Spaces:
Running
on
Zero
Running
on
Zero
#!/usr/bin/env python | |
import os | |
import re | |
import tempfile | |
from collections.abc import Iterator | |
from threading import Thread | |
import cv2 | |
import gradio as gr | |
import spaces | |
import torch | |
from loguru import logger | |
from PIL import Image | |
from transformers import AutoProcessor, Gemma3ForConditionalGeneration, TextIteratorStreamer | |
# CSV/TXT ๋ถ์ | |
import pandas as pd | |
# PDF ํ ์คํธ ์ถ์ถ | |
import PyPDF2 | |
MAX_CONTENT_CHARS = 4000 # ๋๋ฌด ํฐ ํ์ผ์ ๋ง๊ธฐ ์ํด ์ต๋ ํ์ 4000์ | |
model_id = os.getenv("MODEL_ID", "google/gemma-3-27b-it") | |
processor = AutoProcessor.from_pretrained(model_id, padding_side="left") | |
model = Gemma3ForConditionalGeneration.from_pretrained( | |
model_id, | |
device_map="auto", | |
torch_dtype=torch.bfloat16, | |
attn_implementation="eager" | |
) | |
MAX_NUM_IMAGES = int(os.getenv("MAX_NUM_IMAGES", "5")) | |
################################################## | |
# CSV, TXT, PDF ๋ถ์ ํจ์ | |
################################################## | |
def analyze_csv_file(path: str) -> str: | |
""" | |
CSV ํ์ผ์ ์ ์ฒด ๋ฌธ์์ด๋ก ๋ณํ. ๋๋ฌด ๊ธธ ๊ฒฝ์ฐ ์ผ๋ถ๋ง ํ์. | |
""" | |
try: | |
df = pd.read_csv(path) | |
# ๋ฐ์ดํฐ ํ๋ ์ ํฌ๊ธฐ ์ ํ (ํ/์ด ์๊ฐ ๋ง์ ๊ฒฝ์ฐ) | |
if df.shape[0] > 50 or df.shape[1] > 10: | |
df = df.iloc[:50, :10] | |
df_str = df.to_string() | |
if len(df_str) > MAX_CONTENT_CHARS: | |
df_str = df_str[:MAX_CONTENT_CHARS] + "\n...(truncated)..." | |
return f"**[CSV File: {os.path.basename(path)}]**\n\n{df_str}" | |
except Exception as e: | |
return f"Failed to read CSV ({os.path.basename(path)}): {str(e)}" | |
def analyze_txt_file(path: str) -> str: | |
""" | |
TXT ํ์ผ ์ ๋ฌธ ์ฝ๊ธฐ. ๋๋ฌด ๊ธธ๋ฉด ์ผ๋ถ๋ง ํ์. | |
""" | |
try: | |
with open(path, "r", encoding="utf-8") as f: | |
text = f.read() | |
if len(text) > MAX_CONTENT_CHARS: | |
text = text[:MAX_CONTENT_CHARS] + "\n...(truncated)..." | |
return f"**[TXT File: {os.path.basename(path)}]**\n\n{text}" | |
except Exception as e: | |
return f"Failed to read TXT ({os.path.basename(path)}): {str(e)}" | |
def pdf_to_markdown(pdf_path: str) -> str: | |
""" | |
PDF โ Markdown. ํ์ด์ง๋ณ๋ก ๊ฐ๋จํ ํ ์คํธ ์ถ์ถ. | |
""" | |
text_chunks = [] | |
try: | |
with open(pdf_path, "rb") as f: | |
reader = PyPDF2.PdfReader(f) | |
# ์ต๋ 5ํ์ด์ง๋ง ์ฒ๋ฆฌ | |
max_pages = min(5, len(reader.pages)) | |
for page_num in range(max_pages): | |
page = reader.pages[page_num] | |
page_text = page.extract_text() or "" | |
page_text = page_text.strip() | |
if page_text: | |
# ํ์ด์ง๋ณ ํ ์คํธ๋ ์ ํ | |
if len(page_text) > MAX_CONTENT_CHARS // max_pages: | |
