Sfoglia categorie
Esplora
Fiverr Pro
Italiano
$
USD
I will train or fine tune an open source model for you
Marcell Szeles
Informazioni su questo servizio
I will LOCALLY run your AI model training, fine-tuning, or inference on my personal NVIDIA DGX Spark one of the most powerful desktop AI supercomputers available today.
GDPR compliant service, your data does NOT leave Europe!
Hardware: GB10 Grace Blackwell · 1 PFLOP FP4 · 128 GB unified memory · 4 TB NVMe SSD · 1 Gbit/s symmetric internet
What I can run for you:
- LLM fine-tuning (Llama, Mistral, Phi, Qwen, and more)
- Model training from scratch or from a checkpoint
- Batch inference and evaluation pipelines
- PyTorch, TensorFlow, and Hugging Face Transformers
- Custom scripts just send your code and dataset
How it works:
- You share your model, dataset, and configuration, complete job
- I run the job on the DGX Spark
- You receive output files, model weights, and full run logs as per your provided job
Fast 1 Gbit/s transfers mean large datasets and checkpoints move quickly. Results delivered via a secure file transfer link.
Scopri di più su Marcell Szeles
Marcell Szeles
Provider of Choice
- DaUngheria
- Membro daapr 2026
- Tempo di risposta medio1 ora
Lingue
Inglese, Ungherese
I specialize in AI model customization — fine-tuning, training, and inference — running entirely on local EU hardware, fully GDPR compliant.
Based in Hungary, I work with businesses and developers who need powerful AI capabilities without their data leaving Europe. Every job runs on my personal NVIDIA DGX Spark, delivering enterprise-grade performance with the privacy guarantees EU regulations require.
With 6 years in IT project management, I offer a straightforward process: share your model, dataset, and config — I handle the rest and deliver full results, weights, and run logs securely.
FAQ
What do I need to provide to get started?
Just share your model or script, dataset, and any configuration details (hyperparameters, training flags, etc.). You can upload files via Google Drive, Dropbox, or any file sharing link. I will handle the setup and execution on the DGX Spark.
How do I receive my results?
Once the job completes, I will send you a secure download link (Google Drive or similar) containing all output files, model checkpoints, and a full training/inference log. Delivery is typically within the timeframe stated in your chosen package.
What types of models and frameworks are supported?
The DGX Spark supports PyTorch, TensorFlow, JAX, and Hugging Face Transformers out of the box. I can run fine-tuning, LoRA/QLoRA, full training, and inference for most popular LLMs and vision models. If you have a custom environment, just share your requirements.txt or conda env.
