Instructions to use MetaCore-LLM/MetaCore-1-Test-CPT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MetaCore-LLM/MetaCore-1-Test-CPT with Transformers:

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

pipe = pipeline("text-generation", model="MetaCore-LLM/MetaCore-1-Test-CPT")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("MetaCore-LLM/MetaCore-1-Test-CPT")
model = AutoModelForCausalLM.from_pretrained("MetaCore-LLM/MetaCore-1-Test-CPT", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MetaCore-LLM/MetaCore-1-Test-CPT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MetaCore-LLM/MetaCore-1-Test-CPT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MetaCore-LLM/MetaCore-1-Test-CPT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/MetaCore-LLM/MetaCore-1-Test-CPT

SGLang

How to use MetaCore-LLM/MetaCore-1-Test-CPT 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 "MetaCore-LLM/MetaCore-1-Test-CPT" \
    --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": "MetaCore-LLM/MetaCore-1-Test-CPT",
		"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 "MetaCore-LLM/MetaCore-1-Test-CPT" \
        --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": "MetaCore-LLM/MetaCore-1-Test-CPT",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use MetaCore-LLM/MetaCore-1-Test-CPT with Docker Model Runner:
```
docker model run hf.co/MetaCore-LLM/MetaCore-1-Test-CPT
```

MetaCore-1-Test-CPT

File size: 930 Bytes

4602def

---
language:
- ru
license: mit
library_name: transformers
pipeline_tag: text-generation
tags:
- russian
- llama
- causal-lm
- continued-pretraining
- cpt
- tiny
base_model: MetaCore-LLM/MetaCore-1-Test-Base
model-index:
- name: MetaCore-1-Test-CPT
  results: []
---

# MetaCore-1-Test-CPT

**MetaCore-1-Test-CPT** is a lightweight Russian language model obtained by **continued pre‑training (CPT)** the base model [`MetaCore-1-Test-Base`](https://huggingface.co/MetaCore-LLM/MetaCore-1-Test-Base) on a broader and more diverse Russian text corpus.  
This model serves as an intermediate checkpoint, offering improved language understanding over the base model, and is intended to be further fine‑tuned for specific downstream tasks (e.g., instruction tuning, classification).

- **Developer:** MetaCore-LLM
- **Architecture:** Custom LLaMA-style (tiny config)
- **Language:** Russian
- **Parameter count:** ~16.2 million

---