Spaces:

Jaycee766
/

FrontierLab-Meta-Hackathon-2026

Running

FrontierLab-Meta-Hackathon-2026 / NOTICE

Deploy FrontierLab v0.16 source 0c2c50a0595b

2636cac verified 1 day ago

6.23 kB

FrontierLab
Copyright (c) 2026 Pack Kwan Low (Jaycee871)

The original source code and documentation in this repository were written by
Pack Kwan Low (Jaycee871) and are licensed under the MIT License (see LICENSE).

The MIT License does NOT cover the third-party material described below. That
material remains under its own terms.

1. Published excerpts in datasets/real_ph_excerpt_cases_v1.json

   This file contains short, attributed excerpts (at most 25 words each) from
   the abstracts of the published papers listed below. Each excerpt remains
   under its source's original license, exactly as recorded in the dataset.
   No full text is bundled.

   - PMC8686790: "A stepwise mutagenesis approach using histidine and acidic amino acid to engineer highly pH-dependent protein switches" (DOI 10.1007/s13205-021-03079-x)
     License: None listed in the dataset record (license_url: null). Check the original rights before any further reuse.
   - PMC6927761: "Structure-based engineering of pH-dependent antibody binding for selective targeting of solid-tumor microenvironment" (DOI 10.1080/19420862.2019.1682866)
     License: https://creativecommons.org/licenses/by-nc/4.0/
   - PMC4314675: "Construction of pH-sensitive Her2-binding IgG1-Fc by directed evolution" (DOI 10.1002/biot.201300483)
     License: https://creativecommons.org/licenses/by/3.0/
   - PMC4622719: "A generic approach to engineer antibody pH-switches using combinatorial histidine scanning libraries and yeast display" (DOI 10.4161/19420862.2014.985993)
     License: https://creativecommons.org/licenses/by-nc/3.0/

2. Meta Muse Glimmer

   Meta Muse Glimmer (meta/muse-glimmer-30b) is licensed under Apache 2.0 by
   Meta (https://dev.meta.ai/models/muse-glimmer). No model weights are
   bundled in this repository. The model is called as a service through
   NVIDIA NIM, or downloaded at runtime by scripts/start_muse_local.sh when the
   user explicitly opts in.

3. Legacy Facebook Research MUSE

   The archived Facebook Research MUSE project (Multilingual Unsupervised and
   Supervised Embeddings; https://github.com/facebookresearch/MUSE) is licensed
   under CC BY-NC 4.0. It is only referenced in docs/LEGACY_MUSE_MULTILINGUAL.md.
   None of its code, embeddings or data are included in this repository.

4. Liquid AI d1 / LFM models

   Liquid AI's d1 models (for example d1-2B, built on LFM3-VL-2B) and other
   LFM models are licensed under the LFM Open License v1.0 by Liquid AI
   (https://www.liquid.ai/lfm-license). They are used only through the
   optional multimodal evidence gate (frontierlab/multimodal_evidence.py),
   which expects a model the user has already loaded. No model weights are
   bundled in this repository, and nothing is downloaded automatically.
   Users who obtain and run these models must comply with the LFM Open
   License v1.0 and keep its required notices.

5. Meta Llama 3.1 8B Instruct

   Meta Llama 3.1 8B Instruct (meta-llama/Llama-3.1-8B-Instruct;
   https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) is licensed under
   the Llama 3.1 Community License Agreement
   (https://www.llama.com/llama3_1/license/), and its use is subject to the
   Llama 3.1 Acceptable Use Policy (https://www.llama.com/llama3_1/use-policy/).
   It is used only through the optional local baseline worker
   (frontierlab/llama31_baseline.py), which expects a runtime the user has
   already loaded. No Llama Materials (model weights, tokenizer or code) are
   bundled in this repository, none are listed in requirements.txt, and
   nothing is downloaded automatically; the hosted Space does not run Llama.
   Users who obtain and run the model must accept and comply with that licence
   and policy. Anyone who distributes Llama Materials, or a product or service
   that contains them, must provide a copy of the licence, prominently display
   "Built with Llama", and keep the following attribution notice:

     "Llama 3.1 is licensed under the Llama 3.1 Community License,
     Copyright © Meta Platforms, Inc. All Rights Reserved."

