Downloads · 30 days
18
23% of all-time downloads
trocr-noise-bigram2
cyttic/trocr-noise-bigram2
trocr-noise-bigram2 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. --
All-time downloads
77
Public
Parameters
299M
2.4 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1.2 GB · 100%
At a glance
- Task
- Image-Text-to-Text
- Library
- transformers
- License
- mit
- Model type
- vision-encoder-decoder
- Access
- Public
- Created
- Aug 17, 2026
- Updated
- Oct 30, 2026
- SHA
- 8d738d72
Try a prompt
Base models
Model card
From the Hugging Face model README
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
trocr-noise-bigram2
This model is a fine-tuned version of cyttic/exp2-frozen-benyehuda-cont on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5353
- Cer: 0.0311
- Wer: 0.0865
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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: 2
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer | Wer |
|---|---|---|---|---|---|
| 4.4701 | 0.1290 | 2000 | 2.0749 | 0.1892 | 0.3917 |
| 3.4282 | 0.2581 | 4000 | 1.5678 | 0.1288 | 0.2889 |
| 2.6796 | 0.3871 | 6000 | 1.2115 | 0.0937 | 0.2210 |
| 2.3398 | 0.5161 | 8000 | 1.0724 | 0.0778 | 0.1845 |
| 1.9402 | 0.6452 | 10000 | 0.9204 | 0.0661 | 0.1640 |
| 1.8243 | 0.7742 | 12000 | 0.8303 | 0.0570 | 0.1429 |
| 1.6875 | 0.9032 | 14000 | 0.7434 | 0.0509 | 0.1294 |
| 1.2807 | 1.0323 | 16000 | 0.6928 | 0.0457 | 0.1188 |
| 1.2310 | 1.1613 | 18000 | 0.6552 | 0.0419 | 0.1105 |
| 1.1186 | 1.2903 | 20000 | 0.6258 | 0.0380 | 0.1035 |
| 1.0514 | 1.4194 | 22000 | 0.6097 | 0.0380 | 0.1000 |
| 1.1760 | 1.5484 | 24000 | 0.5739 | 0.0346 | 0.0934 |
| 1.0901 | 1.6774 | 26000 | 0.5598 | 0.0339 | 0.0922 |
| 1.0063 | 1.8065 | 28000 | 0.5434 | 0.0328 | 0.0888 |
| 0.9992 | 1.9355 | 30000 | 0.5368 | 0.0309 | 0.0861 |
| 1.0274 | 2.0 | 31000 | 0.5353 | 0.0311 | 0.0865 |
Framework versions
- Transformers 5.9.0
- Pytorch 2.11.0+cu128
- Datasets 5.0.1
- Tokenizers 0.22.2
- Task
- Image-Text-to-Text
- Library
- transformers
- Type
- vision-encoder-decoder
- License
- mit
- Created
- Aug 17, 2026
- Updated
- Oct 30, 2026