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"
File size: 11,248 Bytes
0bc1e1f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 | """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())
|