I’m highly curious of hugging faces role in the new world of huuuge models and few shot learning. Seems to me to be quite an existential thread to them.
All (?) the "open" (to various degrees of openness) LLMs are on HuggingFace and have code to support their unique options. For example here is Llama[1] and here is Opt[2].
All the code I've seen for training them uses Hugging Face. When EleutherAI releases their open foundation LLM then everyone will want to fine tune it and do RLHF on it and they'll all use HuggingFace code to do it.
I think they are in the best position of any company including OpenAI.
They might succeed if they can convince people to subscribe for model updates like we do for security patches and game updates. For example release any new model for paying members in the first 30 days, and for everyone else later. Or some kind of paid API for LLMs, like "Open"AI, because it's too much of a hassle to run your own LLM.
How so? They host a lot of the FOSS and are a huge infrastructure provider in the space. No matter if Stable Diffusion, GPT or some other thing is the hot new thing in AI/ML space, you can be sure that huggingface is hosting some part of it either today already, or tomorrow.
Absolutely, you can say the same for a bunch of different companies. Some did really well (GitHub) while others did less well (Docker). I don't think looking at Docker is representative of the entire "FOSS Hosting" angle, and won't predict the future of huggingface.
I have used them extensively and am a big fan, but none of their models are even close to the type of thing ChatGPT can do (for now). They are built for a world where you take a pre-trained model (say Bert) and fine tune it to your specific task. I feel like ML is moving towards: We have these fantastic model with 50 page context, just give it a few examples and it can do whatever you want out of the box.
If this is the future, then most of HuggingFaces code and infra is obsolete.
But that's models from people uploading models to huggingface, huggingface themselves don't write/maintain those models, it's a community of AI/ML practitioners.
And LLMs are just a small part of the ecosystem, huggingface is hosting much more than just those models. It hosts models for text-to-image, classification, text-to-speech, and everything in-between those. And beyond that, they also host datasets that are being used for training a bunch of models.
As long as there is a FOSS AI/ML ecosystem, huggingface will remain being relevant (granted no other similar platform appear and takes over). For example, if you want to do anything Stable Diffusion today, it's more likely than not that something is being pulled down from huggingface, one way or another.
Even in your own example, where we have one-shot models that don't need fine-tuning, those models still need to be hosted somewhere. Today, that hosting happens most commonly on huggingface.
Pretty interesting discussion - I particularly liked the part about how, in order to better align LLMs, we need better transparency so everyone can help study and align them.
I've seen it used ironically when the guest is not a figurehead trying to score human points, but instead a highly technical person talking technical details.
The words "fireside chat" convey to me a long-ish, casual, relaxed conversation that isn't focused necessarily on a given topic. A "conversation" is broader than that. A fireside chat is a conversation, but so is "we're laying off for poor performance"
I worry that Hugginface is on path to becoming another Docker
Huge contributions to the industry, everyone uses it (or something that came after it) but nobody pays or wants to pay for it.
The first place I look for a new model is Hugginface.
However using them for inference (seemingly their revenue play) basically always ends up in a very unfriendly build vs buy as they don’t eliminate the need for one or more ML engineers.
Paying x% over infra costs for managed models on my choice of cloud only works if I don’t also have to spend to hire ML engineering at the same rate regardless.
Yeah, it's very much a in-and-out experience, like Docker Hub.
My typical workflow involving huggingface:
- Try to find appropriate model, play around with some of them until I find one that is appropriate, pull it down locally and close the huggingface tab.
Yeah, even the name sounds suspicious... sure, it's a nice and friendly emoji, but remember the facehuggers from Alien? Once they got you, you were screwed - which could also be a metaphor for relying too much on any platform that may be here today and gone (or not affordable anymore) tomorrow...
industry insider friends have indicated to me that this isnt true, they are doing well in enterprise. not that i would know but guessing mid 7 digit arr by now.
I really love using HuggingFace, but from what I’ve heard all their revenue comes from consulting on libraries. It’d be really sad if this is only things they put in market, doesn't seem scalable to me. Would be really sad if they will go bust, not sure what happens with all the models and products millions of people are using right now.
> but from what I’ve heard all their revenue comes from consulting on libraries. It’d be really sad if this is only things they put in market
Its not; the inference API and AutoTrain are paid services, and even if lower revenue, scale better than consulting. AutoTrain seems to be what they promote as a commercial offering the most.
28 comments
[ 1.1 ms ] story [ 74.1 ms ] threadI ctrl+f'd for "large" of large language model
All the code I've seen for training them uses Hugging Face. When EleutherAI releases their open foundation LLM then everyone will want to fine tune it and do RLHF on it and they'll all use HuggingFace code to do it.
I think they are in the best position of any company including OpenAI.
[1] https://huggingface.co/docs/transformers/main/model_doc/llam...
[2] https://huggingface.co/docs/transformers/main/model_doc/opt
Everyone uses it but nobody pays for it.
You could say the same about Docker, I think? It seems to have been a mixed blessing for them.
And LLMs are just a small part of the ecosystem, huggingface is hosting much more than just those models. It hosts models for text-to-image, classification, text-to-speech, and everything in-between those. And beyond that, they also host datasets that are being used for training a bunch of models.
As long as there is a FOSS AI/ML ecosystem, huggingface will remain being relevant (granted no other similar platform appear and takes over). For example, if you want to do anything Stable Diffusion today, it's more likely than not that something is being pulled down from huggingface, one way or another.
Even in your own example, where we have one-shot models that don't need fine-tuning, those models still need to be hosted somewhere. Today, that hosting happens most commonly on huggingface.
https://youtube.com/@eladgil3583
Huge contributions to the industry, everyone uses it (or something that came after it) but nobody pays or wants to pay for it.
The first place I look for a new model is Hugginface.
However using them for inference (seemingly their revenue play) basically always ends up in a very unfriendly build vs buy as they don’t eliminate the need for one or more ML engineers.
Paying x% over infra costs for managed models on my choice of cloud only works if I don’t also have to spend to hire ML engineering at the same rate regardless.
I want them to succeed.
My typical workflow involving huggingface:
- Try to find appropriate model, play around with some of them until I find one that is appropriate, pull it down locally and close the huggingface tab.
they are hub first. not a Docker situation.
Its not; the inference API and AutoTrain are paid services, and even if lower revenue, scale better than consulting. AutoTrain seems to be what they promote as a commercial offering the most.