From c2ead634ba832e7a09e1282c5a4b36f550f0ab58 Mon Sep 17 00:00:00 2001 From: phantoslayer21 <156617979+phantoslayer21@users.noreply.github.com> Date: Sat, 14 Nov 2026 20:00:08 -0500 Subject: [PATCH 1/2] docs: Model L is planned for 0.6.5, off by default and under test; its numerical checks are being reworked (Refs #89) Signed-off-by: phantoslayer21 <156617979+phantoslayer21@users.noreply.github.com> --- README.md | 9 +++++---- clean/l/README.md | 8 ++++++-- 2 files changed, 11 insertions(+), 6 deletions(-) diff --git a/README.md b/README.md index 2432f73f..fa54145d 100644 --- a/README.md +++ b/README.md @@ -59,11 +59,12 @@ on thin lines in reflections, blotches in Cyberpunk 2077 with path tracing); in can very occasionally come out far too bright or dark at a scene change. Full details: the release notes ([clean/installer/RELEASE-NOTES.md](clean/installer/RELEASE-NOTES.md)). -### Model L (in development) +### Model L (under test) -Model L is a larger model choice in development for 0.7.0, available from source. Preliminary numbers, Cyberpunk 2077 -at 1080p Quality on an RX 9070 XT: the network takes 2.25 ms with Model L and 1.17 ms with Model M. Speed work is -ongoing and the numbers will change. Details, test setup and records: [clean/l/SPEED.md](clean/l/SPEED.md). +Model L is a larger model choice planned for 0.6.5, off by default and still under test; until then it is available +from source. Preliminary numbers, Cyberpunk 2077 at 1080p Quality on an RX 9070 XT: the network takes 2.25 ms with +Model L and 1.17 ms with Model M. Speed work is ongoing and the numbers will change. Details, test setup and records: +[clean/l/SPEED.md](clean/l/SPEED.md). ### Something broken? diff --git a/clean/l/README.md b/clean/l/README.md index 34ad4e58..89ffd062 100644 --- a/clean/l/README.md +++ b/clean/l/README.md @@ -1,7 +1,11 @@ # Model L (issue #89) -Work in progress for a new model choice, milestone 0.7.0. Nothing here is used by the runtime, the front end or the -launcher yet. +A new model choice, planned for 0.6.5: off by default and still under test. The runtime, the front end and the +launcher use Model L when a game is set to it (`model=L`). + +Status of the numerical checks: the C = 32 block kernels and the oracle below follow Model M's block form, and both are +being reworked to the block form of the specification. Until that is done, no check here is a result against the +specification: none of them shows that Model L matches the original model. | Path | Content | |---|---| From 5f1692cda3073d0512a816e768ab7ab1edbc051e Mon Sep 17 00:00:00 2001 From: phantoslayer21 <156617979+phantoslayer21@users.noreply.github.com> Date: Sat, 14 Nov 2026 22:32:57 -0500 Subject: [PATCH 2/2] docs: README says WrongGPU's arithmetic is its own, so Model M's output is close to DLSS's, not identical (Refs #89) Signed-off-by: phantoslayer21 <156617979+phantoslayer21@users.noreply.github.com> --- README.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index fa54145d..b343cc38 100644 --- a/README.md +++ b/README.md @@ -2,8 +2,10 @@ **Model M upscaling and Ray Reconstruction on AMD Radeon RX 9000 graphics cards.** -WrongGPU runs the neural networks behind DLSS super resolution (Model M) and DLSS Ray Reconstruction on Radeon RX -9000-series cards. Its launcher sets up each of your games and gets the WrongGPU edition of OptiScaler for them. From +WrongGPU runs its own implementation of the neural networks behind DLSS super resolution (Model M) and DLSS Ray +Reconstruction on Radeon RX 9000-series cards. The networks' weights come from your own DLSS files; the arithmetic is +WrongGPU's, so the output is close to DLSS's but not identical to it. Its launcher sets up each of your games and gets +the WrongGPU edition of OptiScaler for them. From this source onward, WrongGPU's source code and its official builds are under the MIT licence (see [Licence](#licence)). **Download:** the [Releases page](https://github.com/phantoslayer21/wronggpu/releases) (latest: 0.6.0, a pre-release).