vit-convnext-mistral-UCM-with-all-captioning

swadhindas324/vit-convnext-mistral-UCM-with-all-captioning

vit-convnext-mistral-UCM-with-all-captioning is a machine learning model from swadhindas324. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for transformers.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

2% of all-time downloads

All-time downloads

284

Public

Parameters

436M

14 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors1.7 GB · 100%

At a glance

Library: transformers
Model type: vision-encoder-decoder
Access: Public
Created: Jul 2, 2026
Updated: Jul 2, 2026
SHA: 2f367ea0

Model card

From the Hugging Face model README

vit-convnext-mistral-UCM-with-all-captioning

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

Loss: 0.4479
Accuracy: 75.19
Bleu-1: 0.8345
Bleu-2: 0.7729
Bleu-3: 0.7221
Bleu-4: 0.6757
Meteor: 0.8199
Rouge-l: 0.8061
Cider: 3.4094

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: 64
eval_batch_size: 64
seed: 50
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 0.1
num_epochs: 128
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Bleu-1	Bleu-2	Bleu-3	Bleu-4	Meteor	Rouge-l	Cider
No log	1.0	148	0.6429	72.42	0.4183	0.2883	0.2137	0.1617	0.3238	0.3520	0.3757
No log	2.0	296	0.4703	71.77	0.2471	0.1637	0.0986	0.0580	0.1748	0.2068	0.1464
No log	3.0	444	0.4219	70.83	0.6354	0.5507	0.4910	0.4486	0.5772	0.5899	2.0520
No log	4.0	592	0.3597	73.98	0.8239	0.7577	0.7051	0.6597	0.7693	0.7665	3.1919
No log	5.0	740	0.3467	73.61	0.8180	0.7503	0.6969	0.6504	0.7961	0.7833	3.2323
No log	6.0	888	0.3542	74.15	0.8323	0.7820	0.7342	0.6899	0.8178	0.8061	3.2944
0.4525	7.0	1036	0.3436	74.75	0.8657	0.8136	0.7657	0.7206	0.8388	0.8196	3.4979
0.4525	8.0	1184	0.3479	74.44	0.8301	0.7710	0.7204	0.6743	0.8021	0.7915	3.3577
0.4525	9.0	1332	0.3497	74.76	0.8618	0.8100	0.7639	0.7219	0.8405	0.8308	3.5613
0.4525	10.0	1480	0.3553	74.36	0.8501	0.7928	0.7446	0.7024	0.8192	0.8060	3.4316
0.4525	11.0	1628	0.3627	73.84	0.8239	0.7630	0.7100	0.6607	0.8100	0.7915	3.2133
0.4525	12.0	1776	0.3686	74.04	0.8643	0.8123	0.7677	0.7264	0.8325	0.8222	3.5114
0.4525	13.0	1924	0.3811	73.95	0.8372	0.7802	0.7330	0.6910	0.8152	0.7995	3.3588
0.1648	14.0	2072	0.3856	74.34	0.8458	0.7917	0.7432	0.6959	0.8355	0.8096	3.4077
0.1648	15.0	2220	0.3780	75.42	0.8701	0.8118	0.7616	0.7139	0.8284	0.8197	3.6168
0.1648	16.0	2368	0.3949	74.79	0.8628	0.8111	0.7671	0.7256	0.8325	0.8243	3.5344
0.1648	17.0	2516	0.4073	75.05	0.8683	0.8189	0.7778	0.7394	0.8494	0.8359	3.6291
0.1648	18.0	2664	0.4107	74.02	0.8495	0.7943	0.7486	0.7075	0.8372	0.8245	3.5079
0.1648	19.0	2812	0.4219	74.35	0.8534	0.7967	0.7472	0.7013	0.8320	0.8194	3.4285
0.1648	20.0	2960	0.4221	74.77	0.8417	0.7871	0.7441	0.7041	0.8271	0.8185	3.5187
0.1393	21.0	3108	0.4146	74.25	0.8521	0.7965	0.7510	0.7081	0.8309	0.8190	3.5273
0.1393	22.0	3256	0.4240	75.07	0.8429	0.7868	0.7419	0.7003	0.8337	0.8149	3.5179
0.1393	23.0	3404	0.4171	74.24	0.8635	0.8146	0.7731	0.7344	0.8309	0.8228	3.5014
0.1393	24.0	3552	0.4516	74.92	0.8456	0.7912	0.7425	0.6998	0.8030	0.7939	3.3774
0.1393	25.0	3700	0.4288	75.28	0.8583	0.8081	0.7640	0.7229	0.8366	0.8228	3.5744
0.1393	26.0	3848	0.4363	74.46	0.8706	0.8162	0.7684	0.7263	0.8358	0.8267	3.5370
0.1393	27.0	3996	0.4479	75.19	0.8345	0.7729	0.7221	0.6757	0.8199	0.8061	3.4094

Framework versions

Transformers 5.12.1
Pytorch 2.12.1+cu130
Datasets 5.0.0
Tokenizers 0.22.2

Library: transformers
Type: vision-encoder-decoder
Created: Jul 2, 2026
Updated: Jul 2, 2026