Difference between revisions of "Free Software Directory talk:Artificial Intelligence Team"

From Free Software Directory
Jump to: navigation, search
m (Text: adding livebench)
m (Opening-up-chatgpt.github.io: saving space)
 
(11 intermediate revisions by 2 users not shown)
Line 1: Line 1:
== Free software replacements that are missing ==
+
== Software classification for FSF assessment ==
* AI Research Assistant
+
To assist the FSF in evaluating the page at https://www.fsf.org/news/fsf-is-working-on-freedom-in-machine-learning-applications, we organized the software categories here accordingly. This structure aims to clarify the criteria for determining when a machine learning application can be considered free, ensuring that users are empowered to control their computing.
** https://elicit.org/ - Elicit uses language models to help you automate research workflows, like parts of literature review.
 
* Voice to instrument: [https://sites.research.google/tonetransfer Tone Transfer-like]
 
* Identification
 
** Photo
 
*** [https://play.google.com/store/apps/details?id=org.plantnet Pl@ntNet for Android] - Pl@ntNet is a citizen science project for automatic plant identification through photographs and based on machine learning. "The observations shared by the community are published with the associated images under a Creative Common CC-BY-SA license (visible author name)." - https://plantnet.org/en/2020/08/06/your-plntnet-data-integrated-into-gbif/
 
** Audio
 
*** Shazam: Shazam is an application that can identify music, movies, advertising, and television shows, based on a short sample played and using the microphone on the device.
 
*** A Shazam-like software that is identifying genres instead of songs.
 
*** A free app that functions like midomi.com -- "You can find songs with midomi and your own voice. Forgot the name of a song? Heard a bit of one on the radio? All you need is your computer's microphone."
 
* http://design.rxnfinder.org/addictedchem/prediction/
 
  
== Potential Freedom issues ==
+
=== Examples ===
 +
==== Opening-up-chatgpt.github.io ====
 +
* '''Availability''': Free code, LLM data, LLM weights, RL data, RL weights, License 
 +
* '''Documentation''': Code, Architecture, Preprint, Paper, Modelcard, Datasheet
 +
* '''Access''': Package, API
  
* Dependencies need to be checked. (e.g. [https://docs.librechat.ai/ LibreChat])
+
==== USA Bipartisan House Task Force Report on AI (Dec '24) ====
* Verify whether a workflow requires non-free GPU or if CPU can be used.
+
[https://republicans-science.house.gov/_cache/files/a/a/aa2ee12f-8f0c-46a3-8ff8-8e4215d6a72b/E4AF21104CB138F3127D8FF7EA71A393.ai-task-force-report-final.pdf Federal AI Governance and Transparency (Page 4)]
* The training data often contains non-free licensed material.
+
* '''Data and Metadata''': information about the data that was used to train, test, or fine-tune the model, including information about the data’s sources or provenance, collection methods, sample size, procedures for cleaning the data, bias and skewness, inclusion of protected characteristics or proxy features, and ultimate integrity.
** According to current copyright laws, this does not impact the license of the model or the output of the model. According to current copyright laws, the output is public domain. [[User:Mmcmahon|Mmcmahon]] ([[User talk:Mmcmahon|talk]]) 11:48, 2 May 2023 (EDT)
+
* '''Software''': information about the software components and their origins.
 +
* '''Model Development''': information about the training, tuning, validation, and testing of the AI system, who requested its development, who developed the model, communities consulted in development, the development process, tools used in development, the model’s intended uses and known limitations, and metrics pertaining to the model’s efficiency, performance, bias, and energy usage.
 +
* '''Model Deployment''': information about model deployment and monitoring, including metrics identified in model development, plans to provide notice and explanation of the use of AI models to members of the public impacted by the model’s use, and any ongoing training, validation, and testing.
 +
* '''Model Use''': information about how the model is used, including the organizational context and design of the entire system in which the model is deployed, specific use case applications of the model, the information that a deployed model utilizes, the types of determinations or decisions the model is intended to inform, meaningful explanations of the model and its outcomes given relevant stakeholders, the policies for how to handle outputs, the risks of harm identified, and risk mitigation plans including human oversight or intervention.
  
