CLIP模型的代码

CLIP的github上给出了几段代码

首先是第一段代码

import torch
import clip
from PIL import Image

device = "cuda" if torch.cuda.is_available() else "cpu"
model, preprocess = clip.load("ViT-B/32", device=device)

image = preprocess(Image.open("CLIP.png")).unsqueeze(0).to(device)
text = clip.tokenize(["a diagram", "a dog", "a cat"]).to(device)

with torch.no_grad():
    image_features = model.encode_image(image)
    text_features = model.encode_text(text)
    
    logits_per_image, logits_per_text = model(image, text)
    probs = logits_per_image.softmax(dim=-1).cpu().numpy()

print("Label probs:", probs)  # prints: [[0.9927937  0.00421068 0.00299572]]

看完代码可以发现，最后的预测，image_features，text_features完全没有参与

第二段代码

import os
import clip
import torch
from torchvision.datasets import CIFAR100

# Load the model
device = "cuda" if torch.cuda.is_available() else "cpu"
model, preprocess = clip.load('ViT-B/32', device)

# Download the dataset
cifar100 = CIFAR100(root=os.path.expanduser("~/.cache"), download=True, train=False)

# Prepare the inputs
image, class_id = cifar100[3637]
image_input = preprocess(image).unsqueeze(0).to(device)
text_inputs = torch.cat([clip.tokenize(f"a photo of a {c}") for c in cifar100.classes]).to(device)

# Calculate features
with torch.no_grad():
    image_features = model.encode_image(image_input)
    text_features = model.encode_text(text_inputs)

# Pick the top 5 most similar labels for the image
image_features /= image_features.norm(dim=-1, keepdim=True)
text_features /= text_features.norm(dim=-1, keepdim=True)
similarity = (100.0 * image_features @ text_features.T).softmax(dim=-1)
values, indices = similarity[0].topk(5)

# Print the result
print("\nTop predictions:\n")
for value, index in zip(values, indices):
    print(f"{cifar100.classes[index]:>16s}: {100 * value.item():.2f}%")

看完可以比较容易的发现，这里直接用image_features和text_features去做矩阵乘法，然后进行softmax操作就得到了最终的预测结果，这里没有经过model计算出logit，所以model所做的事情就是计算两个feature然后使用矩阵乘法乘起来。

最后简单的验证了一下

import os
import clip
import torch
from torchvision.datasets import CIFAR100

# Load the model
device = "cuda" if torch.cuda.is_available() else "cpu"
print(f"use device : {device}")
model, preprocess = clip.load('ViT-B/32', device)
# Download the dataset and the train=False that is mean we will download or load the test set
cifar100 = CIFAR100(root=os.path.expanduser("~/.cache"), download=True, train=False)

print(cifar100)

image, class_id = cifar100[3637]
image_input = preprocess(image).unsqueeze(0).to(device)
text_inputs = torch.cat([clip.tokenize(f"a photo of a {c}") for c in cifar100.classes]).to(device)

with torch.no_grad():
    image_features = model.encode_image(image_input)
    text_features = model.encode_text(text_inputs)
    logits_per_image, logits_per_text = model(image_input, text_inputs)
    print(logits_per_image.softmax(-1))
# doing L2 normalization 
image_features /= image_features.norm(dim=-1, keepdim=True)
text_features /= text_features.norm(dim=-1, keepdim=True)
print(image_features.shape)
similarity = (100.0 * image_features @ text_features.T).softmax(dim=-1)
print(similarity)
print(similarity.shape)
values, indices = similarity[0].topk(5)

# Print the result
print("\nTop predictions:\n")
for value, index in zip(values, indices):
    print(f"{cifar100.classes[index]:>16s}: {100 * value.item():.2f}%")

可以发现两种方法的预测结果确实是一样的。