Instructions to use mkd-dev/Keural-MoE-14B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mkd-dev/Keural-MoE-14B with Transformers:

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

pipe = pipeline("text-generation", model="mkd-dev/Keural-MoE-14B", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("mkd-dev/Keural-MoE-14B", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mkd-dev/Keural-MoE-14B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mkd-dev/Keural-MoE-14B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mkd-dev/Keural-MoE-14B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mkd-dev/Keural-MoE-14B

SGLang

How to use mkd-dev/Keural-MoE-14B 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 "mkd-dev/Keural-MoE-14B" \
    --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": "mkd-dev/Keural-MoE-14B",
		"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 "mkd-dev/Keural-MoE-14B" \
        --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": "mkd-dev/Keural-MoE-14B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mkd-dev/Keural-MoE-14B with Docker Model Runner:
```
docker model run hf.co/mkd-dev/Keural-MoE-14B
```

Keural-MoE-14B / modeling_keural.py

mkd-dev

mkd-hossain

Duplicate from mkd-ai-dev/Keural-MoE-14B

264c832 25 days ago

Raw History Blame Contribute Delete

1.33 kB

"""
KeuralMoECausalLM — Keural Mixture-of-Experts Causal Language Model.
Bilingual Korean-English MoE LLM trained entirely from scratch.
Developed by MKD Corp AI Research, Republic of Korea.
Architecture:
  - 14.83B total parameters (~7.42B active per token)
  - 24 layers, hidden=4096, GQA 32/8 heads
  - 8 experts per layer, top-2 routing
  - Sliding window attention (512, alternating layers)
  - RoPE theta=500,000, context length=4096
  - Vocabulary: 131,074 tokens (131,072 SPM + <|im_start|> + <|im_end|>)
Load with trust_remote_code=True:
  from transformers import AutoModelForCausalLM, AutoTokenizer
  model = AutoModelForCausalLM.from_pretrained(
      "mkd-hossain/keural-sft3-final",
      torch_dtype="bfloat16",
      device_map="auto",
      trust_remote_code=True,
  )
"""

from transformers import MixtralForCausalLM
from transformers.utils import logging

try:
    from configuration_keural import KeuralConfig
except ImportError:
    from .configuration_keural import KeuralConfig

logger = logging.get_logger(__name__)


class KeuralMoECausalLM(MixtralForCausalLM):
    """
    Keural MoE Causal Language Model.
    Bilingual Korean-English 14.83B MoE LLM trained from scratch by MKD Corp AI Research.
    """

    config_class = KeuralConfig
    _keys_to_ignore_on_load_missing = ["lm_head.weight"]

	"""
	KeuralMoECausalLM — Keural Mixture-of-Experts Causal Language Model.

	Bilingual Korean-English MoE LLM trained entirely from scratch.
	Developed by MKD Corp AI Research, Republic of Korea.

	Architecture:
	- 14.83B total parameters (~7.42B active per token)
	- 24 layers, hidden=4096, GQA 32/8 heads
	- 8 experts per layer, top-2 routing
	- Sliding window attention (512, alternating layers)
	- RoPE theta=500,000, context length=4096
	- Vocabulary: 131,074 tokens (131,072 SPM + <\|im_start\|> + <\|im_end\|>)

	Load with trust_remote_code=True:
	from transformers import AutoModelForCausalLM, AutoTokenizer
	model = AutoModelForCausalLM.from_pretrained(
	"mkd-hossain/keural-sft3-final",
	torch_dtype="bfloat16",
	device_map="auto",
	trust_remote_code=True,
	)
	"""

	from transformers import MixtralForCausalLM
	from transformers.utils import logging

	try:
	from configuration_keural import KeuralConfig
	except ImportError:
	from .configuration_keural import KeuralConfig

	logger = logging.get_logger(__name__)


	class KeuralMoECausalLM(MixtralForCausalLM):
	"""
	Keural MoE Causal Language Model.
	Bilingual Korean-English 14.83B MoE LLM trained from scratch by MKD Corp AI Research.
	"""

	config_class = KeuralConfig
	_keys_to_ignore_on_load_missing = ["lm_head.weight"]