Instructions to use acrastt/Marx-3B-V2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use acrastt/Marx-3B-V2 with Transformers:

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

pipe = pipeline("text-generation", model="acrastt/Marx-3B-V2")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("acrastt/Marx-3B-V2")
model = AutoModelForCausalLM.from_pretrained("acrastt/Marx-3B-V2", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use acrastt/Marx-3B-V2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "acrastt/Marx-3B-V2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "acrastt/Marx-3B-V2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/acrastt/Marx-3B-V2

SGLang

How to use acrastt/Marx-3B-V2 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 "acrastt/Marx-3B-V2" \
    --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": "acrastt/Marx-3B-V2",
		"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 "acrastt/Marx-3B-V2" \
        --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": "acrastt/Marx-3B-V2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use acrastt/Marx-3B-V2 with Docker Model Runner:
```
docker model run hf.co/acrastt/Marx-3B-V2
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

This is OpenLLaMA 3B V2 finetuned on EverythingLM Data V2(ShareGPT format) for 2 epochs.

Prompt template:

### HUMAN:
{prompt}

### RESPONSE:
<leave a newline for the model to answer>

q4_1 GGML quant available here.
q4_1 GGUF quant available here.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	42.08
AI2 Reasoning Challenge (25-Shot)	44.03
HellaSwag (10-Shot)	72.92
MMLU (5-Shot)	27.84
TruthfulQA (0-shot)	39.92
Winogrande (5-shot)	66.54
GSM8k (5-shot)	1.21

Downloads last month: 99

Safetensors

Model size

3B params

Tensor type

F32

F16

Model tree for acrastt/Marx-3B-V2

Quantizations

5 models

Dataset used to train acrastt/Marx-3B-V2

Spaces using acrastt/Marx-3B-V2 28

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

44.030
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

72.920
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

27.840
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

39.920
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

66.540
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

1.210