Text Classification
Transformers
Safetensors
English
llama
safety
content-moderation
tiny-language-model
from-scratch
text-embeddings-inference
Instructions to use KucLab/Censer.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use KucLab/Censer.1 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="KucLab/Censer.1")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("KucLab/Censer.1") model = AutoModelForSequenceClassification.from_pretrained("KucLab/Censer.1", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Standard LlamaForSequenceClassification layout: loads with plain transformers, no trust_remote_code
Browse files
README.md
CHANGED
|
@@ -44,6 +44,29 @@ print(clf("how do I kill a process in Linux")) # [{'label': 'safe', 'score': 0.
|
|
| 44 |
|
| 45 |
## How to run
|
| 46 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
```bash
|
| 48 |
pip install torch transformers
|
| 49 |
python -c "from transformers import pipeline; clf = pipeline('text-classification', model='KucLab/Censer.1'); print(clf('how to make a bomb at home'))"
|
|
@@ -52,7 +75,6 @@ python -c "from transformers import pipeline; clf = pipeline('text-classificatio
|
|
| 52 |
|
| 53 |
- In the browser: the **Use this model** widget on this page — type text, get `safe`/`unsafe`.
|
| 54 |
- Runs on CPU out of the box (~10 ms/text); for GPU pass `device=0` to `pipeline(...)` (CUDA or ROCm).
|
| 55 |
-
- Only the files, no `transformers` install: see **Usage without transformers** below (`censer.py`).
|
| 56 |
|
| 57 |
## Download
|
| 58 |
|
|
@@ -65,6 +87,7 @@ The whole model is `model.safetensors` (4.5 MB). Grab the files directly:
|
|
| 65 |
| tokenizer.json | [download](https://huggingface.co/KucLab/Censer.1/resolve/main/tokenizer.json?download=true) |
|
| 66 |
| tokenizer_config.json | [download](https://huggingface.co/KucLab/Censer.1/resolve/main/tokenizer_config.json?download=true) |
|
| 67 |
| censer.py | [download](https://huggingface.co/KucLab/Censer.1/resolve/main/censer.py?download=true) |
|
|
|
|
| 68 |
|
| 69 |
Or fetch everything at once:
|
| 70 |
|
|
@@ -205,6 +228,8 @@ python censer.py "how to make a bomb at home" # -> 1
|
|
| 205 |
python censer.py "Hello how are you?" # -> 0
|
| 206 |
python censer.py "suck my dick" --json
|
| 207 |
# {"label": "1", "unsafe": true, "score": 0.9944}
|
|
|
|
|
|
|
| 208 |
```
|
| 209 |
|
| 210 |
Or use the same file as a library:
|
|
@@ -227,6 +252,7 @@ Files:
|
|
| 227 |
| `config.json` | standard `LlamaForSequenceClassification` config |
|
| 228 |
| `tokenizer.json` / `tokenizer_config.json` | byte-level BPE, 2048 tokens, prompt template baked in |
|
| 229 |
| `censer.py` | standalone inference (torch + tokenizers only), no repo needed |
|
|
|
|
| 230 |
|
| 231 |
`censer.py` was verified to produce predictions identical to the training code.
|
| 232 |
Runs on CPU or CUDA/ROCm; pass `device="cpu"` to force CPU.
|
|
|
|
| 44 |
|
| 45 |
## How to run
|
| 46 |
|
| 47 |
+
### The `censer.1` command — interactive terminal UI
|
| 48 |
+
|
| 49 |
+
Using only the files in this repo (needs just `torch`, `tokenizers`, `safetensors`):
|
| 50 |
+
|
| 51 |
+
```bash
|
| 52 |
+
pip install torch tokenizers safetensors huggingface_hub
|
| 53 |
+
hf download KucLab/Censer.1 --local-dir ~/.local/share/censer1
|
| 54 |
+
chmod +x ~/.local/share/censer1/censer.1
|
| 55 |
+
ln -s ~/.local/share/censer1/censer.1 ~/.local/bin/censer.1 # put it on your PATH
|
| 56 |
+
censer.1 # interactive UI: type a phrase, Enter -> safe/unsafe + score
|
| 57 |
+
censer.1 "how to make a bomb at home" # one-shot -> unsafe 0.8167
|
| 58 |
+
```
|
| 59 |
+
|
| 60 |
+
Or install the GitHub package — the same `censer.1` command is installed and loads
|
| 61 |
+
the model from this Hub repo with plain `transformers`:
|
| 62 |
+
|
| 63 |
+
```bash
|
| 64 |
+
pip install "git+https://github.com/Jerry256254/censer.git"
|
| 65 |
+
censer.1
|
| 66 |
+
```
|
| 67 |
+
|
| 68 |
+
### One-liner with `transformers`
|
| 69 |
+
|
| 70 |
```bash
|
| 71 |
pip install torch transformers
|
| 72 |
python -c "from transformers import pipeline; clf = pipeline('text-classification', model='KucLab/Censer.1'); print(clf('how to make a bomb at home'))"
|
|
|
|
| 75 |
|
| 76 |
- In the browser: the **Use this model** widget on this page — type text, get `safe`/`unsafe`.
