Zero-Shot Image Classification
Transformers
Safetensors
English
Chinese
fgclip2
text-generation
clip
custom_code
Instructions to use qihoo360/fg-clip2-base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use qihoo360/fg-clip2-base with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("zero-shot-image-classification", model="qihoo360/fg-clip2-base", trust_remote_code=True) pipe( "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/parrots.png", candidate_labels=["animals", "humans", "landscape"], )# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("qihoo360/fg-clip2-base", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
demo存在一些问题
#2
by Ahyd - opened
Ahyd changed discussion title from demo似乎拟合严重 to demo存在一些问题
Ahyd changed discussion status to closed
补充:最开始除了上述“两双浅色鞋子”和“两双运动鞋”的差异还有绿植和仙人掌的差异
Ahyd changed discussion status to open
hello,这张demo,是我用chatgpt根据正确描述生成的3个负样本,我们没有在训练中去拟合该图片,我们的宣传稿也有很多测试样例。值得一提的,由于FG-CLIP是一个面向广泛领域,且基于弱监督训练的图文对齐模型,所以不可避免有一些case无法满足需求,恳请你能理解。


