Update app.py

app.py

@@ -1,89 +1,147 @@
 import gradio as gr
 from transformers import AutoProcessor, AutoModelForImageTextToText, TextIteratorStreamer
 from threading import Thread
-import torch
-import spaces
 import re
 import time
+import torch
+import spaces
+import subprocess
+subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
 
-# Utilisation de la version ultra-légère pour garantir la stabilité
-MODEL_ID = "HuggingFaceTB/SmolVLM2-256M-Instruct"
+from io import BytesIO
 
-# Chargement du processeur et du modèle
-processor = AutoProcessor.from_pretrained(MODEL_ID)
-model = AutoModelForImageTextToText.from_pretrained(
-    MODEL_ID,
-    torch_dtype=torch.bfloat16 if torch.cuda.is_available() else torch.float32,
-    device_map="auto"
-)
+processor = AutoProcessor.from_pretrained("HuggingFaceTB/SmolVLM2-2.2B-Instruct")
+model = AutoModelForImageTextToText.from_pretrained("HuggingFaceTB/SmolVLM2-2.2B-Instruct", 
+                                               _attn_implementation="flash_attention_2",
+                                               torch_dtype=torch.bfloat16).to("cuda:0")
 
-@spaces.GPU(duration=60)
-def model_inference(input_dict, history, max_tokens):
+
+@spaces.GPU
+def model_inference(
+    input_dict, history, max_tokens
+): 
     text = input_dict["text"]
-    files = input_dict.get("files", [])
-    
-    # Construction du contenu multimodal
+    images = []
     user_content = []
-    
-    # On ajoute d'abord les médias (images/vidéos)
-    for file in files:
-        if file.lower().endswith((".png", ".jpg", ".jpeg", ".gif", ".bmp")):
-            user_content.append({"type": "image", "path": file})
-        elif file.lower().endswith((".mp4", ".mov", ".avi", ".mkv")):
-            user_content.append({"type": "video", "path": file})
-    
-    # On ajoute le texte
-    if text.strip():
-        user_content.append({"type": "text", "text": text})
-    
-    if not user_content:
-        yield "Veuillez uploader une vidéo de chat ou poser une question."
-        return
+    media_queue = []
+    if history == []:
+        text = input_dict["text"].strip() 
+        
+        for file in input_dict.get("files", []):
+            if file.endswith((".png", ".jpg", ".jpeg", ".gif", ".bmp")):
+                media_queue.append({"type": "image", "path": file})
+            elif file.endswith((".mp4", ".mov", ".avi", ".mkv", ".flv")):
+                media_queue.append({"type": "video", "path": file})
 
