lee101
Text generator is interesting more because it goes too much higher length to understand input content such as crawling the web for any given input links downloading images and analysing the content/ embedded text etc.
Both strategies and more including training new networks especially for harder problems ( like generating 3D content or designing and finishing new projects ) will be required for agi
Text generator is interesting more because it goes too much higher length to understand input content such as crawling the web for any given input links downloading images and analysing the content/ embedded text etc.
Both strategies and more including training new networks especially for harder problems ( like generating 3D content or designing and finishing new projects ) will be required for agi
Checkout https://text-generator.io which is already multi modal (and multi model :), it includes a web crawler to crawl through input links, download and analyse images, do OCR etc to generate better text. As well as Auto picking the right model for the job (multilingual/code/chat/instruct etc)
Normally HN and Reddit are less moderated than other sites that do a lot of shadow banning, condolences for getting stomped on by large co's and also welcome to the internet
Those principles also work well in other AI text generators like https://text-generator.io
https://text-generator.io/blog/embed-images-text-and-code
Your training trick is neat/great innovation but also keep in mind it is likely overfitting meaning when you get a bit of new data you need to index and search that model isn't going to do well at all at embedding it, said differently that training works well if you can cover the types of data you're going to see in production really well at training time. If not there's a big accuracy drop for unseen data due to overfitting
It also analyses linked web pages and image content automatically so helps people build these web integrations or auto img alt tag describers or reiept analysers etc without doing as much of the prompt engineering/crawling themselves.
I've also had customers scared off or investors not want to invest because it is a plain API platform and not B2C...
To answer the question market research is kind of the only way to know, Google adwords keyword planner and Google trends are quite good, as well as researching the companies in the same space well and what gaps there APIs have.
As an example https://text-generator.io distinguishes itself from OpenAI by including automatic analysis for input links, linked images, documents etc and doing speech to text/self hosting/embedding images, text and code in the same space, so isn't just an API clone
You have to be careful with the rate technology progresses and be the one riding the wave of progress and not left behind.
At this point we best believe there will be robots capable of doing near everything we can, but better including sports, trades, war, looking after people, and these are not that far away and will happen in our or kids lifetimes.
Don't listen to the nay sayers, people used to argue hard that computers would never play chess, humans will keep redefining the goalposts including creating new distinctions of what intelligence is like the difference between thinking and writing until machines are just plain better than us at all of it and we will invent a whole bunch of hateful slurs for them
Text-Generator.io is now self hostable, It's priced at $1000 USD per instance per year to self host.
The service runs on a single 24GB VRAM GPU, and runs all services including speech to text, text and code generation for almost all languages and generating embeddings too.
The text generator also downloads and analyses any input with links including documents, images, images with text inside and webpages for better understanding and to generate better text.
It's a great alternative to OpenAI and has a compatible API making switching easy.
Let me know what you think and if there's anything i can do to help!
All the best. Lee Penkman - Founder Text-Generator.io
It also does speech to text over 8x cheaper than Google Cloud, there's a nice free teir, I'm surprised more people aren't taking advantage of it
OpenAI is doing well recently so the next step is to do more things they don't do and let people self Host etc.
It's apparent in the future machines will have a deep level of understanding and it will be easy to extend/build upon and also affordable
try https://text-generator.io It also works with code and in multiple languages and is OpenAI compatible so a one line change. There's a few other things that it does too like linked image understanding.
It supports some things that openAI can't do, it retrieves any linked images of web pages, analyses the images or images with text inside to help the embedding model
I would recommend for vast majority of people to experiment around with these pretrained models first before getting the dedicated GPU/training etc, but cool that your company helps with it
There's companies that offer orders of magnitude of cost savings over OpenAI for text generation too e.g. https://text-generator.io in excited about that going the way of stable diffusion e.g. there's bloom and bloomz for instruct, there should probably be the same excitement when bloom comes out with E.g. ChatBloom, but idk why open source and other co's don't get the press that OpenAi does
Switching is also very easy as text generator is OpenAI compatible/one line change of the base url/API key
Macro examples: cases where there are incentives to use for a period e.g. sustained use or yearly discounts Incentives to use proprietary technology such as S3 or dynamodb being cheap. Situations where migrations are hard (data), expensive (cold storage) or dangerous/slow to recover such as changing the DNS