Instructions to use h34v7/Qwanko3.5-27B-V2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use h34v7/Qwanko3.5-27B-V2 with Transformers:

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

pipe = pipeline("text-generation", model="h34v7/Qwanko3.5-27B-V2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("h34v7/Qwanko3.5-27B-V2")
model = AutoModelForCausalLM.from_pretrained("h34v7/Qwanko3.5-27B-V2", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use h34v7/Qwanko3.5-27B-V2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "h34v7/Qwanko3.5-27B-V2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "h34v7/Qwanko3.5-27B-V2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/h34v7/Qwanko3.5-27B-V2

SGLang

How to use h34v7/Qwanko3.5-27B-V2 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 "h34v7/Qwanko3.5-27B-V2" \
    --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": "h34v7/Qwanko3.5-27B-V2",
		"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 "h34v7/Qwanko3.5-27B-V2" \
        --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": "h34v7/Qwanko3.5-27B-V2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use h34v7/Qwanko3.5-27B-V2 with Docker Model Runner:
```
docker model run hf.co/h34v7/Qwanko3.5-27B-V2
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Qwanko3.5-27B-V2

Qwopus added depth of thinking, Anko added creativity the results is a model that retain 75% ~ 98% of parent skills in both logical and creative task.

Tool calling seems to improve, i'm using hermes tho. 262k context seems stable. sometimes its ponder on the internal chamber for a while but pretty quick once it's done pondering often do rapid tool calling i must say it is an improvement IMO.

My daily driver is using Q6_K with SSM at F32 for stability in long context and first 4 input and output at BF16, check it out here

With ik_llama.cpp -sm graph really helpfull speed things up and lower vram usage fit 262k context with f16 kv cache.

It ponder with itself for a long time dunno why. But it get the job done so i'm not complaining.

Please disable thinking mode for immersive story experience.

I feels like... disable thinking is really good its instantenous compared to chain of thought. Still you gonna need to intervene before the model do some thing but... so far as long as i tell exactly what todo or what happened, non-thinking work like a charm.

Merge Details

This is a merge of pre-trained language models created using mergekit.

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

allura-org/Qwen3.5-27B-Anko
huihui-ai/Huihui-Qwopus3.5-27B-v3-abliterated

Configuration

The following YAML configuration was used to produce this model:

merge_method: slerp
base_model: huihui-ai/Huihui-Qwopus3.5-27B-v3-abliterated
models:
  - model: allura-org/Qwen3.5-27B-Anko
parameters:
  t: 0.5

Downloads last month: 22

Safetensors

Model size

27B params

Tensor type

BF16

Model tree for h34v7/Qwanko3.5-27B-V2

Jackrong/Qwopus3.5-27B-v3

allura-org/Qwen3.5-27B-Anko

huihui-ai/Huihui-Qwopus3.5-27B-v3-abliterated

Merge model

this model

Quantizations

1 model