Instructions to use yukihamada/ElioChat-1.7B-Instruct-v3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use yukihamada/ElioChat-1.7B-Instruct-v3 with Transformers:

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

pipe = pipeline("text-generation", model="yukihamada/ElioChat-1.7B-Instruct-v3")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("yukihamada/ElioChat-1.7B-Instruct-v3", dtype="auto")

llama-cpp-python

How to use yukihamada/ElioChat-1.7B-Instruct-v3 with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="yukihamada/ElioChat-1.7B-Instruct-v3",
	filename="ElioChat-1.7B-Instruct-v3-F16.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use yukihamada/ElioChat-1.7B-Instruct-v3 with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16
# Run inference directly in the terminal:
llama-cli -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16
# Run inference directly in the terminal:
llama-cli -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16
# Run inference directly in the terminal:
./llama-cli -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf yukihamada/ElioChat-1.7B-Instruct-v3:F16

Use Docker

docker model run hf.co/yukihamada/ElioChat-1.7B-Instruct-v3:F16

LM Studio
Jan

vLLM

How to use yukihamada/ElioChat-1.7B-Instruct-v3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "yukihamada/ElioChat-1.7B-Instruct-v3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yukihamada/ElioChat-1.7B-Instruct-v3",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/yukihamada/ElioChat-1.7B-Instruct-v3:F16

SGLang

How to use yukihamada/ElioChat-1.7B-Instruct-v3 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 "yukihamada/ElioChat-1.7B-Instruct-v3" \
    --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": "yukihamada/ElioChat-1.7B-Instruct-v3",
		"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 "yukihamada/ElioChat-1.7B-Instruct-v3" \
        --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": "yukihamada/ElioChat-1.7B-Instruct-v3",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use yukihamada/ElioChat-1.7B-Instruct-v3 with Ollama:
```
ollama run hf.co/yukihamada/ElioChat-1.7B-Instruct-v3:F16
```

Unsloth Studio new

How to use yukihamada/ElioChat-1.7B-Instruct-v3 with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for yukihamada/ElioChat-1.7B-Instruct-v3 to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for yukihamada/ElioChat-1.7B-Instruct-v3 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for yukihamada/ElioChat-1.7B-Instruct-v3 to start chatting

Docker Model Runner
How to use yukihamada/ElioChat-1.7B-Instruct-v3 with Docker Model Runner:
```
docker model run hf.co/yukihamada/ElioChat-1.7B-Instruct-v3:F16
```

Lemonade

How to use yukihamada/ElioChat-1.7B-Instruct-v3 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull yukihamada/ElioChat-1.7B-Instruct-v3:F16

Run and chat with the model

lemonade run user.ElioChat-1.7B-Instruct-v3-F16

List all available models

lemonade list

ElioChat-1.7B-Instruct-v3

iPhoneで動く日本語思考AI。Qwen3-1.7Bベースに日本語思考データでファインチューニング。

特徴

思考過程が見える: <think>...</think> タグで推論プロセスを表示
日本語最適化: 日本語・日本文化に特化した1,166件の学習データ
軽量: Q5_K_M量子化で約1.2GB、iPhone 13以降で快適動作
学習コスト: わずか$0.70（約100円）

ダウンロード

ファイル	サイズ	用途
`ElioChat-1.7B-Instruct-v3-Q5_K_M.gguf`	1.2 GB	推奨（バランス型）
`ElioChat-1.7B-Instruct-v3-F16.gguf`	3.2 GB	高品質版

使用例

llama.cpp

llama-cli -m ElioChat-1.7B-Instruct-v3-Q5_K_M.gguf \
  -p "<|im_start|>user\n2の10乗は？<|im_end|>\n<|im_start|>assistant\n" \
  -n 200 --temp 0.7

iPhoneアプリ

ElioChat - iOS向けローカルLLMアプリ
PocketPal AI - iOS向けLLMアプリ

応答例

数学問題

Q: 2の10乗はいくつ？

A: <think>
2を10回かけます。
2^1 = 2
2^2 = 4
2^3 = 8
...
2^10 = 1024
</think>

答えは1024です。

論理推論

Q: AはBより背が高い。BはCより背が高い。一番背が高いのは誰？

A: <think>
AはBより高い、BはCより高い。
A > B > C の関係。
</think>

一番背が高いのはAです。

ベンチマーク（2025年1月）

モデル	正答率	思考タグ	平均応答時間
ElioChat-1.7B-v3	70%	45%	20.7s
Qwen3-1.7B	65%	20%	16.3s
TinySwallow-1.5B	90%	0%	14.9s
Sarashina2.2-3B	90%	0%	8.8s

ElioChat-v3は思考過程を可視化する唯一のモデル。「なぜその答えなのか」が分かる。

学習データ構成

カテゴリ	件数
コア推論（論理・数学）	515
日本語・文化	193
会話・ユーモア	116
実用・教養	153
その他	189
合計	1,166

学習詳細

項目	値
ベースモデル	Qwen/Qwen3-1.7B
学習方式	LoRA (PEFT)
LoRA Rank	64
GPU	NVIDIA A100-SXM4-40GB
学習時間	32.7分
コスト	$0.70
最終Loss	1.25

オフライン利用

モデルをダウンロードすれば、ネット接続なしで使えます：

飛行機の中
山の中
地下鉄
プライバシーが必要な場面

ライセンス

ベースモデル: Qwen License
学習データ・コード: MIT License

Downloads last month: 100

GGUF

Model size

2B params

Architecture

qwen3

Hardware compatibility

5-bit

16-bit

Model tree for yukihamada/ElioChat-1.7B-Instruct-v3

Base model

Qwen/Qwen3-1.7B-Base

Finetuned

Qwen/Qwen3-1.7B

Quantized

(268)

this model

yukihamada
/

ElioChat-1.7B-Instruct-v3

ElioChat-1.7B-Instruct-v3

特徴

ダウンロード

使用例

llama.cpp

iPhoneアプリ

応答例

数学問題

論理推論

ベンチマーク（2025年1月）

学習データ構成

学習詳細

オフライン利用

関連リンク

ライセンス

Model tree for yukihamada/ElioChat-1.7B-Instruct-v3