=== Legal stuffs ===
+
== Mechanics ==
* [https://documents.un.org/doc/undoc/ltd/n24/065/92/pdf/n2406592.pdf?token=cHeoDbvIBOFdwcPVZ6&fe=true UN Resolution: Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development]
+
It's differentiation on N-dimensional arrays. (n-dimensional means n-amount of axes i.e. x,y,z)
  
==== USA ====
+
Mess with variables: predicted, actual, and error.
* [https://www.whitehouse.gov/briefing-room/statements-releases/2023/10/30/fact-sheet-president-biden-issues-executive-order-on-safe-secure-and-trustworthy-artificial-intelligence/ FACT SHEET: President Biden Issues Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence]
+
=== Python ===
* [https://www.whitehouse.gov/ostp/ai-bill-of-rights/ Blueprint for an AI Bill of Rights]
+
==== Array libraries ====
 +
* [https://numpy.org/doc/stable/index.html Numpy]
 +
==== Differentiation libraries ====
 +
* [https://pytorch.org/ PyTorch]
 +
* [https://www.tensorflow.org/ TensorFlow]
 +
* [https://jax.readthedocs.io/en/latest/index.html JAX]
 +
* [https://github.com/rsokl/MyGrad MyGrad]
  
=== Model licenses ===
+
== Combatting censorship ==
There appears to be a swath of custom model licenses being used independent of the more standardized software licenses used to interact with models. This presents a conflict as to what license is deemed applicable to the files contained in any repo.
+
=== Decentralized LLMS ===
 +
==== Distributed training ====
 +
[https://github.com/PrimeIntellect-ai/prime Prime Intellect] [Apache 2.0]
 +
==== Peer-to-peer generation ====
 +
"[https://github.com/mudler/LocalAI LocalAI] [https://github.com/mudler/LocalAI#MIT-1-ov-file [MIT]] uses https://github.com/libp2p/go-libp2p [MIT] under the hood, the same project powering IPFS. Differently from other frameworks, LocalAI uses peer2peer without a single master server, but rather it uses sub/gossip and ledger functionalities to achieve consensus across different peers.
  
[https://www.reddit.com/r/LocalLLaMA/comments/13t2b67/security_psa_huggingface_models_are_code_not_just/ Reddit - Security PSA: huggingface models are code. not just data.]
+
EdgeVPN [Apache 2.0] is used as a library to establish the network and expose the ledger functionality under a shared token to ease out automatic discovery and have separated, private peer2peer networks.
  
[https://youtube.com/watch?v=W5M-dvzpzSQ?t=16m50s This video (starting at 16:50) illustrates a good argument that model checkpoints may not fall under copyright protection so traditional software licenses that depend on copyright law would be invalid.] The video does illustrate that contract law may try to be used it place of copyright. I would advise not using YouTube directly and instead using yt-dl or Invidious.
+
The weights are split proportional to the memory when running into worker mode, when in federation mode each request is split to every node which have to load the model fully." - [https://localai.io/features/distribute/ source]
  
Worth noting: [https://github.com/openlm-research/open_llama Open LLaMA (out-of-date)] is an example of removing the issue of [https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md model cards] for text generation.
+
=== Isolating a web service like Gradio from internet access ===
 +
Ensures less snooping on local models - [https://rentry.org/IsolatedLinuxWebService Linux Safe Web Service]
  
StabilityAI [https://github.com/Stability-AI/stablediffusion/blob/main/LICENSE-MODEL keeps (Nov '22)] [https://github.com/Stability-AI/generative-models/blob/main/model_licenses/LICENSE-SDXL1.0 updating (July '23)] [https://github.com/Stability-AI/generative-models/blob/main/model_licenses/LICENSE-SDXL-Turbo its licenses (Nov '23)], so I've removed them from the main page until they settle down.
+
=== Decensoring censored models ===
 +
[https://erichartford.com/uncensored-models A guide to decensoring models]; Exercise caution, as it stands to reason an inherently uncensored model would perform better than needing the legwork of decensoring one (and then making mistakes + missing some of the censorship)
  
[https://github.com/gligen/GLIGEN GLIGEN] and [https://github.com/mut-ex/gligen-gui GLIGEN GUI] is quite neat, but states strict model terms and conditions associated with using it.
+
==== Examples of censorship ====
 +
* [https://www.reddit.com/r/LocalLLaMA/comments/1ctiggk/if_you_ask_deepseekv2_through_the_official_site/ If you ask Deepseek-V2 (through the official site) 'What happened at Tienanmen square?', it deletes your question and clears the context.]
  
