Hey, is there a model for text to video generation? A really good one? And if yes, how would I inter

Hey, is there a model for text to video generation? A really good one? And if yes, how would I intergrate it in my java application

Kkaktus___Hey, is there a model for text to video generation? A really good one? And if ye...

Isaac McFadyen•6/24/24, 2:16 AM

There is not one on Workers AI and to be honest you're going to find it challenging to find a good one outside of very experimental models that you'd need to run yourself

Isaac McFadyen•6/24/24, 2:17 AM

The technology for it is just getting to a point where it's usable, and most models are gated behind many security checks because of the potential for abuse; for example, you'd need to reach out to OpenAI (likely as a big enterprise customer) for Sora access and probably go through many different security checks/validations, assuming they'd even want to allow you access yet.

IIsaac McFadyen There is not one on Workers AI and to be honest you're going to find it challeng...

kaktus___OP•6/24/24, 1:33 PM

There is like something called Open-Sora, but it's not possible to get it ran with cloudflare?

kaktus___OP•6/24/24, 1:33 PM

It is the most stable atm for public

Kkaktus___There is like something called Open-Sora, but it's not possible to get it ran wi...

Isaac McFadyen•6/24/24, 1:37 PM

Open Sora is extremely heavy and not currently available to run on Workers AI.

Isaac McFadyen•6/24/24, 1:37 PM

By extremely heavy I mean upwards of 24GB of VRAM.

kaktus___OP•6/24/24, 1:39 PM

Damn. That's a lot. Alr really thank you! :)

LordQ•6/25/24, 2:08 PM

Is it possible to deploy AI models out of the cloudflare model list on workers AI?

LordQ•6/25/24, 2:10 PM

Like Stable Diffusion 3 Medium, I know cf has SD 1.5 to use, but is there any chance to bring new models by ourselves?

LLordQ Like Stable Diffusion 3 Medium, I know cf has SD 1.5 to use, but is there any c...

Isaac McFadyen•6/25/24, 2:59 PM

?workers-ai-models - See below.

Flare•6/25/24, 2:59 PM

Workers AI currently only supports popular open-source models provided by the Cloudflare team, as well as your own LoRAs that can be applied on top of the Cloudflare-provided models. You cannot currently upload your own models or use a model from HuggingFace. See the documentation for the list of Cloudflare-provided models: https://developers.cloudflare.com/workers-ai/models/

Setom29•7/4/24, 11:09 AM

Hi there,

I was wondering, how can I find out when the models are out of beta? Does Cloudflare announce it somewhere? Is there a mailing list or a notification system I can subscribe to? Any info would be great!

�

😈 Donkey 💫•7/4/24, 3:21 PM

translation is GA, but 720req/min is way too low

ulysses•7/5/24, 8:19 PM

When will cloudflare webgpu be available?

Uulysses When will cloudflare webgpu be available?

Victor•7/5/24, 10:23 PM

https://developers.cloudflare.com/durable-objects/api/webgpu/

Cloudflare Docs

WebGPU · Cloudflare Durable Objects docs

The WebGPU API , allows you to use the GPU directly from JavaScript.

VVictor https://developers.cloudflare.com/durable-objects/api/webgpu/

ulysses•7/6/24, 12:38 AM

right, but this is only in local development–when will it be enabled in prod?

Peps•7/6/24, 3:14 AM

I do see the blog post about it back in 2023 said

Over the upcoming weeks, we will merge the workerd WebGPU code in the Cloudflare Workers production environment and make it available globally, on top of our growing GPU nodes fleet

Not sure what halted it's global launch

Peps•7/6/24, 3:16 AM

from what I can see it's just soon

for now. No public timeline right now

Avocadio•7/6/24, 8:07 AM

hey

Avocadio•7/6/24, 8:08 AM

i want to interact with website urls

Avocadio•7/6/24, 8:08 AM

which ai model is able to do it in workers ai do you guys know

Nicklas•7/6/24, 9:21 AM

Hello everyone, unfortunately I am not able to create the LoRa adapter. I am using AutoTrain Advanced in a Huggingface Space Notebook. If I set quantization to none and lora-r to 8 the creation fails (also the Google Colab does not work). Does anyone have a working Google Colab notebook, even the one linked in the CloudFlare tutorial does not work (even if I install torch and torchvision).

