Instructions to use nkasmanoff/qwen3.6-27b-opencode-lora-merged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nkasmanoff/qwen3.6-27b-opencode-lora-merged with Transformers:

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

pipe = pipeline("text-generation", model="nkasmanoff/qwen3.6-27b-opencode-lora-merged")
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)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("nkasmanoff/qwen3.6-27b-opencode-lora-merged")
model = AutoModelForMultimodalLM.from_pretrained("nkasmanoff/qwen3.6-27b-opencode-lora-merged")
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=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nkasmanoff/qwen3.6-27b-opencode-lora-merged with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nkasmanoff/qwen3.6-27b-opencode-lora-merged"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nkasmanoff/qwen3.6-27b-opencode-lora-merged",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/nkasmanoff/qwen3.6-27b-opencode-lora-merged

SGLang

How to use nkasmanoff/qwen3.6-27b-opencode-lora-merged 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 "nkasmanoff/qwen3.6-27b-opencode-lora-merged" \
    --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": "nkasmanoff/qwen3.6-27b-opencode-lora-merged",
		"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 "nkasmanoff/qwen3.6-27b-opencode-lora-merged" \
        --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": "nkasmanoff/qwen3.6-27b-opencode-lora-merged",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use nkasmanoff/qwen3.6-27b-opencode-lora-merged with Docker Model Runner:
```
docker model run hf.co/nkasmanoff/qwen3.6-27b-opencode-lora-merged
```

Qwen3.6-27B-OpenCode-LoRA-Merged

This is Qwen3.6-27B fine-tuned with a LoRA adapter for opencode coding assistant tasks, with the LoRA weights merged into the base model for efficient inference.

Model Details

Base Model: Qwen/Qwen3.6-27B (27B params, BF16)
Fine-tuning: LoRA adapter trained for opencode coding assistant interactions
Status: Weights merged (no separate LoRA adapter needed at inference time)
Architecture: Qwen3.5 architecture with hybrid attention (Mamba-2 linear attention + full attention every 4 layers)
Context Length: 262,144 tokens
Modality: Text + Vision (multimodal)

Usage

The model can be used with transformers and accepts the same chat template as the base Qwen3.6 model.

For best results with opencode, configure it in your opencode.json:

{
  "model": "nkasmanoff/qwen3.6-27b-opencode-lora-merged"
}

Training

This model was fine-tuned using LoRA (Low-Rank Adaptation) on opencode-specific training data to improve performance on AI-assisted coding tasks. The adapter weights have been merged into the base model for seamless deployment.

Downloads last month: 278

Safetensors

Model size

27B params

Tensor type

BF16

Model tree for nkasmanoff/qwen3.6-27b-opencode-lora-merged

Base model

Qwen/Qwen3.6-27B

Adapter

(179)

this model