Spaces:
Running on Zero
Running on Zero
| #!/usr/bin/env python3 | |
| # SPDX-FileCopyrightText: 2026 Shamiur Rashid Sunny | |
| # SPDX-License-Identifier: AGPL-3.0-only | |
| """Probe NVIDIA NIM to verify API key + discover tool-calling models. | |
| Usage: | |
| export NVIDIA_API_KEY=nvapi-... | |
| python scripts/probe_nvidia.py | |
| """ | |
| import os | |
| import sys | |
| import httpx | |
| BASE = "https://integrate.api.nvidia.com/v1" | |
| def main(): | |
| key = os.getenv("NVIDIA_API_KEY", "").strip() | |
| if not key: | |
| print("ERROR: NVIDIA_API_KEY not set") | |
| print("Run: export NVIDIA_API_KEY=nvapi-...") | |
| return 1 | |
| print("=" * 60) | |
| print(" NVIDIA NIM Probe") | |
| print("=" * 60) | |
| print() | |
| print("[1] Fetching model list...") | |
| try: | |
| r = httpx.get( | |
| BASE + "/models", | |
| headers={"Authorization": "Bearer " + key}, | |
| timeout=15.0, | |
| ) | |
| except Exception as e: | |
| print(" FAIL: " + type(e).__name__ + ": " + str(e)) | |
| return 1 | |
| print(" HTTP " + str(r.status_code)) | |
| if r.status_code != 200: | |
| print(" Body: " + r.text[:400]) | |
| return 1 | |
| try: | |
| data = r.json() | |
| models = [m.get("id") for m in data.get("data", [])] | |
| except Exception as e: | |
| print(" FAIL to parse: " + str(e)) | |
| return 1 | |
| print(" Total models: " + str(len(models))) | |
| print() | |
| print("[2] Tool-calling candidates:") | |
| known = [ | |
| "meta/llama-4-maverick-17b-128e-instruct", | |
| "meta/llama-4-scout-17b-16e-instruct", | |
| "mistralai/mistral-small-3.2-24b-instruct-2506", | |
| "nvidia/llama-3.1-nemotron-70b-instruct", | |
| "meta/llama-3.3-70b-instruct", | |
| "meta/llama-3.1-70b-instruct", | |
| ] | |
| found = [] | |
| for k in known: | |
| if k in models: | |
| found.append(k) | |
| print(" [OK] " + k) | |
| else: | |
| print(" [MISSING] " + k) | |
| print() | |
| if found: | |
| print("[3] Recommended first: " + found[0]) | |
| else: | |
| print("[3] No known tool-calling model present in your account.") | |
| print(" Check https://build.nvidia.com/models") | |
| print() | |
| if found: | |
| print("[4] Test completion with " + found[0] + "...") | |
| try: | |
| r = httpx.post( | |
| BASE + "/chat/completions", | |
| headers={ | |
| "Authorization": "Bearer " + key, | |
| "Content-Type": "application/json", | |
| }, | |
| json={ | |
| "model": found[0], | |
| "messages": [{"role": "user", "content": "Say hi in one word."}], | |
| "max_tokens": 30, | |
| }, | |
| timeout=30.0, | |
| ) | |
| print(" HTTP " + str(r.status_code)) | |
| if r.status_code == 200: | |
| d = r.json() | |
| content = d["choices"][0]["message"].get("content", "") | |
| print(" Response: " + content[:100]) | |
| else: | |
| print(" Body: " + r.text[:400]) | |
| except Exception as e: | |
| print(" FAIL: " + type(e).__name__ + ": " + str(e)) | |
| print() | |
| print("=" * 60) | |
| print(" Done.") | |
| return 0 | |
| if __name__ == "__main__": | |
| sys.exit(main()) | |