space code

app.py

@@ -0,0 +1,60 @@
+import os
+import torch
+import clip
+import transformers
+import numpy as np
+import gradio as gr
+from PIL import Image
+from multilingual_clip import pt_multilingual_clip
+from torch.utils.data import DataLoader
+from datasets import load_dataset
+from usearch.index import Index
+
+dataset = load_dataset("dmayboroda/sk-test_1")
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+clipmodel, preprocess = clip.load("ViT-L/14", device=device)
+
+model_name = 'M-CLIP/LABSE-Vit-L-14'
+model = pt_multilingual_clip.MultilingualCLIP.from_pretrained(model_name)
+tokenizer = transformers.AutoTokenizer.from_pretrained(model_name)
+model.to(device)
+
+index = Index(ndim=768, metric='cos', dtype='f32')
+
+img_embeddings = []
+emb_to_img = {}
+print('Encoding images...')
+for img in dataset['train']:
+    image = preprocess(img['image']).unsqueeze(0).to(device)
+    with torch.no_grad():
+      image_features = clipmodel.encode_image(image)
+      img_embeddings.append(image_features)
+      emb_to_img[image_features] = img['image']
+
+for i in range(0, len(img_embeddings)):
+  index.add(i, img_embeddings[i].squeeze(0).cpu().detach().numpy())
+
+def get_similar(text, num_sim):
+    tokens = clip.tokenize(text).to(device)
+    text_features = clipmodel.encode_text(tokens)
+    search = text_features.squeeze(0).cpu().detach().numpy()
+    matches = index.search(search, num_sim)
+    similar = []
+    for match in matches:
+        key = match.key.item()
+        emb = img_embeddings[key]
+        similar.append(emb_to_img[emb])
+    return similar
+
+iface = gr.Interface(
+    fn=get_similar,
+    inputs=[
+        gr.inputs.Textbox(lines=2, placeholder="Enter Text Here..."),
+        gr.inputs.Number(label="Number of Images", default=15)
+    ],
+    outputs="image",
+    title="Model Testing"
+)
+
+iface.launch()

requirements.txt

@@ -0,0 +1,9 @@
+torch
+usearch
+multilingual-clip
+git+https://github.com/openai/CLIP.git
+Pillow
+transformers
+numpy
+datasets
+gradio