Instructions to use ajibawa-2023/carl-llama-2-13b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ajibawa-2023/carl-llama-2-13b with Transformers:

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

pipe = pipeline("text-generation", model="ajibawa-2023/carl-llama-2-13b")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ajibawa-2023/carl-llama-2-13b")
model = AutoModelForCausalLM.from_pretrained("ajibawa-2023/carl-llama-2-13b", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ajibawa-2023/carl-llama-2-13b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ajibawa-2023/carl-llama-2-13b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ajibawa-2023/carl-llama-2-13b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ajibawa-2023/carl-llama-2-13b

SGLang

How to use ajibawa-2023/carl-llama-2-13b 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 "ajibawa-2023/carl-llama-2-13b" \
    --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": "ajibawa-2023/carl-llama-2-13b",
		"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 "ajibawa-2023/carl-llama-2-13b" \
        --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": "ajibawa-2023/carl-llama-2-13b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ajibawa-2023/carl-llama-2-13b with Docker Model Runner:
```
docker model run hf.co/ajibawa-2023/carl-llama-2-13b
```

Upload 15 files

by Feanix - opened Aug 15, 2023

base: refs/heads/main

←

from: refs/pr/3

Discussion Files changed

+45

-0

Feanix

Aug 15, 2023

GGML Quantizations for CPU Inference

Upload 15 files7131e2a9

ajibawa-2023 changed pull request status to merged Aug 15, 2023

ajibawa-2023

Owner Aug 15, 2023

Thanks @Feanix for GGML quantz. Appreciate your work!

Feanix

Aug 15, 2023

The file names MIGHT need to be changed but I needed GGML quantz for myself so I might as well share. I renamed the files on my computer but re-uploading 100+GB isn't really in my books right now, if that's okay? Oobabooga wasn't detecting that they were Llama-2 models because of how I named it. I could alternative upload a yaml (I think?) but that might just be confusing for those who would probably benefit most from access to this. Anywho, I'm on to testing. I'm really optimistic given my quick experience so far. Despite my goals and focus, I could do the other two Carl's and/or Scarlett if there is demand, however little. It's a script pointed at a folder so nbd.

ajibawa-2023

Owner Aug 15, 2023

Ok, I will release few more Scarlett models either today or tomorrow. You can do the GGML quant. I highly appreciate your efforts. Thank you!

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