MichelRosselli/ apertus:70b-instruct-2509-bf16

2,582 5 months ago

Apertus is a 70B and 8B parameter language model designed to push the boundaries of fully-open multilingual and transparent models. 

ollama run MichelRosselli/apertus:70b-instruct-2509-bf16

Details

5 months ago

aec13bc8c1f6 · 141GB ·

model
apertus
·
70.6B
·
BF16
template
{{ if .System }}<s><|system_start|>{{ .System }}<|system_end|><|developer_start|>Deliberation: disab
params
{ "stop": [ "<s>", "<|system_start|>", "<|system_end|>", "<|deve

Readme

I AM NOT THE ORIGINAL AUTHOR – JUST RE-UPLOADING APERTUS FOR OLLAMA 🤓
CREDITS TO THE APERTUS / SWISS AI INITIATIVE TEAM, UNSLOTH, AND LLAMA.CPP, SEE BELOW.

Apertus – Attribution & Provenance

This is not the original release of Apertus.
I am not the author of the model and I do not claim ownership of the weights, training, or original chat template.
This Ollama page is simply a community upload intended to make Apertus easier to run with Ollama.

Recommended Ollama version: >= v0.12.6-rc0

Disclaimer

The chat template provided here is intentionally basic and should be considered a convenience template for Ollama, not a full reproduction of the official Hugging Face chat template.

As a result, this upload may not expose the complete behavior or feature set available in the original Hugging Face releases. Features that depend on the richer official template and the surrounding Hugging Face ecosystem may behave differently or be unavailable here.

For the most complete / reference behavior, please refer to the official sources linked below.

Official Sources

  1. Original Model / Official Releases
    Apertus by the Swiss AI Initiative, developed as a collaborative effort between EPFL, ETH Zurich, and CSCS.
    Website:
    https://APERTVS.ai

    Hugging Face:
    https://huggingface.co/swiss-ai

    Official model repositories:

  2. GGUF Sources
    Quantized GGUF releases by Unsloth:

  3. Merge & Packaging
    The sharded GGUF files were merged into a single GGUF file using the official llama.cpp utilities.

License

The official Apertus instruct releases are published under the Apache 2.0 license.
Please refer to the original license files here:

Notes

  • This is an unofficial Ollama upload by MichelRosselli.
  • All credit for the original model belongs to the Apertus / Swiss AI Initiative team.
  • Credit for the GGUF releases goes to Unsloth.
  • Credit to llama.cpp for the utilities used to merge the GGUF shards into a single file.
  • For the canonical release and the fullest template support, please use the official Hugging Face repositories.