page_text = page_text[:MAX_CONTENT_CHARS // max_pages] + "...(truncated)" | |
text_chunks.append(f"## Page {page_num+1}\n\n{page_text}\n") | |
if len(reader.pages) > max_pages: | |
text_chunks.append(f"\n...(Showing {max_pages} of {len(reader.pages)} pages)...") | |
except Exception as e: | |
return f"Failed to read PDF ({os.path.basename(pdf_path)}): {str(e)}" | |
full_text = "\n".join(text_chunks) | |
if len(full_text) > MAX_CONTENT_CHARS: | |
full_text = full_text[:MAX_CONTENT_CHARS] + "\n...(truncated)..." | |
return f"**[PDF File: {os.path.basename(pdf_path)}]**\n\n{full_text}" | |
################################################## | |
# ์ด๋ฏธ์ง/๋น๋์ค ์ ๋ก๋ ์ ํ ๊ฒ์ฌ | |
################################################## | |
def count_files_in_new_message(paths: list[str]) -> tuple[int, int]: | |
image_count = 0 | |
video_count = 0 | |
for path in paths: | |
if path.endswith(".mp4"): | |
video_count += 1 | |
elif re.search(r"\.(png|jpg|jpeg|gif|webp)$", path, re.IGNORECASE): | |
image_count += 1 | |
return image_count, video_count | |
def count_files_in_history(history: list[dict]) -> tuple[int, int]: | |
image_count = 0 | |
video_count = 0 | |
for item in history: | |
if item["role"] != "user" or isinstance(item["content"], str): | |
continue | |
if isinstance(item["content"], list) and len(item["content"]) > 0: | |
file_path = item["content"][0] | |
if isinstance(file_path, str): | |
if file_path.endswith(".mp4"): | |
video_count += 1 | |
elif re.search(r"\.(png|jpg|jpeg|gif|webp)$", file_path, re.IGNORECASE): | |
image_count += 1 | |
return image_count, video_count | |
def validate_media_constraints(message: dict, history: list[dict]) -> bool: | |
""" | |
- ๋น๋์ค 1๊ฐ ์ด๊ณผ ๋ถ๊ฐ | |
- ๋น๋์ค์ ์ด๋ฏธ์ง ํผํฉ ๋ถ๊ฐ | |
- ์ด๋ฏธ์ง ๊ฐ์ MAX_NUM_IMAGES ์ด๊ณผ ๋ถ๊ฐ | |
- <image> ํ๊ทธ๊ฐ ์์ผ๋ฉด ํ๊ทธ ์์ ์ค์ ์ด๋ฏธ์ง ์ ์ผ์น | |
- CSV, TXT, PDF ๋ฑ์ ์ฌ๊ธฐ์ ์ ํํ์ง ์์ | |
""" | |
# ์ด๋ฏธ์ง์ ๋น๋์ค ํ์ผ๋ง ํํฐ๋ง | |
media_files = [] | |
for f in message["files"]: | |
if re.search(r"\.(png|jpg|jpeg|gif|webp)$", f, re.IGNORECASE) or f.endswith(".mp4"): | |
media_files.append(f) | |
new_image_count, new_video_count = count_files_in_new_message(media_files) | |
history_image_count, history_video_count = count_files_in_history(history) | |
image_count = history_image_count + new_image_count | |
video_count = history_video_count + new_video_count | |
if video_count > 1: | |
gr.Warning("Only one video is supported.") | |
return False | |
if video_count == 1: | |
if image_count > 0: | |
gr.Warning("Mixing images and videos is not allowed.") | |
return False | |
if "<image>" in message["text"]: | |
gr.Warning("Using <image> tags with video files is not supported.") | |
