Instructions to use Johnblick187/SmartCoderMoE with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Johnblick187/SmartCoderMoE with Transformers:

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

pipe = pipeline("text-generation", model="Johnblick187/SmartCoderMoE", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Johnblick187/SmartCoderMoE", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Johnblick187/SmartCoderMoE with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Johnblick187/SmartCoderMoE"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Johnblick187/SmartCoderMoE",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Johnblick187/SmartCoderMoE

SGLang

How to use Johnblick187/SmartCoderMoE 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 "Johnblick187/SmartCoderMoE" \
    --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": "Johnblick187/SmartCoderMoE",
		"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 "Johnblick187/SmartCoderMoE" \
        --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": "Johnblick187/SmartCoderMoE",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Johnblick187/SmartCoderMoE with Docker Model Runner:
```
docker model run hf.co/Johnblick187/SmartCoderMoE
```

Johnblick187 commited on 18 days ago

Commit

4f864d1

verified ·

1 Parent(s): ec57f1e

Update README.md

Browse files

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -25,7 +25,7 @@ SmartCoderMoE is a 4.65B parameter sparse Mixture-of-Experts coding model.
 SmartCoderMoE is not your average fine-tune. He was engineered through a multi-stage weight surgery pipeline:
 1. **Slice Merge** — StarCoder2-15B and StarChat2-15B were each sliced into 3 × 2048-dim pieces and SLERP-merged with deliberate per-slice biases (60/80/90) to preserve coding depth while injecting instruct capability of Starchat2
-1. **MoE Surgery** — Every dense FFN layer was surgically split: The original dim of 245576 was reduced to an intermediate dim of 8192 and kept as a dense FFN, and the remaining 16384 dims were sliced into **32 experts of 512 dim each**, giving Smartcoder an expansive yet tiny network of 1280 total experts.
 1. **Vocab Expansion** — Extended from 49152 to 65536 tokens with multimodal special tokens for code, audio, image, video, and music.
 1. **Zero waste** — Not a single weight was discarded. Every parameter from StarCoder2’s original FFN lives on in either the dense FFN or one of the 1280 expert slots.

 SmartCoderMoE is not your average fine-tune. He was engineered through a multi-stage weight surgery pipeline:
 1. **Slice Merge** — StarCoder2-15B and StarChat2-15B were each sliced into 3 × 2048-dim pieces and SLERP-merged with deliberate per-slice biases (60/80/90) to preserve coding depth while injecting instruct capability of Starchat2
+1. **MoE Surgery** — Every dense FFN layer was surgically split: The original dim of 24576 was reduced to an intermediate dim of 8192 and kept as a dense FFN, and the remaining 16384 dims were sliced into **32 experts of 512 dim each**, giving Smartcoder an expansive yet tiny network of 1280 total experts.
 1. **Vocab Expansion** — Extended from 49152 to 65536 tokens with multimodal special tokens for code, audio, image, video, and music.
 1. **Zero waste** — Not a single weight was discarded. Every parameter from StarCoder2’s original FFN lives on in either the dense FFN or one of the 1280 expert slots.