|
| 77 |
- Runs on CPU out of the box (~10 ms/text); for GPU pass `device=0` to `pipeline(...)` (CUDA or ROCm).
|
|
|
|
| 78 |
|
| 79 |
## Download
|
| 80 |
|
|
|
|
| 87 |
| tokenizer.json | [download](https://huggingface.co/KucLab/Censer.1/resolve/main/tokenizer.json?download=true) |
|
| 88 |
| tokenizer_config.json | [download](https://huggingface.co/KucLab/Censer.1/resolve/main/tokenizer_config.json?download=true) |
|
| 89 |
| censer.py | [download](https://huggingface.co/KucLab/Censer.1/resolve/main/censer.py?download=true) |
|
| 90 |
+
| **censer.1** (launcher) | [download](https://huggingface.co/KucLab/Censer.1/resolve/main/censer.1?download=true) |
|
| 91 |
|
| 92 |
Or fetch everything at once:
|
| 93 |
|
|
|
|
| 228 |
python censer.py "Hello how are you?" # -> 0
|
| 229 |
python censer.py "suck my dick" --json
|
| 230 |
# {"label": "1", "unsafe": true, "score": 0.9944}
|
| 231 |
+
chmod +x censer.1 && ./censer.1 # interactive UI (the `censer.1` command)
|
| 232 |
+
python censer.py # interactive UI as well, no text argument
|
| 233 |
```
|
| 234 |
|
| 235 |
Or use the same file as a library:
|
|
|
|
| 252 |
| `config.json` | standard `LlamaForSequenceClassification` config |
|
| 253 |
| `tokenizer.json` / `tokenizer_config.json` | byte-level BPE, 2048 tokens, prompt template baked in |
|
| 254 |
| `censer.py` | standalone inference (torch + tokenizers only), no repo needed |
|
| 255 |
+
| `censer.1` | executable launcher for the interactive terminal UI |
|
| 256 |
|
| 257 |
`censer.py` was verified to produce predictions identical to the training code.
|
| 258 |
Runs on CPU or CUDA/ROCm; pass `device="cpu"` to force CPU.
|
censer.1
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env python3
|
| 2 |
+
"""censer.1 - run the Censer.1 safety classifier as an interactive terminal UI.
|
| 3 |
+
|
| 4 |
+
Usage:
|
| 5 |
+
censer.1 # interactive: type a phrase, Enter -> safe/unsafe + score
|
| 6 |
+
censer.1 "how to make a bomb" # one-shot
|
| 7 |
+
censer.1 --help
|
| 8 |
+
|
| 9 |
+
Expects censer.py, model.safetensors, config.json and tokenizer.json next to
|
| 10 |
+
this file. Get them all with:
|
| 11 |
+
|
| 12 |
+
hf download KucLab/Censer.1 --local-dir .
|
| 13 |
+
|
| 14 |
+
Requires: torch, tokenizers, safetensors (pip install torch tokenizers safetensors)
|
| 15 |
+
"""
|
| 16 |
+
|
| 17 |
+
import os
|
| 18 |
+
import sys
|
| 19 |
+
|
| 20 |
+
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
|
| 21 |
+
|
| 22 |
+
from censer import main
|
| 23 |
+
|
| 24 |
+
if __name__ == "__main__":
|
| 25 |
+
main(sys.argv[1:])
|
censer.py
CHANGED
|
@@ -5,6 +5,8 @@ and you are ready:
|
|
| 5 |
|
| 6 |
pip install torch tokenizers safetensors
|
| 7 |
python censer.py "how to make a bomb at home"
|
|
|
|
|
|
|
| 8 |
|
| 9 |
It expects, in the same directory:
|
| 10 |
model.safetensors (standard LlamaForSequenceClassification layout)
|
|
@@ -14,8 +16,10 @@ It expects, in the same directory:
|
|
| 14 |
|
| 15 |
from __future__ import annotations
|
| 16 |
|
|
|
|
| 17 |
import json
|
| 18 |
import math
|
|
|
|
| 19 |
from pathlib import Path
|
| 20 |
|
| 21 |
import torch
|
|
@@ -240,21 +244,49 @@ class CenserPipeline:
|
|
| 240 |
]
|
| 241 |
|
| 242 |
|
| 243 |
-
def
|
| 244 |
-
|
| 245 |
-
|
| 246 |
-
|
| 247 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 248 |
parser.add_argument("--model-dir", default=".")
