auto-round
Recurrent Residual Quantization (RRQ) for LLMs
#2308
Open

Recurrent Residual Quantization (RRQ) for LLMs #2308

luoyu-intel wants to merge 60 commits into main from feat/rrq-phase1
luoyu-intel
luoyu-intel
luoyu-intel luoyu-intel requested a review from wenhuach21 wenhuach21 64 days ago
luoyu-intel
AutoRoundBot
azure-pipelines
hshen14 hshen14 requested a review from xin3he xin3he 63 days ago
hshen14 hshen14 requested a review from lvliang-intel lvliang-intel 63 days ago
hshen14 hshen14 requested a review from yiliu30 yiliu30 63 days ago
hshen14 hshen14 requested a review from lkk12014402 lkk12014402 63 days ago
hshen14 hshen14 requested a review from copilot-pull-request-reviewer copilot-pull-request-reviewer 63 days ago
copilot-pull-request-reviewer
copilot-pull-request-reviewer commented on 2026-09-09
hshen14
hshen14 commented on 2026-09-09
hshen14 hshen14 requested a review from Zhenzhong1 Zhenzhong1 63 days ago
lvliang-intel
lvliang-intel commented on 2026-09-09
lvliang-intel
lvliang-intel commented on 2026-09-09
a32543254
a32543254 approved these changes on 2026-09-09
a32543254
a32543254 commented on 2026-09-09
AutoRoundBot
wenhuach21
wenhuach21 commented on 2026-09-09
wenhuach21
wenhuach21 commented on 2026-09-09
wenhuach21
wenhuach21 commented on 2026-09-09
AutoRoundBot
luoyu-intel
luoyu-intel luoyu-intel force pushed from 7006c7a7 to f4aae121 57 days ago
AutoRoundBot
azure-pipelines
luoyu-intel luoyu-intel requested a review from copilot-pull-request-reviewer copilot-pull-request-reviewer 57 days ago
copilot-pull-request-reviewer
copilot-pull-request-reviewer commented on 2026-09-15
wenhuach21
wenhuach21 commented on 2026-09-15
wenhuach21
wenhuach21 commented on 2026-09-15
wenhuach21
wenhuach21 commented on 2026-09-15
wenhuach21
wenhuach21 commented on 2026-09-15
wenhuach21
wenhuach21 commented on 2026-09-15
wenhuach21
wenhuach21 approved these changes on 2026-09-15
yiliu30
yiliu30 approved these changes on 2026-09-19
luoyu-intel luoyu-intel force pushed from 027407b7 to 378a1293 30 days ago
wenhuach21
wenhuach21 commented on 2026-10-12
AutoRoundBot
azure-pipelines
luoyu-intel luoyu-intel force pushed from 6e8ca15d to e66d989a 7 days ago
luoyu-intel luoyu-intel force pushed from e66d989a to 63161b66 7 days ago
luoyu-intel feat(rrq): Phase 1 - Recurrent Residual Quantization (packed INT2 2+2…
9768e6c0
luoyu-intel feat(rrq): Phase 2 - generate_rrq_residual (incremental from existing…
a197b630
luoyu-intel feat(rrq): Phase 3 - add per-plane sign-SGD tuning
5b411e83
luoyu-intel fix(rrq): accumulate optimized residual prefix
40639cf7
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
39967d91
luoyu-intel feat(rrq): match AutoRound per-plane quality + configurable mixed-pre…
c8614b5f
luoyu-intel docs(rrq): add Recurrent Residual Quantization entry to README
da960b21
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
581545d1
luoyu-intel perf(rrq): replace 4-plane matmul accumulation with single dequant+ma…
23fc2b4b
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
aa5bab69
luoyu-intel fix(rrq): declare tunable fields as named __init__ params
827de067
luoyu-intel fix(export): include unquantized vision layers in AWQ modules_to_not_…
99952ab5
luoyu-intel refactor(rrq): minimize common-code changes to preserve main behavior
cfafc4cc
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
52cf8a1c
luoyu-intel revert(export): restore AWQ _collect_modules_to_not_convert to main l…
da8f208f
luoyu-intel Apply batched suggestions from code review
23169c4e
luoyu-intel revert log info
f7204db5
luoyu-intel fix(rrq): raise ValueError for explicit model_free=True with RRQConfig
