HN user
building TransformerLab.ai.
founder of Tulip.io. founder of Well.ca.
Getting ROCm working was... an adventure. We documented the entire (painful) journey in a detailed blog post because honestly, nothing went according to plan. If you've ever wrestled with ROCm setup for ML, you'll probably relate to our struggles.
The good news? Everything works smoothly now! We'd love for you to try it out and see what you think.
Transformer Lab maintainer here. We'd love to build this! Please contact us on our our Discord https://discord.gg/transformerlab or Twitter https://x.com/transformerlab and we'd love to learn more and collaborate.
Thanks for finding this!
Will get this merged as soon as we can test across platforms.
One of the maintainers here. Yes this is a big part of our plans. In addition to our plugin system which allows arbitrary Python scripts, we will soon publish how to add decorators to any existing script which can be run externally but logged into Transformer Lab. So you could do training anywhere but trigger evals in the app, for example.
Yes it all works as webapp, in addition to working as an electron app. we haven’t gotten to publishing how in the docs as this feature is brand new.
Short story, run: “npm run start:cloud” in the git repo.
Will make this more known soon.
Apple MLX is a game changer for what is possible in Local LLM development for everyone. Getting this to work as a single application that "just works" across platforms has been one of the hardest engineering problems we've ever worked on, but we're determined to get it right.
We want this tool to be for everyone that wants to train and tune their own LLMs.
Our vision is that the sphere of people who do these types of tasks will grow as LLMs become more accessible and the tooling becomes more reliable.
Let us know what other types of functionality you'd like to see. I will record Assistant-type functionality for our team to look into!
If you know any we should incorporate, let us know! We really want to bring the best of open source LLM tools together in one UI.
The functionality in Transformer Lab comes from plugins. Plugins are just Python scripts behind the scenes. So anything that can be done in Python can be done as a plugin.
Right now we have export plugins for going to GGUF, MLX and LlamaFile but if you know a good library for exporting to TensorRT, let's make a plugin for this! (Feel free to join our Discord if you want help)
From my digging...
Apple seems to get most of their recycled titanium from IperionX:
https://www.linkedin.com/posts/iperionx_over-20-years-ago-ap...
IperionX announced they get their recycled source Ti from ELG Utica Alloys:
https://www.aero-mag.com/new-partnership-creates-100-recycle...
From ELG we can read that a source of scrap Ti comes from companies like GE, Boeing, etc:
https://cdn.ymaws.com/titanium.org/resource/resmgr/2010_2014...
But this article from 1990 (may be outdated) states that a misunderstanding of the public is thinking that scrap mostly comes from old devices or planes. But these types of things last long and don't get recycled enough to provide enough volume. This seems to say that the bulk of scrap comes as a byproduct of the smelting process:
https://link.springer.com/article/10.1007/s11837-018-3278-1
The following article from 2004 however says that recycled Ti comes from 10% old scrap (old recycled parts we normally think of as recycled metal) and 90% "new scrap" which is partly wasted Ti generated when making parts. For example in 2004, the aerospace industry used 12000t of Ti but wasted 10000t of that.
I’m curious: what was the podcast?
Some of the tools offer a path to doing tool use (fetching URLs and doing things with them) or RAG (searching your documents). I think Oobabooga https://github.com/oobabooga/text-generation-webui offers the latter through plugins.
Our tool, https://github.com/transformerlab/transformerlab-app also supports the latter (document search) using local llms.
You can chat with the models directly in the same way you can chat with GPT 3.5.
Many of the opensource tools that run these models let you also edit the system prompt, which lets you tweak their personality.
The more advanced tools let you train them, but most of the time, people are downloading pre-existing models and using them directly.
If you are training models, it depends what you are doing. Finetuning an existing pre-trained model requires lots of examples but you can often do a lot with, say, 1000 examples in a dataset.
If you are training a large model completely from scratch, then, yes, you need tons of data and very few people are doing that on their local machines.
If you're able to purchase a separate GPU, the most popular option is to get an NVIDIA RTX3090 or RTX4090.
Apple Mac M2 or M3's are becoming a viable option because of MLX https://github.com/ml-explore/mlx . If you are getting an M series Mac for LLMs, I'd recommend getting something with 24GB or more of RAM.
If you are looking for a downloadable app as an alternative to Oobabooga / Sillytavern, we built https://github.com/transformerlab/transformerlab-app as a way to easily download, interact, and even train local llms.
There is some good information on how PCI-Express works here: https://blog.ovhcloud.com/how-pci-express-works-and-why-you-...
For those interested in following the same regimen, the test group took
* 2000 IU of D3 in the form of a softgel and [1] * 1 gram of Omega 3 [1]
But it's important to note that the the Omega 3 was in the form of prescription Omacor [1] which has been renamed to Lovaza [2]
The D3 supplement was donated by Pharmavite [1] which probably means that the brand name was Nature Made https://www.pharmavite.com/what-we-make/nature-made/
1 gram of Lovaza contains 465 mg EPA and 375 mg DHA [4] and note that "Prescription medications may contain up to 90% omega-3 fatty acids, whereas over-the-counter fish oil supplements contain roughly between 30% and 50% omega-3 fatty acids" [3] so you should look at the label on your Omega-3 supplement to see how much EPA and DHA it contains. And finally, note that the ratio of EPA to DHA might matter too [5].
[1] https://clinicaltrials.gov/ct2/show/NCT01169259 [2] http://www.ncbop.org/pdf/omacorbecomeslovazajuly2007.pdf [3] https://www.verywellhealth.com/prescription-and-supplement-o... [4] https://lovaza.com/ [5] https://pubmed.ncbi.nlm.nih.gov/33418065/
There is also TiDB
Favourite part: "Palantir is in no way affiliated with, or endorsed or sponsored by, The Saul Zaentz Company d.b.a. Tolkien Enterprises or the Estate of J.R.R. Tolkien."
Was Zuckerberg thinking about the right time for this in 2009 when he asked this question on Quora:
http://www.quora.com/When-did-Steve-Ballmer-join-Microsofts-...
Once while waiting in the lobby of a Google office I could hear the receptionist answering a frustrating call from someone asking for a search result removal request. She had to read from a script because people call so often -- they call reception at any Google-related business asking for help.
Take a look at startupnorth.ca, Toronto's DemoCamp events, StartupWeekend events, and Extreme VP's Extreme University. A lot of the community connects through these groups.
We created a clone in PHP for anyone to use: https://github.com/well/well-smiley
weekends!
yeah i noticed people doing that too but my guess is that it's mainly on the demo site -- in a real team hopefully your peers won't be jerks.
thanks!
As an alternative, I wrote a port of Vanity for PHP that we use for A/B testing. It uses redis to store data
I could use this copy over static media to my CDN. Cron + rsync is annoying.
You should get the one most of your friends have.