<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <url>
    <loc>https://maosong.website/</loc>
  </url>
  <url>
    <loc>https://maosong.website/blog</loc>
  </url>
  <url>
    <loc>https://maosong.website/browse</loc>
  </url>
  <url>
    <loc>https://maosong.website/about</loc>
  </url>
  <url>
    <loc>https://maosong.website/rss.xml</loc>
  </url>
  <url>
    <loc>https://maosong.website/p/RL4LLM/</loc>
    <lastmod>2026-05-04T00:00:00.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/an-overview-of-adaption-layer-in-multimodal-large-language-models/</loc>
    <lastmod>2024-11-09T01:53:43.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/base-of-rope-bounds-context-length/</loc>
    <lastmod>2025-12-22T03:34:42.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/bleu-bilingual-evaluation-understudy/</loc>
    <lastmod>2024-04-25T14:46:53.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/chinchilla-scaling-law/</loc>
    <lastmod>2025-10-22T06:39:23.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/compression-is-intelligence/</loc>
    <lastmod>2025-07-15T08:36:09.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/data-mixture-in-mllm/</loc>
    <lastmod>2025-04-25T02:25:48.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/distributed-training-basic/</loc>
    <lastmod>2025-05-12T02:15:17.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/distributed-training-computations/</loc>
    <lastmod>2025-05-13T03:26:36.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/distributed-training-pytorch-training/</loc>
    <lastmod>2025-05-13T03:26:36.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/dual-chunk-attention/</loc>
    <lastmod>2025-07-12T02:41:12.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/fix-point-theorem/</loc>
    <lastmod>2026-03-09T09:16:02.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/formal-algorithms-for-transformer/</loc>
    <lastmod>2024-05-02T05:13:12.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/git-authentication-error/</loc>
    <lastmod>2025-02-22T02:51:27.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/gpipe/</loc>
    <lastmod>2025-12-23T08:49:25.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/gshard/</loc>
    <lastmod>2025-10-29T03:22:39.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/hands-on-llm-1-tokenizer/</loc>
    <lastmod>2025-06-29T03:47:22.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/hands-on-llm-2-transformer/</loc>
    <lastmod>2025-06-29T03:40:39.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/how-to-fix-http-error-when-creating-a-new-environment/</loc>
    <lastmod>2023-08-25T00:00:00.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/kaplan-scaling-law/</loc>
    <lastmod>2025-10-22T06:10:52.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/kl_divergence/</loc>
    <lastmod>2026-05-04T00:00:00.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/llm-flops-computation/</loc>
    <lastmod>2025-12-24T07:10:55.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/llm-memory-computation/</loc>
    <lastmod>2026-01-17T02:04:32.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/llm-parameter-computation/</loc>
    <lastmod>2025-12-24T07:12:10.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/load-balancing-tutorial/</loc>
    <lastmod>2025-12-11T08:10:08.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/megatron-lm/</loc>
    <lastmod>2026-01-21T10:04:12.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/minigpt-4-enhancing-vision-language-understanding-with-advanced-large-language-models/</loc>
    <lastmod>2024-05-02T05:13:12.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/mixstral-7b/</loc>
    <lastmod>2025-11-01T07:28:19.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/mixstral-8x7b/</loc>
    <lastmod>2025-11-01T07:32:30.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/moe-tutorial/</loc>
    <lastmod>2025-12-13T08:04:04.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/nextflow-single-branch/</loc>
    <lastmod>2026-01-17T09:31:53.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-adam/</loc>
    <lastmod>2025-09-04T02:11:55.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-adamw/</loc>
    <lastmod>2025-09-04T02:27:03.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-afm2025/</loc>
    <lastmod>2025-07-29T04:36:28.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-alibi/</loc>
    <lastmod>2025-12-24T07:10:55.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-arc-hunyuan-video-7b/</loc>
    <lastmod>2025-08-12T02:57:57.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-attention-bias/</loc>
    <lastmod>2025-05-22T07:25:07.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-aya-vision/</loc>
    <lastmod>2025-07-15T08:36:09.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-comp/</loc>
    <lastmod>2025-12-04T02:58:30.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-dapo/</loc>
    <lastmod>2025-07-15T08:36:09.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-deepseek-llm/</loc>
    <lastmod>2025-10-23T01:46:58.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-deepseek-r1/</loc>
