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"
Download clairic_runtime/media.py from mlx-community/clairic-flash-4bit: direct link, hf CLI and curl.
- Browser
- Download file 3.41 kB
-
https://huggingface.co/mlx-community/clairic-flash-4bit/resolve/main/clairic_runtime/media.py
- Command line
-
hf download hf://mlx-community/clairic-flash-4bit/clairic_runtime/media.py
-
curl -L -o media.py https://huggingface.co/mlx-community/clairic-flash-4bit/resolve/main/clairic_runtime/media.py
3.41 kB
| from __future__ import annotations | |
| from collections.abc import Sequence | |
| from typing import Any | |
| def batch_image_data(media: Sequence[Any]) -> list[Any]: | |
| if all((item is None for item in media)): | |
| return list(media) | |
| return [item if item is not None else [] for item in media] | |
| def _grid_identity(grid: Any) -> Any: | |
| if hasattr(grid, "tolist"): | |
| grid = grid.tolist() | |
| if isinstance(grid, (list, tuple)): | |
| return tuple((_grid_identity(value) for value in grid)) | |
| if isinstance(grid, (str, int, float, bool, type(None))): | |
| return grid | |
| return ("opaque", id(grid)) | |
| def _media_identity(media: Any) -> tuple[Any, ...]: | |
| if media is None: | |
| return ("text",) | |
| if not isinstance(media, list) or not media: | |
| return ("opaque", id(media)) | |
| identities = [] | |
| for item in media: | |
| if ( | |
| not isinstance(item, dict) | |
| or item.get("format") != "processor_output" | |
| or (not isinstance(item.get("clair_media_digest"), str)) | |
| or (item.get("pixel_values") is None) | |
| ): | |
| return ("opaque", id(media)) | |
| identities.append( | |
| ( | |
| item["clair_media_digest"], | |
| id(item["pixel_values"]), | |
| _grid_identity(item.get("image_grid_thw")), | |
| ) | |
| ) | |
| return ("prepared", tuple(identities)) | |
| def _common_prefix(tasks: list[tuple[str, list[int], Any]]) -> list[int]: | |
| (first, last) = (min((task[1] for task in tasks)), max((task[1] for task in tasks))) | |
| length = next( | |
| (i for (i, (left, right)) in enumerate(zip(first, last)) if left != right), | |
| min(len(first), len(last)), | |
| ) | |
| return list(first[:length]) | |
| def media_warmup_groups( | |
| tasks: list[tuple[str, list[int], Any]], | |
| *, | |
| default_salt: str, | |
| task_salts: dict[str, str] | None = None, | |
| annotations: dict[str, dict[str, Any]] | None = None, | |
| min_prefix_tokens: int = 128, | |
| ) -> list[dict[str, Any]]: | |
| (task_salts, annotations) = (task_salts or {}, annotations or {}) | |
| partitions: dict[tuple[str, tuple[Any, ...]], dict[str, Any]] = {} | |
| for task in tasks: | |
| (name, _, media) = task | |
| salt = task_salts.get(name, default_salt) | |
| key = (salt, _media_identity(media)) | |
| partitions.setdefault(key, {"media": media, "salt": salt, "tasks": []})[ | |
| "tasks" | |
| ].append(task) | |
| records = [] | |
| for partition in partitions.values(): | |
| members = partition["tasks"] | |
| if len(members) < 2: | |
| continue | |
| prefix = _common_prefix(members) | |
| if len(prefix) < min_prefix_tokens: | |
| prefix = [] | |
| parents: dict[str, list[tuple[str, list[int], Any]]] = {} | |
| for task in members: | |
| parent = annotations.get(task[0], {}).get("parent") | |
| if parent: | |
| parents.setdefault(parent, []).append(task) | |
| subgroups = [ | |
| _common_prefix(group) for group in parents.values() if len(group) > 1 | |
| ] | |
| subgroups = [group for group in subgroups if group and group != prefix] | |
| if prefix or subgroups: | |
| records.append( | |
| { | |
| "prefix_ids": prefix, | |
| "media": partition["media"], | |
| "salt": partition["salt"], | |
| "group_prefix_ids": subgroups, | |
| } | |
| ) | |
| return records | |