Instructions to use winglian/CaptainZed with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use winglian/CaptainZed with Transformers:

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

pipe = pipeline("text-generation", model="winglian/CaptainZed")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("winglian/CaptainZed")
model = AutoModelForCausalLM.from_pretrained("winglian/CaptainZed")
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 winglian/CaptainZed with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/winglian/CaptainZed

SGLang

How to use winglian/CaptainZed 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 "winglian/CaptainZed" \
    --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": "winglian/CaptainZed",
		"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 "winglian/CaptainZed" \
        --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": "winglian/CaptainZed",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use winglian/CaptainZed with Docker Model Runner:
```
docker model run hf.co/winglian/CaptainZed
```

What's it?

by ehartford - opened Feb 3, 2024

Discussion

ehartford

Feb 3, 2024

What's it?

adarshxs

Feb 3, 2024

They say its a mystery never to be revealed. raw model weights where each forward pass sends a chill down your spine.

lmao even I'm curious

ehartford

Feb 3, 2024

•

edited Feb 3, 2024

Inquisitor-agent @fblgit , your unique skills are needed here!

ehartford

Feb 3, 2024

It's a llama-7b, 32k context

fblgit

Feb 4, 2024

code is out https://github.com/fblgit/model-similarity
Now you have to think on the possible candidates and compare the model against them. Remember, the tool doesn't says whats the base model.. it just say which model is more similar to each other and where.

dustydecapod

Feb 6, 2024

@ehartford more likely a mistral, though in llama's clothing. note the sliding window context, the mistral-sized layers, etc.

ehartford

Feb 6, 2024

Chat template is mistral-instruct

https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/blob/b70aa86578567ba3301b21c8a27bea4e8f6d6d61/tokenizer_config.json#L42

https://huggingface.co/winglian/CaptainZed/blob/b901beeedc76b61a5d7aa2bd6a560451f8154d44/tokenizer_config.json#L32

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