You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/src/content/docs/guides/python/getting-started.mdx
+3-3Lines changed: 3 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -38,11 +38,11 @@ pip install mistralrs
38
38
39
39
### NVIDIA GPU (CUDA)
40
40
41
-
CUDA wheels are published as GitHub release assets, one per compute capability (PyPI has no GPU dimension). Install with `--find-links` pointed at the release and select your GPU's compute capability via the `+smNN` version (replace `0.8.11` / `v0.8.11` with the release you want):
41
+
CUDA wheels are published as GitHub release assets, one per compute capability (PyPI has no GPU dimension). Install with `--find-links` pointed at the release and select your GPU's compute capability via the `+smNN` version (replace `0.8.12` / `v0.8.12` with the release you want):
Look up your GPU's compute capability in [hardware support](/mistral.rs/reference/hardware-support/), which lists the published wheels per architecture. The wheels bundle the CUDA runtime, so no system toolkit is needed; they use the CUTLASS MoE backend (for the faster cuTile path, use the [prebuilt binary](/mistral.rs/quickstart/)).
Copy file name to clipboardExpand all lines: docs/src/content/docs/quickstart.mdx
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -68,7 +68,7 @@ Prebuilt CUDA binaries are published per compute capability for Ampere and newer
68
68
69
69
Install-time environment variables:
70
70
71
-
-`MISTRALRS_INSTALL_TAG=v0.8.11` installs a specific release instead of the latest (downloads that release's prebuilt, or builds that git tag from source).
71
+
-`MISTRALRS_INSTALL_TAG=v0.8.12` installs a specific release instead of the latest (downloads that release's prebuilt, or builds that git tag from source).
72
72
-`MISTRALRS_INSTALL_FROM_SOURCE=1` skips the prebuilt download and builds the latest `master` (bleeding edge) from source. The prebuilt path tracks the latest stable release.
73
73
-`MISTRALRS_INSTALL_NO_NCCL=1` (source builds) skips the `nccl` feature.
Copy file name to clipboardExpand all lines: docs/src/content/docs/reference/environment-variables.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -87,7 +87,7 @@ These are read by build scripts, not at runtime.
87
87
|---|---|
88
88
|`MISTRALRS_METAL_PRECOMPILE`|`MISTRALRS_METAL_PRECOMPILE=0` skips Metal kernel precompilation at build time; kernels are compiled at runtime on first use. |
89
89
|`CUDA_NVCC_FLAGS`| Extra compiler options passed to CUDA builds. |
90
-
|`MISTRALRS_INSTALL_TAG`| Pins the installers to a specific release tag (e.g. `v0.8.11`): the prebuilt is downloaded from that release, and a source build checks out that git tag. Default is the latest stable release (prebuilt) or latest `master` (source). |
90
+
|`MISTRALRS_INSTALL_TAG`| Pins the installers to a specific release tag (e.g. `v0.8.12`): the prebuilt is downloaded from that release, and a source build checks out that git tag. Default is the latest stable release (prebuilt) or latest `master` (source). |
91
91
|`MISTRALRS_INSTALL_FROM_SOURCE`|`MISTRALRS_INSTALL_FROM_SOURCE=1` makes the shell and PowerShell installers skip the prebuilt download and build from the latest `master` (bleeding edge) instead of the latest stable release. |
92
92
|`MISTRALRS_INSTALL_NCCL`|`MISTRALRS_INSTALL_NCCL=1` forces the shell and PowerShell installers to add the `nccl` feature for CUDA builds even if NCCL is not detected. |
93
93
|`MISTRALRS_INSTALL_NO_NCCL`|`MISTRALRS_INSTALL_NO_NCCL=1` makes the shell and PowerShell installers skip the `nccl` feature. |
0 commit comments