The A.I Megathread (LLM , GPT , Development)

[bnew]

 

[levitate]

As a graphic artist who has spent countless hours over the course of over two decades becoming expert level good at doing things like clipping out things/people from backgrounds I feel a kinda way about this. They way they train the AI is by letting it watch humans do it the old fashioned way. Wild...

@3Rivers

 

[bnew]

https://archive.is/pLp12

 

[bnew]

Spoiler: larger image

https://archive.is/Lfx3G

 

[bnew]

Spoiler: thread continued

https://archive.is/mESE1

Forefront Chat: Your new AI assistant

Meet your new AI assistant. Choose from powerful models, chat with files, browse the internet, bring your team, customize assistants, share chats, and much more.

chat.forefront.ai

 

[bnew]

GitHub - vijishmadhavan/UnpromptedControl: Remove unwanted objects and restore images without prompts, powered by ControlNet.

Remove unwanted objects and restore images without prompts, powered by ControlNet. - GitHub - vijishmadhavan/UnpromptedControl: Remove unwanted objects and restore images without prompts, powered b...

github.com

ControlNet is a highly regarded tool for guiding StableDiffusion models, and it has been widely acknowledged for its effectiveness. In this repository, A simple hack that allows for the restoration or removal of objects without requiring user prompts. By leveraging this approach, the workflow can be significantly streamlined, leading to enhanced process efficiency.

No-prompt​

 

[bnew]

HuggingChat

Making the community's best AI chat models available to everyone.

huggingface.co

Making the community's best AI chat models available to everyone.
Current Model
OpenAssistant/oasst-sft-6-llama-30b

 

[bnew]

If I want to make an AI singing voice my MY voice (not Kanye or Drake etc) what's a good AI training model website/app?

GitHub - voicepaw/so-vits-svc-fork: so-vits-svc fork with realtime support, improved interface and more features.

so-vits-svc fork with realtime support, improved interface and more features. - GitHub - voicepaw/so-vits-svc-fork: so-vits-svc fork with realtime support, improved interface and more features.

github.com

 

[bnew]

Natural Language Video Search

Natural Language Movie Scene Search Engine​

The goal of this project was to develop a natural language video search engine that could effectively search through large quantities of video data without relying on metadata like titles, descriptions, or audio transcriptions. The aim was to enable users to search for specific actions or scenes, such as closing and opening a door, and facilitate the comparison of these scenes across different videos.

The dataset for this particular tool is roughly ~30,000 scenes from imdbs top 250 movies.

Features​

• Natural Language Search
• Scene Similarity Search

 

[bnew]

[2304.11267] Speed Is All You Need: On-Device Acceleration of Large Diffusion Models via GPU-Aware Optimizations

Speed Is All You Need: On-Device Acceleration of Large Diffusion Models via GPU-Aware Optimizations​

Yu-Hui Chen, Raman Sarokin, Juhyun Lee, Jiuqiang Tang, Chuo-Ling Chang, Andrei Kulik, Matthias Grundmann

The rapid development and application of foundation models have revolutionized the field of artificial intelligence. Large diffusion models have gained significant attention for their ability to generate photorealistic images and support various tasks. On-device deployment of these models provides benefits such as lower server costs, offline functionality, and improved user privacy. However, common large diffusion models have over 1 billion parameters and pose challenges due to restricted computational and memory resources on devices. We present a series of implementation optimizations for large diffusion models that achieve the fastest reported inference latency to-date (under 12 seconds for Stable Diffusion 1.4 without int8 quantization on Samsung S23 Ultra for a 512x512 image with 20 iterations) on GPU-equipped mobile devices. These enhancements broaden the applicability of generative AI and improve the overall user experience across a wide range of devices.

https://arxiv.org/pdf/2304.11267.pdf

 

[bnew]

 

[TRUEST]

The unfocused invention and proliferation of “AI” technologies by every Tom d1ck and harrry will inevitably lead to a weakness of analysis paralysis. You’ll end up in a position where the general public is expected to do the outright ridiculous…like using a tractor trailer as a commuter vehicle to an office job.

 

[bnew]

The unfocused invention and proliferation of “AI” technologies by every Tom d1ck and harrry will inevitably lead to a weakness of analysis paralysis. You’ll end up in a position where the general public is expected to do the outright ridiculous…like using a tractor trailer as a commuter vehicle to an office job.

not sure why you believe it's unfocused when dozens of focused projects sprout up everyday and thousands of models are being trained in specific knowledge areas.

 

[bnew]

GitHub - X-PLUG/mPLUG-Owl: mPLUG-Owl🦉: Modularization Empowers Large Language Models with Multimodality

mPLUG-Owl🦉: Modularization Empowers Large Language Models with Multimodality - GitHub - X-PLUG/mPLUG-Owl: mPLUG-Owl🦉: Modularization Empowers Large Language Models with Multimodality

github.com

About​

mPLUG-Owl: Modularization Empowers Large Language Models with Multimodality

Examples​

News​

• We provide an online demo on modelscope for the public to experience.
• We released code of mPLUG-Owl with its pre-trained and instruction tuning checkpoints.

Spotlights​

• A new training paradigm with a modularized design for large multi-modal language models.
• Learns visual knowledge while support multi-turn conversation consisting of different modalities.
• Observed abilities such as multi-image correlation and scene text understanding, vision-based document comprehension.
• Release a visually-related instruction evaluation set OwlEval.

Online Demo​

Demo of mPLUG-Owl on Modelscope

 

[bnew]

https://archive.is/sOIUV