exp23-directfit-unfrozen

cyttic/exp23-directfit-unfrozen

exp23-directfit-unfrozen is a image-text-to-text model from cyttic. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as mit.

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

Downloads · 30 days

5% of all-time downloads

All-time downloads

748

Public

Parameters

299M

86.2 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors2.4 GB · 50%

At a glance

Task: Image-Text-to-Text
Library: transformers
License: mit
Model type: vision-encoder-decoder
Access: Public
Created: Jul 11, 2026
Updated: Oct 30, 2026
SHA: 8f34293f

Try a prompt

Base models

cyttic/exp22-exp2warm-directfit-frozen

Model card

From the Hugging Face model README

exp23-directfit-unfrozen

This model is a fine-tuned version of cyttic/exp22-exp2warm-directfit-frozen on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.4588
Cer: 0.0245
Wer: 0.0647

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: 16
seed: 42
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: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
1.2577	0.0424	2500	1.3378	0.1416	0.2986
1.0413	0.0848	5000	1.2441	0.1150	0.2561
0.8436	0.1272	7500	1.0854	0.0917	0.2106
0.7762	0.1696	10000	0.9314	0.0735	0.1700
0.7096	0.2120	12500	0.8954	0.0627	0.1534
0.6205	0.2544	15000	0.7907	0.0550	0.1378
0.5873	0.2968	17500	0.7408	0.0486	0.1215
0.5504	0.3392	20000	0.6894	0.0427	0.1112
0.4826	0.3816	22500	0.6537	0.0406	0.1042
0.5224	0.4240	25000	0.6245	0.0380	0.0965
0.4430	0.4664	27500	0.5866	0.0366	0.0933
0.4219	0.5088	30000	0.5724	0.0340	0.0873
0.3557	0.5512	32500	0.5586	0.0318	0.0849
0.3866	0.5936	35000	0.5399	0.0328	0.0845
0.3586	0.6360	37500	0.5267	0.0305	0.0793
0.4009	0.6784	40000	0.5082	0.0301	0.0769
0.3140	0.7208	42500	0.5011	0.0289	0.0740
0.3435	0.7632	45000	0.4874	0.0276	0.0725
0.3141	0.8056	47500	0.4807	0.0267	0.0691
0.3200	0.8480	50000	0.4738	0.0267	0.0687
0.3201	0.8904	52500	0.4682	0.0250	0.0664
0.3314	0.9328	55000	0.4633	0.0249	0.0652
0.3427	0.9752	57500	0.4604	0.0244	0.0648
0.3020	1.0	58962	0.4588	0.0245	0.0647

Framework versions

Transformers 5.12.1
Pytorch 2.11.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Task: Image-Text-to-Text
Library: transformers
Type: vision-encoder-decoder
License: mit
Created: Jul 11, 2026
Updated: Oct 30, 2026