== Testing model viability ==
+
== Testing models ==
 
Tools are needed to assess the pros/cons of each model.
 
Tools are needed to assess the pros/cons of each model.
  
Line 74: Line 84:
  
 
* [https://livebench.ai/ LiveBench: A Challenging, Contamination-Free LLM Benchmark]
 
* [https://livebench.ai/ LiveBench: A Challenging, Contamination-Free LLM Benchmark]
 +
 +
* [https://opening-up-chatgpt.github.io/ LLM Openness]
  
 
* [https://chat.lmsys.org/?leaderboard LMSYS Chatbot Arena]
 
* [https://chat.lmsys.org/?leaderboard LMSYS Chatbot Arena]
 +
 +
* [https://mixeval.github.io/#leaderboard MixEval]
 +
 +
* [https://novelchallenge.github.io/index.html NoCha]
  
 
* [https://rank.opencompass.org.cn/home OpenCompass (China)]
 
* [https://rank.opencompass.org.cn/home OpenCompass (China)]
Line 82: Line 98:
  
 
* [https://predibase.com/fine-tuning-index Predibase's Open-source Model Fine-Tuning Leaderboard]
 
* [https://predibase.com/fine-tuning-index Predibase's Open-source Model Fine-Tuning Leaderboard]
 +
 +
* [https://scale.com/leaderboard SEAL leaderboards]
 +
 +
* [https://simple-bench.com/index.html Simple Bench - Basic Reasoning]
  
 
* [https://prollm.toqan.ai/leaderboard Toqan's ProLLM Benchmarks: Summarization]
 
* [https://prollm.toqan.ai/leaderboard Toqan's ProLLM Benchmarks: Summarization]
Line 88: Line 108:
  
 
===== Coding =====
 
===== Coding =====
 +
 +
* [https://aider.chat/docs/leaderboards/ Aider]
  
 
* [https://fudanselab-classeval.github.io/leaderboard.html ClassEval]
 
* [https://fudanselab-classeval.github.io/leaderboard.html ClassEval]
Line 98: Line 120:
  
 
* [https://livecodebench.github.io/leaderboard.html LiveCodeBench]
 
* [https://livecodebench.github.io/leaderboard.html LiveCodeBench]
 +
 +
* [https://scale.com/leaderboard/coding SEAL leaderboards' coding prompt set]
  
 
* [https://leaderboard.tabbyml.com/ TabbyML Coding LLMs]
 
* [https://leaderboard.tabbyml.com/ TabbyML Coding LLMs]
Line 104: Line 128:
  
 
==== Voice ====
 
==== Voice ====
 +
* [https://artificialanalysis.ai/speech-to-text Artificial Analysis Speech to Text (ASR) Leaderboard]
 
* [https://huggingface.co/spaces/TTS-AGI/TTS-Arena Huggingface TTS Arena: Benchmarking TTS Models in the Wild]
 
* [https://huggingface.co/spaces/TTS-AGI/TTS-Arena Huggingface TTS Arena: Benchmarking TTS Models in the Wild]
 
* [https://github.com/Vaibhavs10/open-tts-tracker Open TTS Tracker]
 
* [https://github.com/Vaibhavs10/open-tts-tracker Open TTS Tracker]
Line 117: Line 142:
  
 
* Hyena
 
* Hyena
 +
* Mamba/Jamba
  
* Mamba
+
* [https://github.com/vladmandic/automatic/wiki/Models List of popular text-to-image generative models with their respective parameters and architecture overview]
  
 
==== Ordinal value scales could exist for ====
 
==== Ordinal value scales could exist for ====
Line 140: Line 166:
 
* [https://epochai.org/data/epochdb/table?show=compute-intensive Computation costs and datasets for many models]
 