return False | |
if video_count == 0 and image_count > MAX_NUM_IMAGES: | |
gr.Warning(f"You can upload up to {MAX_NUM_IMAGES} images.") | |
return False | |
# ์ด๋ฏธ์ง ํ๊ทธ ๊ฒ์ฆ (์ค์ ์ด๋ฏธ์ง ํ์ผ๋ง ๊ณ์ฐ) | |
if "<image>" in message["text"]: | |
# ์ด๋ฏธ์ง ํ์ผ๋ง ํํฐ๋ง | |
image_files = [f for f in message["files"] if re.search(r"\.(png|jpg|jpeg|gif|webp)$", f, re.IGNORECASE)] | |
image_tag_count = message["text"].count("<image>") | |
if image_tag_count != len(image_files): | |
gr.Warning("The number of <image> tags in the text does not match the number of image files.") | |
return False | |
return True | |
################################################## | |
# ๋น๋์ค ์ฒ๋ฆฌ | |
################################################## | |
def downsample_video(video_path: str) -> list[tuple[Image.Image, float]]: | |
vidcap = cv2.VideoCapture(video_path) | |
fps = vidcap.get(cv2.CAP_PROP_FPS) | |
total_frames = int(vidcap.get(cv2.CAP_PROP_FRAME_COUNT)) | |
# ๋ ์ ์ ํ๋ ์์ ์ถ์ถํ๋๋ก ์กฐ์ | |
frame_interval = max(int(fps), int(total_frames / 10)) # ์ด๋น 1ํ๋ ์ ๋๋ ์ต๋ 10ํ๋ ์ | |
frames = [] | |
for i in range(0, total_frames, frame_interval): | |
vidcap.set(cv2.CAP_PROP_POS_FRAMES, i) | |
success, image = vidcap.read() | |
if success: | |
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) | |
pil_image = Image.fromarray(image) | |
timestamp = round(i / fps, 2) | |
frames.append((pil_image, timestamp)) | |
# ์ต๋ 5ํ๋ ์๋ง ์ฌ์ฉ | |
if len(frames) >= 5: | |
break | |
vidcap.release() | |
return frames | |
def process_video(video_path: str) -> list[dict]: | |
content = [] | |
frames = downsample_video(video_path) | |
for frame in frames: | |
pil_image, timestamp = frame | |
with tempfile.NamedTemporaryFile(delete=False, suffix=".png") as temp_file: | |
pil_image.save(temp_file.name) | |
content.append({"type": "text", "text": f"Frame {timestamp}:"}) | |
content.append({"type": "image", "url": temp_file.name}) | |
logger.debug(f"{content=}") | |
return content | |
################################################## | |
# interleaved <image> ์ฒ๋ฆฌ | |
################################################## | |
def process_interleaved_images(message: dict) -> list[dict]: | |
parts = re.split(r"(<image>)", message["text"]) | |
content = [] | |
image_index = 0 | |
# ์ด๋ฏธ์ง ํ์ผ๋ง ํํฐ๋ง | |
image_files = [f for f in message["files"] if re.search(r"\.(png|jpg|jpeg|gif|webp)$", f, re.IGNORECASE)] | |
for part in parts: | |
if part == "<image>" and image_index < len(image_files): | |
content.append({"type": "image", "url": image_files[image_index]}) | |
image_index += 1 | |
elif part.strip(): | |
