8

app.py

@@ -1,20 +1,9 @@
 import gradio as gr
-import testMch as func
-
-
-def work(imgurl):
-    # output = testMch.describeImage2(img)
-    # # output = testMch.agent(img)
-    # # output = testMch.agent(img)
-    # return output
-    # imgurl = "r" + imgurl
-    # imgurl = imgurl.replace("\\\\\\", "\\")
-    # input = "Describe the following image:\n" + imgurl
-    out = func.agent(f"Please describe the following image:\n{imgurl}")
-    anws = func.toChinese(out['output'])
-    return anws
+import image_feature as func
 
 
+def work(image1, image2):
+    return func.infer1(image1, image2)
 
 
 # with gr.Blocks() as demo:
@@ -25,15 +14,12 @@ def work(imgurl):
 #     submit.click(work, inputs=[image_url, input], outputs=output)
 # demo.launch()
 
-demo = gr.Interface(title="识别图片",
-                    css="",
-                    fn=work,
-                    inputs=[gr.Image(type="filepath", label="请上传图片")],
-                    outputs=[gr.Textbox(lines=3, label="识别结果")])
-#
-# # demo = gr.Interface(fn=work, inputs="image,text", outputs="text")
-#
-demo.launch()
-
-# # interface = gr.Interface(fn=agent, inputs="image", outputs="text")
-# # interface.launch()
+# 定义你的界面
+with gr.Interface(fn=work,
+                  inputs=[gr.Textbox(label='图片1', lines=1), gr.Textbox(label='图片2', lines=1)],  # 两个文本输入框
+                  outputs=[gr.Textbox(lines=3, label="推理结果")],  # 输出为文本
+                  title="图片相似度推理",  # 界面标题
+                  description="输入两张图片链接进行相似度推理",  # 界面描述
+                  examples=[["https://example.com", "https://google.com"],  # 示例输入
+                            ["https://github.com", "https://twitter.com"]]) as demo:  # 更多示例输入
+    demo.launch()  # 启动界面

app_1.py

@@ -0,0 +1,39 @@
+import gradio as gr
+import testMch as func
+
+
+def work(imgurl):
+    # output = testMch.describeImage2(img)
+    # # output = testMch.agent(img)
+    # # output = testMch.agent(img)
+    # return output
+    # imgurl = "r" + imgurl
+    # imgurl = imgurl.replace("\\\\\\", "\\")
+    # input = "Describe the following image:\n" + imgurl
+    out = func.agent(f"Please describe the following image:\n{imgurl}")
+    anws = func.toChinese(out['output'])
+    return anws
+
+
+
+
+# with gr.Blocks() as demo:
+#     image_url = gr.Image(type="filepath", label="请选择一张图片")
+#     input = gr.Textbox(label='请描述您的问题', placeholder="", lines=1)
+#     output = gr.Textbox(label='答案', placeholder="", lines=2, interactive=False)
+#     submit = gr.Button('提问', variant="primary")
+#     submit.click(work, inputs=[image_url, input], outputs=output)
+# demo.launch()
+
+demo = gr.Interface(title="识别图片",
+                    css="",
+                    fn=work,
+                    inputs=[gr.Image(type="filepath", label="请上传图片")],
+                    outputs=[gr.Textbox(lines=3, label="识别结果")])
+#
+# # demo = gr.Interface(fn=work, inputs="image,text", outputs="text")
+#
+demo.launch()
+
+# # interface = gr.Interface(fn=agent, inputs="image", outputs="text")
+# # interface.launch()

image_feature.py

@@ -0,0 +1,75 @@
+# import numpy as np
+# import requests
+import torch
+# from PIL import Image
+from torch.nn.functional import cosine_similarity
+from transformers import AutoImageProcessor, AutoModel
+
+# from transformers import pipeline
+
+# import transformers
+#
+# print(transformers.__version__)
+#
+# img_urls = ["https://img0.baidu.com/it/u=3704428154,2884159591&fm=253&fmt=auto&app=138&f=JPEG?w=889&h=500",
+#             "https://img0.baidu.com/it/u=3704428154,2884159591&fm=253&fmt=auto&app=138&f=JPEG?w=889&h=500"]
+#
+# image_real = Image.open(requests.get(img_urls[0], stream=True).raw).convert("RGB")
+# image_gen = Image.open(requests.get(img_urls[1], stream=True).raw).convert("RGB")
+#
+# # DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+DEVICE = torch.device('cpu')
+# pipe = pipeline(task="image-feature-extraction", model_name="google/vit-base-patch16-384", device=DEVICE, pool=True)
+#
+# # 1提取图片特征向量
+# outputs = pipe([image_real, image_gen])
+#
+# # get the length of a single output
+# print(len(outputs[0][0]))
+# # show outputs
+# print(outputs)
+#
+# # 768
+# # [[[-0.03909236937761307, 0.43381670117378235, -0.06913255900144577,
+#
+# # 2计算图片相似度
+# similarity_score = cosine_similarity(torch.Tensor(outputs[0]),
+#                                      torch.Tensor(outputs[1]), dim=1)
+#
+# print(similarity_score)
+
+# tensor([0.6043])
+
+# pipe = pipeline(task="image-feature-extraction", model_name="google/vit-base-patch16-224", device=DEVICE)
+# output = pipe(image_real)
+#
+# # 其中第一个维度是批量大小，最后两个维度是嵌入形状。
+# print(np.array(outputs).shape)
+# # (1, 197, 768)
+
+
+# 第二种方式推理图片相似度
+processor = AutoImageProcessor.from_pretrained("google/vit-base-patch16-224")
+model = AutoModel.from_pretrained("google/vit-base-patch16-224").to(DEVICE)
+
+
+#  推理
+def infer(image):
+    inputs = processor(image, return_tensors="pt").to(DEVICE)
+    outputs = model(**inputs)
+    return outputs.pooler_output
+
+
+#  推理
+def infer1(image1, image2):
+    embed_real = infer(image1)
+    embed_gen = infer(image2)
+    similarity_score = cosine_similarity(embed_real, embed_gen, dim=1)
+    print(similarity_score)
+    # 如果你想在CPU上操作这个值，你需要先将tensor移动到CPU
+    t_cpu = similarity_score.cpu()
+
+    # 然后提取这个值
+    return t_cpu.item()
+
+# tensor([0.6061], device='cuda:0', grad_fn=<SumBackward1>)

requirements.txt

@@ -1,6 +1,4 @@
-transformers
+transformers==4.39.1
 torch
-langchain
-openai
 gradio
 pillow

requirements_bak.txt

@@ -0,0 +1,6 @@
+transformers
+torch
+langchain
+openai
+gradio
+pillow