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: 5,080 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 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 | from __future__ import annotations
import csv
import io
import re
from collections.abc import Callable
from html.parser import HTMLParser
from typing import Any
_ONE_PER_ROW = re.compile(
"\\bone\\s+(?:[\\w-]+\\s+){0,5}per\\s+(?:[\\w-]+\\s+){0,5}row\\b", re.IGNORECASE
)
class _Table(HTMLParser):
def __init__(self) -> None:
super().__init__(convert_charrefs=False)
self.tables = self.depth = 0
self.rows: list[set[str]] = []
self.closed = False
self.row_header = False
self.in_cell = False
self.ambiguous = False
def handle_starttag(self, tag: str, attrs: list[tuple[str, str | None]]) -> None:
if self.in_cell and tag in ("br", "div", "p", "ul", "ol", "li", "table"):
self.ambiguous = True
if tag == "table":
self.tables += 1
self.depth += 1
if self.depth == 1:
if tag == "tr":
self.rows.append(set())
elif tag in ("td", "th") and self.rows:
self.in_cell = True
if any(
(
key in ("rowspan", "colspan") and value != "1"
for (key, value) in attrs
)
):
self.ambiguous = True
self.rows[-1].add(tag)
if len(self.rows) == 1 and (
dict(attrs).get("scope") or ""
).casefold() in ("row", "rowgroup"):
self.row_header = True
def handle_endtag(self, tag: str) -> None:
if tag in ("td", "th"):
self.in_cell = False
if tag == "table":
self.depth -= 1
self.closed = self.depth == 0
def handle_data(self, data: str) -> None:
if (
self.in_cell
and len([line for line in data.splitlines() if line.strip()]) > 1
):
self.ambiguous = True
def table_rows(source: str) -> int | None:
if re.search("<table(?:\\s|>)", source, re.IGNORECASE):
parser = _Table()
try:
parser.feed(source)
parser.close()
except (ValueError, AssertionError):
return None
if (
parser.tables != 1
or not parser.closed
or parser.depth != 0
or (not parser.rows)
or (parser.rows[0] != {"th"})
or parser.row_header
or parser.ambiguous
):
return None
return len(parser.rows) - 1
lines = source.splitlines()
if len(lines) < 2:
return None
for delimiter in (",", "\t", ";", "|"):
if delimiter not in lines[0]:
continue
try:
rows = list(
csv.reader(io.StringIO(source), delimiter=delimiter, strict=True)
)
except csv.Error:
continue
if (
not rows
or len(rows[0]) < 2
or any((not row or len(row) != len(rows[0]) for row in rows))
or any((not cell.strip() or re.search("\\d", cell) for cell in rows[0]))
or (not any((re.search("\\d", cell) for row in rows[1:] for cell in row)))
):
continue
return len(rows) - 1
return None
def list_bounds(
source: str, fields: dict[str, dict[str, Any]], *, has_images: bool = False
) -> dict[str, int]:
if has_images:
return {}
parents = {}
for name, spec in fields.items():
instructions = spec.get("instructions", "")
if (
spec.get("type") == "list"
and spec.get("items") in ("text", "number")
and (not spec.get("knowledge"))
and _ONE_PER_ROW.search(instructions)
and (
not re.search(
"\\b(?:not|never|multiple|two|three)\\b|don't",
instructions,
re.IGNORECASE,
)
)
):
parents[name] = spec
count = table_rows(source) if parents else None
return {name: count for name in parents} if count is not None else {}
def bounded_predictions(
predictions: dict[str, dict[str, Any]],
fields: dict[str, dict[str, Any]],
bounds: dict[str, int],
item_name: Callable[[str, int], str],
) -> dict[str, dict[str, Any]]:
if not bounds:
return predictions
result = dict(predictions)
for parent, count in bounds.items():
name = item_name(parent, count + 1)
if name not in fields or name not in result:
continue
previous = result[name]
if previous.get("value") is None and previous.get("valid", True):
continue
result[name] = {
**previous,
"value": None,
"valid": True,
"complete": True,
"span": None,
"confidence": 1.0,
"log_probability": 0.0,
"confidence_method": "source_structure",
"grounding_method": "source_row_bound",
}
return result
|