* [https://epochai.org/data/epochdb/table?show=compute-intensive Computation costs and datasets for many models]
 
* [https://arxiv.org/pdf/2404.07413 LLaMA 2 performance reached with $100k]
 
* [https://arxiv.org/pdf/2404.07413 LLaMA 2 performance reached with $100k]
 +
* [https://arxiv.org/abs/2407.15811 Image Stable Diffusion performance reached with $2k]
  
 
=== General trends ===
 
=== General trends ===
Line 145: Line 172:
 
* Quantization (a la GPT-Q) allows consumer hardware to run large models.
 
* Quantization (a la GPT-Q) allows consumer hardware to run large models.
  
== Stable Diffusion ==
+
== Legal ==
 +
* [https://documents.un.org/doc/undoc/ltd/n24/065/92/pdf/n2406592.pdf?token=cHeoDbvIBOFdwcPVZ6&fe=true UN Resolution: Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development]
 +
 
 +
=== Potential Freedom issues ===
 +
 
 +
* Dependencies need to be checked. (e.g. [https://docs.librechat.ai/ LibreChat])
 +
* Verify whether a workflow requires non-free GPU or if CPU can be used.
 +
* The training data often contains non-free licensed material.
 +
** According to current copyright laws, this does not impact the license of the model or the output of the model. According to current copyright laws, the output is public domain. [[User:Mmcmahon|Mmcmahon]] ([[User talk:Mmcmahon|talk]]) 11:48, 2 May 2023 (EDT)
 +
 
 +
==== Model licenses ====
 +
There appears to be a swath of custom model licenses being used independent of the more standardized software licenses used to interact with models. This presents a conflict as to what license is deemed applicable to the files contained in any repo.
 +
 
 +
[https://www.reddit.com/r/LocalLLaMA/comments/13t2b67/security_psa_huggingface_models_are_code_not_just/ Reddit - Security PSA: huggingface models are code. not just data.]
 +
 
 +
[https://youtube.com/watch?v=W5M-dvzpzSQ?t=16m50s This video (starting at 16:50) illustrates a good argument that model checkpoints may not fall under copyright protection so traditional software licenses that depend on copyright law would be invalid.] The video does illustrate that contract law may try to be used it place of copyright. I would advise not using YouTube directly and instead using yt-dl or Invidious.
 +
 
 +
Worth noting: [https://github.com/openlm-research/open_llama Open LLaMA (out-of-date)] is an example of removing the issue of [https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md model cards] for text generation.
 +
 
 +
StabilityAI [https://github.com/Stability-AI/stablediffusion/blob/main/LICENSE-MODEL keeps (Nov '22)] [https://github.com/Stability-AI/generative-models/blob/main/model_licenses/LICENSE-SDXL1.0 updating (July '23)] [https://github.com/Stability-AI/generative-models/blob/main/model_licenses/LICENSE-SDXL-Turbo its licenses (Nov '23)], so I've removed them from the main page until they settle down.
 +
 
 +
[https://github.com/gligen/GLIGEN GLIGEN] and [https://github.com/mut-ex/gligen-gui GLIGEN GUI] is quite neat, but states strict model terms and conditions associated with using it.
 +
 
 +
== Image Models ==
 +
=== Stable Diffusion ===
 
Stable Diffusion model files (.ckpt) are released under a non-free license.
 
Stable Diffusion model files (.ckpt) are released under a non-free license.
  
Line 159: Line 210:
 
*** [https://github.com/deforum-art/deforum-for-automatic1111-webui Deforum]
 
*** [https://github.com/deforum-art/deforum-for-automatic1111-webui Deforum]
  
== Large Language Models ==
+
== Free software replacements that are missing ==
=== Censorship issues ===
+
* AI Research Assistant
[https://erichartford.com/uncensored-models A guide to decensoring models]; I would exercise caution, as it stands to reason an inherently uncensored model would perform better than needing the legwork of decensoring one (and then making mistakes + missing some of the censorship)
+
** https://elicit.org/ - Elicit uses language models to help you automate research workflows, like parts of literature review.
 