6. Muse Research Muse2-230M

   Muse2-230M (Muse-research/Muse2-230M;
   https://huggingface.co/Muse-research/Muse2-230M) is published by Muse
   Research, and its model weights are released under the Apache License 2.0.
   It is used only through the optional edge structured worker
   (frontierlab/muse2_edge.py), which expects a runtime the user has already
   loaded. No Muse2 weights, tokenizer or "muse" runtime package are bundled in
   this repository, none are listed in requirements.txt, and nothing is
   downloaded automatically. Users who obtain and run the model must comply
   with the Apache License 2.0 and keep its notices; the model card also states
   that upstream dataset licences and attribution requirements continue to
   apply to derivative uses.

7. DetectorScope

   DetectorScope (https://github.com/google-deepmind/detectorscope; PyPI
   package "detectorscope") is Copyright 2026 Google LLC. Its software is
   licensed under the Apache License 2.0, and its other materials under the
   Creative Commons Attribution 4.0 International License. The released
   detector checkpoints (for example liamdugan/detector-scope and the
   peer-review variants on Hugging Face) are fine-tuned from Google's
   gemma-2-2b and are published under the Gemma terms of use
   (https://ai.google.dev/gemma/terms). They are used only through the optional
   output-audit lens (frontierlab/detectorscope_audit.py), which expects a
   detector the user has already loaded. No DetectorScope code, Gemma weights
   or sparse-autoencoder weights are bundled in this repository, none are
   listed in requirements.txt, and nothing is downloaded automatically. Users
   who install and run them must comply with those terms, and the DetectorScope
   authors' disclaimer applies: the detectors make mistakes and must not be
   used for decisions about individuals.

8. External services

   NVIDIA NIM, Europe PMC, the Springer Nature Meta API and the Springer Nature
   Open Access API are used as remote services. Their use is governed by
   each provider's own terms. No provider code or data is relicensed here,
   beyond the dataset excerpts listed in section 1.

	FrontierLab
	Copyright (c) 2026 Pack Kwan Low (Jaycee871)

	The original source code and documentation in this repository were written by
	Pack Kwan Low (Jaycee871) and are licensed under the MIT License (see LICENSE).

	The MIT License does NOT cover the third-party material described below. That
	material remains under its own terms.

	1. Published excerpts in datasets/real_ph_excerpt_cases_v1.json

	This file contains short, attributed excerpts (at most 25 words each) from
	the abstracts of the published papers listed below. Each excerpt remains
	under its source's original license, exactly as recorded in the dataset.
	No full text is bundled.

	- PMC8686790: "A stepwise mutagenesis approach using histidine and acidic amino acid to engineer highly pH-dependent protein switches" (DOI 10.1007/s13205-021-03079-x)
	License: None listed in the dataset record (license_url: null). Check the original rights before any further reuse.
	- PMC6927761: "Structure-based engineering of pH-dependent antibody binding for selective targeting of solid-tumor microenvironment" (DOI 10.1080/19420862.2019.1682866)
	License: https://creativecommons.org/licenses/by-nc/4.0/
	- PMC4314675: "Construction of pH-sensitive Her2-binding IgG1-Fc by directed evolution" (DOI 10.1002/biot.201300483)
	License: https://creativecommons.org/licenses/by/3.0/
	- PMC4622719: "A generic approach to engineer antibody pH-switches using combinatorial histidine scanning libraries and yeast display" (DOI 10.4161/19420862.2014.985993)
	License: https://creativecommons.org/licenses/by-nc/3.0/

	2. Meta Muse Glimmer

	Meta Muse Glimmer (meta/muse-glimmer-30b) is licensed under Apache 2.0 by
	Meta (https://dev.meta.ai/models/muse-glimmer). No model weights are
	bundled in this repository. The model is called as a service through
	NVIDIA NIM, or downloaded at runtime by scripts/start_muse_local.sh when the
	user explicitly opts in.

