Skip to content

Commit b6127bf

Browse files
committed
update description of versions for deepcompile
1 parent 93ebac3 commit b6127bf

File tree

1 file changed

+10
-1
lines changed

1 file changed

+10
-1
lines changed

benchmarks/deepcompile/README.md

+10-1
Original file line numberDiff line numberDiff line change
@@ -10,7 +10,10 @@ We tested the scripts with Python 3.10.12 and CUDA 12.4.
1010
In addition, you need to install the following:
1111

1212
- PyTorch v2.6.0
13-
- DeepSpeed (v0.16.6 or newer)
13+
- For the Mixtral model, we recommend using PyTorch v2.7.0 (currently the final release candidate) due to an issue with `torch.where`.
14+
- See details [here](https://github.com/pytorch/pytorch/issues/149278)
15+
- DeepSpeed (v.0.16.6 or newer)
16+
- As DeepCompile is under active development, we recommend using the latest version or installing from source.
1417
- transformers
1518
- accelerate
1619
- datasets v3.1
@@ -24,6 +27,12 @@ pip3 install transformers datasets==3.1 accelerate
2427
# Install DeepSpeed
2528
pip install deepspeed
2629

30+
# Or install the latest revision of DeepSpeed from source
31+
# git clone https://github.com/deepspeedai/DeepSpeed
32+
# cd DeepSpeed
33+
# git pip install .
34+
# cd ..
35+
2736
# Clone this repository
2837
git clone https://github.com/deepspeedai/DeepSpeedExamples
2938
cd benchmarks/deepcompile

0 commit comments

Comments
 (0)