+
* Voice to instrument: [https://sites.research.google/tonetransfer Tone Transfer-like]
==== Examples ====
+
* Identification
* [https://www.reddit.com/r/LocalLLaMA/comments/1ctiggk/if_you_ask_deepseekv2_through_the_official_site/ If you ask Deepseek-V2 (through the official site) 'What happened at Tienanmen square?', it deletes your question and clears the context.]
+
** Photo
 
+
*** [https://play.google.com/store/apps/details?id=org.plantnet Pl@ntNet for Android] - Pl@ntNet is a citizen science project for automatic plant identification through photographs and based on machine learning. "The observations shared by the community are published with the associated images under a Creative Common CC-BY-SA license (visible author name)." - https://plantnet.org/en/2020/08/06/your-plntnet-data-integrated-into-gbif/
== Isolating a web service like Gradio from internet access ==
+
** Audio
 +
*** Shazam: Shazam is an application that can identify music, movies, advertising, and television shows, based on a short sample played and using the microphone on the device.
 +
*** A Shazam-like software that is identifying genres instead of songs.
 +
*** A free app that functions like midomi.com -- "You can find songs with midomi and your own voice. Forgot the name of a song? Heard a bit of one on the radio? All you need is your computer's microphone."
 +
* http://design.rxnfinder.org/addictedchem/prediction/
  
Seems useful to ensure less snooping on local models - [https://rentry.org/IsolatedLinuxWebService Linux Safe Web Service]
 
  
 
== Legacy notable projects ==
 
== Legacy notable projects ==

Latest revision as of 01:06, 21 December 2024

Software classification for FSF assessment

To assist the FSF in evaluating the page at https://www.fsf.org/news/fsf-is-working-on-freedom-in-machine-learning-applications, we organized the software categories here accordingly. This structure aims to clarify the criteria for determining when a machine learning application can be considered free, ensuring that users are empowered to control their computing.

Examples

Opening-up-chatgpt.github.io

  • Availability: Free code, LLM data, LLM weights, RL data, RL weights, License
  • Documentation: Code, Architecture, Preprint, Paper, Modelcard, Datasheet
  • Access: Package, API

USA Bipartisan House Task Force Report on AI (Dec '24)

Federal AI Governance and Transparency (Page 4)

  • Data and Metadata: information about the data that was used to train, test, or fine-tune the model, including information about the data’s sources or provenance, collection methods, sample size, procedures for cleaning the data, bias and skewness, inclusion of protected characteristics or proxy features, and ultimate integrity.
  • Software: information about the software components and their origins.
  • Model Development: information about the training, tuning, validation, and testing of the AI system, who requested its development, who developed the model, communities consulted in development, the development process, tools used in development, the model’s intended uses and known limitations, and metrics pertaining to the model’s efficiency, performance, bias, and energy usage.
  • Model Deployment: information about model deployment and monitoring, including metrics identified in model development, plans to provide notice and explanation of the use of AI models to members of the public impacted by the model’s use, and any ongoing training, validation, and testing.
  • Model Use: information about how the model is used, including the organizational context and design of the entire system in which the model is deployed, specific use case applications of the model, the information that a deployed model utilizes, the types of determinations or decisions the model is intended to inform, meaningful explanations of the model and its outcomes given relevant stakeholders, the policies for how to handle outputs, the risks of harm identified, and risk mitigation plans including human oversight or intervention.

Mechanics

It's differentiation on N-dimensional arrays. (n-dimensional means n-amount of axes i.e. x,y,z)

Mess with variables: predicted, actual, and error.

Python

Array libraries

Differentiation libraries

Combatting censorship

Decentralized LLMS

Distributed training

Prime Intellect [Apache 2.0]

Peer-to-peer generation

"LocalAI [MIT] uses https://github.com/libp2p/go-libp2p [MIT] under the hood, the same project powering IPFS. Differently from other frameworks, LocalAI uses peer2peer without a single master server, but rather it uses sub/gossip and ledger functionalities to achieve consensus across different peers.

EdgeVPN [Apache 2.0] is used as a library to establish the network and expose the ledger functionality under a shared token to ease out automatic discovery and have separated, private peer2peer networks.

