Instructions to use FerrellSyntheticIntelligence/fsi-anomaly 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 FerrellSyntheticIntelligence/fsi-anomaly 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 FerrellSyntheticIntelligence/fsi-anomaly # Run inference directly in the terminal: llama cli -hf FerrellSyntheticIntelligence/fsi-anomaly
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf FerrellSyntheticIntelligence/fsi-anomaly # Run inference directly in the terminal: llama cli -hf FerrellSyntheticIntelligence/fsi-anomaly
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 FerrellSyntheticIntelligence/fsi-anomaly # Run inference directly in the terminal: ./llama-cli -hf FerrellSyntheticIntelligence/fsi-anomaly
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 FerrellSyntheticIntelligence/fsi-anomaly # Run inference directly in the terminal: ./build/bin/llama-cli -hf FerrellSyntheticIntelligence/fsi-anomaly
Use Docker
docker model run hf.co/FerrellSyntheticIntelligence/fsi-anomaly
- LM Studio
- Jan
- Ollama
How to use FerrellSyntheticIntelligence/fsi-anomaly with Ollama:
ollama run hf.co/FerrellSyntheticIntelligence/fsi-anomaly
- Unsloth Desktop
- Docker Model Runner
How to use FerrellSyntheticIntelligence/fsi-anomaly with Docker Model Runner:
docker model run hf.co/FerrellSyntheticIntelligence/fsi-anomaly
- Lemonade
How to use FerrellSyntheticIntelligence/fsi-anomaly with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull FerrellSyntheticIntelligence/fsi-anomaly
Run and chat with the model
lemonade run user.fsi-anomaly-{{QUANT_TAG}}List all available models
lemonade list
- Atomic Chat
File size: 11,194 Bytes
97c39f2 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 | """TinyLiquid Analyst - neon terminal UI (stdlib curses, zero deps).
An OpenCode-style agent terminal themed like Parrot OS meets Matrix:
near-black background, matrix-green neon, pink/blue/purple accents.
Chat with the analyst/skeptic personas, run SOP-driven cases, search and
read the library, manage case files.
Usage:
.venv/bin/python tui/analyst.py --ckpt ckpt/dpo
.venv/bin/python tui/analyst.py --ckpt ckpt/dpo --case mycase.json
"""
import argparse
import curses
import random
import textwrap
from tui.engine import AnalystEngine
from tui.cli import handle_cmd
WRAP = 78
# --- neon palette (256-color) ---
C_BG = 0 # terminal default (black)
C_GREEN = 82 # matrix green
C_PINK = 205
C_BLUE = 39
C_PURPLE = 141
C_CYAN = 51
C_WHITE = 255
C_DIM = 240
PAIR = {
"bg": 1, "green": 2, "pink": 3, "blue": 4, "purple": 5,
"cyan": 6, "white": 7, "dim": 8, "user": 9, "model": 10,
"skeptic": 11, "system": 12, "error": 13, "header": 14,
}
def setup_colors(stdscr):
curses.start_color()
curses.use_default_colors()
try:
if curses.COLORS >= 256:
curses.init_pair(PAIR["bg"], C_WHITE, C_BG)
curses.init_pair(PAIR["green"], C_GREEN, C_BG)
curses.init_pair(PAIR["pink"], C_PINK, C_BG)
curses.init_pair(PAIR["blue"], C_BLUE, C_BG)
curses.init_pair(PAIR["purple"], C_PURPLE, C_BG)
curses.init_pair(PAIR["cyan"], C_CYAN, C_BG)
curses.init_pair(PAIR["white"], C_WHITE, C_BG)
curses.init_pair(PAIR["dim"], C_DIM, C_BG)
curses.init_pair(PAIR["user"], C_CYAN, C_BG)
curses.init_pair(PAIR["model"], C_GREEN, C_BG)
curses.init_pair(PAIR["skeptic"], C_PINK, C_BG)
curses.init_pair(PAIR["system"], C_PURPLE, C_BG)
curses.init_pair(PAIR["error"], C_PINK, C_BG)
curses.init_pair(PAIR["header"], C_GREEN, C_BG)
