A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 22:35 UTC
Update torch_xla pin to 20250730 (#21956)
Signed-off-by: Xiongfei Wei <isaacwxf23@gmail.com>
XiongfeiWei committed
Jul 31, 2025 at 10:26 UTC
53c21e492e0acd140a9984c8ec7cc3a7123efee5
2 files changed
+5
-5
docker/Dockerfile.tpu
+1
-1
@@ -1,4 +1,4 @@
1
-ARG NIGHTLY_DATE="20250724"
1
+ARG NIGHTLY_DATE="20250730"
2
ARG BASE_IMAGE="us-central1-docker.pkg.dev/tpu-pytorch-releases/docker/xla:nightly_3.12_tpuvm_$NIGHTLY_DATE"
3
4
FROM $BASE_IMAGE
requirements/tpu.txt
+4
-4
@@ -19,8 +19,8 @@ nixl==0.3.0
19
--find-links https://storage.googleapis.com/libtpu-releases/index.html
20
--find-links https://storage.googleapis.com/jax-releases/jax_nightly_releases.html
21
--find-links https://storage.googleapis.com/jax-releases/jaxlib_nightly_releases.html
22
-torch==2.9.0.dev20250724
23
-torchvision==0.24.0.dev20250724
24
-torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.9.0.dev20250724-cp311-cp311-linux_x86_64.whl ; python_version == "3.11"
25
-torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.9.0.dev20250724-cp312-cp312-linux_x86_64.whl ; python_version == "3.12"
22
+torch==2.9.0.dev20250730
23
+torchvision==0.24.0.dev20250730
24
+torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.9.0.dev20250730-cp311-cp311-linux_x86_64.whl ; python_version == "3.11"
25
+torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.9.0.dev20250730-cp312-cp312-linux_x86_64.whl ; python_version == "3.12"
26