    <lastmod>2025-12-02T10:21:54.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-deepseek-v2/</loc>
    <lastmod>2025-12-02T10:21:54.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-deepseek-v3-2/</loc>
    <lastmod>2026-01-06T09:30:40.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-deepseek-v3/</loc>
    <lastmod>2025-12-08T03:14:45.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-deepseekmoe/</loc>
    <lastmod>2025-10-28T01:46:09.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-deepstack/</loc>
    <lastmod>2025-12-04T09:32:41.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-dpo/</loc>
    <lastmod>2025-12-09T02:43:11.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-flashattention/</loc>
    <lastmod>2025-08-21T03:32:53.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-gated-attention/</loc>
    <lastmod>2026-01-20T07:41:52.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-gemini2-5/</loc>
    <lastmod>2025-12-06T10:14:15.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-gemini3-0/</loc>
    <lastmod>2026-01-06T02:26:39.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-gemma3/</loc>
    <lastmod>2025-03-15T03:15:29.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-glam/</loc>
    <lastmod>2026-01-06T10:07:29.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-glm-4-1v-thinking/</loc>
    <lastmod>2025-07-14T02:32:04.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-glm-4-5/</loc>
    <lastmod>2025-08-13T04:27:48.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-global-batch-load-balancing/</loc>
    <lastmod>2025-12-11T08:09:34.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-gpt-oss/</loc>
    <lastmod>2025-10-23T01:43:51.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-gqa/</loc>
    <lastmod>2025-10-23T01:46:58.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-gspo/</loc>
    <lastmod>2025-10-23T01:43:51.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-hunyuan-large/</loc>
    <lastmod>2025-08-06T08:46:32.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-internvl3-5/</loc>
    <lastmod>2025-09-01T03:30:50.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-keye-vl-1-5/</loc>
    <lastmod>2025-09-11T03:33:31.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-keye-vl/</loc>
    <lastmod>2025-07-23T03:11:43.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-kimi-k1-5/</loc>
    <lastmod>2025-07-24T02:56:50.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-kimi-k2-5/</loc>
    <lastmod>2026-02-12T03:13:13.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-kimi-k2/</loc>
    <lastmod>2025-07-24T02:56:50.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-kl-divergence/</loc>
    <lastmod>2026-01-24T08:32:14.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-ling-mini-beta/</loc>
    <lastmod>2025-12-13T07:58:51.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-llama3/</loc>
    <lastmod>2024-04-22T08:22:19.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-llama4-blog/</loc>
    <lastmod>2025-04-30T02:44:19.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-loss-free-balancing/</loc>
    <lastmod>2025-11-21T07:38:52.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-magistral/</loc>
    <lastmod>2025-07-16T03:04:04.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-mfa/</loc>
    <lastmod>2025-10-23T01:46:58.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-mimo-vl/</loc>
    <lastmod>2025-06-05T02:51:43.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-minimax-01/</loc>
    <lastmod>2026-01-06T09:38:01.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-mla/</loc>
    <lastmod>2025-12-03T02:13:09.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-moonlight/</loc>
    <lastmod>2025-08-07T02:49:32.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-mqa/</loc>
    <lastmod>2025-08-07T10:06:37.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-muon-blog/</loc>
    <lastmod>2025-08-05T03:10:51.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-mx-format/</loc>
    <lastmod>2025-08-21T10:23:03.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-nope/</loc>
    <lastmod>2025-12-24T07:19:42.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-nsa/</loc>
    <lastmod>2025-12-15T09:39:16.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-olmoe-openmoe/</loc>
    <lastmod>2025-12-06T10:08:11.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-olmoe/</loc>
    <lastmod>2025-11-01T07:23:58.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-openmath-nemotron/</loc>
    <lastmod>2026-04-15T01:50:39.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-phi-4/</loc>
    <lastmod>2024-12-16T09:33:52.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-position-encoding/</loc>
    <lastmod>2025-05-19T02:46:39.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qk-norm/</loc>
    <lastmod>2025-08-13T08:12:11.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen-llm/</loc>
    <lastmod>2025-07-12T02:32:03.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen1-5/</loc>
