Instructions to use sitloboi2012/SEMIKONG-8B-Instruct-GPTQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sitloboi2012/SEMIKONG-8B-Instruct-GPTQ with Transformers:

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

pipe = pipeline("text-generation", model="sitloboi2012/SEMIKONG-8B-Instruct-GPTQ")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("sitloboi2012/SEMIKONG-8B-Instruct-GPTQ")
model = AutoModelForCausalLM.from_pretrained("sitloboi2012/SEMIKONG-8B-Instruct-GPTQ")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use sitloboi2012/SEMIKONG-8B-Instruct-GPTQ with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sitloboi2012/SEMIKONG-8B-Instruct-GPTQ"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sitloboi2012/SEMIKONG-8B-Instruct-GPTQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/sitloboi2012/SEMIKONG-8B-Instruct-GPTQ

SGLang

How to use sitloboi2012/SEMIKONG-8B-Instruct-GPTQ 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 "sitloboi2012/SEMIKONG-8B-Instruct-GPTQ" \
    --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": "sitloboi2012/SEMIKONG-8B-Instruct-GPTQ",
		"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 "sitloboi2012/SEMIKONG-8B-Instruct-GPTQ" \
        --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": "sitloboi2012/SEMIKONG-8B-Instruct-GPTQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use sitloboi2012/SEMIKONG-8B-Instruct-GPTQ with Docker Model Runner:
```
docker model run hf.co/sitloboi2012/SEMIKONG-8B-Instruct-GPTQ
```

sitloboi2012/SEMIKONG-8B-Instruct-GPTQ

SEMIKONG is a Foundation Model in the semiconductor manufacturing process. This model aims to become the baseline for future development LLM in various use case that required domain-specific knowledge in the field of Semiconductor Manufacturing Process.

SEMIKONG is the result of a collaboration between FPT AI Center, AItomatic and Tokyo Electron. With supported from the AI Alliance and IBM.

Model Details

The base model of SEMIKONG was built upon Llama 3 Instruct Model.

We divided the training process into 3 main phases: Pretrained Domain Knowledge -> Self Finetuning (Instruction Dataset) -> Merging and Quantization.

The model got benchmark on domain-expert reviewed dataset which will be release for the community soon !

Model Description

Developed by: FPT AI Center, AItomatic and Tokyo Electron
Funded by: FPT AI Center, AItomatic
Shared by: AI Alliance
Model type: Instruction Model
Language(s): English
License: Apache License 2.0
Finetuned from model: Llama 3

Uses

The model can be use as an endpoint API for extended development use case such as RAG, Chatbot, etc.

You can use either vLLM or NVIDIA NIM to access and setup the infrastructure for the model

Out-of-Scope Use

This is the first version of SEMIKONG which will only be supported Etching Process and General Domain Specific Knowledge in Semiconductor Manufacturing Process

Future version of SEMIKONG will extend it knowledge to end-to-end Semiconductor Manufacturing Process

Model Card Authors

Huy Vo (Henry): huyva2@fpt.com
Christopher Nguyen: ctn@aitomatic.com

Model Card Contact

Huy Vo (Henry): huyva2@fpt.com
Christopher Nguyen: ctn@aitomatic.com

Downloads last month: 6

Safetensors

Model size

8B params

Tensor type

I32

BF16

sitloboi2012
/

SEMIKONG-8B-Instruct-GPTQ