Hmm? AutoTrain’s operation doesn’t change based on whether you’re a Pro subscriber, and duplicating it for use is the standard procedure.
You are misunderstanding the behavior.
AutoTrain is not “disabled” on your PRO account. What you see is the normal, documented flow for AutoTrain Advanced today.
1. There is no separate “AutoTrain activation” for PRO
Hugging Face’s own docs say that no-code users use AutoTrain by creating a new Space with the AutoTrain Docker image. The recommended entry is:
Visit the AutoTrain page → click “Create new project” → you get a setup screen for a new AutoTrain Space. (Hugging Face)
Your PRO subscription changes:
- storage and credits
- ZeroGPU and Spaces privileges
It does not flip an “AutoTrain on/off” flag. The PRO feature list is about storage, inference credits, ZeroGPU quota, Spaces hosting, etc., not about enabling AutoTrain. (Hugging Face)
So there is nothing for anyone to “enable” on your username DRISSI666. The mechanism is self-serve.
2. Why you see “Duplicate this Space” after “Create new project”
AutoTrain Advanced is implemented as a template Space:
- Template:
autotrain-projects/autotrain-advanced
- That template is not meant to be used directly. Its own page explicitly states that it will show an error if the user has not duplicated the Space and that users must duplicate it to use it. (Hugging Face)
The official quickstart spells this out:
- Go to the AutoTrain homepage.
- Click “Create new project”.
- You see a screen like “autotrain-duplicate-space”, where you configure and create a new AutoTrain Space. (Hugging Face)
That “Duplicate this Space” page is just the standard Space duplication UI. The general Spaces docs explain that duplicating a Space is how you take a template and turn it into your own copy, with your hardware, visibility, and settings. (Hugging Face)
So:
- Seeing “Duplicate this Space” after clicking “Create new project”
- Means “you are at the step where you create your own AutoTrain instance”
- Not “AutoTrain is missing from your account”
3. Correct procedure step-by-step (what you should actually do)
Here is the concrete procedure you should follow with your DRISSI666 PRO account.
Step 1 – Go to AutoTrain and start a new project
- Log in on Hugging Face as
DRISSI666.
- Open: https://huggingface.co/autotrain.
- Click “Create new project”.
If you are not logged in, the site will ask you to log in, then forward you to the duplication screen. (Hugging Face)
Step 2 – Use the “Duplicate this Space” screen correctly
You are now on a page that is effectively:
“Create your own Space from template autotrain-projects/autotrain-advanced.”
On this screen:
-
Owner
- Choose
DRISSI666 (your personal account) or an organization, if you belong to one.
-
Space name
- For example:
autotrain-advanced or autotrain-llm.
-
Visibility
- Choose Private if your dataset is not public. The docs explicitly recommend keeping AutoTrain Spaces private for most use cases. (Hugging Face)
-
Hardware
- Pick a CPU only machine for tiny tests.
- Pick a GPU (e.g. T4, A10G or similar) for any real LLM, vision, or serious training. AutoTrain Spaces are billed based on the hardware you attach. (Hugging Face)
Then click “Duplicate Space”.
This creates a new Space under your account (for example https://huggingface.co/spaces/DRISSI666/autotrain-advanced).
Step 3 – Wait for the Space to build and start
After duplication:
- The Space enters a “Building” state while the Docker image is prepared.
- When done, it changes to “Running” and opens the AutoTrain UI inside that Space.
If a Space stays stuck in “Building” or similar, that is a separate issue (there are GitHub issues showing such platform problems), but it is not about AutoTrain being disabled for your account. (GitHub)
Step 4 – Open the Space and sign in inside the UI
When the Space is running:
- Open your new Space URL.
- Inside the embedded AutoTrain page, you will typically see a Sign in with Hugging Face button if the app needs authorization.
- Authorize it so AutoTrain can read your datasets and push models back to your account. (Hugging Face)
Only after this step do you see the “real” AutoTrain UI: project list, “New project” buttons, configuration screens, etc. That UI lives inside your own Space, not on a global central page.
4. What changes with PRO vs free, in this context
Your PRO account matters in these ways:
- More storage and inference credits
- Larger or more frequent use of Spaces and ZeroGPU
- Higher queue priority for some jobs (Hugging Face)
It does not:
- Unlock a secret AutoTrain UI URL
- Gate the “Create new project” button
- Require staff to “turn on” AutoTrain for you
AutoTrain Advanced is designed to run for any user who can create Spaces and attach hardware. The general docs describe it as a no-code system that you use by creating or duplicating a Space with the AutoTrain image. (PyPI)
5. When should you actually contact Hugging Face support?
You only need staff intervention if something is broken, for example:
- Duplicating the Space fails, or the page errors out.
- Your new AutoTrain Space never leaves “Building” or “Preparing Space”, despite multiple attempts. (GitHub)
- You are being billed for a Space you cannot see or control, or you see “Please DUPLICATE to use this Space” on something that is clearly already yours, like in GitHub issue #889. (GitHub)
In such cases, use:
- The AutoTrain FAQ and support section to find the correct channel. (Hugging Face)
- The
huggingface/autotrain-advanced GitHub issues if it looks like a clear bug in the app itself. (GitHub)
For your current situation, you do not need anyone to “enable” AutoTrain; you only need to finish the Space duplication flow.
Short recap
- Your PRO account is not missing any “AutoTrain activation”.
- The “Duplicate this Space” page is the expected next step after clicking “Create new project” on https://huggingface.co/autotrain. (Hugging Face)
- Correct procedure:
AutoTrain page → Create new project → configure and Duplicate Space → wait for it to build → open your new Space → sign in inside the AutoTrain UI → create projects.
- Only contact support if duplication or running the Space fails; not just because you see “Duplicate this Space”.