Instructions to use akameswa/mixtral-4x7b-instruct-code-old with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use akameswa/mixtral-4x7b-instruct-code-old with Transformers:

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

pipe = pipeline("text-generation", model="akameswa/mixtral-4x7b-instruct-code-old")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("akameswa/mixtral-4x7b-instruct-code-old")
model = AutoModelForCausalLM.from_pretrained("akameswa/mixtral-4x7b-instruct-code-old")
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 akameswa/mixtral-4x7b-instruct-code-old with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "akameswa/mixtral-4x7b-instruct-code-old"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "akameswa/mixtral-4x7b-instruct-code-old",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/akameswa/mixtral-4x7b-instruct-code-old

SGLang

How to use akameswa/mixtral-4x7b-instruct-code-old 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 "akameswa/mixtral-4x7b-instruct-code-old" \
    --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": "akameswa/mixtral-4x7b-instruct-code-old",
		"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 "akameswa/mixtral-4x7b-instruct-code-old" \
        --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": "akameswa/mixtral-4x7b-instruct-code-old",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use akameswa/mixtral-4x7b-instruct-code-old with Docker Model Runner:
```
docker model run hf.co/akameswa/mixtral-4x7b-instruct-code-old
```

mixtral-4x7b-instruct-code

mixtral-4x7b-instruct-code is a MoE of the following models using mergekit:

🧩 Configuration

base_model: akameswa/mistral-7b-instruct-v0.2-bnb-16bit
gate_mode: hidden 
dtype: float16 
experts:
  - source_model: akameswa/mistral-7b-instruct-javascript-16bit
    positive_prompts: ["You are helpful a coding assistant good at javascript"]
  - source_model: akameswa/mistral-7b-instruct-java-16bit
    positive_prompts: ["You are helpful a coding assistant good at java"]
  - source_model: akameswa/mistral-7b-instruct-cpp-16bit
    positive_prompts: ["You are helpful a coding assistant good at cpp"]
  - source_model: akameswa/mistral-7b-instruct-python-16bit
    positive_prompts: ["You are helpful a coding assistant good at python"]

Inference

from transformers import AutoTokenizer
import transformers
import torch

model = "akameswa/mixtral-4x7b-instruct-code-trial"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
    model_kwargs={"load_in_4bit": True},
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)

Link to inference notebook

Downloads last month: 9

Safetensors

Model size

24B params

Tensor type

F16

Collection including akameswa/mixtral-4x7b-instruct-code-old

Mistral

Collection

Mistral models finetuned to improve performance in terms of code generation https://github.com/akameswa/CodeGenerationMoE • 13 items • Updated Mar 2