Training

Pinned sources The Llama 3 Herd of Models (arXiv 2407.21783v3), whose results are all for Llama 3.1 · for Llama 3, its model card (meta-llama/llama3 @ a0940f9) and Meta’s post of April 2024 · no training code is published · How to read

  1. Start of training
  3. Initial pre-trainingat 4K, then 8K15.6T8KNTPall
  4. Off this path: domain experts (code, multilingual)2 stages
  5. Branched from the pre-training run; the point is not named
  6. Code expert1T tokens, mostly code1T16KNTPnot stated
  7. Multilingual expert90% multilingual tokensnot statednot statedNTPnot stated
  8. Long-context pre-trainingsix stages, 8K to 128K800B128KNTPallcontext to 128K
  9. Annealinglearning rate to 040M128KNTPall
  10. Pre-trained model
  12. × 6 roundsRejection sampling and SFTthen DPO, each roundnot stated128KSFTRS and synthetic datanot stated
  13. DPOafter SFT, each roundnot statedshortDPO+ NLL 0.2not stated
  14. Llama 3.1 405B Instruct
  • tokens, log scale
  • derived: computed from stated numbers
  • context length
  • all parameters train
  • a part trains
  • all but a part trains
  • NTPdashed: concluded, no source names it
  • not stated by the sources
  • the architecture changes
  • checkpoint
  • checkpoint, weights published
  • a join the sources leave unexplained
  • data from a branch into a stage
  • stages repeated together

Multilingual expert

×

Stage · Llama 3.1 405B

A branch of the pre-training run continues on a mix of 90% multilingual tokens and is post-trained; it collects annotations in non-English languages until pre-training is complete.

Data
90% multilingual tokens
Use
Collecting higher-quality annotations in non-English languages “until pre-training was fully complete”

Derived

Objective. The paper calls this “continuing pre-training” on another data mix and names no objective: next-token prediction is concluded from that, and the paper does not state it.

Not disclosed Tokens · Learning rate

Architecture at this stageA copy of the model, branched from its pre-training run.

All settings3
Data
A mix of 90% multilingual tokens
Post-training
The recipes of §4.1
Use
Collecting higher-quality annotations in non-English languages “until pre-training was fully complete”
Sources1

Llama 3.1

All models, with previews →

Meta · Jul 2024Dense decoder · follows Llama 3

405B