I will fix python bugs in pytorch and hugging face transformers

L
lifatsas
L
lifatsas
Junling G
Sommige informatie wordt in het Engels weergegeven.

Over deze dienst

Stuck on a Python traceback, a PyTorch tensor error, or a Hugging Face model that will not load? I will reproduce the agreed issue, fix the code or configuration, and explain how to verify the result.


I am an AI engineering student and open-source contributor. My contribution to Hugging Face Transformers, PR #45402, fixed registered-buffer loading under DeepSpeed ZeRO-3 and added a regression test. It was merged after maintainer review.


I can help with:

- Python exceptions and dependency conflicts

- PyTorch shape, dtype and device mismatches

- Transformers model, tokenizer and checkpoint loading

- Reproducible inference or training-script errors


You receive a scoped fix, corrected files or patch, and verification notes. Premium adds a regression test and walkthrough.


Please message me before ordering with the full traceback, package versions, hardware, and a minimal code sample. I will confirm feasibility and the right package.


One order covers one agreed issue. New features, full applications, large training runs and hardware upgrades are excluded. Cloud/GPU fees are not included. Distributed or hardware-specific issues require a custom scope.


Maak kennis met Junling G

Junling G

AI Developer and Python Specialist

  • Afkomstig uitChina
  • Lid sindsapr 2025
  • Gem. reactietijd1 uur
  • Talen

    Chinees, Engels
Hi, I’m Junling, an AI engineering student specializing in local LLM deployment, fine-tuning, RAG, and AI agents. I work with Qwen, Llama, Mistral, Ollama, Hugging Face, and PyTorch. I can help you: * Install and optimize local AI models * Create Python automation tools I use a local RTX 40-series workstation, so your data can remain private and you can avoid unnecessary cloud costs. Send me your requirements, hardware specifications, and expected outcome, and I’ll recommend the most practical solution.

Mijn portfolio