else:
for k in PAIR:
curses.init_pair(PAIR[k], curses.COLOR_WHITE, curses.COLOR_BLACK)
curses.init_pair(PAIR["green"], curses.COLOR_GREEN, curses.COLOR_BLACK)
curses.init_pair(PAIR["pink"], curses.COLOR_RED, curses.COLOR_BLACK)
curses.init_pair(PAIR["blue"], curses.COLOR_CYAN, curses.COLOR_BLACK)
curses.init_pair(PAIR["purple"], curses.COLOR_MAGENTA, curses.COLOR_BLACK)
curses.init_pair(PAIR["cyan"], curses.COLOR_CYAN, curses.COLOR_BLACK)
curses.init_pair(PAIR["user"], curses.COLOR_CYAN, curses.COLOR_BLACK)
curses.init_pair(PAIR["model"], curses.COLOR_GREEN, curses.COLOR_BLACK)
curses.init_pair(PAIR["skeptic"], curses.COLOR_RED, curses.COLOR_BLACK)
curses.init_pair(PAIR["system"], curses.COLOR_MAGENTA, curses.COLOR_BLACK)
curses.init_pair(PAIR["error"], curses.COLOR_RED, curses.COLOR_BLACK)
curses.init_pair(PAIR["header"], curses.COLOR_GREEN, curses.COLOR_BLACK)
except curses.error:
pass
def wrap(text, width):
return "\n".join(textwrap.fill(p, width) for p in text.splitlines()) or " "
class Rain:
"""Minimal matrix-rain strip for the header (lightweight, idle-animated)."""
def __init__(self, width, rows=1):
self.width = max(1, width)
self.rows = rows
self.cols = [random.randint(0, self.width - 1) for _ in range(self.rows)]
def frame(self):
chars = [" "] * self.width
for i, c in enumerate(self.cols):
if random.random() < 0.75:
chars[c] = random.choice("01$#&@!?ABCDEF")
self.cols[i] = (c + random.choice([-1, 0, 1])) % self.width
return "".join(chars)
class TerminalUI:
def __init__(self, stdscr, eng):
self.stdscr = stdscr
self.eng = eng
self.lines = [] # list of (kind, text)
self.history = []
self.hist_idx = -1
self.scroll = 0
self.input_text = ""
self.cursor = 0
self.rain = Rain(10)
self.status = f"persona: analyst | case: {eng.case.name}"
def add(self, kind, text):
for ln in wrap(text, WRAP).splitlines() or [""]:
self.lines.append((kind, ln))
self.scroll = 0
def header(self, wd):
title = " TINYLIQUID ANALYST "
left = (wd - len(title)) // 2
top = "═" * wd
line = " " * left + title
return top, line[:wd], self.rain.frame()[:wd]
def draw(self):
h, wd = self.stdscr.getmaxyx()
top, title, rain = self.header(wd)
body_h = max(1, h - 6)
self.stdscr.erase()
self.stdscr.addnstr(0, 0, top[:wd], wd, curses.color_pair(PAIR["purple"]))
self.stdscr.addnstr(1, 0, title[:wd], wd, curses.color_pair(PAIR["header"]))
self.stdscr.addnstr(2, 0, rain[:wd], wd, curses.color_pair(PAIR["green"]))
start = len(self.lines) - body_h - self.scroll
start = max(0, min(start, max(0, len(self.lines) - body_h)))
for i in range(body_h):
idx = start + i
if idx < len(self.lines):
kind, ln = self.lines[idx]
pair = {"user": PAIR["user"], "model": PAIR["model"],
"skeptic": PAIR["skeptic"], "system": PAIR["system"],
"error": PAIR["error"]}.get(kind, PAIR["white"])
self.stdscr.addnstr(3 + i, 0, ln[:wd], wd, curses.color_pair(pair))
self.stdscr.addnstr(h - 3, 0, "─" * wd, wd, curses.color_pair(PAIR["purple"]))
self.stdscr.addnstr(h - 2, 0, self.status[:wd], wd,
curses.color_pair(PAIR["system"]) | curses.A_REVERSE)
prompt = "❯ "
self.stdscr.addnstr(h - 1, 0, prompt + self.input_text[:wd - len(prompt) - 1], wd,
curses.color_pair(PAIR["green"]))
try:
self.stdscr.move(h - 1, min(self.cursor + len(prompt), wd - 1))
except curses.error:
pass
self.stdscr.refresh()
def run_cmd(self, cmd):
if cmd in ("/quit", "/exit", "/q"):
return False
if cmd == "/clear":
self.lines = []
self.add("system", "cleared.")
