Generation completes prematurely but still returns an image

I'm building a Telegram bot for webui_forge_cu121_torch231 that uses the API for all requests. The issue I'm encountering is that the generation process doesn't fully complete, yet it still outputs an image.

Here's my current implementation:
`

async def generate_image(update: Update, context: ContextTypes.DEFAULT_TYPE):
prompt = " ".join(context.args) if context.args else "random cute animal"

payload = {
"prompt": prompt,
"steps": 8,
"width": 512,
"height": 512,
"sampler_name": "DPM++ 2M Karras"
}

await update.message.reply_text(f"🔄 Генерирую: '{prompt}'...")

try:
response = requests.post(SD_API_URL, json=payload)
response.raise_for_status()
result = response.json()

image_data = result["images"][0]
with open("generated_image.png", "wb") as f:
f.write(base64.b64decode(image_data))

with open("generated_image.png", "rb") as photo:
await update.message.reply_photo(photo=photo, caption=f"🎨 Результат: '{prompt}'")

except Exception as e:
await update.message.reply_text(f"❌ Ошибка: {str(e)}")
`

[pic](https://drive.google.com/file/d/1xum1dnmnOwqeope37CuoFE9WSpOYpaU3/view?usp=sharing)

The report provides only an external Python Telegram bot and a POST payload; it names no Forge file, entry point, or test. Start by reproducing the request against the API with the shown settings and determine whether the returned image is produced before generation completes. Done means identifying a Forge-side reproduction and a specific project change or test.

Written by the indexing model from the issue text.