    <lastmod>2025-07-12T02:35:10.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen2-5-1m/</loc>
    <lastmod>2025-07-12T03:00:47.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen2-5-omni/</loc>
    <lastmod>2025-04-01T02:29:00.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen2-5-vl/</loc>
    <lastmod>2025-03-04T02:46:42.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen2-5/</loc>
    <lastmod>2025-07-12T02:51:42.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen2/</loc>
    <lastmod>2025-07-12T02:36:43.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen3-blog/</loc>
    <lastmod>2025-04-29T03:23:04.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen3-next/</loc>
    <lastmod>2026-01-23T02:29:56.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen3-vl/</loc>
    <lastmod>2025-12-05T02:12:01.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwen3/</loc>
    <lastmod>2025-07-12T03:02:17.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-qwq-32b/</loc>
    <lastmod>2025-03-08T01:46:16.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-rag/</loc>
    <lastmod>2024-04-14T04:38:04.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-rnope-swa/</loc>
    <lastmod>2025-09-02T03:24:10.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-roofline-model/</loc>
    <lastmod>2026-02-26T09:23:40.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-sapo/</loc>
    <lastmod>2025-12-05T02:09:06.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-seed1-5-vl/</loc>
    <lastmod>2025-05-14T01:28:07.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-seed1-6/</loc>
    <lastmod>2025-07-18T06:59:35.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-smollm3/</loc>
    <lastmod>2025-07-15T03:01:13.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-softmax/</loc>
    <lastmod>2025-12-27T08:39:53.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-step3-vl-10b/</loc>
    <lastmod>2026-02-13T10:05:47.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-streamingllm/</loc>
    <lastmod>2025-08-20T02:57:03.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-t-sne/</loc>
    <lastmod>2024-05-02T05:13:12.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-t5/</loc>
    <lastmod>2025-12-24T07:07:08.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-v-triune/</loc>
    <lastmod>2025-07-17T01:37:36.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-vapo/</loc>
    <lastmod>2025-04-17T01:41:51.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-vc-ppo/</loc>
    <lastmod>2025-04-14T09:36:15.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-vit/</loc>
    <lastmod>2025-12-04T03:00:44.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-vita/</loc>
    <lastmod>2024-08-13T06:36:45.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/notes-on-yarn/</loc>
    <lastmod>2025-07-15T08:36:09.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/nvidia-gpu-specs/</loc>
    <lastmod>2026-01-14T03:09:19.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/overview-of-qwen-vl-series/</loc>
    <lastmod>2025-03-09T07:11:29.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/ovis-discrete-visual-embedding/</loc>
    <lastmod>2025-08-30T09:32:22.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/ovis2-5-mllm-with-stronger-perception-and-reasoning-capability/</loc>
    <lastmod>2025-08-30T09:34:44.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/performance-and-scalability/</loc>
    <lastmod>2026-03-26T09:44:26.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/practical-advice-for-analysis-of-large-complex-data-sets/</loc>
    <lastmod>2024-04-17T14:40:11.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/regularization-methods-in-deep-learning/</loc>
    <lastmod>2024-04-27T10:02:02.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/relationship-between-mle-and-kl-divergence/</loc>
    <lastmod>2025-06-27T03:35:33.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/rouge-recall-oriented-understudy/</loc>
    <lastmod>2024-05-09T09:35:20.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/rules-of-machine-learning/</loc>
    <lastmod>2024-04-13T11:57:47.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/screen-usage/</loc>
    <lastmod>2025-02-05T07:14:43.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/st-moe/</loc>
    <lastmod>2025-10-29T03:19:37.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/state-of-ai-openrouter-100t-tokenai/</loc>
    <lastmod>2026-01-17T09:04:07.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/switch-transformer/</loc>
    <lastmod>2025-10-28T01:38:12.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/understanding-sigmoid-loss-in-siglip/</loc>
    <lastmod>2025-03-28T06:55:50.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/unified-perspective-on-dllm-and-llm/</loc>
    <lastmod>2025-06-28T07:02:09.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/what-s-next-for-ai-agentic-workflows/</loc>
    <lastmod>2024-04-14T04:38:04.000Z</lastmod>
  </url>
  <url>
    <loc>https://maosong.website/p/%E9%9A%8F%E7%AC%94/</loc>
    <lastmod>2025-04-23T05:24:02.000Z</lastmod>
  </url>
</urlset>