Text Generation
GGUF
Japanese
English
llama.cpp
Mixture of Experts
expert-pruning
intel-mac
cpu
local-agent
imatrix
conversational
Instructions to use miutti/intel-mac-local-llm with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use miutti/intel-mac-local-llm with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf miutti/intel-mac-local-llm:UD-Q2_K_XL # Run inference directly in the terminal: llama cli -hf miutti/intel-mac-local-llm:UD-Q2_K_XL
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf miutti/intel-mac-local-llm:UD-Q2_K_XL # Run inference directly in the terminal: llama cli -hf miutti/intel-mac-local-llm:UD-Q2_K_XL
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf miutti/intel-mac-local-llm:UD-Q2_K_XL # Run inference directly in the terminal: ./build/bin/llama cli -hf miutti/intel-mac-local-llm:UD-Q2_K_XL
Use Docker
docker model run hf.co/miutti/intel-mac-local-llm:UD-Q2_K_XL
- Llama new
- LM Studio
- Jan
- vLLM
How to use miutti/intel-mac-local-llm with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "miutti/intel-mac-local-llm" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "miutti/intel-mac-local-llm", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/miutti/intel-mac-local-llm:UD-Q2_K_XL
- Ollama
How to use miutti/intel-mac-local-llm with Ollama:
ollama run hf.co/miutti/intel-mac-local-llm:UD-Q2_K_XL
- Unsloth Desktop
- Pi
How to use miutti/intel-mac-local-llm with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf miutti/intel-mac-local-llm:UD-Q2_K_XL
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "miutti/intel-mac-local-llm:UD-Q2_K_XL" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use miutti/intel-mac-local-llm with Docker Model Runner:
docker model run hf.co/miutti/intel-mac-local-llm:UD-Q2_K_XL
- Lemonade
How to use miutti/intel-mac-local-llm with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull miutti/intel-mac-local-llm:UD-Q2_K_XL
Run and chat with the model
lemonade run user.intel-mac-local-llm-UD-Q2_K_XL
List all available models
lemonade list
- Hermes Agent
How to use miutti/intel-mac-local-llm with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf miutti/intel-mac-local-llm:UD-Q2_K_XL
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 miutti/intel-mac-local-llm:UD-Q2_K_XL
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use miutti/intel-mac-local-llm with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf miutti/intel-mac-local-llm:UD-Q2_K_XL
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 "miutti/intel-mac-local-llm:UD-Q2_K_XL" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
File size: 10,995 Bytes
df41178 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 | #!/usr/bin/env python3
import io
import json
import sys
import tempfile
import unittest
import urllib.error
from pathlib import Path
from unittest import mock
sys.path.insert(0, str(Path(__file__).resolve().parents[1]))
import public_bench as pb
class Response:
def __init__(self, status, body):
self.status = status
self.body = body
def __enter__(self):
return self
def __exit__(self, *args):
return False
def getcode(self):
return self.status
def read(self):
return self.body
def success_result(content="A", error=None):
return {
"content": content, "reasoning_chars": 0, "elapsed_sec": 0.01,
"timings": {}, "usage": None, "model": "test", "error": error,
"error_type": "test" if error else None, "http_status": 500 if error else 200,
"error_body": "failure" if error else "", "attempts": [], "retry_count": 0,
"prompt_hash": "hash",
}
class PublicBenchTests(unittest.TestCase):
def test_balanced_sample_is_exact_and_stable(self):
rows = [{"id": str(i), "domain": domain}
for i, domain in enumerate("aaabbbcccc")]
