https://llmdeepdive.com/https://llmdeepdive.com/explore/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.1-what-is-a-language-model/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.2-history-ngram-to-gpt/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.3-probability-you-need/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.4-vectors-matrices-dot-products/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.5-matrix-multiplication-transformation/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.6-derivatives-gradients/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.7-python-numpy-pytorch/https://llmdeepdive.com/lessons/0-orientation-and-foundations/0.8-how-to-read-ml-paper/https://llmdeepdive.com/lessons/1-text-to-tensors/1.1-why-computers-cant-read/https://llmdeepdive.com/lessons/1-text-to-tensors/1.2-tokenization-i/https://llmdeepdive.com/lessons/1-text-to-tensors/1.3-bpe-step-by-step/https://llmdeepdive.com/lessons/1-text-to-tensors/1.4-sentencepiece-byte-level-token-counts/https://llmdeepdive.com/lessons/1-text-to-tensors/1.5-embeddings-meaning-as-geometry/https://llmdeepdive.com/lessons/1-text-to-tensors/1.6-word2vec-glove-analogy/https://llmdeepdive.com/lessons/1-text-to-tensors/1.7-entropy-perplexity/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.1-perceptron/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.2-multilayer-perceptrons-nonlinearity/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.3-activation-functions/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.4-loss-functions-cross-entropy/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.5-backpropagation/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.6-gradient-descent-loss-landscape/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.7-optimizers/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.8-initialization-normalization-residuals/https://llmdeepdive.com/lessons/2-neural-network-fundamentals/2.9-overfitting-regularization-bitter-lesson/https://llmdeepdive.com/lessons/3-sequence-models/3.1-modelling-sequences/https://llmdeepdive.com/lessons/3-sequence-models/3.2-rnns-vanishing-gradient/https://llmdeepdive.com/lessons/3-sequence-models/3.3-lstm-gru-gates/https://llmdeepdive.com/lessons/3-sequence-models/3.4-seq2seq-encoder-decoder-bottleneck/https://llmdeepdive.com/lessons/3-sequence-models/3.5-bahdanau-attention/https://llmdeepdive.com/lessons/4-transformer/4.1-attention-is-all-you-need-in-context/https://llmdeepdive.com/lessons/4-transformer/4.2-self-attention/https://llmdeepdive.com/lessons/4-transformer/4.3-scaled-dot-product-attention/https://llmdeepdive.com/lessons/4-transformer/4.4-causal-masking/https://llmdeepdive.com/lessons/4-transformer/4.5-multi-head-attention/https://llmdeepdive.com/lessons/4-transformer/4.6-positional-encoding-sinusoidal-learned/https://llmdeepdive.com/lessons/4-transformer/4.7-rope/https://llmdeepdive.com/lessons/4-transformer/4.8-alibi-relative-bias/https://llmdeepdive.com/lessons/4-transformer/4.9-feed-forward-block/https://llmdeepdive.com/lessons/4-transformer/4.10-residuals-pre-norm-post-norm/https://llmdeepdive.com/lessons/4-transformer/4.11-full-transformer-block/https://llmdeepdive.com/lessons/4-transformer/4.12-transformer-families/https://llmdeepdive.com/lessons/4-transformer/4.13-build-gpt-from-scratch/https://llmdeepdive.com/lessons/4-transformer/4.14-reading-real-weights/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.1-pretraining-objectives/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.2-data-pipeline/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.3-training-tokenizer/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.4-kaplan-scaling-laws/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.5-chinchilla-compute-optimality/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.6-inference-aware-scaling/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.7-data-parallelism-zero-fsdp/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.8-model-parallelism/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.9-mixed-precision/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.10-gradient-checkpointing/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.11-learning-rate-schedules/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.12-training-instabilities/https://llmdeepdive.com/lessons/5-pretraining-at-scale/5.13-pretraining-costs/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.1-why-base-models-arent-assistants/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.2-supervised-fine-tuning-and-instruction-data/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.3-lora-qlora-and-peft/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.4-reward-models-and-human-preference-data/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.5-rlhf-with-ppo/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.6-dpo-skipping-the-reward-model/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.7-orpo-kto-simpo-and-the-alignment-zoo/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.8-grpo-and-rl-on-verifiable-rewards/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.9-reasoning-models-test-time-compute-and-long-cot/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.10-constitutional-ai-and-rlaif/https://llmdeepdive.com/lessons/6-post-training-and-alignment/6.11-distillation-making-small-models-punch-up/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.1-what-actually-happens-when-you-hit-send/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.2-the-kv-cache/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.3-prefill-vs-decode-two-different-machines/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.4-sampling/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.5-beam-search-speculative-decoding-and-medusa/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.6-mqa-gqa-and-mla/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.7-flashattention-and-io-awareness/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.8-pagedattention-and-continuous-batching/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.9-quantization-i-int8-int4-the-basics/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.10-quantization-ii-gptq-awq-gguf-qat/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.11-serving-stacks-vllm-sglang-tensorrt-llm/https://llmdeepdive.com/lessons/7-inference-and-efficiency/7.12-running-models-locally/https://llmdeepdive.com/pt-br/https://llmdeepdive.com/pt-br/explore/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.1-what-is-a-language-model/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.2-history-ngram-to-gpt/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.3-probability-you-need/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.4-vectors-matrices-dot-products/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.5-matrix-multiplication-transformation/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.6-derivatives-gradients/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