content.append({"type": "text", "text": part.strip()}) | |
else: | |
# ๊ณต๋ฐฑ์ด๊ฑฐ๋ \n ๊ฐ์ ๊ฒฝ์ฐ | |
if isinstance(part, str) and part != "<image>": | |
content.append({"type": "text", "text": part}) | |
return content | |
################################################## | |
# PDF + CSV + TXT + ์ด๋ฏธ์ง/๋น๋์ค | |
################################################## | |
def is_image_file(file_path: str) -> bool: | |
"""์ด๋ฏธ์ง ํ์ผ์ธ์ง ํ์ธ""" | |
return bool(re.search(r"\.(png|jpg|jpeg|gif|webp)$", file_path, re.IGNORECASE)) | |
def is_video_file(file_path: str) -> bool: | |
"""๋น๋์ค ํ์ผ์ธ์ง ํ์ธ""" | |
return file_path.endswith(".mp4") | |
def is_document_file(file_path: str) -> bool: | |
"""๋ฌธ์ ํ์ผ์ธ์ง ํ์ธ (PDF, CSV, TXT)""" | |
return (file_path.lower().endswith(".pdf") or | |
file_path.lower().endswith(".csv") or | |
file_path.lower().endswith(".txt")) | |
def process_new_user_message(message: dict) -> list[dict]: | |
if not message["files"]: | |
return [{"type": "text", "text": message["text"]}] | |
# 1) ํ์ผ ๋ถ๋ฅ | |
video_files = [f for f in message["files"] if is_video_file(f)] | |
image_files = [f for f in message["files"] if is_image_file(f)] | |
csv_files = [f for f in message["files"] if f.lower().endswith(".csv")] | |
txt_files = [f for f in message["files"] if f.lower().endswith(".txt")] | |
pdf_files = [f for f in message["files"] if f.lower().endswith(".pdf")] | |
# 2) ์ฌ์ฉ์ ์๋ณธ text ์ถ๊ฐ | |
content_list = [{"type": "text", "text": message["text"]}] | |
# 3) CSV | |
for csv_path in csv_files: | |
csv_analysis = analyze_csv_file(csv_path) | |
content_list.append({"type": "text", "text": csv_analysis}) | |
# 4) TXT | |
for txt_path in txt_files: | |
txt_analysis = analyze_txt_file(txt_path) | |
content_list.append({"type": "text", "text": txt_analysis}) | |
# 5) PDF | |
for pdf_path in pdf_files: | |
pdf_markdown = pdf_to_markdown(pdf_path) | |
content_list.append({"type": "text", "text": pdf_markdown}) | |
# 6) ๋น๋์ค (ํ ๊ฐ๋ง ํ์ฉ) | |
if video_files: | |
content_list += process_video(video_files[0]) | |
return content_list | |
# 7) ์ด๋ฏธ์ง ์ฒ๋ฆฌ | |
if "<image>" in message["text"] and image_files: | |
# interleaved | |
interleaved_content = process_interleaved_images({"text": message["text"], "files": image_files}) | |
# ์๋ณธ content_list ์๋ถ๋ถ(ํ ์คํธ)์ ์ ๊ฑฐํ๊ณ interleaved๋ก ๋์ฒด | |
if content_list[0]["type"] == "text": | |
content_list = content_list[1:] # ์๋ณธ ํ ์คํธ ์ ๊ฑฐ | |
return interleaved_content + content_list # interleaved + ๋๋จธ์ง ๋ฌธ์ ๋ถ์ ๋ด์ฉ | |
else: | |
# ์ผ๋ฐ ์ฌ๋ฌ ์ฅ | |
for img_path in image_files: | |
content_list.append({"type": "image", "url": img_path}) | |
