Text Generation
Transformers
Safetensors
English
boris2
base-model
small-language-model
causal-lm
custom_code
Instructions to use opencerebral/Boris-2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use opencerebral/Boris-2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="opencerebral/Boris-2", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("opencerebral/Boris-2", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use opencerebral/Boris-2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "opencerebral/Boris-2" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "opencerebral/Boris-2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/opencerebral/Boris-2
- SGLang
How to use opencerebral/Boris-2 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" \ --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": "opencerebral/Boris-2", "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 "opencerebral/Boris-2" \ --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": "opencerebral/Boris-2", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use opencerebral/Boris-2 with Docker Model Runner:
docker model run hf.co/opencerebral/Boris-2
Download benchmarks.csv from opencerebral/Boris-2: direct link, hf CLI and curl.
- Browser
- Download file 2.82 kB
-
https://huggingface.co/opencerebral/Boris-2/resolve/main/benchmarks.csv
- Command line
-
hf download hf://opencerebral/Boris-2/benchmarks.csv
-
curl -L -o benchmarks.csv https://huggingface.co/opencerebral/Boris-2/resolve/main/benchmarks.csv
2.82 kB
| model,checkpoint,tokens_B,phase,hellaswag,arc_easy,arc_challenge,piqa,arithmark3,arithmark_scorer,open_slm_index,val_loss | |
| Boris-2 (main run),export_0010B,10.0,stable,37.12,52.15,27.05,67.95,35.9,boris eval.py,22.19,2.8393 | |
| Boris-2 (main run),export_0020B,20.0,stable,38.56,52.9,28.58,67.85,37.6,boris eval.py,23.47,2.8092 | |
| Boris-2 (main run),export_0030B,30.0,stable,39.42,54.42,29.78,68.39,38.6,boris eval.py,24.82,2.7919 | |
| Boris-2 (main run),export_0040B,40.0,stable,39.32,52.86,27.73,67.25,36.7,boris eval.py,23.05,2.7832 | |
| Boris-2 (main run),export_0050B,50.0,stable,39.8,55.64,29.86,68.17,38.7,boris eval.py,25.1,2.7829 | |
| Boris-2 (main run),export_0060B,60.0,stable,39.9,56.78,28.07,67.9,35.8,boris eval.py,24.18,2.7774 | |
| Boris-2 (main run),export_0070B,70.0,stable,40.05,55.01,29.61,67.52,37.3,boris eval.py,24.34,2.7744 | |
| Boris-2 (main run),export_0080B,80.0,stable,40.37,55.72,28.24,68.01,35.7,boris eval.py,24.23,2.7755 | |
| Boris-2 (main run),export_0090B,90.0,stable,40.54,56.14,28.16,67.52,36.3,boris eval.py,24.22,2.7758 | |
| Boris-2 (main run),export_0100B,100.0,stable,40.75,56.19,29.18,69.37,38.0,boris eval.py,25.91,2.7713 | |
| Boris-2 (main run),export_0110B,110.0,stable,40.51,56.06,29.52,69.75,38.5,boris eval.py,26.19,2.7756 | |
| Boris-2 (main run),export_0120B,120.0,stable,40.65,56.44,30.03,68.28,38.6,boris eval.py,25.62,2.7692 | |
| Boris-2 (main run),export_0130B,130.0,stable,40.95,55.18,28.84,67.95,38.6,boris eval.py,25.11,2.7659 | |
| Boris-2 (main run),export_0140B,140.0,stable,41.07,54.88,28.75,68.12,36.7,boris eval.py,24.72,2.7662 | |
| Boris-2 (main run),export_0150B,150.0,stable,40.92,55.01,30.03,68.39,38.5,boris eval.py,25.5,2.7644 | |
| Boris-2 (main run),export_0160B,160.0,stable,41.81,53.54,29.01,68.77,39.5,boris eval.py,25.81,2.7655 | |
| Boris-2 (main run),export_0170B,170.0,stable,41.8,55.56,29.86,69.7,39.9,boris eval.py,26.94,2.764 | |
| Boris-2 (main run),export_0180B,180.0,decay,46.7,58.46,30.97,71.65,42.0,boris eval.py,31.03,2.6499 | |
| Boris-2 (main run),export_0190B,190.0,decay,48.5,58.8,32.08,71.76,43.1,boris eval.py,32.27,2.594 | |
| Boris-2-Preview-1,final (50B + 2.5B decay),52.5,preview,45.23,58.38,29.95,69.21,37.6,official,27.91,2.6252 | |
| Boris-2-Preview-2,final (100B + 2.5B decay),102.5,preview,46.59,59.13,30.2,70.84,42.2,official,30.57,2.6074 | |
| Boris-2-Preview-3 (internal),final (150B + 2.5B decay),152.5,preview,46.76,58.88,30.46,70.51,43.7,official,30.81,2.5917 | |
| Boris-2-Preview-3b (internal),"final (150B + 2.5B decay, Preview-2 mix)",152.5,preview,46.91,57.79,32.17,71.06,41.1,official,30.66,2.6021 | |
| Boris-2,stable phase end (pre-decay),174.0,stable,44.32,56.78,29.95,69.59,41.7,boris eval.py,28.47,2.7092 | |
| Boris-2,main run final (200B),200.0,final,50.36,60.1,32.25,72.91,42.0,official,33.59,2.5383 | |
| Boris-2,release (200B + 40M final stages),200.04,release,50.26,59.81,32.0,72.91,45.1,official,34.19, | |