Instructions to use carlyou/DeepSeek-V2-Lite-FP8-Group with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use carlyou/DeepSeek-V2-Lite-FP8-Group with Transformers:

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

pipe = pipeline("text-generation", model="carlyou/DeepSeek-V2-Lite-FP8-Group", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("carlyou/DeepSeek-V2-Lite-FP8-Group", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("carlyou/DeepSeek-V2-Lite-FP8-Group", trust_remote_code=True)
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 Settings

vLLM

How to use carlyou/DeepSeek-V2-Lite-FP8-Group with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "carlyou/DeepSeek-V2-Lite-FP8-Group"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "carlyou/DeepSeek-V2-Lite-FP8-Group",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/carlyou/DeepSeek-V2-Lite-FP8-Group

SGLang

How to use carlyou/DeepSeek-V2-Lite-FP8-Group 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 "carlyou/DeepSeek-V2-Lite-FP8-Group" \
    --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": "carlyou/DeepSeek-V2-Lite-FP8-Group",
		"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 "carlyou/DeepSeek-V2-Lite-FP8-Group" \
        --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": "carlyou/DeepSeek-V2-Lite-FP8-Group",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use carlyou/DeepSeek-V2-Lite-FP8-Group with Docker Model Runner:
```
docker model run hf.co/carlyou/DeepSeek-V2-Lite-FP8-Group
```

DeepSeek-V2-Lite-FP8-Group / README.md

carlyou

Add DeepSeek-V2-Lite quantized to per-group FP8 (group_size=64) via llm-compressor

68c5c3a verified 3 months ago

2.19 kB

license: other
license_name: deepseek-license
license_link: LICENSE
base_model: deepseek-ai/DeepSeek-V2-Lite
tags:
  - deepseek
  - mla
  - moe
  - fp8
  - group-quantization
  - compressed-tensors
library_name: transformers

DeepSeek-V2-Lite-FP8-Group

Per-group FP8 quantized version of deepseek-ai/DeepSeek-V2-Lite, created with llm-compressor.

Quantization Details

Property	Value
Base model	deepseek-ai/DeepSeek-V2-Lite
Parameters	15.7B total (2.4B active)
Architecture	DeepSeek-V2 (MLA + MoE, 64 experts, top-6)
Quantization	Per-group FP8 (E4M3), dynamic activations
Weight strategy	Group, group_size=64
Activation strategy	Per-token, dynamic
Format	compressed-tensors (float-quantized)
Ignored layers	lm_head
Model size	~16 GB
Tool	llm-compressor 0.10.0

This model uses the same per-group FP8 quantization scheme as DeepSeek-V3 (weight_block_size: [1, 64]), making it useful for testing and validating group FP8 inference paths (e.g., MLA attention + group FP8 fusion in vLLM) without needing a 671B model.

Evaluation

GSM8K accuracy (100 samples, via lm_eval harness):

Model	exact_match
Baseline (BF16)	0.300
FP8-Group (this model)	0.330

No precision degradation observed from group FP8 quantization.

Usage

With vLLM

vllm serve carlyou/DeepSeek-V2-Lite-FP8-Group --trust-remote-code

With Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    "carlyou/DeepSeek-V2-Lite-FP8-Group",
    torch_dtype="auto",
    trust_remote_code=True,
    device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained(
    "carlyou/DeepSeek-V2-Lite-FP8-Group",
    trust_remote_code=True,
)

Reproduction

pip install llmcompressor transformers
python quantize.py --model deepseek-ai/DeepSeek-V2-Lite --scheme fp8-group

See carlyou/llm-quant for the quantization script.