return True
self.add("user", cmd if cmd.startswith("/") else "you> " + cmd)
if not cmd.startswith("/"):
# streamed chat: redraw as tokens arrive
buf = []
self.lines.append(("model", "…"))
try:
def on_token(tok_id):
buf.append(tok_id)
self.lines[-1] = ("model", self.eng.tok.decode(buf))
self.draw()
self.eng.chat(cmd, on_token=on_token)
except Exception as e:
self.lines[-1] = ("error", f"error: {e}")
self.draw()
return True
self.lines[-1] = ("model", self.eng.tok.decode(buf))
self.draw()
return True
try:
out = handle_cmd(self.eng, cmd)
except Exception as e:
out = f"error: {e}"
self.add("error", out)
return True
kind = "skeptic" if cmd.startswith(("/skeptic",)) else "model"
# render JSON reports nicely
if out.strip().startswith("{") or out.strip().startswith("["):
import json as _json
try:
data = _json.loads(out)
out = _json.dumps(data, indent=2, ensure_ascii=False)
except Exception:
pass
for ln in out.splitlines():
self.add(kind, ln)
return True
def loop(self):
setup_colors(self.stdscr)
curses.curs_set(1)
self.stdscr.timeout(80)
self.add("system", "TinyLiquid Analyst — on-device forensic research terminal")
self.add("dim", "Authorized research/OSINT only. Outputs are decision support, never a verdict.")
self.add("dim", "type /help for commands | /quit to exit")
while True:
self.draw()
try:
ch = self.stdscr.get_wch()
except curses.error:
continue
if ch in ("\n", "\r", "KEY_ENTER"):
cmd = self.input_text.strip()
self.input_text, self.cursor = "", 0
self.history.append(cmd)
self.hist_idx = len(self.history)
if cmd and not self.run_cmd(cmd):
break
continue
if ch in ("\x7f", "KEY_BACKSPACE"):
if self.cursor > 0:
self.input_text = self.input_text[:self.cursor - 1] + self.input_text[self.cursor:]
self.cursor -= 1
continue
if ch == "KEY_DC":
self.input_text = self.input_text[:self.cursor] + self.input_text[self.cursor + 1:]
continue
if ch == "KEY_LEFT":
self.cursor = max(0, self.cursor - 1)
continue
if ch == "KEY_RIGHT":
self.cursor = min(len(self.input_text), self.cursor + 1)
continue
if ch == "KEY_HOME":
self.cursor = 0
continue
if ch == "KEY_END":
self.cursor = len(self.input_text)
continue
if ch in ("KEY_PPAGE",):
self.scroll = min(self.scroll + 5, max(0, len(self.lines) - 1))
continue
if ch in ("KEY_NPAGE",):
self.scroll = max(0, self.scroll - 5)
continue
if ch == "KEY_UP":
if self.history and self.hist_idx > 0:
self.hist_idx -= 1
self.input_text = self.history[self.hist_idx]
self.cursor = len(self.input_text)
continue
if ch == "KEY_DOWN":
if self.history and self.hist_idx < len(self.history) - 1:
self.hist_idx += 1
self.input_text = self.history[self.hist_idx]
self.cursor = len(self.input_text)
else:
self.input_text, self.hist_idx = "", len(self.history)
continue
if isinstance(ch, str) and ch.isprintable():
self.input_text = self.input_text[:self.cursor] + ch + self.input_text[self.cursor:]
self.cursor += 1
def main():
ap = argparse.ArgumentParser()
ap.add_argument("--ckpt", default="ckpt/v8_lora/best.pt")
ap.add_argument("--tok", default="data/tokenizer.json")
ap.add_argument("--case", default=None)
ap.add_argument("--threads", type=int, default=4)
args = ap.parse_args()
eng = AnalystEngine(ckpt=args.ckpt, tok_path=args.tok, threads=args.threads)
if args.case:
eng.load_case(args.case)
def go(stdscr):
ui = TerminalUI(stdscr, eng)
ui.status = f"persona: {eng.persona} | case: {eng.case.name} | {eng.ckpt}"
ui.loop()
try:
curses.wrapper(go)
except KeyboardInterrupt:
pass
print("session ended.")
if __name__ == "__main__":
main()
|