I will train custom flux and sdxl lora models on a100 gpu

Einige Informationen werden in englischer Sprache angezeigt.

Vietnam

Ich spreche Vietnamesisch, Englisch

Generative AI and ComfyUI Engineer

Generative AI & ML Engineer (M.Sc. Control & Automation, HUST) building production ComfyUI, Python & NVIDIA A100 80GB GPU pipelines. I deliver 4K/6K commercial product visuals, virtual try-on & virtua...
Über diesen Service

Stop struggling with random AI faces and inconsistent brand visuals. I train production-grade custom LoRA models on dedicated NVIDIA A100 80GB GPUs for lock-solid character, virtual KOL, product, and art-style consistency.


Pricing reflects dedicated A100 80GB compute, manual dataset curation, and rigorous multi-epoch checkpoint evaluation.


What You Get:

- Custom Flux, SDXL, or video model (Wan / HunyuanVideo) LoRA weights (.safetensors)

- Dataset cleaning, deduplication, resolution bucketing, and precision captioning

- Hyperparameter sweep & multi-epoch comparison grid (zero overfitting)

- Consistent Virtual Brand KOL / AI Influencer & virtual try-on compatibility

- Plug-and-play ComfyUI JSON workflow + optimal trigger words, CFG, and sampler guide


Send 15-50 reference images and your target use case to plan the training run.

Programmiersprache:

Python

KI-Engine:

Stable Diffusion

•

PyTorch

Mein Portfolio

Meine weiteren Dienstleistungen im Bereich Datenwissenschaft & ML