Instructions to use Khetterman/Multilingual-SaigaSuzume-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Khetterman/Multilingual-SaigaSuzume-8B with Transformers:

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

pipe = pipeline("text-generation", model="Khetterman/Multilingual-SaigaSuzume-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Khetterman/Multilingual-SaigaSuzume-8B")
model = AutoModelForCausalLM.from_pretrained("Khetterman/Multilingual-SaigaSuzume-8B")
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 Khetterman/Multilingual-SaigaSuzume-8B with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Khetterman/Multilingual-SaigaSuzume-8B

SGLang

How to use Khetterman/Multilingual-SaigaSuzume-8B 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 "Khetterman/Multilingual-SaigaSuzume-8B" \
    --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": "Khetterman/Multilingual-SaigaSuzume-8B",
		"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 "Khetterman/Multilingual-SaigaSuzume-8B" \
        --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": "Khetterman/Multilingual-SaigaSuzume-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Khetterman/Multilingual-SaigaSuzume-8B with Docker Model Runner:
```
docker model run hf.co/Khetterman/Multilingual-SaigaSuzume-8B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Multilingual-SaigaSuzume-8B

Your words are like rain falling from heaven on a tower in a sinful land; can anyone in Babylon understand them?

This model was created as the basis of multilingual abilities for other models. I think it will be very useful as an integral part of your model. There is some censorship, keep this in mind.

Merge Details

Method

This is a simple, but usefull merge of 7 cool models, created using mergekit.

Models

The following models were included in the merge:

Configuration

The following YAML configurations was used to produce this model:

# Multilingual-SaigaSuzume-8B-BFH
models:
  - model: lightblue/suzume-llama-3-8B-multilingual-orpo-borda-full
  - model: IlyaGusev/saiga_llama3_8b
  - model: lightblue/suzume-llama-3-8B-multilingual-orpo-borda-half
merge_method: model_stock
base_model: huihui-ai/Meta-Llama-3.1-8B-Instruct-abliterated
dtype: bfloat16

# Multilingual-SaigaSuzume-8B-BTP
models:
  - model: lightblue/suzume-llama-3-8B-multilingual-orpo-borda-top75
  - model: IlyaGusev/saiga_llama3_8b
  - model: lightblue/suzume-llama-3-8B-multilingual-orpo-borda-top25
merge_method: model_stock
base_model: huihui-ai/Meta-Llama-3.1-8B-Instruct-abliterated
dtype: bfloat16

# Multilingual-SaigaSuzume-8B-Classic
models:
  - model: IlyaGusev/saiga_llama3_8b
  - model: lightblue/suzume-llama-3-8B-multilingual
merge_method: model_stock
base_model: huihui-ai/Meta-Llama-3.1-8B-Instruct-abliterated
dtype: bfloat16

# Multilingual-SaigaSuzume-8B
models:
  - model: Multilingual-SaigaSuzume-8B-BFH
  - model: Multilingual-SaigaSuzume-8B-BTP
merge_method: model_stock
base_model: Multilingual-SaigaSuzume-8B-Classic
dtype: bfloat16