ai_coder/vast_coder_spawner_ollama
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: increase disk space to 128GB and update OLLAMA_MODEL and MIN_VRAM_MB for improved performance

Browse Source
This commit is contained in:
stephen schuresko
2026-05-14 11:22:55 -04:00
parent 246c2ce102
commit 984a331a21
2 changed files with 17 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
search_params.json
View File

@@ -8,7 +8,7 @@
"gpu_ram": {"gt": 50000},
"reliability": {"gt": ".99"},
"inet_down": {"gt": 500},
"disk_space": {"gt": 70},
"disk_space": {"gt": 200},
"order": [
[
"price",

22
vast_start.py
View File

@@ -17,10 +17,10 @@ import requests
# -----------------------------------------------------------------------------
# Configuration
# -----------------------------------------------------------------------------
OLLAMA_MODEL = "qwen3-coder-next:latest"
MIN_VRAM_MB = 20000
OLLAMA_MODEL = "deepseek-v4-flash:cloud"
MIN_VRAM_MB = 30000
MAX_BID_HOURLY = 1.5
DISK_SPACE_GB = int(MIN_VRAM_MB / 700 + 20)
DISK_SPACE_GB = int(MIN_VRAM_MB * 2 / 700 + 20)
VAST_API_BASE = "https://console.vast.ai/api/v0"
@@ -55,7 +55,8 @@ def vast_request(method: str, endpoint: str, params: Dict = None, json_data: Dic
url = f"{VAST_API_BASE}/{endpoint}"
headers = {"Authorization": f"Bearer {API_KEY}"}
print(f"[{method}] {url}")
#print(f"[{method}] {url}")
response = requests.request(method, url, headers=headers, params=params, json=json_data)
if not response.ok:
@@ -169,18 +170,27 @@ def wait_until_ready(instance_id: int, poll_interval: int = 15) -> Dict[str, Any
if tag in data and data[tag] is not None:
info.append(f"{tag}={data[tag]}")
print(f"Status: {status} | {' | '.join(info)}")
#print(f"Status: {status} | {' | '.join(info)}")
if status == "running":
# Extra check to ensure SSH port is populated
if data.get("ssh_port") and data.get("ssh_host"):
print("Instance is running and network is configured.")
time.sleep(5) # Brief pause to ensure services are up
return data
else:
print("Instance running, but waiting for SSH details to populate...")
print("Instance not ready yet. Fetching logs for insights...")
#print("Instance not ready yet. Fetching logs for insights...")
try:
logs = get_instance_demon_logs(instance_id)
print(chr(27) + "[2J")
print(f"--- Logs for Instance {instance_id} ---")
print(f"Status: {status} | {' | '.join(info)}")
#print("instanxce_details:"+json.dumps(data))
#print(f"Current Status: {status}")
print(f"Status MSG: {data['status_msg']}")
print(f"instance_id: {instance_id}")
print(f"Date: {time.strftime('%Y-%m-%d %H:%M:%S')}\n\n")
print("Retrieved logs:")
print('\n'.join(logs.splitlines()[-30:])) # Print last 10 lines for brevity
except Exception as e: