Model Downloads are painfully SLOW

After contact support about how slow it is to download models, all I get back is that it’s my system. And to use a direct Ethernet connection. OMG, is support really that bad? I am using a direct Ethernet connection and sent them a speedtest of over 900Mbps from my Gigspeed internet.

Wow, I really expect more from Topaz Labs, maybe something like we know or we will investigate, etc. I have read other post from people saying the same thing. The main programs download with over 200mbps, but all of the models are under 6mbps.

I did go back and download missing modules and just shrank the window until done, But it doesn’t make sense why modules are so slow to download?


Agreed. So many updates, such big files and so slow. It’s not my connection either - it’s the source server can’t serve data fast enough. It’s a massive bottleneck to my workflow. Do better Topaz.

1 Like

Has there really been no progress on this? I can confirm the above poster’s comment, as I’m on a fiber optic connection, and my model download from Topaz is always only 10Mbps, no matter what I do. Every minor tweak in video AI requires a new model, which takes forever to download.
If the speed can’t improve, at least let us download them all overnight, like it used to do!

Did you find any solution for that problem if yes then please help me

Still facing this issue. If anyone has managed to solve it please guide

It’s even worse for me, the download never completes, it gets stuck after a ridiculous amount of MBs!
Is ther someone at Topaz support who could answer with a REAL fix.
Obviously it would be way simpler to provide an IP@ to download the latest models!!!

I didn’t notice the slow model download until recently, around Sep-Oct 2023. It used to be fine.

1 Like

This is just so unacceptable. There are many products that perform the same functions at least partly as good. It’s time we all chose to spend our money elsewhere.

I’ve had similar problems now, could be because the upgrade license has expired or because I switched to wifi instead of cable from the router. It starts a download message even though you have already downloaded the model 1min(1hour) ago.

Yes, PhotoAI models are like molasses, definitely their servers. I was curious if the recent slash in product pricing caused a bottleneck, but I see that this has been going on for awhile now :frowning:

1 Like

Indeed, same here. Downloads are just painfully slow and it seems like they don’t even try. :frowning:

In March 2024, I understood at a first glance, that it’s not about my router’s top notch speed, then immediately I plugged in 4G wired connection. But even then whole download process for the segments (bits & pieces) took hours to install. Meaning thereby, it’s not about our connection speed, it’s Topaz’s server connection or Load, which hinders or delivers the data with painfully slow rate. Maybe with time Topaz Inc. will handle this grave issue with their forelock to eradicate the probable hiccups on the way. Because not every single person can have the high speed wired connection in the first place.

It’s MAY 2024 and the downloads are still painfully slow. Gah.

June 11/2024 … CanuckBC: I purchased Topaz Photo AI (Windows 11) June 8th … I am still downloading the models! I have been downloading during the day and overnight for the past two days and so far today (the 11th). I still have 21 models to download. By my calculations , the current model download (1 of 21) of 1721mb should take 78hrs! My internet speed is not the greatest but it is definltely not this bad. is there a fix on the Topaz end? At this rate I cannot even use the product. 2024-06-11 11:33:00 Vancouver

Currently staring at 80-350 Kbps…Glad its not just me. And I just spent 4 hours thinking maybe it was DNS or something on my end. Nope. Just horrific transfer rates it looks like. I can be patient, but the problem like the commenter above I cannot use the product… Is there at least a zip file or something I can download and apply later so I can use the software while I wait for this??