About this fork, licensing, and benchmark policies
Quick post to document the background of this fork, open-source licensing compliance, and our general benchmark policies.
1. Why this fork exists
We originally submitted these features (native architecture filters, URL query syncing, schema checks, and UI fixes) directly to AxiomicLabs/Open_SLM_Leaderboard in PR #97.
When the PR was closed without merging, we forked the repository to make those features available to the community. For developers deploying SLMs on edge runtimes (vLLM, GGUF, Ollama, WebGPU), being able to filter between native models and custom code requiring trust_remote_code=True is an essential tool.
2. Licensing and Apache 2.0
This project is an open-source derivative work created when the upstream repository was publicly distributed under the Apache License 2.0.
- Under Section 2 of Apache 2.0, the open-source license grant is perpetual and irrevocable. Relicensing an upstream repository after a fork has already been created applies only to future commits, not existing forks.
- Original attribution to Axiomic Labs is preserved in the footer and source files in accordance with Section 4.
3. Model listings and benchmark data
Benchmark scores, evaluation metrics, and parameter counts are public factual data.
Listing publicly available open-weights models does not imply any endorsement, sponsorship, or formal partnership with model authors. All model entries link directly to their official Hugging Face repositories for full transparency.
4. Submissions and moderation
We welcome new model submissions, benchmark verifications, and metadata corrections through Pull Requests or constructive discussions.
To keep this Space focused and useful for developers, off-topic trolling, duplicate threads, and personal attacks will be closed and removed.