Instructions to use NeuralNovel/Gecko-7B-v0.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NeuralNovel/Gecko-7B-v0.1 with Transformers:

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

pipe = pipeline("text-generation", model="NeuralNovel/Gecko-7B-v0.1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NeuralNovel/Gecko-7B-v0.1")
model = AutoModelForCausalLM.from_pretrained("NeuralNovel/Gecko-7B-v0.1")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use NeuralNovel/Gecko-7B-v0.1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NeuralNovel/Gecko-7B-v0.1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NeuralNovel/Gecko-7B-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/NeuralNovel/Gecko-7B-v0.1

SGLang

How to use NeuralNovel/Gecko-7B-v0.1 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 "NeuralNovel/Gecko-7B-v0.1" \
    --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": "NeuralNovel/Gecko-7B-v0.1",
		"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 "NeuralNovel/Gecko-7B-v0.1" \
        --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": "NeuralNovel/Gecko-7B-v0.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use NeuralNovel/Gecko-7B-v0.1 with Docker Model Runner:
```
docker model run hf.co/NeuralNovel/Gecko-7B-v0.1
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Gecko-7B-v0.1

Designed to generate instructive and narrative text, with a focus on mathematics & numeracy.

Full-parameter fine-tune (FFT) of Mistral-7B-Instruct-v0.2, with apache-2.0 license.

You may download and use this model for research, training and commercial purposes.

Data-set

The model was finetuned using the Neural-Mini-Math dataset (Currently Private)

Summary

Fine-tuned with the intention of following all prompt directions, making it more suitable for roleplay and problem solving.

Out-of-Scope Use

The model may not perform well in scenarios unrelated to instructive and narrative text generation. Misuse or applications outside its designed scope may result in suboptimal outcomes.

Bias, Risks, and Limitations

This model may not work as intended. As such all users are encouraged to use this model with caution and respect.

This model is for testing and research purposes only, it has reduced levels of alignment and as a result may produce NSFW or harmful content. The user is responsible for their output and must use this model responsibly.

Hardware and Training


  n_epochs = 3,
  n_checkpoints = 3,
  batch_size = 12,
  learning_rate = 1e-5,

Sincere appreciation to Techmind for their generous sponsorship.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	64.58
AI2 Reasoning Challenge (25-Shot)	61.35
HellaSwag (10-Shot)	83.36
MMLU (5-Shot)	61.05
TruthfulQA (0-shot)	62.60
Winogrande (5-shot)	77.58
GSM8k (5-shot)	41.55