Update README.md

README.md

@@ -53,16 +53,103 @@ Model Performance on Visual Dialog Benchmark
 ```python
 import torch
 from PIL import Image
-from clip_encoder import CLIPVisionTower
 from transformers import AutoTokenizer, AutoModelForCausalLM
-
+from urllib.request import urlopen
+import torch.nn as nn
+from transformers import CLIPVisionModel, CLIPImageProcessor, CLIPVisionConfig
 
 DEVICE = "cuda:0"
 PROMPT = "This is a dialog with AI assistant.\n"
 tokenizer = AutoTokenizer.from_pretrained("OmniMistral-tokenizer", use_fast=False)
 model = AutoModelForCausalLM.from_pretrained("OmniMistral-model", torch_dtype=torch.bfloat16, device_map=DEVICE)
 
-clip = CLIPVisionTower("openai/clip-vit-large-patch14-336")
+projection = torch.load("projection", map_location=DEVICE)
+special_embs = torch.load("special_embeddings.pt", map_location=DEVICE)
+
+
+
+
+
+class CLIPVisionTower(nn.Module):
+    def __init__(self, vision_tower, args, delay_load=False):
+        super().__init__()
+
+        self.is_loaded = False
+
+        self.vision_tower_name = vision_tower
+        self.select_layer = args.mm_vision_select_layer
+        self.select_feature = getattr(args, 'mm_vision_select_feature', 'patch')
+
+        if not delay_load:
+            self.load_model()
+        else:
+            self.cfg_only = CLIPVisionConfig.from_pretrained(self.vision_tower_name)
+
+    def load_model(self):
+        self.image_processor = CLIPImageProcessor.from_pretrained(self.vision_tower_name)
+        self.vision_tower = CLIPVisionModel.from_pretrained(self.vision_tower_name)
+        self.vision_tower.requires_grad_(False)
+
+        self.is_loaded = True
+
+    def feature_select(self, image_forward_outs):
+        image_features = image_forward_outs.hidden_states[self.select_layer]
+        if self.select_feature == 'patch':
+            image_features = image_features[:, 1:]
+        elif self.select_feature == 'cls_patch':
+            image_features = image_features
+        else:
+            raise ValueError(f'Unexpected select feature: {self.select_feature}')
+        return image_features
+
+    @torch.no_grad()
+    def forward(self, images):
+        if type(images) is list:
+            image_features = []
+            for image in images:
+                image_forward_out = self.vision_tower(image.to(device=self.device, dtype=self.dtype).unsqueeze(0), output_hidden_states=True)
+                image_feature = self.feature_select(image_forward_out).to(image.dtype)
+                image_features.append(image_feature)
+        else:
+            image_forward_outs = self.vision_tower(images.to(device=self.device, dtype=self.dtype), output_hidden_states=True)
+            image_features = self.feature_select(image_forward_outs).to(images.dtype)
+
+        return image_features
+
+    @property
+    def dummy_feature(self):
+        return torch.zeros(1, self.hidden_size, device=self.device, dtype=self.dtype)
+
+    @property
+    def dtype(self):
+        return self.vision_tower.dtype
+
+    @property
+    def device(self):
+        return self.vision_tower.device
+
+    @property
+    def config(self):
+        if self.is_loaded:
+            return self.vision_tower.config
+        else:
+            return self.cfg_only
+
+    @property
+    def hidden_size(self):
+        return self.config.hidden_size
+
+    @property
+    def num_patches(self):
+        return (self.config.image_size // self.config.patch_size) ** 2
+
+
+class ClipTowerCfg:
+    def __init__(self):
+        self.mm_vision_select_feature = 'patch'
+        self.mm_vision_select_layer =  -2
+
+clip = CLIPVisionTower("openai/clip-vit-large-patch14-336", ClipTowerCfg())
 clip.load_model()
 clip = clip.to(device=DEVICE, dtype=torch.bfloat16)
 
@@ -84,7 +171,7 @@ def gen_answer(model, tokenizer, clip, projection, query, special_embs, image=No
         "num_return_sequences": 1,
     }
     with torch.no_grad():
-        image_features = clip.module.image_processor(image, return_tensors='pt')
+        image_features = clip.image_processor(image, return_tensors='pt')
         image_embedding = clip(image_features['pixel_values']).to(device=DEVICE, dtype=torch.bfloat16)
         
         projected_vision_embeddings = projection(image_embedding).to(device=DEVICE, dtype=torch.bfloat16)
@@ -111,17 +198,22 @@ def gen_answer(model, tokenizer, clip, projection, query, special_embs, image=No
     generated_texts = tokenizer.batch_decode(out)[0]
     return generated_texts
 
+img_url = "https://i.pinimg.com/originals/32/c7/81/32c78115cb47fd4825e6907a83b7afff.jpg"
+question = "who is the author?"
+img = Image.open(urlopen(img_url))
 
 answer = gen_answer(
     model,
     tokenizer, 
     clip, 
     projection, 
-    query = "who is the author?", 
-    special_embs, 
-    Image.open("https://i.pinimg.com/originals/32/c7/81/32c78115cb47fd4825e6907a83b7afff.jpg")
+    query=question, 
+    special_embs=special_embs, 
+    image=img
 )
 
+img.show()
+print(question)
 print(answer)
 ```