Instructions to use mixtao/MixTAO-7Bx2-MoE-v8.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mixtao/MixTAO-7Bx2-MoE-v8.1 with Transformers:

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

pipe = pipeline("text-generation", model="mixtao/MixTAO-7Bx2-MoE-v8.1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mixtao/MixTAO-7Bx2-MoE-v8.1")
model = AutoModelForCausalLM.from_pretrained("mixtao/MixTAO-7Bx2-MoE-v8.1")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mixtao/MixTAO-7Bx2-MoE-v8.1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mixtao/MixTAO-7Bx2-MoE-v8.1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mixtao/MixTAO-7Bx2-MoE-v8.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mixtao/MixTAO-7Bx2-MoE-v8.1

SGLang

How to use mixtao/MixTAO-7Bx2-MoE-v8.1 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 "mixtao/MixTAO-7Bx2-MoE-v8.1" \
    --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": "mixtao/MixTAO-7Bx2-MoE-v8.1",
		"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 "mixtao/MixTAO-7Bx2-MoE-v8.1" \
        --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": "mixtao/MixTAO-7Bx2-MoE-v8.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mixtao/MixTAO-7Bx2-MoE-v8.1 with Docker Model Runner:
```
docker model run hf.co/mixtao/MixTAO-7Bx2-MoE-v8.1
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

MixTAO-7Bx2-MoE

MixTAO-7Bx2-MoE is a Mixture of Experts (MoE). This model is mainly used for large model technology experiments, and increasingly perfect iterations will eventually create high-level large language models.

Prompt Template (Alpaca)

### Instruction:
<prompt> (without the <>)
### Response:

🦒 Colab

Link	Info - Model Name
	MixTAO-7Bx2-MoE-v8.1
mixtao-7bx2-moe-v8.1.Q4_K_M.gguf	GGUF of MixTAO-7Bx2-MoE-v8.1 Only Q4_K_M in https://huggingface.co/zhengr/MixTAO-7Bx2-MoE-v8.1-GGUF
Demo Space	https://huggingface.co/spaces/zhengr/MixTAO-7Bx2-MoE-v8.1/

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	77.50
AI2 Reasoning Challenge (25-Shot)	73.81
HellaSwag (10-Shot)	89.22
MMLU (5-Shot)	64.92
TruthfulQA (0-shot)	78.57
Winogrande (5-shot)	87.37
GSM8k (5-shot)	71.11

Downloads last month: 16,063

Safetensors

Model size

13B params

Tensor type

BF16

Model tree for mixtao/MixTAO-7Bx2-MoE-v8.1

Merges

3 models

Quantizations

4 models

Spaces using mixtao/MixTAO-7Bx2-MoE-v8.1 2

Evaluation results

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

73.810
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

89.220
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

64.920
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

78.570
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

87.370
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

71.110