1b3f7da3
luoyu-intel fix(rrq): enforce disable_opt_rtn=True in RRQConfig.check_config()
3711adc5
luoyu-intel Refactor RRQ exports to sub-package entry points and add eval scripts
dc9601d2
luoyu-intel test(rrq): add lm-eval accuracy test for RRQ W2A16 quantization
2e9febaa
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
349f8124
luoyu-intel docs(README): note that RRQ does not support production-level deploym…
a0a13de6
luoyu-intel feat(rrq): hide RRQ from user-facing listings, auto-select via format
051606dc
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
1ca218a0
luoyu-intel fix(rrq): auto-select W2A16 scheme and force disable_opt_rtn for RRQ
bf772b38
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
ab09a52f
luoyu-intel refactor(rrq): use quant_method='auto-round' + packing_format='auto_r…
41526ee4
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
4746d58b
luoyu-intel fix(test): add proper category prefix to skip_ci reason
47211d71
luoyu-intel fix: normalize Conv1D weight orientation in RRQ quantizer
9030d9c4
luoyu-intel fix: allow save_quantized(format=...) to override constructor-set format
d660c1e4
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
679baf93
luoyu-intel fix(rrq): implement RTN quantization for layers outside transformer b…
c7c07e91
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
69e46f19
luoyu-intel fix(rrq): ensure base+residual are both exported via a single save_rr…
2131b6b2
luoyu-intel Potential fix for pull request finding
d333e09c
luoyu-intel fix(rrq): default active_bits to max supported by checkpoint
5ed9c9f9
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
44492379
luoyu-intel Potential fix for pull request finding
9ca5b698
luoyu-intel fix(rrq): address PR #2308 review comments
9139c25f
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
ddcc2117
pre-commit-ci[bot] [pre-commit.ci] auto fixes from pre-commit.com hooks
0a501b95
luoyu-intel docs(rrq): move RRQ entry to 2026/10 in What's New
6024cbe7
luoyu-intel fix(rrq): accept enable_neuqi in RRQConfig constructor
7296e738
luoyu-intel fix(rrq): resolve ruff F821 + RUF013 pre-commit failures
072912b2
chensuyue fix: don't let MLLM save_quantized default override resolved format
63161b66
AutoRoundBot
AutoRoundBot
azure-pipelines
luoyu-intel merge: resolve conflicts from main (PR #2308)
ef130026
luoyu-intel fix: add inplace param to save_rrq_base_model to match save_checkpoin…
7985c1b0
AutoRoundBot
AutoRoundBot
luoyu-intel test: include auto_round:rrq in public-format contract set
f9ab7def
luoyu-intel chore: remove stale backends/rrq.py file still in the index
4dfe2324
luoyu-intel Merge branch 'main' into feat/rrq-phase1
1dfbda6d
AutoRoundBot
azure-pipelines
AutoRoundBot
luoyu-intel Merge branch 'main' into feat/rrq-phase1
d99b2a26
AutoRoundBot
azure-pipelines
luoyu-intel ci: drop 24GB MIG slice from CUDA UT GPU fallback list
6c643737
AutoRoundBot
azure-pipelines
luoyu-intel docs: update RRQ What's New entry date to 2026/11
a37976ed
luoyu-intel docs: move RRQ entry above W2A16 entry (reverse-chronological order)
c24dd437
luoyu-intel revert: restore runpod_manager.py to main (CI infra change not part o…
cac231ef
AutoRoundBot
azure-pipelines
AutoRoundBot
chensuyue chensuyue added this to the 0.17.0 milestone 5 hours ago
luoyu-intel merge: resolve conflicts from main (PR #2308)
9ababfb0
luoyu-intel docs(rrq): remove outdated Phase 1/2/3 references
34a21c9f
AutoRoundBot
luoyu-intel fix(rrq): pop scheme fields from kwargs before super().__init__
ec2ccb34
AutoRoundBot
azure-pipelines

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone