# Can I do finetuning on a HF space? **URL:** \ **Category:** Beginners\ **Created:** [April 7, 2025, 6:59am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213 "2025-04-07T06:59:53Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![n094t23g](https://avatars.discourse-cdn.com/v4/letter/n/8dc957/32.png) [@n094t23g](https://discuss.huggingface.co/u/n094t23g)\ **Post date:** [April 7, 2025, 6:59am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213/1 "2025-04-07T06:59:53Z")
I tried doing that using [HF NLP course unsloth tutorial](https://huggingface.co/learn/llm-course/en/chapter12/6?fw=pt) but it can’t find the gpu.. I tried using `spaces.GPU`. I didn’t want to use google collab. I can’t find anything on search or using a llm; unless im doing it wrongly ---
**Author:** ![LPX55](https://avatars.discourse-cdn.com/v4/letter/l/a8b319/32.png) [@LPX55](https://discuss.huggingface.co/u/LPX55)\ **Post date:** [April 7, 2025, 7:19am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213/2 "2025-04-07T07:19:08Z")
Plenty of alternatives around the block. Some official HF-recognized ones off the top of my head: - [Studio Lab](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/transformers_doc/en/training.ipynb) since Google is not your preference - HF Autotrain: [AutoTrain – Hugging Face](https://huggingface.co/autotrain) ---
**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](https://discuss.huggingface.co/u/John6666)\ **Post date:** [April 7, 2025, 7:22am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213/3 "2025-04-07T07:22:33Z")
The free space on Hugging Face does not have a GPU. I think that if the script can run on the CPU, it will run… In fact, if the model is very small, it can be trained in CPU space. If you don’t do **.to(“cuda”)** or so, it will work more or less. It’s slow, though… I think that the Colab Free environment is often used as an example because it can use a GPU, although there is a time limit. ---
**Author:** ![LPX55](https://avatars.discourse-cdn.com/v4/letter/l/a8b319/32.png) [@LPX55](https://discuss.huggingface.co/u/LPX55)\ **Post date:** [April 7, 2025, 7:28am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213/4 "2025-04-07T07:28:40Z")
[Modal.com](http://Modal.com) is also worth checking out if you’re in for a bit of learning journey (the best kind!) (Unaffiliated disclaimer, just had a playdate with their infra yesterday and had a blast) ---
**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](https://discuss.huggingface.co/u/John6666)\ **Post date:** [April 7, 2025, 7:32am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213/5 "2025-04-07T07:32:32Z")
There are other free environments where you can use GPUs other than Colab, but it seems that some applications are accepted and others are not. I was accepted, so I don’t think it’s very strict… > **[Lightning AI | Turn ideas into AI, Lightning fast](https://lightning.ai)** > > The all-in-one platform for AI development. Code together. Prototype. Train. Scale. Serve. From your browser - with zero setup. From the creators of PyTorch Lightning. ---
**Author:** ![LPX55](https://avatars.discourse-cdn.com/v4/letter/l/a8b319/32.png) [@LPX55](https://discuss.huggingface.co/u/LPX55)\ **Post date:** [April 7, 2025, 7:46am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213/6 "2025-04-07T07:46:13Z")
Lightning AI seemed pretty cool, but rejected me for whatever reason earlier last year 🤷‍♂️ Perhaps I shall give it another try… ---
**Author:** ![n094t23g](https://avatars.discourse-cdn.com/v4/letter/n/8dc957/32.png) [@n094t23g](https://discuss.huggingface.co/u/n094t23g)\ **Post date:** [April 7, 2025, 10:28am UTC](https://discuss.huggingface.co/t/can-i-do-finetuning-on-a-hf-space/149213/7 "2025-04-07T10:28:40Z")
welp i guess I may need to spend more mulah