Blog
6 blog available in the DeepSeek directory
Articles, tutorials, and deep dives on this AI platform — from getting-started guides to advanced techniques and real-world use cases.
machinelearningPorting Gemma-4 (2B / 4B / 12B) to AWS Inferentia2
A field report on running Google's Gemma-4 on AWS Inferentia2: mixed attention heads, the vLLM / optimum-neuron / NxD dead-ends, and the neuronx-cc compiler limits.
X
xbill
machinelearningHow Model Distillation Actually Works (and What the 'China Distilled Our Model' Headlines Really Mean)
A practical, no-hype explainer of knowledge distillation in LLMs — the actual mechanics, why distilling from a closed API is different, and what the OpenAI/Anthropic vs DeepSeek allegations are really about.
S
Sergei Parfenov
machinelearningI Rebuilt Karpathy's NanoChat in JAX. Here's What XLA Gets Right and What It Gets Dead Wrong.
AI GDE TPU Sprint 2026 · Google TPU Research Cloud Quick summary: We ported Andrej Karpathy's...
O
Omotayo Aina
machinelearningFine-Tune Any HuggingFace Model like Gemma on TPUs with TorchAX
Learn how to fine-tune PyTorch HuggingFace models on Google TPUs using torchax and LoRA — no JAX rewrite needed. Includes evaluation, save/reload, and a Colab notebook.
A
Ahmed Elnaggar
machinelearningRun Any HuggingFace Model on TPUs: A Beginner's Guide to TorchAX
Learn how to run PyTorch HuggingFace models on Google TPUs using torchax — no JAX rewrite needed. Includes benchmarks, text classification, text generation, and a Colab notebook.
A
Ahmed Elnaggar
machinelearningImage Augmentation in Practice — Lessons from 10 Years of Training CV Models and Building Albumentations
TL;DR Image augmentation is usually explained as “flip, rotate, color jitter”. In practice it...
V
Vladimir Iglovikov