Instructions to use DeepSeekOracle/lygo-console-models with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use DeepSeekOracle/lygo-console-models with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf DeepSeekOracle/lygo-console-models # Run inference directly in the terminal: llama cli -hf DeepSeekOracle/lygo-console-models
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf DeepSeekOracle/lygo-console-models # Run inference directly in the terminal: llama cli -hf DeepSeekOracle/lygo-console-models
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf DeepSeekOracle/lygo-console-models # Run inference directly in the terminal: ./llama-cli -hf DeepSeekOracle/lygo-console-models
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf DeepSeekOracle/lygo-console-models # Run inference directly in the terminal: ./build/bin/llama-cli -hf DeepSeekOracle/lygo-console-models
Use Docker
docker model run hf.co/DeepSeekOracle/lygo-console-models
- LM Studio
- Jan
- Ollama
How to use DeepSeekOracle/lygo-console-models with Ollama:
ollama run hf.co/DeepSeekOracle/lygo-console-models
- Unsloth Desktop
- Pi
How to use DeepSeekOracle/lygo-console-models with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf DeepSeekOracle/lygo-console-models
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "DeepSeekOracle/lygo-console-models" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use DeepSeekOracle/lygo-console-models with Docker Model Runner:
docker model run hf.co/DeepSeekOracle/lygo-console-models
- Lemonade
How to use DeepSeekOracle/lygo-console-models with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull DeepSeekOracle/lygo-console-models
Run and chat with the model
lemonade run user.lygo-console-models-{{QUANT_TAG}}List all available models
lemonade list
- Hermes Agent
How to use DeepSeekOracle/lygo-console-models with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf DeepSeekOracle/lygo-console-models
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default DeepSeekOracle/lygo-console-models
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use DeepSeekOracle/lygo-console-models with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf DeepSeekOracle/lygo-console-models
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "DeepSeekOracle/lygo-console-models" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Download fetch_models.py from DeepSeekOracle/lygo-console-models: direct link, hf CLI and curl.
- Browser
- Download file 11.2 kB
-
https://huggingface.co/DeepSeekOracle/lygo-console-models/resolve/main/fetch_models.py
- Command line
-
hf download hf://DeepSeekOracle/lygo-console-models/fetch_models.py
-
curl -L -o fetch_models.py https://huggingface.co/DeepSeekOracle/lygo-console-models/resolve/main/fetch_models.py
11.2 kB
| """Get the basic models from OUR vault - nobody else's site. | |
| The console ships without weights (they are 4-8 GB each and every one carries | |
| its author's licence). This fetcher pulls them from the steward's own store, | |
| verifies each SHA-256 from models.lock.json, and drops them where the console | |
| scans. Standard library only, so the bundled Python can run it anywhere. | |
| python fetch_models.py --list | |
| python fetch_models.py --profile basic | |
| python fetch_models.py --profile full --dest "I:/LYGO_MODELS" | |
| python fetch_models.py --check # verify what is already here | |
| python fetch_models.py --profile basic --mirror | |
| Nothing is downloaded from a URL that is not in models.lock.json, and no file | |
| is accepted unless its SHA-256 matches the lock. Re-running is safe: files | |
| already correct are skipped (and a half-finished download resumes). | |
| """ | |
| from __future__ import annotations | |
| import argparse | |
| import hashlib | |
| import json | |
| import os | |
| import shutil | |
| import sys | |
| import time | |
| import urllib.error | |
| import urllib.request | |
| from pathlib import Path | |
| HERE = Path(__file__).resolve().parent | |
| DEFAULT_LOCK = HERE / "models.lock.json" | |
| CHUNK = 1024 * 1024 | |
| def human(n: float) -> str: | |
| for unit in ("B", "KB", "MB", "GB", "TB"): | |
| if n < 1024 or unit == "TB": | |
| return "%.1f %s" % (n, unit) | |
| n /= 1024.0 | |
| return "%.1f TB" % n | |
| def load_lock(path: Path) -> dict: | |
| if not path.exists(): | |
| sys.exit("no lock file at %s - the console install is incomplete" % path) | |
| lock = json.loads(path.read_text(encoding="utf-8")) | |
| if not lock.get("models"): | |
| sys.exit("lock file %s lists no models" % path) | |
