Instructions to use jstarks/Qwen3.5-9B-Weight-Space-Programming with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="jstarks/Qwen3.5-9B-Weight-Space-Programming")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("jstarks/Qwen3.5-9B-Weight-Space-Programming")
model = AutoModelForMultimodalLM.from_pretrained("jstarks/Qwen3.5-9B-Weight-Space-Programming", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming 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 jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

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 jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

Use Docker

docker model run hf.co/jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jstarks/Qwen3.5-9B-Weight-Space-Programming"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jstarks/Qwen3.5-9B-Weight-Space-Programming",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

SGLang

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "jstarks/Qwen3.5-9B-Weight-Space-Programming" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jstarks/Qwen3.5-9B-Weight-Space-Programming",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "jstarks/Qwen3.5-9B-Weight-Space-Programming" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jstarks/Qwen3.5-9B-Weight-Space-Programming",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with Ollama:
```
ollama run hf.co/jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M
```
Unsloth Desktop

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

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": "jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with Docker Model Runner:
```
docker model run hf.co/jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M
```

Lemonade

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.5-9B-Weight-Space-Programming-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

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 jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use jstarks/Qwen3.5-9B-Weight-Space-Programming with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M

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 "jstarks/Qwen3.5-9B-Weight-Space-Programming:Q4_K_M" \
  --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"

Qwen3.5-9B-Weight-Space-Programming / scripts /model_registry.py

jstarks

Upload WSP v3.0 qwen35_9b GGUF/BF16 models, 9-subspace adapters, safetensors patches, runner, and 100-task results

88132b4 verified about 6 hours ago

Raw History Blame Contribute Delete

7.34 kB

"""Centralized registry of the 5 candidate models for WSP v3.0 (9 Subspaces)."""

from __future__ import annotations

from dataclasses import dataclass
from pathlib import Path
from typing import List, Optional

PIPELINE_DIR = Path(r"C:\Users\jhjwi\.gemini\antigravity\scratch\wsp-5model-pipeline")
MODELS_DIR = PIPELINE_DIR / "models"


@dataclass(frozen=True)
class ModelSpec:
    slug: str
    display_name: str
    hf_repo: str
    gguf_repo: Optional[str]
    gguf_file: str
    gguf_arch: str
    num_layers: int
    hidden_size: int
    intermediate_size: int
    tied_embeddings: bool
    rmsnorm_convention: str  # "1+w" or "w"
    mlp_activation: str  # "swiglu" | "relu2" | "geglu"
    mid_layers: List[int]
    late_layers: List[int]
    embed_key: str
    lm_head_key: Optional[str]
    norm_f_key: str
    layer_norm_tmpl: str
    post_ffn_norm_tmpl: Optional[str]
    gate_proj_tmpl: Optional[str]
    up_proj_tmpl: str
    down_proj_tmpl: str
    reasoning_format: Optional[str]
    enable_thinking_kwargs: dict
    control_family: str  # "qwen35_xml" | "gemma4"
    local_hf_dir: Optional[Path] = None
    local_gguf_path: Optional[Path] = None

    @property
    def work_dir(self) -> Path:
        return MODELS_DIR / self.slug

    @property
    def gguf_path(self) -> Path:
        return self.local_gguf_path if self.local_gguf_path else (self.work_dir / self.gguf_file)

    @property
    def adapter_dir(self) -> Path:
        return self.work_dir / "adapters"