-    # Structure du message pour SmolVLM2
-    messages = [{"role": "user", "content": user_content}]
-    
-    # Préparation des inputs via le template
+        if "<image>" in text or "<video>" in text:
+            parts = re.split(r'(<image>|<video>)', text)  
+            for part in parts:
+                if part == "<image>" and media_queue:
+                    user_content.append(media_queue.pop(0)) 
+                elif part == "<video>" and media_queue:
+                    user_content.append(media_queue.pop(0))  
+                elif part.strip():  
+                    user_content.append({"type": "text", "text": part.strip()})
+        else:
+            user_content.append({"type": "text", "text": text})
+            
+            for media in media_queue:
+                user_content.append(media)
+
+        resulting_messages = [{"role": "user", "content": user_content}]
+
+    elif len(history) > 0:
+        resulting_messages = []
+        user_content = []
+        media_queue = []
+        for hist in history:
+            if hist["role"] == "user" and isinstance(hist["content"], tuple): 
+                file_name = hist["content"][0]
+            if file_name.endswith((".png", ".jpg", ".jpeg")):
+                media_queue.append({"type": "image", "path": file_name})
+            elif file_name.endswith(".mp4"):
+                media_queue.append({"type": "video", "path": file_name})
+
+
+        for hist in history:
+            if hist["role"] == "user" and isinstance(hist["content"], str): 
+                text = hist["content"]
+                parts = re.split(r'(<image>|<video>)', text)  
+                
+                for part in parts:
+                    if part == "<image>" and media_queue:
+                        user_content.append(media_queue.pop(0)) 
+                    elif part == "<video>" and media_queue:
+                        user_content.append(media_queue.pop(0))  
+                    elif part.strip(): 
+                        user_content.append({"type": "text", "text": part.strip()})
+            
+            elif hist["role"] == "assistant":
+                resulting_messages.append({
+                    "role": "user",
+                    "content": user_content
+                })
+                resulting_messages.append({
+                    "role": "assistant",
+                    "content": [{"type": "text", "text": hist["content"]}]
+                })
+                user_content = [] 
+
+
+    if text == "" and not images:
+        gr.Error("Please input a query and optionally image(s).")
+
+    if text == "" and images:
+        gr.Error("Please input a text query along the images(s).")
+    print("resulting_messages", resulting_messages)
     inputs = processor.apply_chat_template(
-        messages,
-        add_generation_prompt=True,
-        tokenize=True,
-        return_dict=True,
-        return_tensors="pt"
-    ).to(model.device)
-
-    # Configuration du streamer pour l'affichage en direct
+    resulting_messages,
+    add_generation_prompt=True,
+    tokenize=True,
+    return_dict=True,
+    return_tensors="pt",
+    )
+
+    inputs = inputs.to(model.device)
+    
+
+    # Generate
     streamer = TextIteratorStreamer(processor, skip_prompt=True, skip_special_tokens=True)
     generation_args = dict(inputs, streamer=streamer, max_new_tokens=max_tokens)
-    
-    # Lancement de la génération dans un thread séparé
+    generated_text = ""
+
     thread = Thread(target=model.generate, kwargs=generation_args)
     thread.start()
-    
+
+    yield "..."
     buffer = ""
+    
+      
     for new_text in streamer:
-        buffer += new_text
-        yield buffer
-
-# Création de l'interface Chat
-demo = gr.ChatInterface(
-    fn=model_inference,
-    title="SmolVLM2: The Smollest Video Model Ever 📺",
-    description="Analyse de vidéos et d'images avec SmolVLM2-256M.",
-    textbox=gr.MultimodalTextbox(
-        label="Query Input", 
-        file_types=["image", ".mp4"], 
-        file_count="multiple"
-    ), 
-    stop_btn="Stop Generation",
-    # On SUPPRIME 'multimodal=True' ici, car gr.MultimodalTextbox l'est par défaut
-    cache_examples=False,
-    additional_inputs=[
-        gr.Slider(minimum=100, maximum=500, step=50, value=200, label="Max Tokens")
-    ],
-    type="messages"
-)
-
-if __name__ == "__main__":
-    # ss=False pour éviter les bugs de schéma JSON vus précédemment
-    demo.launch(debug=True)
+    
+      buffer += new_text
+      generated_text_without_prompt = buffer#[len(ext_buffer):]
+      time.sleep(0.01)
+      yield buffer
+
+
+examples=[
+              [{"text": "Where do the severe droughts happen according to this diagram?", "files": ["example_images/examples_weather_events.png"]}],
+              [{"text": "What art era this artpiece <image> and this artpiece <image> belong to?", "files": ["example_images/rococo.jpg", "example_images/rococo_1.jpg"]}],
+              [{"text": "Describe this image.", "files": ["example_images/mosque.jpg"]}],
+              [{"text": "When was this purchase made and how much did it cost?", "files": ["example_images/fiche.jpg"]}],
+              [{"text": "What is the date in this document?", "files": ["example_images/document.jpg"]}],
+              [{"text": "What is happening in the video?", "files": ["example_images/short.mp4"]}],
+      ]
+demo = gr.ChatInterface(fn=model_inference, title="SmolVLM2: The Smollest Video Model Ever 📺", 
+                description="Play with [SmolVLM2-2.2B-Instruct](https://huggingface.co/HuggingFaceTB/SmolVLM2-2.2B-Instruct) in this demo. To get started, upload an image and text or try one of the examples. This demo doesn't use history for the chat, so every chat you start is a new conversation.",
+                examples=examples,
+                textbox=gr.MultimodalTextbox(label="Query Input", file_types=["image", ".mp4"], file_count="multiple"), stop_btn="Stop Generation", multimodal=True,
+                cache_examples=False,
+                additional_inputs=[gr.Slider(minimum=100, maximum=500, step=50, value=200, label="Max Tokens")],
+                type="messages"
+                )
+      
+      
+
+demo.launch(debug=True)