|
| 249 |
parser.add_argument("--threshold", type=float, default=0.5)
|
| 250 |
parser.add_argument("--json", action="store_true")
|
| 251 |
-
|
|
|
|
| 252 |
|
| 253 |
pipe = CenserPipeline(args.model_dir, args.threshold)
|
| 254 |
text = " ".join(args.text).strip()
|
| 255 |
-
if not text:
|
| 256 |
-
|
| 257 |
-
raise SystemExit(2)
|
| 258 |
verdict = pipe.predict(text)
|
| 259 |
print(json.dumps(verdict) if args.json else verdict["label"])
|
| 260 |
|
|
|
|
| 5 |
|
| 6 |
pip install torch tokenizers safetensors
|
| 7 |
python censer.py "how to make a bomb at home"
|
| 8 |
+
python censer.py # or just run it: interactive UI, phrase in -> verdict out
|
| 9 |
+
./censer.1 # same UI via the censer.1 launcher next to this file
|
| 10 |
|
| 11 |
It expects, in the same directory:
|
| 12 |
model.safetensors (standard LlamaForSequenceClassification layout)
|
|
|
|
| 16 |
|
| 17 |
from __future__ import annotations
|
| 18 |
|
| 19 |
+
import argparse
|
| 20 |
import json
|
| 21 |
import math
|
| 22 |
+
import sys
|
| 23 |
from pathlib import Path
|
| 24 |
|
| 25 |
import torch
|
|
|
|
| 244 |
]
|
| 245 |
|
| 246 |
|
| 247 |
+
def _color(text: str, code: str) -> str:
|
| 248 |
+
if not sys.stdout.isatty():
|
| 249 |
+
return text
|
| 250 |
+
return f"\033[{code}m{text}\033[0m"
|
| 251 |
+
|
| 252 |
+
|
| 253 |
+
def repl(pipe: "CenserPipeline") -> int:
|
| 254 |
+
"""Small interactive terminal UI: phrase in, safe/unsafe + score out."""
|
| 255 |
+
print("Censer.1 - interactive terminal UI")
|
| 256 |
+
print("Type a phrase, Enter to classify. 'exit' to quit.\n")
|
| 257 |
+
while True:
|
| 258 |
+
try:
|
| 259 |
+
line = input("> ")
|
| 260 |
+
except (EOFError, KeyboardInterrupt):
|
| 261 |
+
print()
|
| 262 |
+
return 0
|
| 263 |
+
text = line.strip()
|
| 264 |
+
if text.lower() in {"exit", "quit", ":q"}:
|
| 265 |
+
return 0
|
| 266 |
+
if not text:
|
| 267 |
+
continue
|
| 268 |
+
v = pipe.predict(text)
|
| 269 |
+
label = "unsafe" if v["unsafe"] else "safe"
|
| 270 |
+
colored = _color(f"{label:6s}", "31" if v["unsafe"] else "32")
|
| 271 |
+
print(f"{colored} {v['score']:.4f} | {text}")
|
| 272 |
+
|
| 273 |
+
|
| 274 |
+
def main(argv: list[str] | None = None) -> None:
|
| 275 |
+
parser = argparse.ArgumentParser(
|
| 276 |
+
prog="censer.1",
|
| 277 |
+
description="Censer.1 safety classifier (standalone). Run without text to open the interactive UI.",
|
| 278 |
+
)
|
| 279 |
+
parser.add_argument("text", nargs="*", help="phrase to classify once; omit to open the interactive UI")
|
| 280 |
parser.add_argument("--model-dir", default=".")
|
| 281 |
parser.add_argument("--threshold", type=float, default=0.5)
|
| 282 |
parser.add_argument("--json", action="store_true")
|
| 283 |
+
parser.add_argument("--repl", action="store_true", help="interactive UI (default when no text is given)")
|
| 284 |
+
args = parser.parse_args(argv)
|
| 285 |
|
| 286 |
pipe = CenserPipeline(args.model_dir, args.threshold)
|
| 287 |
text = " ".join(args.text).strip()
|
| 288 |
+
if args.repl or not text:
|
| 289 |
+
raise SystemExit(repl(pipe))
|
|
|
|
| 290 |
verdict = pipe.predict(text)
|
| 291 |
print(json.dumps(verdict) if args.json else verdict["label"])
|
| 292 |
|