return content_list | |
################################################## | |
# history -> LLM ๋ฉ์์ง ๋ณํ | |
################################################## | |
def process_history(history: list[dict]) -> list[dict]: | |
messages = [] | |
current_user_content: list[dict] = [] | |
for item in history: | |
if item["role"] == "assistant": | |
# user_content๊ฐ ์์ฌ์๋ค๋ฉด user ๋ฉ์์ง๋ก ์ ์ฅ | |
if current_user_content: | |
messages.append({"role": "user", "content": current_user_content}) | |
current_user_content = [] | |
# ๊ทธ ๋ค item์ assistant | |
messages.append({"role": "assistant", "content": [{"type": "text", "text": item["content"]}]}) | |
else: | |
# user | |
content = item["content"] | |
if isinstance(content, str): | |
current_user_content.append({"type": "text", "text": content}) | |
elif isinstance(content, list) and len(content) > 0: | |
file_path = content[0] | |
if is_image_file(file_path): | |
current_user_content.append({"type": "image", "url": file_path}) | |
else: | |
# ๋น์ด๋ฏธ์ง ํ์ผ์ ํ ์คํธ๋ก ์ฒ๋ฆฌ | |
current_user_content.append({"type": "text", "text": f"[File: {os.path.basename(file_path)}]"}) | |
# ๋ง์ง๋ง ์ฌ์ฉ์ ๋ฉ์์ง๊ฐ ์ฒ๋ฆฌ๋์ง ์์ ๊ฒฝ์ฐ ์ถ๊ฐ | |
if current_user_content: | |
messages.append({"role": "user", "content": current_user_content}) | |
return messages | |
################################################## | |
# ๋ฉ์ธ ์ถ๋ก ํจ์ | |
################################################## | |
def run(message: dict, history: list[dict], system_prompt: str = "", max_new_tokens: int = 512) -> Iterator[str]: | |
if not validate_media_constraints(message, history): | |
yield "" | |
return | |
try: | |
messages = [] | |
if system_prompt: | |
messages.append({"role": "system", "content": [{"type": "text", "text": system_prompt}]}) | |
messages.extend(process_history(history)) | |
# ์ฌ์ฉ์ ๋ฉ์์ง ์ฒ๋ฆฌ | |
user_content = process_new_user_message(message) | |
# ํ ํฐ ์๋ฅผ ์ค์ด๊ธฐ ์ํด ๋๋ฌด ๊ธด ํ ์คํธ๋ ์๋ผ๋ด๊ธฐ | |
for item in user_content: | |
if item["type"] == "text" and len(item["text"]) > MAX_CONTENT_CHARS: | |
item["text"] = item["text"][:MAX_CONTENT_CHARS] + "\n...(truncated)..." | |
messages.append({"role": "user", "content": user_content}) | |
# ๋ชจ๋ธ ์ ๋ ฅ ์์ฑ ์ ์ต์ข ํ์ธ | |
# ์ด๋ฏธ์ง๋ ๋น๋์ค๊ฐ ์๋ ํ์ผ๋ค์ ๋ชจ๋ธ์ "image" ํ์ดํ๋ผ์ธ์ผ๋ก ์ ๋ฌ๋์ง ์๋๋ก ํํฐ๋ง | |
for msg in messages: | |
if msg["role"] != "user": | |
continue | |
filtered_content = [] | |
for item in msg["content"]: | |
if item["type"] == "image": | |
if is_image_file(item["url"]): | |
filtered_content.append(item) | |
else: | |
# ์ด๋ฏธ์ง ํ์ผ์ด ์๋ ๊ฒฝ์ฐ ํ ์คํธ๋ก ๋ณํ | |
filtered_content.append({ | |
"type": "text", | |
