Image-Text-to-Text
MLX
Safetensors
qwen3_5
clairic
cloudflare
systemone
structured-output
information-extraction
multimodal
custom-code
conversational
4-bit precision
Instructions to use mlx-community/clairic-flash-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/clairic-flash-4bit with MLX:
# Make sure mlx-vlm is installed # pip install --upgrade mlx-vlm from mlx_vlm import load, generate from mlx_vlm.prompt_utils import apply_chat_template from mlx_vlm.utils import load_config # Load the model model, processor = load("mlx-community/clairic-flash-4bit") config = load_config("mlx-community/clairic-flash-4bit") # Prepare input image = ["http://images.cocodataset.org/val2017/000000039769.jpg"] prompt = "Describe this image." # Apply chat template formatted_prompt = apply_chat_template( processor, config, prompt, num_images=1 ) # Generate output output = generate(model, processor, formatted_prompt, image) print(output) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Pi
How to use mlx-community/clairic-flash-4bit with Pi:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "mlx-community/clairic-flash-4bit"
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "mlx-lm": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "mlx-community/clairic-flash-4bit" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent
How to use mlx-community/clairic-flash-4bit with Hermes Agent:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "mlx-community/clairic-flash-4bit"
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 mlx-community/clairic-flash-4bit
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use mlx-community/clairic-flash-4bit with OpenClaw:
Start the MLX server
# Install MLX LM: uv tool install mlx-lm # Start a local OpenAI-compatible server: mlx_lm.server --model "mlx-community/clairic-flash-4bit"
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 "mlx-community/clairic-flash-4bit" \ --custom-provider-id mlx-lm \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
File size: 3,178 Bytes
ae97c5d | 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 | """Conservative page orientation from detached dots beside Latin letter stems.
This examines glyph geometry only; it performs no text readout. Pages without
enough consistent evidence retain their supplied orientation.
"""
from __future__ import annotations
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from PIL import Image
def document_rotation(image: Image.Image) -> tuple[int, list[int]]:
import cv2
import numpy as np
page = image.convert("L")
page.thumbnail((1600, 1600))
pixels = np.asarray(page)
_, ink = cv2.threshold(pixels, 0, 255, cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)
_, _, stats, _ = cv2.connectedComponentsWithStats(ink, connectivity=8)
stats = stats[1:]
votes = [0, 0, 0, 0]
if not len(stats) or len(stats) > 12_000:
return 0, votes
width, height, area = stats[:, 2:].T
short, long = np.minimum(width, height), np.maximum(width, height)
stems = stats[(long >= 9) & (long <= 60) & (long >= 2.5 * short) & (area >= 8)]
dots = stats[(width <= 12) & (height <= 12) & (area >= 2) & (area <= 100)]
if not len(stems) or not len(dots) or len(stems) > 1200 or len(dots) > 2400:
return 0, votes
dx, dy, dw, dh, da = dots.T
for sx, sy, sw, sh, sa in stems:
vertical = sh > sw
across = dx + dw / 2 if vertical else dy + dh / 2
center = sx + sw / 2 if vertical else sy + sh / 2
start, length = (sy, sh) if vertical else (sx, sw)
left, size = (dy, dh) if vertical else (dx, dw)
thin = sw if vertical else sh
aligned = (abs(across - center) <= max(1.5, thin * 0.6)) & (
size <= length * 0.4
)
aligned &= (dw <= max(3, thin * 2)) if vertical else (dh <= max(3, thin * 2))
aligned &= da < sa * 0.7
before = (
aligned & (left + size <= start) & (start - left - size <= length * 0.45)
)
after = (
aligned
& (left >= start + length)
& (left - start - length <= length * 0.45)
)
if vertical:
votes[0] += int(np.any(before))
votes[2] += int(np.any(after))
else:
votes[1] += int(np.any(before))
votes[3] += int(np.any(after))
letters = (area >= 10) & (short >= 3) & (long <= 60) & (long <= 4 * short)
if int(letters.sum()) < 80:
return 0, votes
aspect = float(np.median(width[letters] / height[letters]))
if aspect < 0.8:
directions = (0, 2)
elif aspect > 1.25:
directions = (1, 3)
else:
return 0, votes
# Text-line direction excludes accidental dot/stem pairs in the other axis.
winner = max(directions, key=votes.__getitem__)
runner = votes[directions[0] if winner == directions[1] else directions[1]]
if (
votes[winner] < 12
or votes[winner] < 3 * max(1, runner)
or votes[winner] < 0.7 * (votes[winner] + runner)
):
return 0, votes
return winner * 90, votes
def orient_document(image: Image.Image) -> Image.Image:
rotation, _ = document_rotation(image)
return image.rotate(-rotation, expand=True) if rotation else image
|