MODEL_SPECS: List[ModelSpec] = [
    ModelSpec(
        slug="qwen35_2b",
        display_name="Qwen/Qwen3.5-2B",
        hf_repo="Qwen/Qwen3.5-2B",
        gguf_repo="unsloth/Qwen3.5-2B-GGUF",
        gguf_file="Qwen3.5-2B-Q4_K_M.gguf",
        gguf_arch="qwen35",
        num_layers=24,
        hidden_size=2048,
        intermediate_size=6144,
        tied_embeddings=True,
        rmsnorm_convention="1+w",
        mlp_activation="swiglu",
        mid_layers=[3, 7, 11, 15],
        late_layers=[18, 19, 21, 22, 23],
        embed_key="model.language_model.embed_tokens.weight",
        lm_head_key=None,
        norm_f_key="model.language_model.norm.weight",
        layer_norm_tmpl="model.language_model.layers.{l}.post_attention_layernorm.weight",
        post_ffn_norm_tmpl=None,
        gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
        up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
        down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
        reasoning_format="deepseek",
        enable_thinking_kwargs={},
        control_family="qwen35_xml",
    ),
    ModelSpec(
        slug="qwen35_4b",
        display_name="Qwen/Qwen3.5-4B",
        hf_repo="Qwen/Qwen3.5-4B",
        gguf_repo=None,
        gguf_file="Qwen3.5-4B-Q4_K_M.gguf",
        gguf_arch="qwen35",
        num_layers=32,
        hidden_size=2560,
        intermediate_size=9216,
        tied_embeddings=True,
        rmsnorm_convention="1+w",
        mlp_activation="swiglu",
        mid_layers=[7, 11, 15, 19],
        late_layers=[23, 27, 29, 30, 31],
        embed_key="model.language_model.embed_tokens.weight",
        lm_head_key=None,
        norm_f_key="model.language_model.norm.weight",
        layer_norm_tmpl="model.language_model.layers.{l}.post_attention_layernorm.weight",
        post_ffn_norm_tmpl=None,
        gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
        up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
        down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
        reasoning_format="deepseek",
        enable_thinking_kwargs={},
        control_family="qwen35_xml",
        local_hf_dir=Path(r"C:\Users\jhjwi\.gemini\antigravity\scratch\qwen35-4b-wsp\hf_model"),
        local_gguf_path=Path(r"C:\Users\jhjwi\.gemini\antigravity\scratch\qwen35-4b-wsp\gguf\Qwen3.5-4B-Q4_K_M.gguf"),
    ),
    ModelSpec(
        slug="nemotron3_4b",
        display_name="nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16",
        hf_repo="nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16",
        gguf_repo="nvidia/NVIDIA-Nemotron-3-Nano-4B-GGUF",
        gguf_file="NVIDIA-Nemotron3-Nano-4B-Q4_K_M.gguf",
        gguf_arch="nemotron_h",
        num_layers=42,
        hidden_size=3136,
        intermediate_size=12544,
        tied_embeddings=False,
        rmsnorm_convention="w",
        mlp_activation="relu2",
        mid_layers=[13, 18, 22, 25],
        late_layers=[29, 33, 37, 39, 41],
        embed_key="backbone.embeddings.weight",
        lm_head_key="lm_head.weight",
        norm_f_key="backbone.norm_f.weight",
        layer_norm_tmpl="backbone.layers.{l}.norm.weight",
        post_ffn_norm_tmpl=None,
        gate_proj_tmpl=None,
        up_proj_tmpl="backbone.layers.{l}.mixer.up_proj.weight",
        down_proj_tmpl="backbone.layers.{l}.mixer.down_proj.weight",
        reasoning_format="deepseek",
        enable_thinking_kwargs={"enable_thinking": True},
        control_family="qwen35_xml",
    ),
    ModelSpec(
        slug="qwen35_9b",
        display_name="Qwen/Qwen3.5-9B",
        hf_repo="Qwen/Qwen3.5-9B",
        gguf_repo="unsloth/Qwen3.5-9B-GGUF",
        gguf_file="Qwen3.5-9B-Q4_K_M.gguf",
        gguf_arch="qwen35",
        num_layers=32,
        hidden_size=4096,
        intermediate_size=12288,
        tied_embeddings=False,
        rmsnorm_convention="1+w",
        mlp_activation="swiglu",
        mid_layers=[7, 11, 15, 19],
        late_layers=[23, 27, 29, 30, 31],
        embed_key="model.language_model.embed_tokens.weight",
        lm_head_key="lm_head.weight",
        norm_f_key="model.language_model.norm.weight",
        layer_norm_tmpl="model.language_model.layers.{l}.post_attention_layernorm.weight",
        post_ffn_norm_tmpl=None,
        gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
        up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
        down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
        reasoning_format="deepseek",
        enable_thinking_kwargs={},
        control_family="qwen35_xml",
    ),
    ModelSpec(
        slug="gemma4_12b",
        display_name="google/gemma-4-12B-it",
        hf_repo="google/gemma-4-12B-it",
        gguf_repo="unsloth/gemma-4-12b-it-GGUF",
        gguf_file="gemma-4-12b-it-Q4_K_M.gguf",
        gguf_arch="gemma4",
        num_layers=48,
        hidden_size=3840,
        intermediate_size=15360,
        tied_embeddings=True,
        rmsnorm_convention="w",
        mlp_activation="geglu",
        mid_layers=[11, 17, 23, 29],
        late_layers=[35, 41, 45, 46, 47],
        embed_key="model.language_model.embed_tokens.weight",
        lm_head_key=None,
        norm_f_key="model.language_model.norm.weight",
        layer_norm_tmpl="model.language_model.layers.{l}.pre_feedforward_layernorm.weight",
        post_ffn_norm_tmpl="model.language_model.layers.{l}.post_feedforward_layernorm.weight",
        gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
        up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
        down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
        reasoning_format=None,
        enable_thinking_kwargs={"enable_thinking": True},
        control_family="gemma4",
    ),
]

