.0	3440	1.5456	0.5
1.5343	173.0	3460	1.5355	0.5
1.5343	174.0	3480	1.5306	0.5
1.4643	175.0	3500	1.5436

📄 License

This model 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