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 16, 02:00 UTC
Remove pytorch comments for outlines + compressed-tensors (#12260)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
Thomas Parnell committed
Jan 21, 2025 at 14:49 UTC
9a7c3a0042bb7f44b6c2784220cc3d99391eb4c2
1 file changed
+2
-2
requirements-common.txt
+2
-2
@@ -19,7 +19,7 @@ pillow # Required for image processing
19
prometheus-fastapi-instrumentator >= 7.0.0
20
tiktoken >= 0.6.0 # Required for DBRX tokenizer
21
lm-format-enforcer >= 0.10.9, < 0.11
22
-outlines == 0.1.11 # Requires pytorch
22
+outlines == 0.1.11
23
lark == 1.2.2
24
xgrammar >= 0.1.6; platform_machine == "x86_64"
25
typing_extensions >= 4.10
@@ -34,6 +34,6 @@ pyyaml
34
six>=1.16.0; python_version > '3.11' # transitive dependency of pandas that needs to be the latest version for python 3.12
35
setuptools>=74.1.1; python_version > '3.11' # Setuptools is used by triton, we need to ensure a modern version is installed for 3.12+ so that it does not try to import distutils, which was removed in 3.12
36
einops # Required for Qwen2-VL.
37
-compressed-tensors == 0.8.1 # required for compressed-tensors, requires pytorch
37
+compressed-tensors == 0.8.1 # required for compressed-tensors
38
depyf==0.18.0 # required for profiling and debugging with compilation config
39
cloudpickle # allows pickling lambda functions in model_executor/models/registry.py