Instructions to use arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX 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("arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX")
config = load_config("arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX")

# 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

How to use arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX"

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": "arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX 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 "arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX"

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 arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX"

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 "arsis-dev/ukisai-Swift-Qwen3.8-27b-MTPLX" \
  --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"

arsis-dev commited on 13 days ago

Commit

c754731

verified ·

1 Parent(s): dc020d2

Publish mtplx_runtime.json without local paths

Browse files

Files changed (1) hide show

mtplx_runtime.json +147 -0

mtplx_runtime.json ADDED Viewed

	@@ -0,0 +1,147 @@

+{
+  "arch_id": "qwen3-next-mtp",
+  "artifact_role": "forge-local",
+  "base_trunk": "ukisai/Swift-Qwen3.8-27b",
+  "exactness_baseline": {},
+  "forge_provenance": {
+    "forge_inputs": {
+      "mtp_source_path": "<redacted>/ukisai--Swift-Qwen3.8-27b",
+      "trunk_path": "<redacted>/ukisai-Swift-Qwen3.8-27b-MTPLX"
+    },
+    "forge_recipe": {
+      "body_bits": 4,
+      "body_dtype": "auto",
+      "body_group_size": 64,
+      "body_mode": "affine",
+      "mtp_policy": "keep_bf16"
+    },
+    "forged_at": "2026-09-14T16:09:09+02:00",
+    "forged_locally": true,
+    "mtp_contract": {
+      "base_hidden_variant": "post_norm",
+      "concat_order": "embedding_hidden",
+      "hidden_variant": "post_norm",
+      "mtp_position_mode": "local",
+      "mtp_quant_group_size": 64,
+      "mtp_quant_mode": "affine"
+    },
+    "mtplx_version": "2.11.2",
+    "published_to_hf": null,
+    "source_format": "bf16_native",
+    "source_repo": "ukisai/Swift-Qwen3.8-27b",
+    "source_sha": "1b30aaaf753fe5c1cb51ada2ea0367a53445359c"
+  },
+  "mtp_contract": {
+    "base_hidden_variant": "post_norm",
+    "concat_order": "embedding_hidden",
+    "hidden_variant": "post_norm",
+    "mtp_position_mode": "local",
+    "mtp_quant_group_size": 64,
+    "mtp_quant_mode": "affine"
+  },
+  "mtp_depth_max": 3,
+  "mtp_sidecar": "bf16",
+  "mtplx_version": "2.11.2",
+  "recommended_profile": "sustained",
+  "sampler": {
+    "temperature": 1.0,
+    "top_k": 20,
+    "top_p": 0.95
+  },
+  "speed_evidence": {
+    "acceptance_by_depth": [
+      0.9669811320754716,
+      0.9339622641509434,
+      0.8773584905660378
+    ],
+    "acceptance_collapsed": [],
+    "artifact_fingerprint": "sha256:33122592b7c324a965b4bb8f745f6d8e41e0a550b509d2e8882c0c7b194edfe3",
+    "depth": 3,
+    "failure_reasons": [],
+    "forge_verify_rows": [
+      {
+        "acceptance_by_position": [],
+        "depth": 0,
+        "finish_reasons": {
+          "stop": 1
+        },
+        "hit_token_budget": false,
+        "hit_token_budget_count": 0,
+        "multiplier_vs_ar": 1.0,
+        "quality_passed": true,
+        "tok_s": 29.377510075912657,
+        "verify_time_s": 25.18232329195598
+      },
+      {
+        "acceptance_by_position": [
+          0.9644194756554307
+        ],
+        "depth": 1,
+        "finish_reasons": {
+          "stop": 1
+        },
+        "hit_token_budget": false,
+        "hit_token_budget_count": 0,
+        "multiplier_vs_ar": 1.5828654412370624,
+        "quality_passed": true,
+        "tok_s": 46.50064544875573,
+        "verify_time_s": 21.02465662942268
+      },
+      {
+        "acceptance_by_position": [
+          1.0,
+          0.9548387096774194
+        ],
+        "depth": 2,
+        "finish_reasons": {
+          "stop": 1
+        },
+        "hit_token_budget": false,
+        "hit_token_budget_count": 0,
+        "multiplier_vs_ar": 2.107085140405572,
+        "quality_passed": true,
+        "tok_s": 61.900914943070525,
+        "verify_time_s": 8.45314194925595
+      },
+      {
+        "acceptance_by_position": [
+          0.9669811320754716,
+          0.9339622641509434,
+          0.8773584905660378
+        ],
+        "depth": 3,
+        "finish_reasons": {
+          "stop": 1
+        },
+        "hit_token_budget": false,
+        "hit_token_budget_count": 0,
+        "multiplier_vs_ar": 2.220396896668368,
+        "quality_passed": true,
+        "tok_s": 65.22973220440018,
+        "verify_time_s": 10.251036165573169
+      }
+    ],
+    "greedy_diagnostic": {
+      "tok_s": 29.377510075912657
+    },
+    "quality_rejected": [],
+    "tok_s": [
+      65.22973220440018
+    ],
+    "verdict": "mtp_depth_wins"
+  },
+  "verified_on": {
+    "hardware": "macOS-26.6.2-arm64-arm-64bit-Mach-O",
+    "machine_arch": "arm64",
+    "macos": "26.6.2",
+    "model": "ukisai-Swift-Qwen3.8-27b-MTPLX",
+    "timestamp": "2026-09-14T16:09:09+02:00"
+  },
+  "vision": {
+    "prefix": "vision_tower.",
+    "shards": [
+      "model-vision.safetensors"
+    ],
+    "tensor_count": 333
+  }
+}