koko210Serve
1fc3d74a5b
Features: - Built custom ROCm container for AMD RX 6800 GPU - Added GPU selection toggle in web UI (NVIDIA/AMD) - Unified model names across both GPUs for seamless switching - Vision model always uses NVIDIA GPU (optimal performance) - Text models (llama3.1, darkidol) can use either GPU - Added /gpu-status and /gpu-select API endpoints - Implemented GPU state persistence in memory/gpu_state.json Technical details: - Multi-stage Dockerfile.llamaswap-rocm with ROCm 6.2.4 - llama.cpp compiled with GGML_HIP=ON for gfx1030 (RX 6800) - Proper GPU permissions without root (groups 187/989) - AMD container on port 8091, NVIDIA on port 8090 - Updated bot/utils/llm.py with get_current_gpu_url() and get_vision_gpu_url() - Modified bot/utils/image_handling.py to always use NVIDIA for vision - Enhanced web UI with GPU selector button (blue=NVIDIA, red=AMD) Files modified: - docker-compose.yml (added llama-swap-amd service) - bot/globals.py (added LLAMA_AMD_URL) - bot/api.py (added GPU selection endpoints and helper function) - bot/utils/llm.py (GPU routing for text models) - bot/utils/image_handling.py (GPU routing for vision models) - bot/static/index.html (GPU selector UI) - llama-swap-rocm-config.yaml (unified model names) New files: - Dockerfile.llamaswap-rocm - bot/memory/gpu_state.json - bot/utils/gpu_router.py (load balancing utility) - setup-dual-gpu.sh (setup verification script) - DUAL_GPU_*.md (documentation files)
70 lines
2.5 KiB
Python
70 lines
2.5 KiB
Python
"""
|
|
GPU Model Preloading Utility
|
|
Preloads models on AMD GPU to take advantage of 16GB VRAM
|
|
"""
|
|
import aiohttp
|
|
import asyncio
|
|
import json
|
|
import globals
|
|
|
|
async def preload_amd_models():
|
|
"""
|
|
Preload both text and vision models on AMD GPU
|
|
Since AMD RX 6800 has 16GB VRAM, we can keep both loaded simultaneously
|
|
"""
|
|
print("🔧 Preloading models on AMD GPU...")
|
|
|
|
# Preload text model
|
|
try:
|
|
async with aiohttp.ClientSession() as session:
|
|
payload = {
|
|
"model": "llama3.1",
|
|
"messages": [{"role": "user", "content": "Hi"}],
|
|
"max_tokens": 1
|
|
}
|
|
async with session.post(
|
|
f"{globals.LLAMA_AMD_URL}/v1/chat/completions",
|
|
json=payload,
|
|
timeout=aiohttp.ClientTimeout(total=60)
|
|
) as response:
|
|
if response.status == 200:
|
|
print("✅ Text model (llama3.1) preloaded on AMD GPU")
|
|
else:
|
|
print(f"⚠️ Text model preload returned status {response.status}")
|
|
except Exception as e:
|
|
print(f"⚠️ Failed to preload text model on AMD: {e}")
|
|
|
|
# Preload vision model
|
|
try:
|
|
async with aiohttp.ClientSession() as session:
|
|
# Create a minimal test image (1x1 white pixel)
|
|
import base64
|
|
test_image = "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mP8z8DwHwAFBQIAX8jx0gAAAABJRU5ErkJggg=="
|
|
|
|
payload = {
|
|
"model": "vision",
|
|
"messages": [
|
|
{
|
|
"role": "user",
|
|
"content": [
|
|
{"type": "text", "text": "What do you see?"},
|
|
{"type": "image_url", "image_url": {"url": f"data:image/png;base64,{test_image}"}}
|
|
]
|
|
}
|
|
],
|
|
"max_tokens": 1
|
|
}
|
|
async with session.post(
|
|
f"{globals.LLAMA_AMD_URL}/v1/chat/completions",
|
|
json=payload,
|
|
timeout=aiohttp.ClientTimeout(total=120)
|
|
) as response:
|
|
if response.status == 200:
|
|
print("✅ Vision model preloaded on AMD GPU")
|
|
else:
|
|
print(f"⚠️ Vision model preload returned status {response.status}")
|
|
except Exception as e:
|
|
print(f"⚠️ Failed to preload vision model on AMD: {e}")
|
|
|
|
print("✅ AMD GPU preload complete - both models ready")
|