Instructions to use iamtatsuki05/Llama-JP-0.5B-PT-stage2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use iamtatsuki05/Llama-JP-0.5B-PT-stage2 with Transformers:

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

pipe = pipeline("text-generation", model="iamtatsuki05/Llama-JP-0.5B-PT-stage2")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("iamtatsuki05/Llama-JP-0.5B-PT-stage2")
model = AutoModelForCausalLM.from_pretrained("iamtatsuki05/Llama-JP-0.5B-PT-stage2")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use iamtatsuki05/Llama-JP-0.5B-PT-stage2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "iamtatsuki05/Llama-JP-0.5B-PT-stage2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "iamtatsuki05/Llama-JP-0.5B-PT-stage2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/iamtatsuki05/Llama-JP-0.5B-PT-stage2

SGLang

How to use iamtatsuki05/Llama-JP-0.5B-PT-stage2 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 "iamtatsuki05/Llama-JP-0.5B-PT-stage2" \
    --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": "iamtatsuki05/Llama-JP-0.5B-PT-stage2",
		"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 "iamtatsuki05/Llama-JP-0.5B-PT-stage2" \
        --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": "iamtatsuki05/Llama-JP-0.5B-PT-stage2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use iamtatsuki05/Llama-JP-0.5B-PT-stage2 with Docker Model Runner:
```
docker model run hf.co/iamtatsuki05/Llama-JP-0.5B-PT-stage2
```

Llama-JP-0.5B-PT-stage2

English / Japanese

Overview

Llama-JP-0.5B-PT-stage2 extends iamtatsuki05/Llama-JP-0.5B-PT-stage1 with additional training on fujiki/wiki40b_ja. The model observes roughly 1B tokens with 8,192-token contexts to strengthen long-form Japanese generation and encyclopedic coverage.

Hugging Face Collection
GitHub

Usage

Requirements

transformers>=4.51.0
accelerate>=1.6.0
sentencepiece>=0.2.0
flash-attn>=2.7.3

Sample Code

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "iamtatsuki05/Llama-JP-0.5B-PT-stage2"
model_kwargs = {
  "torch_dtype": torch.bfloat16,
  "attn_implementation": "flash_attention_2",
  "device_map": "auto",
}
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, **model_kwargs)

prompt = "ちいかわのハチワレは"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
output = model.generate(
    **inputs,
    max_new_tokens=256,
    temperature=0.8,
    top_p=0.9,
    do_sample=True,
)
print(tokenizer.decode(output[0], skip_special_tokens=True))

Model Details

Base model: iamtatsuki05/Llama-JP-0.5B-PT-stage1
Architecture: Llama
Maximum sequence length: 8,192 tokens
Embedding dimension: 1280
Tokenizer: SentencePiece / vocabulary size 102,400
Positional encoding: RoPE
Supported languages: Japanese

Model Series

These checkpoints were further pre-trained on fujiki/wiki40b_ja for roughly 1B tokens with 8,192-token contexts to enhance encyclopedic coverage.

ID	Architecture	#Param.	#Param. w/o Emb.	JGLUE-Avg	JGLUE-JSTS	JGLUE-JNLI	JGLUE-JCoLA
iamtatsuki05/ModernBERT-JP-0.5B-PT-stage2	ModernBERT	679M	548M	86.32	89.29	86.68	83.00
iamtatsuki05/Llama-JP-0.5B-PT-stage2 (this model)	Llama	661M	530M	81.88	82.91	78.67	84.06

Licence

This model is distributed under the MIT License.

How to Cite

@article{MIREI
  title={同一条件下における Encoder/Decoder アーキテクチャによる文埋め込みの性能分析},
  author={岡田 龍樹 and 杉本 徹},
  journal={言語処理学会第 32 回年次大会 (NLP2026)},
  year={2026}
}