Instructions to use Ismantic/Summer-0.5B-S0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Ismantic/Summer-0.5B-S0 with Transformers:

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

pipe = pipeline("text-generation", model="Ismantic/Summer-0.5B-S0")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Ismantic/Summer-0.5B-S0")
model = AutoModelForCausalLM.from_pretrained("Ismantic/Summer-0.5B-S0", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Ismantic/Summer-0.5B-S0 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Ismantic/Summer-0.5B-S0"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ismantic/Summer-0.5B-S0",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Ismantic/Summer-0.5B-S0

SGLang

How to use Ismantic/Summer-0.5B-S0 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 "Ismantic/Summer-0.5B-S0" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ismantic/Summer-0.5B-S0",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Ismantic/Summer-0.5B-S0" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ismantic/Summer-0.5B-S0",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Ismantic/Summer-0.5B-S0 with Docker Model Runner:
```
docker model run hf.co/Ismantic/Summer-0.5B-S0
```

Summer-0.5B-S0 / example_vllm.py

tf-bao

Upload Summer-0.5B-S0

16cf0c6 verified 2 months ago

Raw History Blame Contribute Delete

1.62 kB

"""用 vLLM 跑这个模型。
    pip install vllm
    pip install git+https://github.com/Ismantic/PieceTokenizer
    python example_vllm.py
## 为什么要 skip_tokenizer_init
vLLM 能加载**权重** —— config.json 里 architectures 是 Qwen3ForCausalLM,
它用自己那份 Qwen3 实现,按 state_dict 的 key 名灌进去。
但它**认不了这个词表**:81903 的 piece 词表不是 HF 标准格式,AutoTokenizer
走不通。所以要 skip_tokenizer_init=True,自己编码好 token id 用 TokensPrompt
喂进去,拿回来的 id 自己解码。
推论:`vllm serve` 那种 OpenAI 兼容服务**开箱用不了** —— 它要把文本转成
token,而它转不了。调用方得自己传 token id。
"""
import torch
from vllm import LLM, SamplingParams
from vllm.inputs import TokensPrompt

from tokenizer import PieceTokenizerWrapper

HERE = "."

tok = PieceTokenizerWrapper(HERE)
llm = LLM(model=HERE, skip_tokenizer_init=True, dtype="bfloat16",
          gpu_memory_utilization=0.85, max_model_len=4096,
          trust_remote_code=True)

prompts = [
    "机器翻译的基本任务是",
    "中文分词的目标是",
]

def encode(p):
    return [tok.bos_token_id] + tok.encode(p, add_special_tokens=False)  # 训练时每行都以 <bos> 开头,不加是分布外输入

outputs = llm.generate(
    [TokensPrompt(prompt_token_ids=encode(p)) for p in prompts],
    SamplingParams(temperature=0.0, max_tokens=64,
                   stop_token_ids=tok.stop_token_ids),
)

for prompt, out in zip(prompts, outputs):
    print(f"\n{prompt}", end="")
    print(tok.decode(list(out.outputs[0].token_ids)))

	"""用 vLLM 跑这个模型。

	pip install vllm
	pip install git+https://github.com/Ismantic/PieceTokenizer
	python example_vllm.py

	## 为什么要 skip_tokenizer_init

	vLLM 能加载权重 —— config.json 里 architectures 是 Qwen3ForCausalLM,
	它用自己那份 Qwen3 实现,按 state_dict 的 key 名灌进去。

	但它认不了这个词表:81903 的 piece 词表不是 HF 标准格式,AutoTokenizer
	走不通。所以要 skip_tokenizer_init=True,自己编码好 token id 用 TokensPrompt
	喂进去,拿回来的 id 自己解码。

	推论:`vllm serve` 那种 OpenAI 兼容服务开箱用不了 —— 它要把文本转成
	token,而它转不了。调用方得自己传 token id。
	"""
	import torch
	from vllm import LLM, SamplingParams
	from vllm.inputs import TokensPrompt

	from tokenizer import PieceTokenizerWrapper

	HERE = "."

	tok = PieceTokenizerWrapper(HERE)
	llm = LLM(model=HERE, skip_tokenizer_init=True, dtype="bfloat16",
	gpu_memory_utilization=0.85, max_model_len=4096,
	trust_remote_code=True)

	prompts = [
	"机器翻译的基本任务是",
	"中文分词的目标是",
	]

	def encode(p):
	return [tok.bos_token_id] + tok.encode(p, add_special_tokens=False) # 训练时每行都以 <bos> 开头,不加是分布外输入

	outputs = llm.generate(
	[TokensPrompt(prompt_token_ids=encode(p)) for p in prompts],
	SamplingParams(temperature=0.0, max_tokens=64,
	stop_token_ids=tok.stop_token_ids),
	)

	for prompt, out in zip(prompts, outputs):
	print(f"\n{prompt}", end="")
	print(tok.decode(list(out.outputs[0].token_ids)))