Pith. sign in

REVIEW 6 cited by

What Language Model to Train if You Have One Million GPU Hours?

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2210.15424 v2 pith:4CYQ5DEN submitted 2022-10-27 cs.CL cs.AIcs.LG

classification cs.CLcs.AIcs.LG
keywords impactmodelinglanguagemodelmodelsscalearchitecturecapabilities
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

The crystallization of modeling methods around the Transformer architecture has been a boon for practitioners. Simple, well-motivated architectural variations can transfer across tasks and scale, increasing the impact of modeling research. However, with the emergence of state-of-the-art 100B+ parameters models, large language models are increasingly expensive to accurately design and train. Notably, it can be difficult to evaluate how modeling decisions may impact emergent capabilities, given that these capabilities arise mainly from sheer scale alone. In the process of building BLOOM--the Big Science Large Open-science Open-access Multilingual language model--our goal is to identify an architecture and training setup that makes the best use of our 1,000,000 A100-GPU-hours budget. Specifically, we perform an ablation study at the billion-parameter scale comparing different modeling practices and their impact on zero-shot generalization. In addition, we study the impact of various popular pre-training corpora on zero-shot generalization. We also study the performance of a multilingual model and how it compares to the English-only one. Finally, we consider the scaling behaviour of Transformers to choose the target model size, shape, and training setup. All our models and code are open-sourced at https://huggingface.co/bigscience .

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Mix-LN: Unleashing the Power of Deeper Layers by Combining Pre-LN and Post-LN

    cs.LG 2024-12 conditional novelty 6.0 of 10

    Mix-LN, which uses Post-LN in early layers and Pre-LN in later layers, gives more uniform layer gradients and better LLM pretraining and fine-tuning results than Pre-LN or Post-LN alone.

  2. SoftSignSGD(S3): An Enhanced Optimizer for Practical DNN Training and Loss Spikes Minimization Beyond Adam

    cs.LG 2025-07 reject novelty 5.0 of 10

    S3, an optimizer with a p-th order momentum denominator, equal EMA coefficients, and Nesterov acceleration, is claimed to match AdamW's 100k-step perplexity at 50k steps while avoiding loss spikes.

  3. GreenMind: A Next-Generation Vietnamese Large Language Model for Structured and Logical Reasoning

    cs.CL 2025-04 conditional novelty 5.0 of 10

    GreenMind-Medium-14B-R1, a Vietnamese reasoning model trained with GRPO plus language and semantic-similarity rewards, reports accuracy gains on VLSP 2023, SeaExam, and VMLU.

  4. LLM360 K2: Building a 65B 360-Open-Source Large Language Model from Scratch

    cs.LG 2025-01 conditional novelty 5.0 of 10

    K2 Diamond is a fully open 65B-parameter LLM that reaches Llama 2 70B-level performance on standard benchmarks.

  5. SPAM: Spike-Aware Adam with Momentum Reset for Stable LLM Training

    cs.LG 2025-01 conditional novelty 5.0 of 10

    SPAM, an Adam variant with periodic momentum reset and ratio-based spike clipping, reports better validation perplexity and lower memory use than Adam, Adafactor, GaLore, and Adam-mini in LLM pretraining and fine-tuning.

  6. YuLan-Mini: An Open Data-efficient Language Model

    cs.CL 2024-12 conditional novelty 4.0 of 10

    A 2.42B-parameter base model trained on 1.08T tokens matches or beats several industry baselines trained on 7T to 18T tokens across math, code, and general benchmarks.

Pith tools