added minimum example

mwe.ipynb

@@ -0,0 +1,208 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "6942ccac",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'cuda'"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "from transformers import CLIPModel, CLIPVisionModel, CLIPProcessor\n",
+    "from transformers import logging\n",
+    "# Supress some unnecessary warnings when loading the CLIPTextModel\n",
+    "logging.set_verbosity_error()\n",
+    "\n",
+    "from diffusers import AutoencoderKL, UNet2DConditionModel, LMSDiscreteScheduler\n",
+    "from tqdm.auto import tqdm\n",
+    "from PIL import Image\n",
+    "from matplotlib import pyplot as plt\n",
+    "import numpy as np\n",
+    "\n",
+    "from torchvision import transforms as tfms\n",
+    "import requests\n",
+    "\n",
+    "\n",
+    "torch_device = \"cuda\" if torch.cuda.is_available() else \"cpu\"; torch_device"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "6591cd09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "\n",
+    "model = CLIPModel.from_pretrained(\"openai/clip-vit-large-patch14\")\n",
+    "processor = CLIPProcessor.from_pretrained(\"openai/clip-vit-large-patch14\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "0a701777",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "url = \"http://images.cocodataset.org/val2017/000000039769.jpg\"\n",
+    "image = Image.open(requests.get(url, stream=True).raw)\n",
+    "inputs = processor(text=[\"a photo of two cats sleeping in a pink sofa\"], images=image, return_tensors=\"pt\", padding=True)\n",
+    "inputs"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "e148125e",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "BaseModelOutputWithPooling(last_hidden_state=tensor([[[-0.5297, -0.7713,  0.4655,  ..., -0.3993, -0.0721, -0.3703],\n",
+       "         [ 0.8688,  0.1690,  0.6678,  ...,  0.5126, -1.1465, -0.1258],\n",
+       "         [ 1.1742, -0.7551,  0.0396,  ...,  0.7166, -0.5458,  0.0031],\n",
+       "         ...,\n",
+       "         [ 0.8636,  0.2223,  0.6411,  ...,  0.5242, -0.8104,  0.0170],\n",
+       "         [ 0.6842, -1.1056, -0.2486,  ...,  0.7901,  0.4862, -0.0949],\n",
+       "         [ 0.8934,  0.0066,  0.9235,  ...,  0.5707, -0.8436, -0.2182]]]), pooler_output=tensor([[-0.9326, -1.3289,  0.7919,  ..., -0.3337, -0.0479, -0.7106]]), hidden_states=None, attentions=None)"
+      ]
+     },
+     "execution_count": 19,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "with torch.no_grad():\n",
+    "    img_emb = model.vision_model(inputs.pixel_values)[0]\n",
+    "    txt_emb = model.text_model(inputs.input_ids)[0]\n",
+    "img_emb.shape, txt_emb.shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "f28bb4b6",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "CLIPVisionConfig {\n",
+       "  \"attention_dropout\": 0.0,\n",
+       "  \"dropout\": 0.0,\n",
+       "  \"hidden_act\": \"quick_gelu\",\n",
+       "  \"hidden_size\": 1024,\n",
+       "  \"image_size\": 224,\n",
+       "  \"initializer_factor\": 1.0,\n",
+       "  \"initializer_range\": 0.02,\n",
+       "  \"intermediate_size\": 4096,\n",
+       "  \"layer_norm_eps\": 1e-05,\n",
+       "  \"model_type\": \"clip_vision_model\",\n",
+       "  \"num_attention_heads\": 16,\n",
+       "  \"num_channels\": 3,\n",
+       "  \"num_hidden_layers\": 24,\n",
+       "  \"patch_size\": 14,\n",
+       "  \"projection_dim\": 768,\n",
+       "  \"transformers_version\": \"4.23.1\"\n",
+       "}"
+      ]
+     },
+     "execution_count": 22,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "model.vision_model.config"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "6726b263",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "CLIPTextConfig {\n",
+       "  \"attention_dropout\": 0.0,\n",
+       "  \"bos_token_id\": 0,\n",
+       "  \"dropout\": 0.0,\n",
+       "  \"eos_token_id\": 2,\n",
+       "  \"hidden_act\": \"quick_gelu\",\n",
+       "  \"hidden_size\": 768,\n",
+       "  \"initializer_factor\": 1.0,\n",
+       "  \"initializer_range\": 0.02,\n",
+       "  \"intermediate_size\": 3072,\n",
+       "  \"layer_norm_eps\": 1e-05,\n",
+       "  \"max_position_embeddings\": 77,\n",
+       "  \"model_type\": \"clip_text_model\",\n",
+       "  \"num_attention_heads\": 12,\n",
+       "  \"num_hidden_layers\": 12,\n",
+       "  \"pad_token_id\": 1,\n",
+       "  \"projection_dim\": 768,\n",
+       "  \"transformers_version\": \"4.23.1\",\n",
+       "  \"vocab_size\": 49408\n",
+       "}"
+      ]
+     },
+     "execution_count": 23,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "model.text_model.config"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d000675d",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3.9.13 ('py39')",
+   "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": "3.9.13"
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "8b806adfb64333d0ca5c14ed2dbf613d5d551ec856d702e8a01588c05fb48e2e"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}