Instructions to use DiscoResearch/DiscoLM-mixtral-8x7b-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DiscoResearch/DiscoLM-mixtral-8x7b-v2 with Transformers:

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

pipe = pipeline("text-generation", model="DiscoResearch/DiscoLM-mixtral-8x7b-v2", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("DiscoResearch/DiscoLM-mixtral-8x7b-v2", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("DiscoResearch/DiscoLM-mixtral-8x7b-v2", trust_remote_code=True, device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use DiscoResearch/DiscoLM-mixtral-8x7b-v2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DiscoResearch/DiscoLM-mixtral-8x7b-v2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DiscoResearch/DiscoLM-mixtral-8x7b-v2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/DiscoResearch/DiscoLM-mixtral-8x7b-v2

SGLang

How to use DiscoResearch/DiscoLM-mixtral-8x7b-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 "DiscoResearch/DiscoLM-mixtral-8x7b-v2" \
    --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": "DiscoResearch/DiscoLM-mixtral-8x7b-v2",
		"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 "DiscoResearch/DiscoLM-mixtral-8x7b-v2" \
        --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": "DiscoResearch/DiscoLM-mixtral-8x7b-v2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use DiscoResearch/DiscoLM-mixtral-8x7b-v2 with Docker Model Runner:
```
docker model run hf.co/DiscoResearch/DiscoLM-mixtral-8x7b-v2
```

Trying to quantize. Running into the issue below. Any suggestions?

by BigDeeper - opened Dec 10, 2023

Discussion

BigDeeper

Dec 10, 2023

Permuting layer 30
Permuting layer 31
model.embed_tokens.weight -> token_embd.weight | F16 | [32002, 4096]
model.layers.0.input_layernorm.weight -> blk.0.attn_norm.weight | F16 | [4096]
Traceback (most recent call last):
File "/home/developer/llama.cpp/convert.py", line 1228, in
main()
File "/home/developer/llama.cpp/convert.py", line 1215, in main
model = convert_model_names(model, params)
File "/home/developer/llama.cpp/convert.py", line 1004, in convert_model_names
raise Exception(f"Unexpected tensor name: {name}")
Exception: Unexpected tensor name: model.layers.0.mlp.experts.0.w1.weight

bjoernp

Disco Research org Dec 10, 2023

Llama.cpp doesn't have support for this architecture. They'll probably wait until official arch is released before implementing

bjoernp changed discussion status to closed Dec 10, 2023

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment