Instructions to use opencerebral/Boris-2-Think-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use opencerebral/Boris-2-Think-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="opencerebral/Boris-2-Think-Instruct", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("opencerebral/Boris-2-Think-Instruct", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use opencerebral/Boris-2-Think-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "opencerebral/Boris-2-Think-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "opencerebral/Boris-2-Think-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/opencerebral/Boris-2-Think-Instruct

SGLang

How to use opencerebral/Boris-2-Think-Instruct 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 "opencerebral/Boris-2-Think-Instruct" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "opencerebral/Boris-2-Think-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "opencerebral/Boris-2-Think-Instruct" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "opencerebral/Boris-2-Think-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use opencerebral/Boris-2-Think-Instruct with Docker Model Runner:
```
docker model run hf.co/opencerebral/Boris-2-Think-Instruct
```

Boris-2-Think-Instruct

File size: 1,181 Bytes

45c3b6a

model,params,hellaswag,arc_easy,arc_challenge,piqa,mmlu_continuation,winogrande,openbookqa,sciq,commonsense_qa,average_9,source
Boris-2-Think-Instruct,144M,48.56,59.76,32.0,71.6,28.61,51.93,33.0,84.5,19.9,47.76,lm-eval (all 9)
Boris-2-Think,144M,50.13,63.93,34.81,72.31,29.3,53.67,34.4,87.1,20.15,49.53,lm-eval (all 9)
Boris-2-Instruct,144M,48.84,57.74,31.48,71.06,28.41,52.09,34.6,83.6,23.1,47.88,lm-eval (all 9)
Boris-2,144M,50.27,59.76,32.0,72.85,28.96,52.57,36.6,85.1,20.88,48.78,lm-eval (all 9)
Boris-125M,125M,29.25,40.7,23.81,59.58,25.08,49.25,28.2,58.8,19.49,37.13,lm-eval (all 9)
GPT-X3-150M,149M,46.16,59.64,31.31,71.33,28.87,54.62,34.4,80.3,23.1,47.75,leaderboard (HS/ARC/PIQA) + lm-eval (other 5)
SmolLM2-135M,135M,43.22,58.63,29.69,68.44,29.9,52.96,33.4,78.2,19.49,45.99,leaderboard (HS/ARC/PIQA) + lm-eval (other 5)
KLM-1.6,99M,45.28,61.07,33.53,70.18,29.26,52.33,32.8,89.3,18.92,48.07,leaderboard (HS/ARC/PIQA) + lm-eval (other 5)
BananaMind-2.5-Pro,139M,43.6,62.79,31.14,68.12,29.24,52.57,34.0,85.1,20.39,47.44,lm-eval (all 9)
LFM-2.5-350M-Base,354M,45.16,68.86,41.13,67.63,33.32,55.25,36.2,84.9,41.28,52.64,leaderboard (HS/ARC/PIQA) + lm-eval (other 5)