| return lock | |
| def pick_dest(explicit: str | None) -> Path: | |
| if explicit: | |
| return Path(explicit).expanduser().resolve() | |
| env = os.environ.get("LYGO_MODELS") | |
| if env: | |
| return Path(env).expanduser().resolve() | |
| for cand in (HERE.parent / "models", HERE / "models"): | |
| if cand.is_dir(): | |
| return cand.resolve() | |
| return (HERE.parent / "models").resolve() | |
| def sha256_of(path: Path) -> str: | |
| h = hashlib.sha256() | |
| with path.open("rb") as fh: | |
| for block in iter(lambda: fh.read(CHUNK), b""): | |
| h.update(block) | |
| return h.hexdigest() | |
| def url_for(lock: dict, name: str, use_mirror: bool) -> str: | |
| hosts = lock.get("hosts") or {} | |
| key = "mirror" if use_mirror else "primary" | |
| base = hosts.get(key) or hosts.get("primary") | |
| if not base: | |
| sys.exit("lock file has no %s host" % key) | |
| if not base.endswith("/"): | |
| base += "/" | |
| return base + name | |
| def host_allowed(lock: dict, url: str) -> bool: | |
| allowed = [h for h in (lock.get("hosts") or {}).values() if isinstance(h, str) and h.startswith("http")] | |
| return any(url.startswith(a if a.endswith("/") else a + "/") or url.startswith(a) for a in allowed) | |
| def download(url: str, dest: Path, entry: dict, lock: dict) -> tuple[bool, str]: | |
| if not host_allowed(lock, url): | |
| return False, "refused: %s is not one of our hosts" % url | |
| want_bytes = int(entry["bytes"]) | |
| want_sha = entry["sha256"].lower() | |
| part = dest.with_suffix(dest.suffix + ".part") | |
| have = part.stat().st_size if part.exists() else 0 | |
| if have > want_bytes: | |
| part.unlink() | |
| have = 0 | |
| req = urllib.request.Request(url, headers={"User-Agent": "lygo-console-fetch/1"}) | |
| if have: | |
| req.add_header("Range", "bytes=%d-" % have) | |
| print(" resuming at %s" % human(have)) | |
| mode = "ab" if have else "wb" | |
| t0 = time.time() | |
| try: | |
| with urllib.request.urlopen(req, timeout=120) as r, part.open(mode) as out: | |
| if have and r.status != 206: | |
| out.close() | |
| part.unlink(missing_ok=True) | |
| return download(url, dest, entry, lock) | |
| done = have | |
| while True: | |
| block = r.read(CHUNK) | |
| if not block: | |
| break | |
| out.write(block) | |
| done += len(block) | |
| pct = 100.0 * done / want_bytes if want_bytes else 0.0 | |
| speed = done / max(1e-6, time.time() - t0) | |
| sys.stdout.write("\r %5.1f%% %s / %s (%s/s) " % (pct, human(done), human(want_bytes), human(speed))) | |
| sys.stdout.flush() | |
| except urllib.error.HTTPError as e: | |
| return False, "http %s from %s" % (e.code, url) | |
| except Exception as e: # noqa: BLE001 - report anything the network throws | |
| print() | |
| return False, "%s: %s" % (e.__class__.__name__, e) | |
| print() | |
| got = part.stat().st_size | |
| if got != want_bytes: | |
| return False, "wrong size: got %s, lock says %s (kept as %s)" % (got, want_bytes, part.name) | |
| digest = sha256_of(part) | |
| if digest != want_sha: | |
| return False, "SHA-256 mismatch - refusing it (got %s..., lock says %s...)" % (digest[:16], want_sha[:16]) | |
| shutil.move(str(part), str(dest)) | |
| return True, "sha256 ok" | |
| def write_vault_manifest(dest: Path, lock: dict) -> None: | |
| """Keep the vault manifest the console reads in sync (same shape as before).""" | |
| path = dest / "manifest.json" | |
| data: dict = {} | |
| if path.exists(): | |
| try: | |
| data = json.loads(path.read_text(encoding="utf-8")) | |
| except Exception: # noqa: BLE001 - a broken manifest must not block a fetch | |
| data = {} | |
| for entry in lock["models"]: | |
| local = dest / entry["file"] | |
| if entry.get("kind") == "mmproj" or not local.exists(): | |
| continue | |
| vid = entry["vault_id"] | |
| rec = data.get(vid) or {"id": vid} | |
| rec["path"] = str(local) | |
| rec["bytes"] = local.stat().st_size | |
| rec["sha256"] = sha256_of(local) | |
| rec["ok"] = rec["sha256"] == entry["sha256"].lower() | |
| rec.setdefault("mmproj", None) | |
| for other in lock["models"]: | |
| if other.get("kind") == "mmproj" and other.get("vault_id") == vid: | |
| p = dest / other["file"] | |
| if p.exists(): | |
| rec["mmproj"] = str(p) | |
| rec["mmproj_sha256"] = sha256_of(p) | |
| data[vid] = rec | |
| path.write_text(json.dumps(data, indent=1), encoding="utf-8") | |
| def check_only(dest: Path, lock: dict) -> int: | |
| bad = 0 | |
| for entry in lock["models"]: | |
| local = dest / entry["file"] | |
| if not local.exists(): | |
| print(" MISSING %s" % entry["file"]) | |
| bad += 1 | |
| continue | |