The weights are split proportional to the memory when running into worker mode, when in federation mode each request is split to every node which have to load the model fully." - source

Isolating a web service like Gradio from internet access

Ensures less snooping on local models - Linux Safe Web Service

Decensoring censored models

A guide to decensoring models; Exercise caution, as it stands to reason an inherently uncensored model would perform better than needing the legwork of decensoring one (and then making mistakes + missing some of the censorship)

Examples of censorship

Testing models

Tools are needed to assess the pros/cons of each model.

Meta-Leaderboards

Leaderboards

Due to the issue of merely training a model to become good at whatever tests are on a leaderboard, multiple leaderboards are preferential (hence not putting HuggingFace on the main page). A more comprehensive evaluation would be a meta-analysis of existing leaderboards.

Text

Coding

Voice

Benchmarks

Architectures

  • Hyena
  • Mamba/Jamba

Ordinal value scales could exist for

Source of model training data

  • amount of data
  • date range (e.g. distinguishing old science from new science for smaller scale models)
  • level of censorship (important to make personal+research use distinct from business use)

Problem solving

  • math
  • creative problem solving (there exists methodology for testing this in humans)

"The present findings suggest that the current state of AI language models demonstrate higher creative potential than human respondents." Nature (Feb 10, '24)

Computation Costs

General trends

  • Larger models are more prone to human superstition[1], but also generate more human-like readability.
  • Quantization (a la GPT-Q) allows consumer hardware to run large models.

Legal

Potential Freedom issues

  • Dependencies need to be checked. (e.g. LibreChat)
  • Verify whether a workflow requires non-free GPU or if CPU can be used.
  • The training data often contains non-free licensed material.
    • According to current copyright laws, this does not impact the license of the model or the output of the model. According to current copyright laws, the output is public domain. Mmcmahon (talk) 11:48, 2 May 2023 (EDT)

Model licenses

There appears to be a swath of custom model licenses being used independent of the more standardized software licenses used to interact with models. This presents a conflict as to what license is deemed applicable to the files contained in any repo.

Reddit - Security PSA: huggingface models are code. not just data.

This video (starting at 16:50) illustrates a good argument that model checkpoints may not fall under copyright protection so traditional software licenses that depend on copyright law would be invalid. The video does illustrate that contract law may try to be used it place of copyright. I would advise not using YouTube directly and instead using yt-dl or Invidious.

Worth noting: Open LLaMA (out-of-date) is an example of removing the issue of model cards for text generation.

StabilityAI keeps (Nov '22) updating (July '23) its licenses (Nov '23), so I've removed them from the main page until they settle down.

GLIGEN and GLIGEN GUI is quite neat, but states strict model terms and conditions associated with using it.

Image Models

Stable Diffusion

Stable Diffusion model files (.ckpt) are released under a non-free license.

Here's the stable diffusion beginning point: https://huggingface.co/CompVis/stable-diffusion-v1-4 https://huggingface.co/spaces/CompVis/stable-diffusion-license

stable-diffusion-webui

Free software replacements that are missing

  • AI Research Assistant
    • https://elicit.org/ - Elicit uses language models to help you automate research workflows, like parts of literature review.
  • Voice to instrument: Tone Transfer-like
  • Identification
    • Photo
    • Audio
      • Shazam: Shazam is an application that can identify music, movies, advertising, and television shows, based on a short sample played and using the microphone on the device.
      • A Shazam-like software that is identifying genres instead of songs.
      • A free app that functions like midomi.com -- "You can find songs with midomi and your own voice. Forgot the name of a song? Heard a bit of one on the radio? All you need is your computer's microphone."
  • http://design.rxnfinder.org/addictedchem/prediction/


Legacy notable projects

Project Credit License Description
DALL-E Mini borisdayma (Boris Dayma) Apache 2.0 Generate images from a text prompt
neural-style anishathalye GPLv3 An implementation of neural style in TensorFlow

External links



Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license is included in the page “GNU Free Documentation License”.

The copyright and license notices on this page only apply to the text on this page. Any software or copyright-licenses or other similar notices described in this text has its own copyright notice and license, which can usually be found in the distribution or license text itself.