	"""Centralized registry of the 5 candidate models for WSP v3.0 (9 Subspaces)."""

	from __future__ import annotations

	from dataclasses import dataclass
	from pathlib import Path
	from typing import List, Optional

	PIPELINE_DIR = Path(r"C:\Users\jhjwi\.gemini\antigravity\scratch\wsp-5model-pipeline")
	MODELS_DIR = PIPELINE_DIR / "models"


	@dataclass(frozen=True)
	class ModelSpec:
	slug: str
	display_name: str
	hf_repo: str
	gguf_repo: Optional[str]
	gguf_file: str
	gguf_arch: str
	num_layers: int
	hidden_size: int
	intermediate_size: int
	tied_embeddings: bool
	rmsnorm_convention: str # "1+w" or "w"
	mlp_activation: str # "swiglu" \| "relu2" \| "geglu"
	mid_layers: List[int]
	late_layers: List[int]
	embed_key: str
	lm_head_key: Optional[str]
	norm_f_key: str
	layer_norm_tmpl: str
	post_ffn_norm_tmpl: Optional[str]
	gate_proj_tmpl: Optional[str]
	up_proj_tmpl: str
	down_proj_tmpl: str
	reasoning_format: Optional[str]
	enable_thinking_kwargs: dict
	control_family: str # "qwen35_xml" \| "gemma4"
	local_hf_dir: Optional[Path] = None
	local_gguf_path: Optional[Path] = None

	@property
	def work_dir(self) -> Path:
	return MODELS_DIR / self.slug

	@property
	def gguf_path(self) -> Path:
	return self.local_gguf_path if self.local_gguf_path else (self.work_dir / self.gguf_file)

	@property
	def adapter_dir(self) -> Path:
	return self.work_dir / "adapters"


