-
Notifications
You must be signed in to change notification settings - Fork 2.4k
Pull requests: openai/parameter-golf
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Submit 1x A100 QAT Fix - 1.4078 BPB (Non-Record) [v2]
#707
opened Mar 25, 2026 by
Shuvam-Banerji-Seal
Loading…
Podracing: 1.0461 BPB (3-seed mean) — 5-gram eval + LeakyReLU²
#706
opened Mar 25, 2026 by
newjordan
Loading…
Byte-Level Tokenizer-Free Transformer: 1.2151 BPB (beats baseline 1.2244)
#705
opened Mar 25, 2026 by
seanward
Loading…
Record: PR549 + MiLe decay + 8-bit Muon + 1.04x LR + Cache+Backout — val_bpb 1.1176
#703
opened Mar 25, 2026 by
Gusanidas
Loading…
3 tasks
Record: 1.0240 BPB — Multi-Order N-gram Backoff + Entropy-Adaptive Alpha (100% autonomous research via goldfish)
#702
opened Mar 25, 2026 by
lukacf
Loading…
Record Submission: 1.0541 BPB - 5-expert Hedge Mixer + CROWN-Q + stride=64
#700
opened Mar 25, 2026 by
RoyiRa
Loading…
11L EMA LeakyReLU2 Int6 XSA4 PartialRoPE submission
#699
opened Mar 25, 2026 by
RohanMulay1
Loading…
Add MergedTop3_v3 clean 8xH100 record-track submission
#698
opened Mar 25, 2026 by
hesong0222-dev
Loading…
Add non-record JEPA byte-level encoder-decoder submission
#696
opened Mar 25, 2026 by
gravelBridge
Loading…
Record: 11L XSA6 + Warmdown3000 + QAT@0.30 (val_bpb=1.1352, 2-seed mean)
#695
opened Mar 25, 2026 by
0xNoramiya
Loading…
6 tasks done
Record: CROWN-Q + Full GPTQ + SWA/EMA Blend — val_bpb 1.1186 (3-seed mean)
#693
opened Mar 25, 2026 by
EthanYangTW
Loading…
Record: 5-expert Hedge Mixer + TTT (3-seed mean val_bpb=1.0745)
#688
opened Mar 25, 2026 by
RoyiRa
Loading…
Record: Depth Recurrence (layers 4 and 5 repeated): val_bpb 1.1182
#686
opened Mar 25, 2026 by
msisovic
Loading…
[WIP] Non-record: Local Ablation Pipeline — EMA + Int6 + Partial RoPE (GTX 1650)
#682
opened Mar 25, 2026 by
gthgomez
Loading…
Non-record: BigramHash(4096) + Cosine EMA + LZMA-9
#681
opened Mar 25, 2026 by
Alfaxad
Loading…
4 of 5 tasks
Add non-record 10min/16MB submission: Wavelet-Lite PR549 Parallel Muon (1.1483)
#680
opened Mar 25, 2026 by
bro4all
Loading…
Non-record: ASQU activation, Mixture of Convolutions, BankedLinear
#679
opened Mar 25, 2026 by
andrewmouldon
Loading…
Attention Warm-Start: Initializing Q/K from Bigram Co-occurrence SVD
#678
opened Mar 25, 2026 by
SPThole
Loading…
4 tasks done
SwiGLU MLP: parameter-neutral gated activation over LeakyReLU^2
#676
opened Mar 25, 2026 by
they-call-me-god
Loading…
Previous Next
ProTip!
Filter pull requests by the default branch with base:main.