feat: 初始提交

.gitignore

+.idea

Dockerfile

+FROM image.sourcefind.cn:5000/gpu/admin/base/jupyterlab-pytorch:2.2.0-python3.10-cuda12.1-ubuntu22.04 as base
+ARG IMAGE=joytag
+ARG IMAGE_UPPER=joytag
+ARG BRANCH=gpu
+RUN cd /root && git clone -b $BRANCH http://developer.hpccube.com/codes/chenpangpang/$IMAGE.git
+WORKDIR /root/$IMAGE/$IMAGE_UPPER
+RUN pip install -r requirements.txt
+
+
+#########
+# Prod  #
+#########
+FROM image.sourcefind.cn:5000/gpu/admin/base/jupyterlab-pytorch:2.2.0-python3.10-cuda12.1-ubuntu22.04
+ARG IMAGE=joytag
+ARG IMAGE_UPPER=joytag
+COPY chenyh/$IMAGE/frpc_linux_amd64_v0.2 /opt/conda/lib/python3.10/site-packages/gradio/
+RUN chmod +x /opt/conda/lib/python3.10/site-packages/gradio/frpc_linux_amd64_v0.2
+COPY chenyh/$IMAGE/fancyfeast/joytag /root/$IMAGE_UPPER/fancyfeast/joytag
+COPY --from=base /opt/conda/lib/python3.10/site-packages /opt/conda/lib/python3.10/site-packages
+COPY --from=base /root/$IMAGE/$IMAGE_UPPER /root/$IMAGE_UPPER
+COPY --from=base /root/$IMAGE/启动器.ipynb /root/$IMAGE/start.sh /root/

hf_down.py

+# pip install huggingface-cli
+import os
+
+os.environ['HF_ENDPOINT'] = 'https://hf-mirror.com'
+
+model_list = [
+    "fancyfeast/joytag"
+]
+
+for model_path in model_list:
+    os.system(
+        f"huggingface-cli download --resume-download  {model_path} --local-dir ./{model_path} --local-dir-use-symlinks False")

joytag/README.md

+---
+title: Joytag
+emoji: 📊
+colorFrom: yellow
+colorTo: gray
+sdk: gradio
+sdk_version: 4.36.0
+app_file: app.py
+pinned: false
+license: apache-2.0
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

joytag/app.py

+import gradio as gr
+from Models import VisionModel
+from PIL import Image
+import torch.amp.autocast_mode
+from pathlib import Path
+import torch
+import torchvision.transforms.functional as TVF
+
+
+MODEL_REPO = "fancyfeast/joytag"
+THRESHOLD = 0.4
+DESCRIPTION = """
+joytag：一款图像多分类打标签工具，预测标签种类多达5000，可生成多标签及相应的概率预测
+"""
+
+
+def prepare_image(image: Image.Image, target_size: int) -> torch.Tensor:
+	# Pad image to square
+	image_shape = image.size
+	max_dim = max(image_shape)
+	pad_left = (max_dim - image_shape[0]) // 2
+	pad_top = (max_dim - image_shape[1]) // 2
+
+	padded_image = Image.new('RGB', (max_dim, max_dim), (255, 255, 255))
+	padded_image.paste(image, (pad_left, pad_top))
+
+	# Resize image
+	if max_dim != target_size:
+		padded_image = padded_image.resize((target_size, target_size), Image.BICUBIC)
+	
+	# Convert to tensor
+	image_tensor = TVF.pil_to_tensor(padded_image) / 255.0
+
+	# Normalize
+	image_tensor = TVF.normalize(image_tensor, mean=[0.48145466, 0.4578275, 0.40821073], std=[0.26862954, 0.26130258, 0.27577711])
+
+	return image_tensor
+
+
+@torch.no_grad()
+def predict(image: Image.Image):
+	image_tensor = prepare_image(image, model.image_size)
+	batch = {
+		'image': image_tensor.unsqueeze(0),
+	}
+
+	with torch.amp.autocast_mode.autocast('cpu', enabled=True):
+		preds = model(batch)
+		tag_preds = preds['tags'].sigmoid().cpu()
+	
+	scores = {top_tags[i]: tag_preds[0][i] for i in range(len(top_tags))}
+	predicted_tags = [tag for tag, score in scores.items() if score > THRESHOLD]
+	tag_string = ', '.join(predicted_tags)
+
+	return tag_string, scores
+
+
+print("Loading model...")
+path = MODEL_REPO
+model = VisionModel.load_model(path)
+model.eval()
+
+with open(Path(path) / 'top_tags.txt', 'r') as f:
+	top_tags = [line.strip() for line in f.readlines() if line.strip()]
+
+print("Starting server...")
+
+gradio_app = gr.Interface(
+	predict,
+	inputs=gr.Image(label="Source", sources=['upload', 'webcam'], type='pil'),
+	outputs=[
+		gr.Textbox(label="Tag String"),
+		gr.Label(label="Tag Predictions", num_top_classes=100),
+	],
+	title="JoyTag",
+	description=DESCRIPTION,
+	allow_flagging="never",
+)
+
+
+if __name__ == '__main__':
+	gradio_app.launch()

joytag/requirements.txt

+torch
+transformers
+torchvision
+einops
+safetensors
\ No newline at end of file

start.sh

+#!/bin/bash
+cd /root/joytag
+python app.py

启动器.ipynb

+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "e5c5a211-2ccd-4341-af10-ac546484b91f",
+   "metadata": {
+    "tags": []
+   },
+   "source": [
+    "## 说明\n",
+    "- 启动需要加载模型，需要2分钟左右的时间\n",
+    "- 启动和重启 Notebook 点上方工具栏中的「重启并运行所有单元格」。出现如下内容就算成功了：\n",
+    "    - `Running on local URL:  http://0.0.0.0:7860`\n",
+    "    - `Running on public URL: https://xxxxxxxxxxxxxxx.gradio.live`\n",
+    "- 通过以下方式开启页面：\n",
+    "    - 控制台打开「自定义服务」了，访问自定义服务端口号设置为7860\n",
+    "    - 直接打开显示的公开链接`public URL`\n",
+    "\n",
+    "## 功能介绍\n",
+    "- 原项目地址：https://github.com/fpgaminer/joytag\n",
+    "- joytag：一款图像多分类打标签工具，预测标签种类多达5000，可生成多标签及相应的概率预测"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "53a96614-e2d2-4710-a82b-0d5ca9cb9872",
+   "metadata": {
+    "tags": [],
+    "is_executing": true
+   },
+   "outputs": [],
+   "source": [
+    "# 启动\n",
+    "!sh start.sh"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9e81ae9d-3a34-43a0-943a-ff5e9d6ce961",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}