Instructions to use BlueNipples/TimeCrystal-l2-13B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use BlueNipples/TimeCrystal-l2-13B with Transformers:

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

pipe = pipeline("text-generation", model="BlueNipples/TimeCrystal-l2-13B")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("BlueNipples/TimeCrystal-l2-13B")
model = AutoModelForCausalLM.from_pretrained("BlueNipples/TimeCrystal-l2-13B")

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use BlueNipples/TimeCrystal-l2-13B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "BlueNipples/TimeCrystal-l2-13B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "BlueNipples/TimeCrystal-l2-13B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/BlueNipples/TimeCrystal-l2-13B

SGLang

How to use BlueNipples/TimeCrystal-l2-13B 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 "BlueNipples/TimeCrystal-l2-13B" \
    --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": "BlueNipples/TimeCrystal-l2-13B",
		"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 "BlueNipples/TimeCrystal-l2-13B" \
        --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": "BlueNipples/TimeCrystal-l2-13B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use BlueNipples/TimeCrystal-l2-13B with Docker Model Runner:
```
docker model run hf.co/BlueNipples/TimeCrystal-l2-13B
```

Q8_0 GGUF

#14

by BRP - opened Nov 15, 2023

Discussion

BRP

Nov 15, 2023

Dear Matthew Andrews,
Can you please share a q8 file of this model?

BlueNipples

Owner Nov 16, 2023

•

edited Nov 16, 2023

I don't actually think I'm going to be able to do that. My internet is a bit crappy, and I doubt I'll be able to upload a file that large. If you notice the HF bin files are all 2gb each, I broke it up that way so my bad wireless internet could handle it. I was hoping thebloke would get to this one like he did trion. I'll try and get up a q5_k for you.

If you really need q8, download the full model, and use llamacpp to convert.

It runs something like:

py convert.py ./models/TimeCrystal-l2-13B --outfile ./models/TimeCrystal-l2-13B-f16.gguf --outtype f16
./quantize ./models/TimeCrystal-l2-13B-f16.gguf 7

7=q8. It's not a very gpu intensive task at all, and runs quite quickly, creating a gguf

BlueNipples

Owner Nov 16, 2023

Right so there's this. I might add a smaller one too. Just internet being a bit dodgy I try to avoid bigger uploads. But you can do it yourself pretty easy.

https://huggingface.co/BlueNipples/TimeCrystal-l2-13B-GGUF

BRP

Nov 16, 2023

Thank you so much for all your attention and for the instructions on the conversion as well.
For now, I'll download the version you uploaded (5_K_S) and test TimeCrystal l2 Model.

All the best and thanks again!

TheBloke

Nov 16, 2023

I only just saw this. I'm doing it now

TheBloke

Nov 16, 2023

(Feel free to ping me on future model uploads)

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