| size_ok = local.stat().st_size == int(entry["bytes"]) | |
| digest = sha256_of(local) if size_ok else "" | |
| if size_ok and digest == entry["sha256"].lower(): | |
| print(" ok %s (%s, %s)" % (entry["file"], human(local.stat().st_size), entry["licence"])) | |
| else: | |
| print(" BAD %s (%s, sha256 %s)" % (entry["file"], human(local.stat().st_size), digest[:16] or "size")) | |
| bad += 1 | |
| print() | |
| print(" %s" % ("all present and verified" if not bad else "%d file(s) not right" % bad)) | |
| return 1 if bad else 0 | |
| def main(argv: list[str] | None = None) -> int: | |
| ap = argparse.ArgumentParser(description="Fetch the basic LYGO models from our own vault.") | |
| ap.add_argument("--lock", default=str(DEFAULT_LOCK)) | |
| ap.add_argument("--dest", default=None, help="where to put them (default: LYGO_MODELS, else the console's models\\)") | |
| ap.add_argument("--profile", default=None, help="core / basic / full / coder / embed (default: basic)") | |
| ap.add_argument("--mirror", action="store_true", help="use the mirror host instead of the primary") | |
| ap.add_argument("--list", action="store_true", help="show what the vault offers and exit") | |
| ap.add_argument("--check", action="store_true", help="verify the models already on this machine, download nothing") | |
| ap.add_argument("--yes", action="store_true", help="do not stop for confirmation") | |
| args = ap.parse_args(argv) | |
| lock = load_lock(Path(args.lock)) | |
| by_id = {m["id"]: m for m in lock["models"]} | |
| hosts = lock.get("hosts") or {} | |
| print("LYGO model vault · %s" % lock.get("product", "console")) | |
| print(" store : %s" % hosts.get("repo_page", hosts.get("primary", "?"))) | |
| print(" mirror : %s" % hosts.get("mirror_page", "?")) | |
| print(" revision: %s" % hosts.get("revision", "main")) | |
| print() | |
| if args.list: | |
| for pid, ids in sorted((lock.get("profiles") or {}).items()): | |
| total = sum(int(by_id[i]["bytes"]) for i in ids if i in by_id) | |
| print(" profile %-6s %s" % (pid, human(total))) | |
| for i in ids: | |
| e = by_id.get(i) | |
| if e: | |
| print(" %-22s %9s %-11s %s" % (e["file"], human(e["bytes"]), e["licence"], ",".join(e.get("modalities") or ["text"]))) | |
| print() | |
| return 0 | |
| dest = pick_dest(args.dest) | |
| dest.mkdir(parents=True, exist_ok=True) | |
| if args.check: | |
| print(" checking %s" % dest) | |
| return check_only(dest, lock) | |
| profile = (args.profile or "basic").lower() | |
| if profile not in (lock.get("profiles") or {}): | |
| sys.exit("unknown profile %r - try --list (known: %s)" % (profile, ", ".join(lock.get("profiles") or {}))) | |
| wanted = [by_id[i] for i in lock["profiles"][profile] if i in by_id] | |
| total = sum(int(m["bytes"]) for m in wanted) | |
| print(" profile : %s (%d file(s), %s)" % (profile, len(wanted), human(total))) | |
| print(" into : %s" % dest) | |
| for m in wanted: | |
| print(" %-24s %9s %s %s" % (m["file"], human(m["bytes"]), m["licence"], m["author"])) | |
| print() | |
| if not args.yes: | |
| try: | |
| if input(" download now? [y/N] ").strip().lower() not in ("y", "yes"): | |
| print(" stopped - nothing downloaded") | |
| return 0 | |
| except EOFError: | |
| print(" no answer available - stopping (use --yes for unattended runs)") | |
| return 0 | |
| failures = [] | |
| for m in wanted: | |
| target = dest / m["file"] | |
| if target.exists() and target.stat().st_size == int(m["bytes"]) and sha256_of(target) == m["sha256"].lower(): | |
| print(" have %s (already verified)" % m["file"]) | |
| continue | |
| print(" get %s %s %s" % (m["file"], human(m["bytes"]), m["licence"])) | |
| print(" from %s" % url_for(lock, m["file"], args.mirror)) | |
| ok, why = download(url_for(lock, m["file"], args.mirror), target, m, lock) | |
| print(" %s: %s" % ("done" if ok else "FAILED", why)) | |
| if not ok: | |
| failures.append((m["file"], why)) | |
| write_vault_manifest(dest, lock) | |
| print() | |
| print(" licence/attribution for what was fetched:") | |
| for m in wanted: | |
| print(" %-24s %-11s %s - %s" % (m["file"], m["licence"], m["author"], m["upstream"])) | |
| print(" full Apache-2.0 text: %s" % hosts.get("license_text", "https://www.apache.org/licenses/LICENSE-2.0.txt")) | |
| print() | |
| if failures: | |
| print(" %d of %d failed - nothing else was touched:" % (len(failures), len(wanted))) | |
| for name, why in failures: | |
| print(" %s: %s" % (name, why)) | |
| return 2 | |
| print(" all set - start the console and it will pick the model up") | |
| return 0 | |
| if __name__ == "__main__": | |
| raise SystemExit(main()) | |