Instructions to use M4-ai/Hercules-Mini-1.8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use M4-ai/Hercules-Mini-1.8B with Transformers:

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

pipe = pipeline("text-generation", model="M4-ai/Hercules-Mini-1.8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("M4-ai/Hercules-Mini-1.8B")
model = AutoModelForCausalLM.from_pretrained("M4-ai/Hercules-Mini-1.8B")
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 Settings

vLLM

How to use M4-ai/Hercules-Mini-1.8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "M4-ai/Hercules-Mini-1.8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "M4-ai/Hercules-Mini-1.8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/M4-ai/Hercules-Mini-1.8B

SGLang

How to use M4-ai/Hercules-Mini-1.8B 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 "M4-ai/Hercules-Mini-1.8B" \
    --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": "M4-ai/Hercules-Mini-1.8B",
		"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 "M4-ai/Hercules-Mini-1.8B" \
        --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": "M4-ai/Hercules-Mini-1.8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use M4-ai/Hercules-Mini-1.8B with Docker Model Runner:
```
docker model run hf.co/M4-ai/Hercules-Mini-1.8B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Hercules-Mini-1.8B

We fine-tuned Qwen1.5-1.8B on Locutusque's Hercules-v4.

Model Details

Model Description

This model has capabilities in math, coding, function calling, roleplay, and more. We fine-tuned it using 700,000 examples of Hercules-v4.

Developed by: M4-ai
Language(s) (NLP): English and maybe Chinese
License: tongyi-qianwen license
Finetuned from model: Qwen1.5-1.8B

Uses

General purpose assistant, question answering, chain-of-thought, etc..

Bias, Risks, and Limitations

The eos token was not setup properly, so to prevent infinite generation you'll need to implement a stopping criteria when the model generates the <|im_end|> token.

Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

Evaluation

Coming soon

Training Details

Training Data

https://huggingface.co/datasets/Locutusque/hercules-v4.0

Training Hyperparameters

Training regime: bf16 non-mixed precision

Technical Specifications

Hardware

We used 8 Kaggle TPUs, and we trained at a global batch size of 256 and sequence length of 1536

Contributions

Thanks to @Tonic, @aloobun, @fhai50032, and @Locutusque for their contributions to this model.

Downloads last month: 101

Safetensors

Model size

2B params

Tensor type

BF16

Model tree for M4-ai/Hercules-Mini-1.8B

Quantizations

3 models

M4-ai
/

Hercules-Mini-1.8B