Instructions to use virtuous7373/Lambent-Mira-Iris-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use virtuous7373/Lambent-Mira-Iris-v2 with Transformers:

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

pipe = pipeline("image-text-to-text", model="virtuous7373/Lambent-Mira-Iris-v2")
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("virtuous7373/Lambent-Mira-Iris-v2")
model = AutoModelForMultimodalLM.from_pretrained("virtuous7373/Lambent-Mira-Iris-v2", 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=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use virtuous7373/Lambent-Mira-Iris-v2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "virtuous7373/Lambent-Mira-Iris-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": "virtuous7373/Lambent-Mira-Iris-v2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/virtuous7373/Lambent-Mira-Iris-v2

SGLang

How to use virtuous7373/Lambent-Mira-Iris-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 "virtuous7373/Lambent-Mira-Iris-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": "virtuous7373/Lambent-Mira-Iris-v2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "virtuous7373/Lambent-Mira-Iris-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": "virtuous7373/Lambent-Mira-Iris-v2",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use virtuous7373/Lambent-Mira-Iris-v2 with Docker Model Runner:
```
docker model run hf.co/virtuous7373/Lambent-Mira-Iris-v2
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

All credits go to Lambent for the wonderful model series.
They did all the work, I just ran a few commands.

ϻira Iʀιs-v2-27β

In ancient Greek religion and mythology, Iris is a daughter of the gods Thaumas and Electra, and the personification of the rainbow. She is depicted as a beautiful golden goddess, with wings (for flight), sandals, a staff, and a tunic. There is often much art, such as vases and murals, shown of her because she is worshiped and very beautiful. She was also seen briefly in Homer's Legacies. Iris is generally seen as Hera's personal servant and messenger. Iris is very swift and fast and she generally travels by rainbow. It is even said that the reason the rainbow comes and goes so fast is because Iris only uses it for travel and since she is so fast, it is only in the sky for a small amount of time. Iris

Samplers [barebones lcpp webui]
Temp: 1 | Top-K: 0 | Top-P: 1.0 | Min-P: 0.025

Notes:

- I wanted creativity & uniqueness over general intelligence, maybe good standalone or as merge food.

- This is an extremely aggressive 5-part multistage merge [Magic > Magic > Varcross > RCZ > RCZ]

- Intelligence and overall continuity may be low.

- Possibly whips out a single Russian or Gibberish word in the middle of responses. I've only noticed this in the Koboldcpp backend/frontend. I haven't tested it a whole lot, though.

2026-03-02:23:19:03 INFO     [loggers.evaluation_tracker:119] Saving per-task samples to eval_results/Lambent-Mira-Iris-v2-Q5_K_X/*.jsonl
local-chat-completions ({'base_url': 'http://localhost:1234/v1/chat/completions', 'model': 'Lambent-Mira-Iris-v2-Q5_K_X', 'tokenizer': '$MODELS/google_gemma-3-27b-it', 'num_concurrent': 1}), gen_kwargs: ({}), limit: None, num_fewshot: None, batch_size: 1
| Tasks  |Version|Filter|n-shot|     Metric      |   | Value  |   |Stderr|
|--------|------:|------|-----:|-----------------|---|-------:|---|-----:|
|eq_bench|    2.1|none  |     0|eqbench          |↑  | 80.4738|±  |1.5788|
|        |       |none  |     0|percent_parseable|↑  |100.0000|±  |0.0000|

--- Technical Details ---

View Merge Details & Configuration

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

Merge Method

This model was merged using the RCZ method, updated standalone Magic method, Sparse Complementary Fusion (RKL), and Varcross method (I will of course release the code for these here soon.)

As promised, here is the code for the various methods. Have fun! :D https://github.com/HuggingMerger/mergekitter/tree/main/mergekit/merge_methods

Configuration

The following YAML configuration was used to produce this model:

merge_method: rcz

base_model: $MODELS/kitchen/SF-Foundation_gemma3_27b_mulitmodal_sft_01-31-2026

models:
    - model: $MODELS/kitchen/Mira-LA-SCF_RKM-v2a
    - model: $MODELS/kitchen/Mira-LA-Magic-v2c
    - model: $MODELS/kitchen/Mira-LA-Varcross-v2a
    - model: $MODELS/kitchen/Mira-LA-Magic-v2d
    - model: $MODELS/kitchen/Mira-MZ-RCZ-v2a

dtype: float32
out_dtype: bfloat16

tokenizer:
  source: union

parameters:
  lambda_scale: 0.75
  max_rho: 0.15
  tol: 1e-6

Downloads last month: 24

Safetensors

Model size

27B params

Tensor type

BF16

Model tree for virtuous7373/Lambent-Mira-Iris-v2

Base model

Lambent/Mira-v1.22.2-27B

Finetuned

Lambent/Mira-v1.23-27B-rlvr

Finetuned

(4)

this model

Quantizations

2 models