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-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/clairic-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-4bit") config = load_config("mlx-community/clairic-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-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-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-4bit" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent
How to use mlx-community/clairic-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-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-4bit
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use mlx-community/clairic-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-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-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: 6,505 Bytes
1a48d6b | 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 | from __future__ import annotations
import asyncio
import json
import uuid
from typing import TYPE_CHECKING
from . import (
ar,
batch_completion,
grounding,
image_rows,
printed_numbers,
printed_temporal,
)
from .backend import MAX_OUTPUT_TOKENS, ClairInferenceError, scalar_result
from .media import batch_image_data
from . import text_list_completion
if TYPE_CHECKING:
from typing import Any
from .large import Clairic27BModel
async def text_ar_all(
self: Clairic27BModel, record: dict[str, Any], eos: int
) -> dict[str, dict[str, Any]]:
tasks = record["_ar_tasks"]
salt = record["_cache_salt"]
row_gates = set(record.get("_source_row_gates", ()))
false_id = self.tokenizer.convert_tokens_to_ids("false") if row_gates else None
score_ids = [self.tokenizer.convert_tokens_to_ids("null")]
if false_id is not None:
score_ids.append(false_id)
results = await text_list_completion.generate(
self.engine,
record,
tasks,
eos,
self.tokenizer,
parse_result=scalar_result,
normalize=lambda record, predictions: printed_numbers.normalize_predictions(
record, printed_temporal.normalize_predictions(record, predictions)
),
input_ids=[ids for (_, ids, _) in tasks],
image_data=batch_image_data([media for (_, _, media) in tasks]),
rid=["ar-" + uuid.uuid4().hex for _ in tasks],
cache_salt=[
record.get("_ar_task_salts", {}).get(name, salt) for (name, _, _) in tasks
],
return_logprob=True,
logprob_start_len=-1,
token_ids_logprob=score_ids,
sampling_params=[
{
"temperature": 0,
"max_new_tokens": image_rows.output_limit(
record, name, len(ids), self.settings.max_length
)
if image_rows.is_inventory(record["fields"][name])
else record.get("_ar_max_output_tokens", MAX_OUTPUT_TOKENS),
"stop_token_ids": [eos],
**(
record["_ar_constraints"][name]
if name in record.get("_ar_constraints", {})
else {"json_schema": json.dumps(ar.schema(record["fields"][name]))}
),
"skip_special_tokens": False,
}
for (name, ids, _) in tasks
],
)
if usage := record.get("_completed_generation_usage"):
usage["tasks"].extend(tasks)
usage["readings"].extend(
(
{"tokens": int(result.get("meta_info", {}).get("completion_tokens", 0))}
for result in results
)
)
predictions = {
name: batch_completion.unused_prediction(result)
if result.get("_clair_unused_row_proposal")
else scalar_result(
result,
record["fields"][name],
self.tokenizer,
record["input"],
record.get("_ar_sources", {}).get(name, record.get("_ar_source"))
if grounding.applies(record["fields"][name], media)
else None,
excluded_token_id=false_id if name in row_gates else None,
provenance=record.get("_ar_sources", {}).get(name, record.get("_ar_source"))
if media is None and (not record["fields"][name].get("knowledge"))
else None,
)
for ((name, _, media), result) in zip(tasks, results, strict=True)
}
if any(
(
"confidence" not in prediction
for prediction in predictions.values()
if not prediction.get("_unused_row_proposal")
)
):
raise ClairInferenceError("Transformers omitted generation log probabilities")
if any(
(
"p_null" not in prediction
for prediction in predictions.values()
if not prediction.get("_unused_row_proposal")
)
):
raise ClairInferenceError("Transformers omitted null-token log probabilities")
if self.RECOVER_NULLS and (not record.get("_null_recovery")):
fields = {
name
for (name, prediction) in predictions.items()
if (
prediction["valid"]
and prediction["value"] is None
or prediction.get("failure_reason") == "not_in_source"
)
and (not record["fields"][name].get("knowledge"))
and (name not in record.get("_meta", {}).get("printed_temporal", {}))
and (name not in record.get("_meta", {}).get("printed_numbers", {}))
and (
record["fields"][name]["type"] == "number"
or (record["fields"][name]["type"] == "text" and record["input"])
)
and (
not record["_meta"]["extraction_contract"]["fields"]
.get(name, {})
.get("parent")
)
}
if fields:
repair = await asyncio.get_running_loop().run_in_executor(
self._encoder, self.null_recovery_record, record, fields
)
if repair["_ar_tasks"]:
alternatives = await self._ar_all(repair, eos)
retry_lengths = {
name: len(ids) for (name, ids, _) in repair["_ar_tasks"]
}
for name, alternative in alternatives.items():
original = predictions[name]
tokens = original["tokens"] + alternative["tokens"]
if (
alternative["valid"]
and alternative["value"] is not None
and (
alternative.get("grounding_method")
in ("literal_source", "whitespace_normalized_source")
or record["fields"][name]["type"] == "number"
or (
alternative.get("span") is not None
and alternative.get("confidence", 0.0)
> alternative.get("p_null", 1.0)
)
)
):
predictions[name] = {**alternative, "null_recovered": True}
predictions[name]["tokens"] = tokens
predictions[name]["retry_input_tokens"] = retry_lengths[name]
return predictions
|