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 15, 04:45 UTC
[TPU] Update torch-xla version to include paged attention tuned block change (#19813)
Signed-off-by: Qiliang Cui <derrhein@gmail.com>
QiliangCui committed
Jun 18, 2025 at 15:41 UTC
04fefe7c9a7947390885bac4f31f8c83a8aefd73
1 file changed
+5
-5
requirements/tpu.txt
+5
-5
@@ -18,9 +18,9 @@ setuptools==78.1.0
18
--find-links https://storage.googleapis.com/libtpu-releases/index.html
19
--find-links https://storage.googleapis.com/jax-releases/jax_nightly_releases.html
20
--find-links https://storage.googleapis.com/jax-releases/jaxlib_nightly_releases.html
21
-torch==2.8.0.dev20250617
22
-torchvision==0.23.0.dev20250617
23
-torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.8.0.dev20250617-cp39-cp39-linux_x86_64.whl ; python_version == "3.9"
24
-torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.8.0.dev20250617-cp310-cp310-linux_x86_64.whl ; python_version == "3.10"
25
-torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.8.0.dev20250617-cp311-cp311-linux_x86_64.whl ; python_version == "3.11"
21
+torch==2.8.0.dev20250618
22
+torchvision==0.23.0.dev20250618
23
+torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.8.0.dev20250618-cp39-cp39-linux_x86_64.whl ; python_version == "3.9"
24
+torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.8.0.dev20250618-cp310-cp310-linux_x86_64.whl ; python_version == "3.10"
25
+torch_xla[tpu, pallas] @ https://storage.googleapis.com/pytorch-xla-releases/wheels/tpuvm/torch_xla-2.8.0.dev20250618-cp311-cp311-linux_x86_64.whl ; python_version == "3.11"
26