	MODEL_SPECS: List[ModelSpec] = [
	ModelSpec(
	slug="qwen35_2b",
	display_name="Qwen/Qwen3.5-2B",
	hf_repo="Qwen/Qwen3.5-2B",
	gguf_repo="unsloth/Qwen3.5-2B-GGUF",
	gguf_file="Qwen3.5-2B-Q4_K_M.gguf",
	gguf_arch="qwen35",
	num_layers=24,
	hidden_size=2048,
	intermediate_size=6144,
	tied_embeddings=True,
	rmsnorm_convention="1+w",
	mlp_activation="swiglu",
	mid_layers=[3, 7, 11, 15],
	late_layers=[18, 19, 21, 22, 23],
	embed_key="model.language_model.embed_tokens.weight",
	lm_head_key=None,
	norm_f_key="model.language_model.norm.weight",
	layer_norm_tmpl="model.language_model.layers.{l}.post_attention_layernorm.weight",
	post_ffn_norm_tmpl=None,
	gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
	up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
	down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
	reasoning_format="deepseek",
	enable_thinking_kwargs={},
	control_family="qwen35_xml",
	),
	ModelSpec(
	slug="qwen35_4b",
	display_name="Qwen/Qwen3.5-4B",
	hf_repo="Qwen/Qwen3.5-4B",
	gguf_repo=None,
	gguf_file="Qwen3.5-4B-Q4_K_M.gguf",
	gguf_arch="qwen35",
	num_layers=32,
	hidden_size=2560,
	intermediate_size=9216,
	tied_embeddings=True,
	rmsnorm_convention="1+w",
	mlp_activation="swiglu",
	mid_layers=[7, 11, 15, 19],
	late_layers=[23, 27, 29, 30, 31],
	embed_key="model.language_model.embed_tokens.weight",
	lm_head_key=None,
	norm_f_key="model.language_model.norm.weight",
	layer_norm_tmpl="model.language_model.layers.{l}.post_attention_layernorm.weight",
	post_ffn_norm_tmpl=None,
	gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
	up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
	down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
	reasoning_format="deepseek",
	enable_thinking_kwargs={},
	control_family="qwen35_xml",
	local_hf_dir=Path(r"C:\Users\jhjwi\.gemini\antigravity\scratch\qwen35-4b-wsp\hf_model"),
	local_gguf_path=Path(r"C:\Users\jhjwi\.gemini\antigravity\scratch\qwen35-4b-wsp\gguf\Qwen3.5-4B-Q4_K_M.gguf"),
	),
	ModelSpec(
	slug="nemotron3_4b",
	display_name="nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16",
	hf_repo="nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16",
	gguf_repo="nvidia/NVIDIA-Nemotron-3-Nano-4B-GGUF",
	gguf_file="NVIDIA-Nemotron3-Nano-4B-Q4_K_M.gguf",
	gguf_arch="nemotron_h",
	num_layers=42,
	hidden_size=3136,
	intermediate_size=12544,
	tied_embeddings=False,
	rmsnorm_convention="w",
	mlp_activation="relu2",
	mid_layers=[13, 18, 22, 25],
	late_layers=[29, 33, 37, 39, 41],
	embed_key="backbone.embeddings.weight",
	lm_head_key="lm_head.weight",
	norm_f_key="backbone.norm_f.weight",
	layer_norm_tmpl="backbone.layers.{l}.norm.weight",
	post_ffn_norm_tmpl=None,
	gate_proj_tmpl=None,
	up_proj_tmpl="backbone.layers.{l}.mixer.up_proj.weight",
	down_proj_tmpl="backbone.layers.{l}.mixer.down_proj.weight",
	reasoning_format="deepseek",
	enable_thinking_kwargs={"enable_thinking": True},
	control_family="qwen35_xml",
	),
	ModelSpec(
	slug="qwen35_9b",
	display_name="Qwen/Qwen3.5-9B",
	hf_repo="Qwen/Qwen3.5-9B",
	gguf_repo="unsloth/Qwen3.5-9B-GGUF",
	gguf_file="Qwen3.5-9B-Q4_K_M.gguf",
	gguf_arch="qwen35",
	num_layers=32,
	hidden_size=4096,
	intermediate_size=12288,
	tied_embeddings=False,
	rmsnorm_convention="1+w",
	mlp_activation="swiglu",
	mid_layers=[7, 11, 15, 19],
	late_layers=[23, 27, 29, 30, 31],
	embed_key="model.language_model.embed_tokens.weight",
	lm_head_key="lm_head.weight",
	norm_f_key="model.language_model.norm.weight",
	layer_norm_tmpl="model.language_model.layers.{l}.post_attention_layernorm.weight",
	post_ffn_norm_tmpl=None,
	gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
	up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
	down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
	reasoning_format="deepseek",
	enable_thinking_kwargs={},
	control_family="qwen35_xml",
	),
	ModelSpec(
	slug="gemma4_12b",
	display_name="google/gemma-4-12B-it",
	hf_repo="google/gemma-4-12B-it",
	gguf_repo="unsloth/gemma-4-12b-it-GGUF",
	gguf_file="gemma-4-12b-it-Q4_K_M.gguf",
	gguf_arch="gemma4",
	num_layers=48,
	hidden_size=3840,
	intermediate_size=15360,
	tied_embeddings=True,
	rmsnorm_convention="w",
	mlp_activation="geglu",
	mid_layers=[11, 17, 23, 29],
	late_layers=[35, 41, 45, 46, 47],
	embed_key="model.language_model.embed_tokens.weight",
	lm_head_key=None,
	norm_f_key="model.language_model.norm.weight",
	layer_norm_tmpl="model.language_model.layers.{l}.pre_feedforward_layernorm.weight",
	post_ffn_norm_tmpl="model.language_model.layers.{l}.post_feedforward_layernorm.weight",
	gate_proj_tmpl="model.language_model.layers.{l}.mlp.gate_proj.weight",
	up_proj_tmpl="model.language_model.layers.{l}.mlp.up_proj.weight",
	down_proj_tmpl="model.language_model.layers.{l}.mlp.down_proj.weight",
	reasoning_format=None,
	enable_thinking_kwargs={"enable_thinking": True},
	control_family="gemma4",
	),
	]