Skip to content
This repository was archived by the owner on May 21, 2025. It is now read-only.

Commit f7638ba

Browse files
add disclaimer that docker setup needs fixing, but other functionality works
1 parent a3dd409 commit f7638ba

File tree

1 file changed

+10
-4
lines changed

1 file changed

+10
-4
lines changed

README.md

Lines changed: 10 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -5,10 +5,7 @@ Registration for Intel optimized bitsandbytes operators.
55
## Quick Start
66

77
```
8-
# Build and enter container
9-
docker compose run --build --rm bnb-intel-dev /bin/bash
10-
11-
# Run validation (inside container):
8+
# Run validation:
129
python -m bitsandbytes_intel
1310
```
1411

@@ -48,6 +45,15 @@ int8_linear_matmul_xpu called with tensors of shape: torch.Size([32, 64]) torch.
4845
new kernel: registered at /build/intel-pytorch-extension/build/Release/csrc/gpu/csrc/aten/generated/ATen/RegisterXPU.cpp:468 (function operator())
4946
```
5047

48+
# Docker
49+
50+
**TODO**: the docker compose setup needs debugging, as we got everything to work outside the container, but in the container it's weirdly still looking for `bnb_intel` as a torch extension. Sth to fix for another day.
51+
52+
```
53+
# Build and enter container
54+
docker compose run --build --rm bnb-intel-dev /bin/bash
55+
```
56+
5157
## Technical Implementation
5258

5359
Key files:

0 commit comments

Comments
 (0)