Instructions to use nkpz/llama2-22b-blocktriangular-alpaca with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nkpz/llama2-22b-blocktriangular-alpaca with Transformers:

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

pipe = pipeline("text-generation", model="nkpz/llama2-22b-blocktriangular-alpaca")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("nkpz/llama2-22b-blocktriangular-alpaca")
model = AutoModelForCausalLM.from_pretrained("nkpz/llama2-22b-blocktriangular-alpaca")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use nkpz/llama2-22b-blocktriangular-alpaca with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nkpz/llama2-22b-blocktriangular-alpaca"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nkpz/llama2-22b-blocktriangular-alpaca",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/nkpz/llama2-22b-blocktriangular-alpaca

SGLang

How to use nkpz/llama2-22b-blocktriangular-alpaca 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 "nkpz/llama2-22b-blocktriangular-alpaca" \
    --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": "nkpz/llama2-22b-blocktriangular-alpaca",
		"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 "nkpz/llama2-22b-blocktriangular-alpaca" \
        --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": "nkpz/llama2-22b-blocktriangular-alpaca",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use nkpz/llama2-22b-blocktriangular-alpaca with Docker Model Runner:
```
docker model run hf.co/nkpz/llama2-22b-blocktriangular-alpaca
```

There is no official 22b model, this is just a weird experiment, and any potential benefits of doing this have not been validated

https://huggingface.co/chargoddard/llama2-22b-blocktriangular trained one one epoch of 52k rows of Stanford Alpaca. About 11 hours on a 3090.

I had trouble with training using the other 22b method with BLOCK_DIAGONAL=True as done in https://huggingface.co/chargoddard/llama2-22b, but with this method, this is the first time I've been able to target all modules without breaking the output.

target_modules = ["q_proj", "k_proj", "v_proj", "o_proj", "up_proj", "gate_proj", "down_proj"]

Trained at 5e-5 with r=32. For more info see https://wandb.ai/nkpz/huggingface/runs/3oy5nbtv/workspace?workspace=user-nkpz

It's been responding coherently enough that I would need to run some objective benchmarks to determine if this is better/worse than stock llama 13b

Downloads last month: 18

Safetensors

Model size

22B params

Tensor type

F32

F16