repoGitHubTrust 82 · PrimaryPublished 2h agoLive · 2h ago
DaoyuanLi2816/mini-verl
Run a documented subset of verl-style OPD on one consumer GPU—typed config, Parquet prompts, and PEFT scale-out artifacts.
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 50%We'll benchmark an Open weights LLM on any GPU you choose — drop your model + hardware and we'll run it. [D] →
- PossiblePossibly related (embedding) · 46%Open Memory Protocol – One Memory Store for Claude, ChatGPT, Curso →
- PossiblePossibly related (embedding) · 46%DeepSeek-V4-Flash (MXFP4): compute buffer scales ~3x just from KV cache quant type (f16 vs q8_0) — anyone else seeing this? Llama.cpp →
- PossiblePossibly related (embedding) · 46%Open-sourcing a two-stage prompt-injection detector (regex gate + quantised DeBERTa-v3 ONNX), trained partly on real attacks from a game I ran [P] →
- PossiblePossibly related (embedding) · 45%Pokee-Isaac 28B: 10M-Token Context on a Single GPU - Pasquale Pillitteri →
Covers
newsWe'll benchmark an Open weights LLM on any GPU you choose — drop your model + hardware and we'll run it. [D]newsOpen Memory Protocol – One Memory Store for Claude, ChatGPT, CursonewsDeepSeek-V4-Flash (MXFP4): compute buffer scales ~3x just from KV cache quant type (f16 vs q8_0) — anyone else seeing this? Llama.cppnewsOpen-sourcing a two-stage prompt-injection detector (regex gate + quantised DeBERTa-v3 ONNX), trained partly on real attacks from a game I ran [P]
Covers (incoming)
Related across the graph
newsDeepSeek-V4-Flash (MXFP4): compute buffer scales ~3x just from KV cache quant type (f16 vs q8_0) — anyone else seeing this? Llama.cppnewsOpen-sourcing a two-stage prompt-injection detector (regex gate + quantised DeBERTa-v3 ONNX), trained partly on real attacks from a game I ran [P]newsWe'll benchmark an Open weights LLM on any GPU you choose — drop your model + hardware and we'll run it. [D]newsPokee-Isaac 28B: 10M-Token Context on a Single GPU - Pasquale PillitterinewsOpen Memory Protocol – One Memory Store for Claude, ChatGPT, Curso