Instructions to use MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0 with Transformers:

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

pipe = pipeline("text-generation", model="MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0", trust_remote_code=True)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0

SGLang

How to use MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0 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 "MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0" \
    --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": "MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0",
		"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 "MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0" \
        --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": "MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0 with Docker Model Runner:
```
docker model run hf.co/MagistrTheOne/NULLXES-L-TEX-100M-Stage0a-v0
```

NULLXES-L-TEX-100M-Stage0a-v0 / meta.json

MagistrTheOne

NULLXES-LÆTEX-100M-Stage0a-v0 identity bootstrap

786df5b verified about 2 months ago

Raw

History Blame Contribute Delete

380 Bytes

	{
	"name": "latex-tokenizer-v0.1",
	"version": "0.1",
	"vocab_size_config": 131072,
	"vocab_size_trained": 4144,
	"vocab_size_export": 131072,
	"vocab_padded": true,
	"smoke": false,
	"byte_fallback": true,
	"pretrained_load": false,
	"seed": 42,
	"note": "SP pieces from corpus; unused_* pads to 131072 for embedding table. Expand corpus later to fill real pieces."
	}