Instructions to use DaehanKim/korjev-26b4a-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DaehanKim/korjev-26b4a-v1 with Transformers:

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

pipe = pipeline("text-generation", model="DaehanKim/korjev-26b4a-v1")
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("DaehanKim/korjev-26b4a-v1")
model = AutoModelForMultimodalLM.from_pretrained("DaehanKim/korjev-26b4a-v1", 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

vLLM

How to use DaehanKim/korjev-26b4a-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DaehanKim/korjev-26b4a-v1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DaehanKim/korjev-26b4a-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/DaehanKim/korjev-26b4a-v1

SGLang

How to use DaehanKim/korjev-26b4a-v1 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 "DaehanKim/korjev-26b4a-v1" \
    --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": "DaehanKim/korjev-26b4a-v1",
		"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 "DaehanKim/korjev-26b4a-v1" \
        --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": "DaehanKim/korjev-26b4a-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use DaehanKim/korjev-26b4a-v1 with Docker Model Runner:
```
docker model run hf.co/DaehanKim/korjev-26b4a-v1
```

korjev-26b4a-v1 / precision_audit.json

DaehanKim

Release merged korjev-26b4a-v1 FP8 weights, evaluation and serving instructions

8041088 verified 25 days ago

Raw History Blame Contribute Delete

2.15 kB

{
  "checked_at": "2026-10-09",
  "conclusion": "No FP8 serving recommendation found in the official documentation checked. Keep the measured BF16 specialist configurations.",
  "models": {
    "cygnet": {
      "precision": "BF16",
      "fp8_recommended": false,
      "reason": "The official README explicitly identifies the frozen Gemma checkpoint as BF16.",
      "url": "https://github.com/blockbrain-ai/cygnet-recipe#disclosures"
    },
    "mica": {
      "precision": "BF16",
      "fp8_recommended": false,
      "reason": "BF16 is the quick-start default. Optional GGUF integer quantizations are offered; these are not FP8.",
      "url": "https://huggingface.co/sky7350/Mica-v0.1-4B#quick-start"
    },
    "kdecision": {
      "precision": "BF16 backbone / FP32 pointer head",
      "fp8_recommended": false,
      "reason": "The model card delegates serving to Kev, whose GPU serving default is BF16.",
      "urls": [
        "https://huggingface.co/ThakiCloud/kd-4b-ko-v0#how-to-use",
        "https://github.com/jaredpalmer/kev#serving-performance"
      ]
    }
  },
  "sources": [
    {
      "url": "https://raw.githubusercontent.com/blockbrain-ai/cygnet-recipe/10b4099a61e7dca6a82032b54c74dc0a8c496d4b/README.md",
      "sha256": "ec82a0772cdc3e21df65b9b708e8c569d19d7b1b1209c3ec0c13057f35903323"
    },
    {
      "url": "https://huggingface.co/sky7350/Mica-v0.1-4B/raw/9b6428b272f4708ef54a4702b52bb453a4059363/README.md",
      "sha256": "367f2ae50e514ee9b4224c4bb167c3a9dc28cac13b1a404171495358907f1d45"
    },
    {
      "url": "https://huggingface.co/ThakiCloud/kd-4b-ko-v0/raw/9e7d29789e421b7796453d57dbe5252735e33128/README.md",
      "sha256": "945fa65a3cf809491d24577139c86131fd3893a4c460d3c3e676cae5b46e08d1"
    },
    {
      "url": "https://raw.githubusercontent.com/jaredpalmer/kev/fe64b1274ea7f80d4095866df90666abb03e9cf6/README.md",
      "sha256": "19c9d7d30cd723b9f112f1f68d0b415a28c5ee616dbc21b358fc29bd07a66311"
    },
    {
      "url": "https://raw.githubusercontent.com/akivet/Mica-v0.1-4B/main/README.md",
      "sha256": "dcdc154a6b338c13b4089ce596f60235733b61151fb640b320b4713a509c97da"
    }
  ]
}

	{
	"checked_at": "2026-10-09",
	"conclusion": "No FP8 serving recommendation found in the official documentation checked. Keep the measured BF16 specialist configurations.",
	"models": {
	"cygnet": {
	"precision": "BF16",
	"fp8_recommended": false,
	"reason": "The official README explicitly identifies the frozen Gemma checkpoint as BF16.",
	"url": "https://github.com/blockbrain-ai/cygnet-recipe#disclosures"
	},
	"mica": {
	"precision": "BF16",
	"fp8_recommended": false,
	"reason": "BF16 is the quick-start default. Optional GGUF integer quantizations are offered; these are not FP8.",
	"url": "https://huggingface.co/sky7350/Mica-v0.1-4B#quick-start"
	},
	"kdecision": {
	"precision": "BF16 backbone / FP32 pointer head",
	"fp8_recommended": false,
	"reason": "The model card delegates serving to Kev, whose GPU serving default is BF16.",
	"urls": [
	"https://huggingface.co/ThakiCloud/kd-4b-ko-v0#how-to-use",
	"https://github.com/jaredpalmer/kev#serving-performance"
	]
	}
	},
	"sources": [
	{
	"url": "https://raw.githubusercontent.com/blockbrain-ai/cygnet-recipe/10b4099a61e7dca6a82032b54c74dc0a8c496d4b/README.md",
	"sha256": "ec82a0772cdc3e21df65b9b708e8c569d19d7b1b1209c3ec0c13057f35903323"
	},
	{
	"url": "https://huggingface.co/sky7350/Mica-v0.1-4B/raw/9b6428b272f4708ef54a4702b52bb453a4059363/README.md",
	"sha256": "367f2ae50e514ee9b4224c4bb167c3a9dc28cac13b1a404171495358907f1d45"
	},
	{
	"url": "https://huggingface.co/ThakiCloud/kd-4b-ko-v0/raw/9e7d29789e421b7796453d57dbe5252735e33128/README.md",
	"sha256": "945fa65a3cf809491d24577139c86131fd3893a4c460d3c3e676cae5b46e08d1"
	},
	{
	"url": "https://raw.githubusercontent.com/jaredpalmer/kev/fe64b1274ea7f80d4095866df90666abb03e9cf6/README.md",
	"sha256": "19c9d7d30cd723b9f112f1f68d0b415a28c5ee616dbc21b358fc29bd07a66311"
	},
	{
	"url": "https://raw.githubusercontent.com/akivet/Mica-v0.1-4B/main/README.md",
	"sha256": "dcdc154a6b338c13b4089ce596f60235733b61151fb640b320b4713a509c97da"
	}
	]
	}