SohoBook selection of external ML, AI and system design labs: model visualizations, algorithms, LLM practice, deep learning lectures and tool catalogs.
Shows Hugging Face model architecture as a graph: layers, repeated blocks, branches and the overall data flow. It is useful when a learner has read a model card and wants to understand the internal structure.
Open the resource, paste a Hugging Face URL or repo id such as `gpt2`, and inspect the graph from overview to detail. For quick comparison, start with the popular models on the home page.
Builds step-by-step intuition for algorithms and data structures: sorting, graphs, dynamic programming, search and trees. A learner sees not only code, but also how the algorithm state changes.
Choose a topic in the left menu, press Play and follow the visualization. Then change the input or code to test edge cases and see why the algorithm behaves that way.
Shows how a small GPT model reads data, tokenizes characters, learns patterns and generates new names. It is a compact bridge from the abstract idea of a language model to an observable learning experiment.
Start with the Start from data button and move through the panels in order: data, tokens, training and generation. Change parameters when the interface offers them and compare generated results.
Trains systems thinking: distributed architecture components, request flows, latency, overload and fault tolerance. It is useful for moving from an ML model to the real service around it.
Choose a template or build a diagram from components, connect flows and run the simulation. Then enable chaos scenarios to see where the design breaks and which nodes need reinforcement.
Provides an LLM roadmap: foundational math and Python, the LLM Scientist role, fine-tuning, quantization, datasets, evaluation and application engineering. It is a strong guide for steady progression.
Open the README and choose a learning path: Fundamentals, Scientist or Engineer. For practice, run the Colab notebooks and track which techniques you have reproduced by hand.
Gives an academic deep learning frame: neural network basics, training, approximation, architectures, graphs and generalization. The material connects practical tools with university-level theory.
Watch lectures in order while keeping Lecture Notes and Homework open. After each lecture, write down the key idea and test it with a small code example.
Introduces the ecosystem of ML libraries, frameworks and software by language and task. It helps learners understand available tools around Python, TensorFlow, data mining, NLP and production ML.
Open the README, find the category you need with browser search and compare several mature projects by documentation, activity and fit for your learning task.