first = pb.balanced_sample(rows, "domain", 7, 123,
lambda row: row["id"])
second = pb.balanced_sample(list(reversed(rows)), "domain", 7, 123,
lambda row: row["id"])
self.assertEqual([x["id"] for x in first], [x["id"] for x in second])
self.assertEqual(len(first), 7)
counts = {domain: sum(x["domain"] == domain for x in first)
for domain in "abc"}
self.assertEqual(sum(counts.values()), 7)
self.assertLessEqual(max(counts.values()) - min(counts.values()), 1)
def test_client_retries_and_records_http_body(self):
calls = []
def opener(request, timeout):
calls.append(request)
if len(calls) == 1:
raise urllib.error.HTTPError(
request.full_url, 500, "server", {}, io.BytesIO(b"overloaded"))
return Response(200, b'{"choices":[{"message":{"content":"A"}}]}')
client = pb.LocalClient("http://example.test/v1/chat/completions", False,
8, 1, retries=2, retry_backoff=0, opener=opener)
result = client.call("question", "item-1")
self.assertEqual(result["content"], "A")
self.assertEqual(len(calls), 2)
self.assertEqual(result["retry_count"], 1)
self.assertEqual(result["attempts"][0]["status"], 500)
self.assertEqual(result["attempts"][0]["body_excerpt"], "overloaded")
self.assertIsNone(result["error"])
def test_client_does_not_repeat_reasoning_only_truncation(self):
calls = []
def opener(request, timeout):
calls.append(request)
return Response(
200,
b'{"choices":[{"finish_reason":"length","message":'
b'{"content":"","reasoning_content":"long thought"}}]}',
)
client = pb.LocalClient("http://example.test/v1/chat/completions", True,
8, 1, retries=3, retry_backoff=0, opener=opener)
result = client.call("question", "item-1")
self.assertEqual(len(calls), 1)
self.assertEqual(result["error_type"], "OutputTruncatedError")
self.assertEqual(result["finish_reason"], "length")
self.assertEqual(result["reasoning_chars"], len("long thought"))
def test_adaptive_client_votes_and_assigns_max_profile_to_aime(self):
class FakeClient:
seed = 100
def __init__(self):
self.calls = []
def call(self, prompt, item_id, **kwargs):
self.calls.append((prompt, item_id, kwargs))
answers = {"sample-1": "7", "sample-2": "8", "sample-3": "7"}
content = next((value for key, value in answers.items()
if key in item_id), "7")
return {
"content": content, "error": None, "error_type": None,
"http_status": 200, "error_body": "", "attempts": [],
"retry_count": 0, "elapsed_sec": 0.01, "timings": {},
"usage": None, "model": "test", "reasoning_chars": 10,
"request": kwargs,
}
profiles = {
"fixed": pb.RequestProfile("fixed", False, 8, 1),
"fast": pb.RequestProfile("fast", False, 8, 1),
"deep": pb.RequestProfile("deep", True, 16, 2),
"max": pb.RequestProfile("max", True, 32, 3),
}
fake = FakeClient()
client = pb.AdaptiveClient(
fake, "aime", "adaptive", profiles, pb.extract_integer,
samples=3, sample_temperature=0.35,
)
result = client.call("Solve this problem", "aime/1")
self.assertEqual(result["winner"], "7")
self.assertEqual(result["vote_counts"], {"7": 2, "8": 1})
self.assertEqual(result["profile"], "max")
self.assertEqual(result["request_count"], 3)
self.assertTrue(all(call[2]["think"] for call in fake.calls))
self.assertTrue(all(call[2]["max_tokens"] == 32 for call in fake.calls))
def test_adaptive_client_retries_fast_when_answer_format_is_invalid(self):
class FakeClient:
seed = 0
def __init__(self):
self.calls = []
def call(self, prompt, item_id, **kwargs):
self.calls.append((item_id, kwargs))
content = "not an answer" if "fallback" not in item_id else "B"
return {
"content": content, "error": None, "error_type": None,
"http_status": 200, "error_body": "", "attempts": [],
"retry_count": 0, "elapsed_sec": 0.01, "timings": {},
"usage": None, "model": "test", "reasoning_chars": 0,
"request": kwargs,
