dongho/Munich-news
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

gpu config

Browse Source
This commit is contained in:
Dongho Kim
2025-11-12 15:21:58 +01:00
parent ec281f9440
commit d09d96ca4b
2 changed files with 63 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
backend/routes/ollama_routes.py
View File

@@ -196,16 +196,45 @@ def get_gpu_status():
# Check for GPU usage in loaded models
for model in models_loaded:
if 'gpu' in str(model).lower() or model.get('gpu_layers', 0) > 0:
# Check various GPU indicators
gpu_layers = model.get('gpu_layers', 0)
details = model.get('details', {})
# Check if GPU is mentioned in any field
if (gpu_layers > 0 or
'gpu' in str(model).lower() or
'cuda' in str(model).lower() or
details.get('families', []) and 'gpu' in str(details.get('families', [])).lower()):
gpu_info['gpu_in_use'] = True
gpu_info['gpu_available'] = True
gpu_info['gpu_details'] = {
'model': model.get('name', 'unknown'),
'gpu_layers': model.get('gpu_layers', 0),
'size': model.get('size', 0)
'gpu_layers': gpu_layers,
'size': model.get('size', 0),
'size_vram': model.get('size_vram', 0)
}
break
# If no models loaded, check Docker container for GPU
if not gpu_info['gpu_in_use']:
try:
import subprocess
# Check if nvidia-smi works in ollama container
result = subprocess.run(
['docker', 'exec', 'munich-news-ollama', 'nvidia-smi', '--query-gpu=name', '--format=csv,noheader'],
capture_output=True,
text=True,
timeout=5
)
if result.returncode == 0 and result.stdout.strip():
gpu_info['gpu_available'] = True
gpu_info['gpu_details'] = {
'gpu_name': result.stdout.strip(),
'note': 'GPU available but no model currently loaded'
}
except:
pass
# Try to get system info
try:
tags_response = requests.get(