"text": f"[Non-image file: {os.path.basename(item['url'])}]" | |
}) | |
else: | |
filtered_content.append(item) | |
msg["content"] = filtered_content | |
# ๋ชจ๋ธ ์ ๋ ฅ ์์ฑ | |
inputs = processor.apply_chat_template( | |
messages, | |
add_generation_prompt=True, | |
tokenize=True, | |
return_dict=True, | |
return_tensors="pt", | |
).to(device=model.device, dtype=torch.bfloat16) | |
# ํ ์คํธ ์์ฑ ์คํธ๋ฆฌ๋จธ ์ค์ | |
streamer = TextIteratorStreamer(processor, timeout=30.0, skip_prompt=True, skip_special_tokens=True) | |
gen_kwargs = dict( | |
inputs, | |
streamer=streamer, | |
max_new_tokens=max_new_tokens, | |
) | |
# ๋ณ๋ ์ค๋ ๋์์ ํ ์คํธ ์์ฑ | |
t = Thread(target=model.generate, kwargs=gen_kwargs) | |
t.start() | |
# ๊ฒฐ๊ณผ ์คํธ๋ฆฌ๋ฐ | |
output = "" | |
for new_text in streamer: | |
output += new_text | |
yield output | |
except Exception as e: | |
logger.error(f"Error in run: {str(e)}") | |
yield f"์ฃ์กํฉ๋๋ค. ์ค๋ฅ๊ฐ ๋ฐ์ํ์ต๋๋ค: {str(e)}" | |
################################################## | |
# ์์๋ค (ํ๊ธํ ๋ฒ์ ) | |
################################################## | |
examples = [ | |
[ | |
{ | |
"text": "PDF ํ์ผ ๋ด์ฉ์ ์์ฝ, ๋ถ์ํ๋ผ.", | |
"files": ["assets/additional-examples/pdf.pdf"], | |
} | |
], | |
[ | |
{ | |
"text": "CSV ํ์ผ ๋ด์ฉ์ ์์ฝ, ๋ถ์ํ๋ผ", | |
"files": ["assets/additional-examples/sample-csv.csv"], | |
} | |
], | |
[ | |
{ | |
"text": "๋์ผํ ๋ง๋ ๊ทธ๋ํ๋ฅผ ๊ทธ๋ฆฌ๋ matplotlib ์ฝ๋๋ฅผ ์์ฑํด์ฃผ์ธ์.", | |
"files": ["assets/additional-examples/barchart.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด ์์์์ ์ด์ํ ์ ์ด ๋ฌด์์ธ๊ฐ์?", | |
"files": ["assets/additional-examples/tmp.mp4"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด๋ฏธ ์ด ์์์ ๋ฅผ <image> ๊ฐ์ง๊ณ ์๊ณ , ์ด ์ ํ <image>์ ์๋ก ์ฌ๋ ค ํฉ๋๋ค. ํจ๊ป ์ญ์ทจํ ๋ ์ฃผ์ํด์ผ ํ ์ ์ด ์์๊น์?", | |
"files": ["assets/additional-examples/pill1.png", "assets/additional-examples/pill2.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด๋ฏธ์ง์ ์๊ฐ์ ์์์์ ์๊ฐ์ ๋ฐ์ ์๋ฅผ ์์ฑํด์ฃผ์ธ์.", | |
"files": ["assets/sample-images/06-1.png", "assets/sample-images/06-2.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด๋ฏธ์ง์ ์๊ฐ์ ์์๋ฅผ ํ ๋๋ก ์งง์ ์ ๊ณก์ ์๊ณกํด์ฃผ์ธ์.", | |
"files": [ | |
"assets/sample-images/07-1.png", | |
"assets/sample-images/07-2.png", | |
"assets/sample-images/07-3.png", | |
"assets/sample-images/07-4.png", | |
], | |
} | |
], | |
[ | |
{ | |
"text": "์ด ์ง์์ ๋ฌด์จ ์ผ์ด ์์์์ง ์งง์ ์ด์ผ๊ธฐ๋ฅผ ์ง์ด๋ณด์ธ์.", | |
"files": ["assets/sample-images/08.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด๋ฏธ์ง๋ค์ ์์๋ฅผ ๋ฐํ์ผ๋ก ์งง์ ์ด์ผ๊ธฐ๋ฅผ ๋ง๋ค์ด ์ฃผ์ธ์.", | |
"files": [ | |
"assets/sample-images/09-1.png", | |
"assets/sample-images/09-2.png", | |