}
profiles = {
"fixed": pb.RequestProfile("fixed", False, 8, 1),
"fast": pb.RequestProfile("fast", False, 8, 1),
"deep": pb.RequestProfile("deep", True, 16, 2),
}
profiles["max"] = pb.RequestProfile("max", True, 32, 3)
fake = FakeClient()
client = pb.AdaptiveClient(
fake, "mmlu_pro", "adaptive", profiles,
lambda text: pb.extract_letter(text, "ABCD"),
)
result = client.call("Choose one", "mmlu/1")
self.assertEqual(result["content"], "B")
self.assertEqual(len(fake.calls), 2)
self.assertEqual(fake.calls[-1][1]["think"], False)
self.assertEqual(fake.calls[-1][1]["max_tokens"], 264)
def test_max_client_rescues_reasoning_only_response_with_fast_call(self):
class FakeClient:
seed = 0
def __init__(self):
self.calls = []
def call(self, prompt, item_id, **kwargs):
self.calls.append((item_id, kwargs))
if "fallback-fast" not in item_id:
return {
"item_id": item_id, "content": "", "error": "cut",
"error_type": "OutputTruncatedError", "elapsed_sec": 1,
"timings": {}, "request": kwargs,
}
return {
"item_id": item_id, "content": "C", "error": None,
"error_type": None, "elapsed_sec": 0.1, "timings": {},
"request": kwargs,
}
profiles = {
"fixed": pb.RequestProfile("fixed", False, 8, 1),
"fast": pb.RequestProfile("fast", False, 8, 1),
"deep": pb.RequestProfile("deep", True, 16, 2),
"max": pb.RequestProfile("max", True, 32, 3),
}
fake = FakeClient()
client = pb.AdaptiveClient(
fake, "gpqa_diamond", "max", profiles,
lambda text: pb.extract_letter(text, "ABCD"),
)
result = client.call("Choose one", "gpqa/1")
self.assertEqual(result["content"], "C")
self.assertTrue(result["fallback_used"])
self.assertEqual([kwargs["think"] for _, kwargs in fake.calls], [True, False])
def test_corrupt_cache_is_quarantined_and_refetched(self):
with tempfile.TemporaryDirectory() as directory:
out = Path(directory)
cache = out / "datasets" / "x__c__test.jsonl"
cache.parent.mkdir()
cache.write_text('{broken\n', encoding="utf-8")
with mock.patch.object(pb, "get_json", return_value={
"num_rows_total": 1, "rows": [{"row": {"id": 1}}]
}):
rows = pb.fetch_rows("x", "c", "test", out)
self.assertEqual(rows, [{"id": 1}])
self.assertTrue(list(cache.parent.glob(cache.name + ".corrupt-*")))
self.assertEqual(pb.read_jsonl(cache), [{"id": 1}])
def test_run_rows_retries_failed_existing_item_only(self):
class FakeClient:
def __init__(self, result):
self.result = result
self.calls = []
def call(self, prompt, item_id):
self.calls.append(item_id)
return dict(self.result)
rows = [{"id": "one", "answer": "A"}, {"id": "two", "answer": "A"}]
with tempfile.TemporaryDirectory() as directory:
out = Path(directory)
first_client = FakeClient(success_result("", "HTTPError"))
first = pb.run_rows("toy", rows, lambda row: row["id"],
lambda content: content or None, first_client, out,
lambda row: row["id"], gold_fn=lambda row: "A")
self.assertEqual(first["failed_n"], 2)
self.assertEqual(first_client.calls, ["one", "two"])
second_client = FakeClient(success_result("A"))
second = pb.run_rows("toy", rows, lambda row: row["id"],
lambda content: content, second_client, out,
lambda row: row["id"], gold_fn=lambda row: "A")
self.assertEqual(second["failed_n"], 0)
self.assertEqual(second["correct"], 2)
self.assertEqual(second_client.calls, ["one", "two"])
items = pb.read_jsonl(out / "toy.jsonl")
self.assertEqual(len(items), 2)
self.assertTrue(all(item["row_sha256"] for item in items))
def test_atomic_jsonl_is_valid_after_write(self):
with tempfile.TemporaryDirectory() as directory:
path = Path(directory) / "rows.jsonl"
pb.jsonl_dump(path, [{"x": 1}, {"x": 2}])
self.assertEqual(pb.read_jsonl(path), [{"x": 1}, {"x": 2}])
self.assertFalse(list(Path(directory).glob("*.tmp")))
if __name__ == "__main__":
unittest.main()
|