Streaming output

app.py

@@ -1,34 +1,81 @@
 import gradio as gr
 import spaces
-from transformers import Qwen2VLForConditionalGeneration, AutoProcessor
+from transformers import Qwen2VLForConditionalGeneration, AutoProcessor, TextIteratorStreamer
 from qwen_vl_utils import process_vision_info
 import torch
 from PIL import Image
 import subprocess
 import numpy as np
 import os
-
-# Install flash-attn
-subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
+from threading import Thread
+import uuid
+import io
 
 # Model and Processor Loading (Done once at startup)
 MODEL_ID = "Qwen/Qwen2-VL-7B-Instruct"
-model = Qwen2VLForConditionalGeneration.from_pretrained(MODEL_ID, trust_remote_code=True, torch_dtype=torch.float16).to("cuda").eval()
+model = Qwen2VLForConditionalGeneration.from_pretrained(
+    MODEL_ID,
+    trust_remote_code=True,
+    torch_dtype=torch.float16
+).to("cuda").eval()
 processor = AutoProcessor.from_pretrained(MODEL_ID, trust_remote_code=True)
 
 DESCRIPTION = "[Qwen2-VL-7B Demo](https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct)"
 
+image_extensions = Image.registered_extensions()
+video_extensions = ("avi", "mp4", "mov", "mkv", "flv", "wmv", "mjpeg", "wav", "gif", "webm", "m4v", "3gp")
+
+
+def identify_and_save_blob(blob_path):
+    """Identifies if the blob is an image or video and saves it accordingly."""
+    try:
+        with open(blob_path, 'rb') as file:
+            blob_content = file.read()
+            
+            # Try to identify if it's an image
+            try:
+                Image.open(io.BytesIO(blob_content)).verify()  # Check if it's a valid image
+                extension = ".png"  # Default to PNG for saving
+                media_type = "image"
+            except (IOError, SyntaxError):
+                # If it's not a valid image, assume it's a video
+                extension = ".mp4"  # Default to MP4 for saving
+                media_type = "video"
+            
+            # Create a unique filename
+            filename = f"temp_{uuid.uuid4()}_media{extension}"
+            with open(filename, "wb") as f:
+                f.write(blob_content)
+                
+            return filename, media_type
+            
+    except FileNotFoundError:
+        raise ValueError(f"The file {blob_path} was not found.")
+    except Exception as e:
+        raise ValueError(f"An error occurred while processing the file: {e}")
+
+
 @spaces.GPU
-def qwen_inference(media_path, text_input=None):
-
-    image_extensions = Image.registered_extensions()
-    if media_path.endswith(tuple([i for i, f in image_extensions.items()])):
-        media_type = "image"
-    elif media_path.endswith(("avi", "mp4", "mov", "mkv", "flv", "wmv", "mjpeg", "wav", "gif", "webm", "m4v", "3gp")):  # Check if it's a video path
-        media_type = "video"
-    else:
-        raise ValueError("Unsupported media type. Please upload an image or video.")
-    
+def qwen_inference(media_input, text_input=None):
+    if isinstance(media_input, str):  # If it's a filepath
+        media_path = media_input
+        if media_path.endswith(tuple([i for i, f in image_extensions.items()])):
+            media_type = "image"
+        elif media_path.endswith(video_extensions): 
+            media_type = "video"
+        else:
+            try:
+                media_path, media_type = identify_and_save_blob(media_input)
+                print(media_path, media_type)
+            except Exception as e:
+                print(e)
+                raise ValueError(
+                    "Unsupported media type. Please upload an image or video."
+                )
+        
+
+    print(media_path)
+
     messages = [
         {
             "role": "user",
@@ -36,15 +83,17 @@ def qwen_inference(media_path, text_input=None):
                 {
                     "type": media_type,
                     media_type: media_path,
-                    **({"fps": 8.0} if media_type == "video" else {}), 
+                    **({"fps": 8.0} if media_type == "video" else {}),
                 },
                 {"type": "text", "text": text_input},
             ],
         }
     ]
 
-    text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
-    image_inputs, video_inputs = process_vision_info(messages) 
+    text = processor.apply_chat_template(
+        messages, tokenize=False, add_generation_prompt=True
+    )
+    image_inputs, video_inputs = process_vision_info(messages)
     inputs = processor(
         text=[text],
         images=image_inputs,
@@ -53,11 +102,18 @@ def qwen_inference(media_path, text_input=None):
         return_tensors="pt",
     ).to("cuda")
 
-    generated_ids = model.generate(**inputs, max_new_tokens=1024)
-    generated_ids_trimmed = [out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)]
-    output_text = processor.batch_decode(generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
-    
-    return output_text
+    streamer = TextIteratorStreamer(
+        processor, skip_prompt=True, **{"skip_special_tokens": True}
+    )
+    generation_kwargs = dict(inputs, streamer=streamer, max_new_tokens=1024)
+
+    thread = Thread(target=model.generate, kwargs=generation_kwargs)
+    thread.start()
+
+    buffer = ""
+    for new_text in streamer:
+        buffer += new_text
+        yield buffer
 
 css = """
   #output {
@@ -73,12 +129,16 @@ with gr.Blocks(css=css) as demo:
     with gr.Tab(label="Image/Video Input"):
         with gr.Row():
             with gr.Column():
-                input_media = gr.File(label="Upload Image or Video", type="filepath") 
+                input_media = gr.File(
+                    label="Upload Image or Video", type="filepath" 
+                )
                 text_input = gr.Textbox(label="Question")
                 submit_btn = gr.Button(value="Submit")
             with gr.Column():
                 output_text = gr.Textbox(label="Output Text")
 
-        submit_btn.click(qwen_inference, [input_media, text_input], [output_text])
+        submit_btn.click(
+            qwen_inference, [input_media, text_input], [output_text]
+        )
 
 demo.launch(debug=True)