HN

Comments (116)

• maxdo
  Interesting :Q: Did you just way over-optimize for a specific CPU and tokenizer? How is it so fast? No, I way over-optimized for every combination of these! The results are very consistent across CPUs (modern x86 and ARM), and across specific tokenizers.The major improvements are in optimizing heavily an implementation that usually is outsourced to a Regex engine (pretokenization) using SIMD, minimizing branching and other tricks, as well as heavily optimizing caching of pretoken mappings (if a word has been seen before, look it up its encoded tokens efficiently). Caching is a very hard problem in this domain since the cache grows very quickly, and pretoken distributions are very long-tailed.Finally, interactions with Python are minimized, and threads have minimal interactions with each other.
• Menu

  cschmidt
  Can I say this seems to be fantastic work. I cloned your repo earlier today after seeing it on the tokenization discord. I know everyone in the tokenization community wants to absorb the lessons of how you got such a speedup. The caching and replacing the regex for pretokenization seem like generally useful ideas.And screw all the 0.1% haters on here, this is great stuff.
• Menu

  XCSme
  Congrats, I love performance optimizations!Hardware nowadays is so powerful, but our code so inefficient... I think most libraries/apps could easily be 10x-100x faster if we really try to optimize them.The good thing is, that now with AI, we'll probably have the time to implement those optimizations rather quickly.
• Menu

  ubedan
  Spectacular... Reminds me of the SimdJson algorithm in terms of jaw dropping nearly unbelievable speeds through creative programming. I hope this code get popular, as it will save tons of electricity, money, CO2, etc.Have you considered publishing a rust crate as well? (If not, I volunteer.)
• Menu

  onlyrealcuzzo
  This is awesome, but tokenization is typically <0.1% of total inference time.Presumably there's a host of applications that just need to tokenize, though, and this would be great for those!
• Menu

  apollopower
  Cool stuff. From my understanding, this is less valuable at inference time and more useful when running offline pre-training data prep.When tokenizing terabytes of text for your training corpus, the speedup here is probably doing real work in saving you time (and money?). You get a faster iteration cycle when figuring out and adjusting your datasets.
• Menu

  Antibabelic
  "AI Use Disclosure: A majority of this code base was crafted by hand without any use of AI (which can be seen from the project's Git history)."So much for "human programming is obsolete".
• zX41ZdbW
  This is exactly what we need! Will try: https://github.com/ClickHouse/ClickHouse/issues/108247It will be nicer if the README focuses more on per-core performance.About the actual algorithm - will something like matching in a perfect hash table help?
• Menu

  luciana1u
  engineering effort to make something 1000x faster that accounts for 0.1% of total runtime is the most software developer thing imaginable
• 0xnyn
  I had to stare at that chart for a minute just to let the numbers sink in. It's genuinely mind-bending, incredible ship OP
• Menu

  swiftcoder
  So the question becomes, how many other parts of the inference pipeline have left 1000x optimization opportunities lying on the table?
• Menu

  fwip
  What sort of setups do people have that are bounded by the speed of the tokenizer?
• chocrates
  Practically I would need to wait for hugging face models to adopt this? My harness tokenizer is just an estimate since the model tokenizes on my api calls?
• hansvm
  For the lazy among us (not me of course), is there a small number of core techniques which enabled this for even a single architecture and single CPU core?
• mcpindex-ai
  Tokenization is one of the most under appreciated and under optimized part of the agentic stack- not sure if this is truly production grade, and applicable across all hardware+stack combo but this for sure can help inspire a lot of that work. Good work!
• sashank_1509
  This is really cool, great work!
• nxpnsv
  Surely it should be kilotoken
• Menu

  phplovesong
  Is tokenizing really the bottleneck? If we go from 20ms to 15ms does it really matter?
• anonymousmoos
  Quality software here.
• Menu

  dmezzetti
  Very interesting project! Are there benchmarks for the "compatibility mode" or are all the numbers for the Gigatoken API?
• zerolines
  wow, best release all week.
• semiinfinitely
  quite excellent software
• fenestella
  [flagged]
• Timmyzzz
  [flagged]
• wacjiomv
  [dead]
• saschag
  [flagged]
• tim_tihub
  [dead]
• luck7710
  [dead]
• histiq
  [dead]
• kiaansaraiya
  [dead]
• sghn
  [flagged]
• anon
  undefined
• Menu

  wacjiomv
  [flagged]
• Menu

  vmware508
  We should just rewrite everything in Rust, especially bloated Python code, and the world would be a better place. ;) Disclosure: I'm a Rust advocate!