Repository navigation

feat(rocm): enable bitwise Qwen3 Dense VIME alignment - #388

Merged

Flink-ddd merged 32 commits into

testfrom

rocm-vime-qwen3-bitwise

Sep 5, 2026

Commits on Aug 31, 2026
- Add VIME Qwen3-8B 200-round consistency experiment
  frank-2077
  committed
  577d51f
- Show single-step ablation points in plots
  frank-2077
  committed
  f12b154
- Make GRPO explicit in VIME experiments
  frank-2077
  committed
  7a93b1f
- Enable activation recompute for long GRPO responses
  frank-2077
  committed
  b4fc7e8
- Keep deterministic collective staging IPC compatible
  frank-2077
  committed
  64a41f3
- Tune GRPO experiment for long math responses
  frank-2077
  committed
  c4040cd
- Cap GRPO responses below CP attention OOM threshold
  frank-2077
  committed
  0818028
- Run VIME consistency matrix with TP4 CP2
  inaniloquentee
  committed
  9caf032
- Keep colocated TP4 actor resident
  inaniloquentee
  committed
  bcb5e69
- Align Megatron TP4 attention reduction tree
  inaniloquentee
  committed
  5403df6
Commits on Sep 1, 2026
- Pin VIME Qwen3 production attention backend
  frank-2077
  committed
  77c9d77
- Use fused attention for TP4 CP2 production arms
  frank-2077
  committed
  07aede0
- fix-native-production-logp-route
  frank-2077
  committed
  576741f
- decouple-cuda-graph-from-logp-route
  frank-2077
  committed
  d2173e8
- docs: publish VIME TP4 CP2 consistency results
  frank-2077
  committed
  09695b0
- docs: add VIME TP4 CP2 reproduction runbook
  frank-2077
  committed
  e2f6279
- optimize TP4 rollout scheduling
  
  frank-2077
  authored and
  inaniloquentee
  committed
  ea46f22
- docs: clarify VIME mismatch comparison
  inaniloquentee
  committed
  a463e77
- perf: write deterministic GEMMs into collective staging
  inaniloquentee
  committed
  f50c367
- feat: configure reference-model comparison runs
  inaniloquentee
  committed
  b3bfd2b
Commits on Sep 4, 2026
- perf: optimize strict VIME TP4/CP2 execution
  frank-2077
  committed
  6a7646c
- docs: publish optimized G11 200-step comparison
  frank-2077
  committed
  40db4d3
- style: format strict VIME integration changes
  frank-2077
  committed
  c663a71
- feat: add TP4 VIME supplement suites
  frank-2077
  committed
  acdf7ae
- test: shorten VIME precision supplements to eight rounds
  inaniloquentee
  committed
  0ae2c7b
Commits on Sep 5, 2026
- Merge pull request #377 from RL-Align/vime-qwen3-8b-tp4-cp2-200
  Flink-ddd
  authored
  ccb70e3
- feat(rocm): add end-to-end attention ablation matrix
  
  codex
  authored and
  inaniloquentee
  committed
  65f28d2
- feat(rocm): enable bitwise Qwen3 VIME alignment
  inaniloquentee
  committed
  ef18a08
- refactor(rocm): isolate strict VIME platform paths
  inaniloquentee
  committed
  0bc2085
- refactor(rocm): move deterministic RoPE to ROCm
  inaniloquentee
  committed
  999d666
- feat(rocm): complete deterministic VIME runtime
  inaniloquentee
  committed
  6152f16
- Merge remote-tracking branch 'origin/test' into pr388-conflict-resolve
  inaniloquentee
  committed
  872780c

Navigation Menu

Repository navigation

feat(rocm): enable bitwise Qwen3 Dense VIME alignment - #388

feat(rocm): enable bitwise Qwen3 Dense VIME alignment#388
Flink-ddd merged 32 commits into
testfrom
rocm-vime-qwen3-bitwise

Commits on Aug 31, 2026

Add VIME Qwen3-8B 200-round consistency experiment

Show single-step ablation points in plots

Make GRPO explicit in VIME experiments

Enable activation recompute for long GRPO responses

Keep deterministic collective staging IPC compatible

Tune GRPO experiment for long math responses

Cap GRPO responses below CP attention OOM threshold

Run VIME consistency matrix with TP4 CP2

Keep colocated TP4 actor resident

Align Megatron TP4 attention reduction tree

Commits on Sep 1, 2026

Pin VIME Qwen3 production attention backend

Use fused attention for TP4 CP2 production arms

fix-native-production-logp-route

decouple-cuda-graph-from-logp-route

docs: publish VIME TP4 CP2 consistency results

docs: add VIME TP4 CP2 reproduction runbook

optimize TP4 rollout scheduling

docs: clarify VIME mismatch comparison

perf: write deterministic GEMMs into collective staging

feat: configure reference-model comparison runs

Commits on Sep 4, 2026

perf: optimize strict VIME TP4/CP2 execution

docs: publish optimized G11 200-step comparison

style: format strict VIME integration changes

feat: add TP4 VIME supplement suites

test: shorten VIME precision supplements to eight rounds

Commits on Sep 5, 2026

Merge pull request #377 from RL-Align/vime-qwen3-8b-tp4-cp2-200

feat(rocm): add end-to-end attention ablation matrix

feat(rocm): enable bitwise Qwen3 VIME alignment

refactor(rocm): isolate strict VIME platform paths

refactor(rocm): move deterministic RoPE to ROCm

feat(rocm): complete deterministic VIME runtime

Merge remote-tracking branch 'origin/test' into pr388-conflict-resolve