Visual Question Answering
Transformers
Safetensors
Vietnamese
openvivqa
scene-text-vqa
vietnamese
ocr
t5
Instructions to use Kus669/ViSceT5-OpenViVQA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Kus669/ViSceT5-OpenViVQA with Transformers:
# Use a pipeline as a high-level helper # Warning: Pipeline type "visual-question-answering" is no longer supported in transformers v5. # You must load the model directly (see below) or downgrade to v4.x with: # pip install "transformers<5.0.0" from transformers import pipeline pipe = pipeline("visual-question-answering", model="Kus669/ViSceT5-OpenViVQA")# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Kus669/ViSceT5-OpenViVQA", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download predictions/eval_test.json from Kus669/ViSceT5-OpenViVQA: direct link, hf CLI and curl.
- Browser
- Download file 472 Bytes
-
https://huggingface.co/Kus669/ViSceT5-OpenViVQA/resolve/main/predictions/eval_test.json
- Command line
-
hf download hf://Kus669/ViSceT5-OpenViVQA/predictions/eval_test.json
-
curl -L -o eval_test.json https://huggingface.co/Kus669/ViSceT5-OpenViVQA/resolve/main/predictions/eval_test.json
472 Bytes
| { | |
| "Exact Match (EM)": 0.13207013574660634, | |
| "F1 (Token-level)": 0.58759340489632, | |
| "F1 (Char-level)": 0.6936284312263501, | |
| "BLEU-1": 0.5551767097239343, | |
| "BLEU-2": 0.4964673680534233, | |
| "BLEU-3": 0.44992158882951094, | |
| "BLEU-4": 0.41233142716421745, | |
| "BLEU-avg": 0.4784742734427715, | |
| "CIDEr": 3.7202144011530436, | |
| "BERTScore_F1": 0.8585119033443038, | |
| "BERTScore_P": 0.8521413641109455, | |
| "BERTScore_R": 0.8666205599027521, | |
| "n_evaluated": 3536, | |
| "n_gt": 3536 | |
| } |