Hugging Face's logo

flax-community
/
clip-reply

TensorBoard
Model card Files
xet
 Metrics Community
clip-reply
File size: 3,848 Bytes
c636b54
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60bfd82
 
 
 
 
 
 
 
c636b54
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "source": [
    "import jsonlines\n",
    "import os"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "source": [
    "import cv2"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "source": [
    "# extract first frame from mp4 file and save it as image\n",
    "\n",
    "def extract_frame(mp4,file_name):\n",
    "    save_path='./imgs/'+file_name[:-4]+\".jpg\"\n",
    "    # print(save_path)\n",
    "    if os.path.exists(save_path):\n",
    "        # print(\"file already exists\")\n",
    "        return\n",
    "    else:\n",
    "        try:\n",
    "            # extract first frame\n",
    "            cap = cv2.VideoCapture(mp4)\n",
    "            ret, frame = cap.read()\n",
    "            cv2.imwrite(save_path, frame)\n",
    "        except:\n",
    "            print(f\"error in {save_path} file\")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "source": [
    "# read jsonlines\n",
    "im_un=[]\n",
    "base_dir=\"./gifs\"\n",
    "data=jsonlines.open('./ReactionGIF.json')\n",
    "\n",
    "train_writer=jsonlines.open(\"train.json\", mode='w')\n",
    "val_writer=jsonlines.open(\"val.json\", mode='w')\n",
    "writer=train_writer\n",
    "\n",
    "for x in data:\n",
    "    text=x[\"text\"]\n",
    "    sentiment=x[\"label\"]\n",
    "    image_name=x[\"reply\"]\n",
    "    if image_name is None:\n",
    "        continue\n",
    "    else:\n",
    "        image_name=image_name.split(\"/\")[-1]\n",
    "    jpg_name='/home/ceyda/data/ReactionGIF/imgs/'+image_name[:-4]+\".jpg\"\n",
    "\n",
    "    # image_path=os.path.join(base_dir,image_name)\n",
    "    # extract_frame(image_path,image_name)\n",
    "    if os.path.exists(jpg_name):\n",
    "        dic={\n",
    "            \"image_path\":jpg_name,\n",
    "            \"captions\":[text,sentiment]\n",
    "        }\n",
    "        im_un.append(jpg_name)\n",
    "        if len(im_un)>=19000:\n",
    "            writer=val_writer\n",
    "        writer.write(dic)\n",
    "        # text,jpg_name,sentiment"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "source": [
    "len(im_un)"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "19387"
      ]
     },
     "metadata": {},
     "execution_count": 13
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "source": [
    "len(set(im_un))"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "19385"
      ]
     },
     "metadata": {},
     "execution_count": 14
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "source": [
    "import jax"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "source": [
    "jax.device_count()"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "metadata": {},
     "execution_count": 18
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "source": [
    "import shutil\n",
    "from pathlib import Path\n",
    "val=jsonlines.open('./val.json')\n",
    "for v in val:\n",
    "\n",
    "    shutil.copy(v['image_path'],\"/home/ceyda/code/clip-reply-demo/imgs/\" +Path(v['image_path']).name)\n"
   ],
   "outputs": [],
   "metadata": {}
  }
 ],
 "metadata": {
  "orig_nbformat": 4,
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}