Instructions to use hs4167/pal4vst_food with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use hs4167/pal4vst_food with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-segmentation", model="hs4167/pal4vst_food")# pip install -U transformers accelerate # Load model directly from transformers import AutoImageProcessor, UperNetForSemanticSegmentation processor = AutoImageProcessor.from_pretrained("hs4167/pal4vst_food") model = UperNetForSemanticSegmentation.from_pretrained("hs4167/pal4vst_food", device_map="auto") - Notebooks
- Google Colab
- Kaggle
pal4vst_food
AI μμ± μμ μ΄λ―Έμ§μμ μ΄μν μμ(μν°ν©νΈ) ν보λ₯Ό ν½μ λ¨μλ‘ μΆμ νλ λΆν λͺ¨λΈμ λλ€. Lingzhi/PAL4VST(MIT, ICCV 2023)λ₯Ό μ§μ λΌλ²¨λ§ν AI μμ± μμ μ΄λ―Έμ§λ‘ λ―ΈμΈμ‘°μ νμ΅λλ€.
- ꡬ쑰: Swin-B + UperNet (
transformers.UperNetForSemanticSegmentation), 2ν΄λμ€ (0 background, 1 bad_region) - νμ΅: AI μμ± μμ μ΄λ―Έμ§ 12μ₯μ μ΄μν μμ λ§μ€ν¬(μ§μ λΌλ²¨λ§), 1250 iter
- μ μ²λ¦¬: λΉμ¨ μ μ§λ‘ κΈ΄ λ³ 512, RGB νκ· (123.675, 116.28, 103.53) / νμ€νΈμ°¨ (58.395, 57.12, 57.375)
λ²μ
| μ»€λ° | λ΄μ© |
|---|---|
da449b1 |
v1 |
deed190 |
v2: λ‘λ³Άμ΄ μ΄λ―Έμ§ λΌλ²¨ μμ ν μ¬νμ΅ (νμ΅ μ€μ λμΌ) |
μ¬μ©
from transformers import UperNetForSemanticSegmentation
model = UperNetForSemanticSegmentation.from_pretrained("hs4167/pal4vst_food", revision="deed190ca7f734adaa275624744143c005f2e3e8")
νκ³
- νμ΅ λ°μ΄ν°κ° 12μ₯λΏμ΄κ³ μ€μ μ¬μ§(μ€λ₯ μμ)μ νμ΅μ μ°μ§ μμ, μ€μ μμ μ¬μ§μμλ λμ μμμ νλ³΄λ‘ μ‘λ μ€νμ΄ μμ΅λλ€.
- κ²°κ³Όλ μ€λ₯ ν보μ μΆμ μ΄λ©°, μ΄λ―Έμ§κ° AIλ‘ μμ±λμλμ§ νλ³νκ±°λ νμ§μ 보μ¦νμ§ μμ΅λλ€.
λΌμ΄μ μ€
Apache-2.0 (LICENSE). μλ³Έ PAL4VST κ°μ€μΉμ MIT κ³ μ§λ NOTICEμ μμ΅λλ€.
- Downloads last month
- 47
Model tree for hs4167/pal4vst_food
Base model
Lingzhi/PAL4VST