"assets/sample-images/09-3.png", | |
"assets/sample-images/09-4.png", | |
"assets/sample-images/09-5.png", | |
], | |
} | |
], | |
[ | |
{ | |
"text": "์ด ์ธ๊ณ์์ ์ด๊ณ ์์ ์๋ฌผ๋ค์ ์์ํด์ ๋ฌ์ฌํด์ฃผ์ธ์.", | |
"files": ["assets/sample-images/10.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด๋ฏธ์ง์ ์ ํ ํ ์คํธ๋ฅผ ์ฝ์ด์ฃผ์ธ์.", | |
"files": ["assets/additional-examples/1.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด ํฐ์ผ์ ์ธ์ ๋ฐ๊ธ๋ ๊ฒ์ด๊ณ , ๊ฐ๊ฒฉ์ ์ผ๋ง์ธ๊ฐ์?", | |
"files": ["assets/additional-examples/2.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด๋ฏธ์ง์ ์๋ ํ ์คํธ๋ฅผ ๊ทธ๋๋ก ์ฝ์ด์ ๋งํฌ๋ค์ด ํํ๋ก ์ ์ด์ฃผ์ธ์.", | |
"files": ["assets/additional-examples/3.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด ์ ๋ถ์ ํ์ด์ฃผ์ธ์.", | |
"files": ["assets/additional-examples/4.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด ์ด๋ฏธ์ง๋ฅผ ๊ฐ๋จํ ์บก์ ์ผ๋ก ์ค๋ช ํด์ฃผ์ธ์.", | |
"files": ["assets/sample-images/01.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด ํ์งํ์๋ ๋ฌด์จ ๋ฌธ๊ตฌ๊ฐ ์ ํ ์๋์?", | |
"files": ["assets/sample-images/02.png"], | |
} | |
], | |
[ | |
{ | |
"text": "๋ ์ด๋ฏธ์ง๋ฅผ ๋น๊ตํด์ ๊ณตํต์ ๊ณผ ์ฐจ์ด์ ์ ๋งํด์ฃผ์ธ์.", | |
"files": ["assets/sample-images/03.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ด๋ฏธ์ง์ ๋ณด์ด๋ ๋ชจ๋ ์ฌ๋ฌผ๊ณผ ๊ทธ ์์์ ๋์ดํด์ฃผ์ธ์.", | |
"files": ["assets/sample-images/04.png"], | |
} | |
], | |
[ | |
{ | |
"text": "์ฅ๋ฉด์ ๋ถ์๊ธฐ๋ฅผ ๋ฌ์ฌํด์ฃผ์ธ์.", | |
"files": ["assets/sample-images/05.png"], | |
} | |
], | |
] | |
################################################## | |
# Gradio ์ธํฐํ์ด์ค ์ค์ | |
################################################## | |
demo = gr.ChatInterface( | |
fn=run, | |
type="messages", | |
chatbot=gr.Chatbot(type="messages", scale=1, allow_tags=["image"]), | |
# .webp, .png, .jpg, .jpeg, .gif, .mp4, .csv, .txt, .pdf ๋ชจ๋ ํ์ฉ | |
textbox=gr.MultimodalTextbox( | |
file_types=[ | |
".webp", ".png", ".jpg", ".jpeg", ".gif", | |
".mp4", ".csv", ".txt", ".pdf" | |
], | |
file_count="multiple", | |
autofocus=True | |
), | |
multimodal=True, | |
additional_inputs=[ | |
gr.Textbox( | |
label="System Prompt", | |
value=( | |
"You are a deeply thoughtful AI. Consider problems thoroughly and derive " | |
"correct solutions through systematic reasoning. Please answer in korean." | |
) | |
), | |
gr.Slider(label="Max New Tokens", minimum=100, maximum=8000, step=50, value=2000), | |
], | |
stop_btn=False, | |
title="Vidraft-Gemma-3-27B", | |
examples=examples, | |
run_examples_on_click=False, | |
cache_examples=False, | |
css_paths="style.css", | |
delete_cache=(1800, 1800), | |
) | |
if __name__ == "__main__": | |
demo.launch() |