.7-python-numpy-pytorch/https://llmdeepdive.com/pt-br/lessons/0-orientation-and-foundations/0.8-how-to-read-ml-paper/https://llmdeepdive.com/pt-br/lessons/1-text-to-tensors/1.1-why-computers-cant-read/https://llmdeepdive.com/pt-br/lessons/1-text-to-tensors/1.2-tokenization-i/https://llmdeepdive.com/pt-br/lessons/1-text-to-tensors/1.3-bpe-step-by-step/https://llmdeepdive.com/pt-br/lessons/1-text-to-tensors/1.4-sentencepiece-byte-level-token-counts/https://llmdeepdive.com/pt-br/lessons/1-text-to-tensors/1.5-embeddings-meaning-as-geometry/https://llmdeepdive.com/pt-br/lessons/1-text-to-tensors/1.6-word2vec-glove-analogy/https://llmdeepdive.com/pt-br/lessons/1-text-to-tensors/1.7-entropy-perplexity/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.1-perceptron/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.2-multilayer-perceptrons-nonlinearity/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.3-activation-functions/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.4-loss-functions-cross-entropy/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.5-backpropagation/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.6-gradient-descent-loss-landscape/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.7-optimizers/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.8-initialization-normalization-residuals/https://llmdeepdive.com/pt-br/lessons/2-neural-network-fundamentals/2.9-overfitting-regularization-bitter-lesson/https://llmdeepdive.com/pt-br/lessons/3-sequence-models/3.1-modelling-sequences/https://llmdeepdive.com/pt-br/lessons/3-sequence-models/3.2-rnns-vanishing-gradient/https://llmdeepdive.com/pt-br/lessons/3-sequence-models/3.3-lstm-gru-gates/https://llmdeepdive.com/pt-br/lessons/3-sequence-models/3.4-seq2seq-encoder-decoder-bottleneck/https://llmdeepdive.com/pt-br/lessons/3-sequence-models/3.5-bahdanau-attention/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.1-attention-is-all-you-need-in-context/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.2-self-attention/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.3-scaled-dot-product-attention/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.4-causal-masking/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.5-multi-head-attention/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.6-positional-encoding-sinusoidal-learned/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.7-rope/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.8-alibi-relative-bias/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.9-feed-forward-block/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.10-residuals-pre-norm-post-norm/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.11-full-transformer-block/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.12-transformer-families/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.13-build-gpt-from-scratch/https://llmdeepdive.com/pt-br/lessons/4-transformer/4.14-reading-real-weights/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.1-pretraining-objectives/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.2-data-pipeline/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.3-training-tokenizer/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.4-kaplan-scaling-laws/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.5-chinchilla-compute-optimality/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.6-inference-aware-scaling/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.7-data-parallelism-zero-fsdp/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.8-model-parallelism/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.9-mixed-precision/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.10-gradient-checkpointing/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.11-learning-rate-schedules/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.12-training-instabilities/https://llmdeepdive.com/pt-br/lessons/5-pretraining-at-scale/5.13-pretraining-costs/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.1-why-base-models-arent-assistants/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.2-supervised-fine-tuning-and-instruction-data/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.3-lora-qlora-and-peft/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.4-reward-models-and-human-preference-data/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.5-rlhf-with-ppo/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.6-dpo-skipping-the-reward-model/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.7-orpo-kto-simpo-and-the-alignment-zoo/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.8-grpo-and-rl-on-verifiable-rewards/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.9-reasoning-models-test-time-compute-and-long-cot/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.10-constitutional-ai-and-rlaif/https://llmdeepdive.com/pt-br/lessons/6-post-training-and-alignment/6.11-distillation-making-small-models-punch-up/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.1-what-actually-happens-when-you-hit-send/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.2-the-kv-cache/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.3-prefill-vs-decode-two-different-machines/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.4-sampling/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.5-beam-search-speculative-decoding-and-medusa/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.6-mqa-gqa-and-mla/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.7-flashattention-and-io-awareness/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.8-pagedattention-and-continuous-batching/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.9-quantization-i-int8-int4-the-basics/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.10-quantization-ii-gptq-awq-gguf-qat/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.11-serving-stacks-vllm-sglang-tensorrt-llm/https://llmdeepdive.com/pt-br/lessons/7-inference-and-efficiency/7.12-running-models-locally/https://llmdeepdive.com/pt-br/tracks/https://llmdeepdive.com/pt-br/tracks/0-orientation-and-foundations/https://llmdeepdive.com/pt-br/tracks/1-text-to-tensors/https://llmdeepdive.com/pt-br/tracks/2-neural-network-fundamentals/https://llmdeepdive.com/pt-br/tracks/3-sequence-models/https://llmdeepdive.com/pt-br/tracks/4-transformer/https://llmdeepdive.com/pt-br/tracks/5-pretraining-at-scale/https://llmdeepdive.com/pt-br/tracks/6-post-training-and-alignment/https://llmdeepdive.com/pt-br/tracks/7-inference-and-efficiency/https://llmdeepdive.com/tracks/https://llmdeepdive.com/tracks/0-orientation-and-foundations/https://llmdeepdive.com/tracks/1-text-to-tensors/https://llmdeepdive.com/tracks/2-neural-network-fundamentals/https://llmdeepdive.com/tracks/3-sequence-models/https://llmdeepdive.com/tracks/4-transformer/https://llmdeepdive.com/tracks/5-pretraining-at-scale/https://llmdeepdive.com/tracks/6-post-training-and-alignment/https://llmdeepdive.com/tracks/7-inference-and-efficiency/