ounce100m-code / kernels

Commit History

p7 probe: prove --ckpt-repo chooses the publisher's weight source
50154fa
verified

Cion-lab commited on

mirror kernels/ so the kept code is self-contained
17135b7
verified

Cion-lab commited on

add p6 launcher with a Hub heartbeat: a kernel that cannot report is a kernel you cannot debug (E-056)
03388e8
verified

Cion-lab commited on

p6_cli_probe v3: build metric keys from the YAML's own metric/aggregation fields and rank candidate config files -- the first pass guessed metric_uri and picked a mmlu subject file
5f5b188
verified

Cion-lab commited on

p6_cli_probe: locate lm_eval via a child process -- a user-site dir created mid-run is not on the parent's sys.path
8ec129e
verified

Cion-lab commited on

p6_cli_probe: check run_benchmarks' flag list and PRIMARY_METRIC against the installed 0.4.13 CLI and its shipped YAMLs
48949db
verified

Cion-lab commited on

p6_harness_probe v2: measure what a resolving lm-eval install moves, instead of assuming --no-deps is safe (it is not: sacrebleu import failure)
1a8377a
verified

Cion-lab commited on

p6_harness_probe: does lm-eval 0.4.13 resolve on transformers 5.0.0, and what do the 9 task configs declare (configs only, no eval data)
1742ac7
verified

Cion-lab commited on

bootstrap: QUOTA_LEFT_HOURS 24.9 from the 11:10Z live read -- session 1 launch card
bb1c552
verified

Cion-lab commited on

phase4_bootstrap: REV must be at or after the launcher it sha-pins -- point at c237b478
a59346c
verified

Cion-lab commited on

phase4_prep: REV must be at or after the launcher it sha-pins -- point at c237b478
6a668e6
verified

Cion-lab commited on

p4_plan_sweep: repin to the D-018 launcher at c237b478
6655304
verified

Cion-lab commited on

phase4_bootstrap: repin to the D-018 launcher at c237b478
736d0dd
verified

Cion-lab commited on

phase4_prep: repin to the D-018 launcher at c237b478
bba8c7f
verified

Cion-lab commited on

launcher: adopt D-018 -- no gradient checkpointing, push every 127 steps, plan at the sustained 12,300 tok/s
c237b47
verified

Cion-lab commited on

p4_plan_sweep: drive the shipped plan() at each candidate rate to settle the session shape before D-018
09a421f
verified

Cion-lab commited on

p4_probe2_soak_runner: the wrapper as submitted for probe chain v8 / soak / rehearsal v4
7fa7ceb
verified

Cion-lab commited on

p4_prep_runner: the wrapper as submitted for probe chain v8 / soak / rehearsal v4
93e54ec
verified

Cion-lab commited on

p4_probe_runner: the wrapper as submitted for probe chain v8 / soak / rehearsal v4
f408637
verified

Cion-lab commited on

phase4_bootstrap: E-044 repin (rank-0 stop assertions; memory gate moved onto the training-only leg)
add04ae
verified

Cion-lab commited on

phase4_prep: E-044 repin (rank-0 stop assertions; memory gate moved onto the training-only leg)
b91f46c
verified

Cion-lab commited on

p4_stop_probe: E-044 repin (rank-0 stop assertions; memory gate moved onto the training-only leg)
14e982b
verified

Cion-lab commited on

launcher: repin the E-044 trainer
55b8fc4
verified

Cion-lab commited on

p4_prep_runner: re-pin to the body actually submitted
576e5ae
verified

Cion-lab commited on

p4_probe_runner: re-pin to the body actually submitted
39360ce
verified

Cion-lab commited on

phase4_bootstrap: repin to b6210a29 (per-rank post-train marks + readable probe failures)
b92088f
verified

Cion-lab commited on

phase4_prep: repin to b6210a29 (per-rank post-train marks + readable probe failures)
40f3bf5
verified

Cion-lab commited on

p4_stop_probe: repin to b6210a29 (per-rank post-train marks + readable probe failures)
98922ef
verified

Cion-lab commited on

launcher: repin the instrumented trainer
b6210a2
verified

Cion-lab commited on

phase4_bootstrap: E-042 repin + rehearsal of the launcher's real argv and the prune guard
619c5a6
verified

Cion-lab commited on

phase4_prep: E-042 repin + rehearsal of the launcher's real argv and the prune guard
7995d55
verified

Cion-lab commited on

p4_stop_probe: E-042 repin + rehearsal of the launcher's real argv and the prune guard
a4158df
verified

Cion-lab commited on

launcher: repin train files; print TRAIN_ARGV under PREP_ONLY so the rehearsal can parse the real flag list
5889e40
verified

Cion-lab commited on

prep runner: re-pin to the corrected stage-4 assertion
9fc366b
verified

Cion-lab commited on

prep stage 4: assert on n_hub, not on the log-capped missing list
b856632
verified

Cion-lab commited on

kaggle wrappers: pinned runner for the stop/resume probe
59b8c35
verified

Cion-lab commited on

kaggle wrappers: pinned runner for the CPU rehearsal
9648848
verified

Cion-lab commited on

phase4_bootstrap: pins at a813e5c7 + bounded-readback rehearsal stage / tee-prefix fix
5399519
verified

Cion-lab commited on

phase4_prep: pins at a813e5c7 + bounded-readback rehearsal stage / tee-prefix fix
1f22abe
verified

Cion-lab commited on

p4_stop_probe: pins at a813e5c7 + bounded-readback rehearsal stage / tee-prefix fix
bc34e7d
verified

Cion-lab commited on

launcher: repin hubckpt to the bounded read-back
a813e5c
verified

Cion-lab commited on

pin 7570b7b2: forced-stop legs skip the collective validation pass; probe expectations split leg1 (skip by design) from leg2 (must run it)
41ec03f
verified

Cion-lab commited on

pin 7570b7b2: forced-stop legs skip the collective validation pass; probe expectations split leg1 (skip by design) from leg2 (must run it)
dd31402
verified

Cion-lab commited on

pin 7570b7b2: forced-stop legs skip the collective validation pass; probe expectations split leg1 (skip by design) from leg2 (must run it)
0b3be64
verified

Cion-lab commited on

phase4_session: repin the trainer
7570b7b
verified

Cion-lab commited on

p4_stop_probe: repair the soak branch's timeout tuple (an earlier edit dropped it, which would have been a NameError only in soak mode) and right-size the leg ceilings under the notebook timeout
e7fecad
verified

Cion-lab commited on

phase4_session: repin the trainer (rank-visible eval failure, cross-rank peak memory)
850729d
verified

Cion-lab commited on