Log: https://just-paste.it/Yfd536Pfco

AAvocadio i want to interact with website urls

rob•7/6/24, 11:34 PM

How so

Avocadio•7/8/24, 10:01 AM

for example i want ai to summarize new by giving url of the new in prompt

AAvocadio for example i want ai to summarize new by giving url of the new in prompt

Ryan•7/8/24, 4:16 PM

do you mean news?

RRyan do you mean news?

Avocadio•7/8/24, 4:23 PM

yup. or an article

Avocadio•7/8/24, 4:24 PM

I used to use mistral and it was able to do web scraping

Ryan•7/8/24, 4:57 PM

generally it would be sending a request to the url provided by the user, extracting the text of the article, feeding it into a llm, then sending it back in the worker's response

Avocadio•7/8/24, 5:19 PM

Yes i am wondering is there any ai model that is able to do that

Avocadio•7/8/24, 5:20 PM

That cloudflare supports

Avocadio•7/8/24, 5:21 PM

If not i am gonna scrape content in the url and feed the model but if there is a model that is able to do it then its better choice for me

AAvocadio If not i am gonna scrape content in the url and feed the model but if there is a...

Isaac McFadyen•7/8/24, 7:44 PM

No, there are no models that auto-scrape URLs - it's outside of the scope of an LLM.

Avocadio•7/8/24, 8:54 PM

i thought there were llms that able to web scrape

rob•7/8/24, 9:34 PM

https://github.com/jina-ai/reader

GitHub

GitHub - jina-ai/reader: Convert any URL to an LLM-friendly input w...

Convert any URL to an LLM-friendly input with a simple prefix https://r.jina.ai/

jina-ai/reader

rob•7/8/24, 9:34 PM

Check this

DDhravya Shah if you wanna do it on cloudflare, https://github.com/dhravya/markdowner

Avocadio•7/9/24, 9:07 PM

Yep i was about to build the same thing

Avocadio•7/9/24, 9:08 PM

Thx a lot

user•7/10/24, 9:14 AM

Does cloudflare AI's qwen model support tools?

user•7/10/24, 2:58 PM

according to qwen's docs, it supports function calling tools & uses openAI syntax

user•7/10/24, 2:58 PM

https://qwen.readthedocs.io/en/stable/framework/function_call.html

Fokke•7/11/24, 5:55 PM

We are building a state-of-the-art retrieval augmented generation chatbot completely on Cloudflare. We are eventually going to publish the whole source to the community, but for starters, we have two blogs that you all might like if you are thinking about building something like this yourself.

Intro - https://liquidmetal.ai/blog/sota-rag-intro/
Data Stores - https://liquidmetal.ai/blog/sota-rag-datastores/
Indexer (coming soon)

LiquidMetal AI - We build AI on Cloudflare

kufeng.•7/12/24, 6:18 AM

Uuser Does cloudflare AI's qwen model support tools?

user•7/12/24, 11:41 AM

any plans on this?

akazwz•7/14/24, 4:07 PM

Hello, I build a LLMs/AI chat platform build and design for Cloudflare worker AI. with many features like function calling, image to text, text to image, text to speech. all build with Cloudflare worker. https://github.com/akazwz/antonai

GitHub

GitHub - akazwz/antonai: a LLMs/AI chat platform build and design f...

a LLMs/AI chat platform build and design for Cloudflare worker AI. with many features like function calling, image to text, text to image, text to speech - akazwz/antonai

Thiago•7/14/24, 7:10 PM

What happened to the quality of the dreamshaper-8-lcmdreamshaper-8-lcm model? Its outputs have been of extremely low quality for the past ~3 days, to the point that they are completely unusable. Below are two examples, 1 month apart. Same prompt.

TThiago What happened to the quality of the `dreamshaper-8-lcm` model? Its outputs have ...

Raylight•7/14/24, 8:56 PM

Looks like the "guidance" parameter has been exposed/activated and it's possible that it doesn't behave as intended. The default of 7.5 will give results like the image on the right. If you add "guidance: 0.0" (or maybe 1.0) you'll get the same or similar results as you got before. You'll get better prompt following with higher numbers, but as of now there's a narrow range before images become degraded to the point of being unusable.

Hey, is there a model for text to video generation? A really good one? And if yes, how would I inter

Similar Threads