Model L: propagate a NaN through enc0's ReLU + FP8 step as the original does; converter check for non-finite weight codes
メンテナーはふだん 1 日以内に返信
まだ誰も着手していません。
評価
- 難易度
- 5/5
- 見積もり時間
- 1週間以上
- 初心者へのやさしさ
- 35/100
- issue の種類
- バグ
- 明瞭さ
- 明確に書かれている
- 活発さ
- 活発
調査の方向性
Start with clean/l/reference/oracle.py and its _out_round relu_q8 mode, then compare the listed HIP and D3D12 kernels and the encoder block paths. Trace how the NaN-input case reaches the second linear and pack12 coefficients, and inspect clean/l/runner/l_hyb_stage_check.py for the separately reported leg. Done when the NaN cases match the oracle across implementations and the converter rejects non-finite weight codes with the stage and blob named.
索引モデルが issue の本文から書いたものです。
説明
Problem
Model L's first input linear ends in a ReLU and an FP8 conversion: x1 = q8(max(0, RN16(b + W q))). When the
accumulator is NaN (a NaN input code in q, or a NaN weight code), the original keeps the NaN through the fused
ReLU + FP8 conversion. Our kernels give 0: max(0, NaN) returns 0 on the hardware and in the HIP code.
This affects every implementation of that step alike:
- main's kernels:
clean/l/kernels/l_ops.hip(linear_32_32<RELU = true>) and its D3D12 ports in
clean/dx12/kernels/l_ops_linear.hlsl(l_linear_32_32_relu,l_linear_32_32_fused, via
clean/dx12/hlsl/l_lin_px.hlsli); - the encoder block with the linears as its prologue (
clean/dx12/kernels/swin_block.hlslSW_SRC == 3,
clean/dx12/kernels/fused_swin.hlsl); - the hybrid twins of the same step (D3D12 and HIP).
The reference oracle (clean/l/reference/oracle.py, _out_round mode relu_q8) keeps the NaN, so the hybrid stage
check's random NaN-input cases differ from it (the second linear's chain and the pack12 coefficients downstream).
Finite inputs are unaffected.
Spec references: L P3 item 13; spec 05 edge cases. Spec 03 maps NaN inputs to 0, and real converted models contain no
FP8 NaN weight codes, so this is not visible on real data. It was kept out of 0.6.5.
Until it is fixed, the random-weight legs of clean/l/runner/l_hyb_stage_check.py draw finite codes only, and one
dedicated NaN-input leg is reported separately (listed, neither pass nor fail).
Proposed fix
- In every implementation listed above, make the ReLU + FP8 step keep a NaN as the original does (one form shared by
main's kernels, the hybrid twins and HIP, so the byte checks between them stay identical). - Fold the dedicated NaN-input leg back into the pass/fail random legs of
l_hyb_stage_check.pyonce the kernels
match the oracle.
Converter validation
Add a check to the model converter's validation: a real converted model must contain no non-finite weight codes (no FP8
NaN codes, no FP16 NaN or infinity in the converted blobs). A model that fails is refused with one line naming the stage
and blob.
Tests
l_hyb_stage_check.py: the NaN-input leg passes under the normal rule (then merged into the random legs).- The byte checks (HIP == D3D12) for the linears and the fused encoder block stay identical, NaN cases included.
- The converter check: one synthetic model with a NaN weight code is refused; the real models pass.
- 主要言語
- C++
- スター
- 12
- フォーク
- 2
- 平均マージ
- 3時間 10分
- マージ済み PR(30日)
- 171
環境構築
- Dockerfile・Docker Compose ファイルなし
- プルリクエストのテンプレートあり
- コントリビューションガイドを読む
はじめの一歩
- issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
- 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
- リポジトリをフォークし、ブランチを切って変更します。
- issue 番号を参照したプルリクエストを送ります。
phantoslayer21/wronggpu のほかの issue
-
難易度 2/5 1〜3時間 初心者へのやさしさ 88/100
phantoslayer21/wronggpu#275 ·
メンテナーはふだん 1 日以内に返信
-
難易度 2/5 1〜3時間 初心者へのやさしさ 82/100
phantoslayer21/wronggpu#274 ·
メンテナーはふだん 1 日以内に返信
-
bug
難易度 2/5 1〜3時間 初心者へのやさしさ 88/100
phantoslayer21/wronggpu#241 ·
メンテナーはふだん 1 日以内に返信
-
bug build
難易度 2/5 1〜3時間 初心者へのやさしさ 74/100
phantoslayer21/wronggpu#234 ·
メンテナーはふだん 1 日以内に返信
-
bug build
難易度 2/5 1〜3時間 初心者へのやさしさ 80/100
phantoslayer21/wronggpu#233 ·
メンテナーはふだん 1 日以内に返信
phantoslayer21/wronggpu の issue をすべて見る
似ている issue
-
難易度 1/5 1〜3時間 初心者へのやさしさ 77/100
メンテナーはふだん 1 日以内に返信
-
bug good first issue
難易度 2/5 1〜3時間 初心者へのやさしさ 75/100
ReyeMe/SaturnRingLib#155 ·
-
bug
難易度 2/5 1〜3時間 初心者へのやさしさ 62/100
Nheko-Reborn/nheko#2092 ·
-
難易度 1/5 1〜3時間 初心者へのやさしさ 78/100
isocpp/CppCoreGuidelines#2345 ·
-
難易度 2/5 1〜3時間 初心者へのやさしさ 71/100
ROCm/rocm-libraries#16108 ·
メンテナーはふだん 2 日以内に返信