	3. Legacy Facebook Research MUSE

	The archived Facebook Research MUSE project (Multilingual Unsupervised and
	Supervised Embeddings; https://github.com/facebookresearch/MUSE) is licensed
	under CC BY-NC 4.0. It is only referenced in docs/LEGACY_MUSE_MULTILINGUAL.md.
	None of its code, embeddings or data are included in this repository.

	4. Liquid AI d1 / LFM models

	Liquid AI's d1 models (for example d1-2B, built on LFM3-VL-2B) and other
	LFM models are licensed under the LFM Open License v1.0 by Liquid AI
	(https://www.liquid.ai/lfm-license). They are used only through the
	optional multimodal evidence gate (frontierlab/multimodal_evidence.py),
	which expects a model the user has already loaded. No model weights are
	bundled in this repository, and nothing is downloaded automatically.
	Users who obtain and run these models must comply with the LFM Open
	License v1.0 and keep its required notices.

	5. Meta Llama 3.1 8B Instruct

	Meta Llama 3.1 8B Instruct (meta-llama/Llama-3.1-8B-Instruct;
	https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) is licensed under
	the Llama 3.1 Community License Agreement
	(https://www.llama.com/llama3_1/license/), and its use is subject to the
	Llama 3.1 Acceptable Use Policy (https://www.llama.com/llama3_1/use-policy/).
	It is used only through the optional local baseline worker
	(frontierlab/llama31_baseline.py), which expects a runtime the user has
	already loaded. No Llama Materials (model weights, tokenizer or code) are
	bundled in this repository, none are listed in requirements.txt, and
	nothing is downloaded automatically; the hosted Space does not run Llama.
	Users who obtain and run the model must accept and comply with that licence
	and policy. Anyone who distributes Llama Materials, or a product or service
	that contains them, must provide a copy of the licence, prominently display
	"Built with Llama", and keep the following attribution notice:

	"Llama 3.1 is licensed under the Llama 3.1 Community License,
	Copyright © Meta Platforms, Inc. All Rights Reserved."

	6. Muse Research Muse2-230M

	Muse2-230M (Muse-research/Muse2-230M;
	https://huggingface.co/Muse-research/Muse2-230M) is published by Muse
	Research, and its model weights are released under the Apache License 2.0.
	It is used only through the optional edge structured worker
	(frontierlab/muse2_edge.py), which expects a runtime the user has already
	loaded. No Muse2 weights, tokenizer or "muse" runtime package are bundled in
	this repository, none are listed in requirements.txt, and nothing is
	downloaded automatically. Users who obtain and run the model must comply
	with the Apache License 2.0 and keep its notices; the model card also states
	that upstream dataset licences and attribution requirements continue to
	apply to derivative uses.

	7. DetectorScope

	DetectorScope (https://github.com/google-deepmind/detectorscope; PyPI
	package "detectorscope") is Copyright 2026 Google LLC. Its software is
	licensed under the Apache License 2.0, and its other materials under the
	Creative Commons Attribution 4.0 International License. The released
	detector checkpoints (for example liamdugan/detector-scope and the
	peer-review variants on Hugging Face) are fine-tuned from Google's
	gemma-2-2b and are published under the Gemma terms of use
	(https://ai.google.dev/gemma/terms). They are used only through the optional
	output-audit lens (frontierlab/detectorscope_audit.py), which expects a
	detector the user has already loaded. No DetectorScope code, Gemma weights
	or sparse-autoencoder weights are bundled in this repository, none are
	listed in requirements.txt, and nothing is downloaded automatically. Users
	who install and run them must comply with those terms, and the DetectorScope
	authors' disclaimer applies: the detectors make mistakes and must not be
	used for decisions about individuals.

	8. External services

	NVIDIA NIM, Europe PMC, the Springer Nature Meta API and the Springer Nature
	Open Access API are used as remote services. Their use is governed by
	each provider's own terms. No provider code or data is relicensed here,
	beyond the dataset excerpts listed in section 1.