Instructions to use Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp with Transformers:

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

pipe = pipeline("text-generation", model="Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp")
model = AutoModelForCausalLM.from_pretrained("Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp

SGLang

How to use Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp 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 "Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp with Docker Model Runner:
```
docker model run hf.co/Nitral-Archive/Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp
```

Irix-12B-Qwen3-Tokenizer-Homunculus-Slerp / README.md

Nitral-AI

Update README.md

9e55096 verified about 1 year ago

preview code

Raw

History Blame Contribute Delete

1.07 kB

metadata

base_model:
  - arcee-ai/Homunculus
  - Nitral-Archive/Irix-12B-MS-Qwen3-Tokenizer
library_name: transformers
tags:
  - mergekit
  - merge

I would like to train this out of personal curiosity, based on some qlora test runs for healing. However, due to financial constraints this will come at a later date.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: Nitral-Archive/Irix-12B-MS-Qwen3-Tokenizer
        layer_range: [0, 40]
      - model: arcee-ai/Homunculus
        layer_range: [0, 40]
merge_method: slerp
base_model: Nitral-Archive/Irix-12B-MS-Qwen3-Tokenizer
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.69
dtype: bfloat16