Article
From GRPO to DAPO and GSPO: What, Why, and How
NormalUhr
• • 128Joseph Lee
jiosephlee
AI & ML interests
None yet
Recent Activity
updated a dataset about 21 hours ago
jiosephlee/starling-transfer-shared-eval-same-species-v2-no-source-value published a dataset about 22 hours ago
jiosephlee/starling-transfer-shared-eval-same-species-v2-no-source-value updated a dataset about 22 hours ago
jiosephlee/starling-transfer-shared-eval-same-species-v2-source-valueOrganizations
None yet
upvoted an article 4 months ago
upvoted 2
articles 5 months ago
Article
SFT with vLLM Downstream Evaluation: A VRAM-Efficient Pipeline (arm64)
AlioLeuchtmann
• • 3 Article
Unlocking Agentic RL Training for GPT-OSS: A Practical Retrospective
LinkedIn
• • 80 upvoted 2
articles 6 months ago
Article
No GPU left behind: Unlocking Efficiency with Co-located vLLM in TRL
- +4
toslali-ibm, mirinflim, qgallouedec, esnible, rganti, mudhakar
• • 101 Article
Tool Use, Unified
Rocketknight1
• • 121 upvoted 3
articles 12 months ago
Article
Fixing Gradient Accumulation
- +4
lysandre, ArthurZ, muellerzr, ydshieh, BenjaminB, pcuenq
• • 66 Article
Improving Hugging Face Training Efficiency Through Packing with Flash Attention 2
- +4
RQlee, ArthurZ, achikundu, lwtr, rganti, mayank-mishra
• • 41 Article
Efficient LLM Pretraining: Packed Sequences and Masked Attention
sirluk
• • 71 upvoted 2
articles about 1 year ago
Article
Good answers are not necessarily factual answers: an analysis of hallucination in leading LLMs
davidberenstein1957
• • 42 Article
Saving Memory Using Padding-Free Transformer Layers during Finetuning
mayank-mishra
• • 21