@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / b422d4961a
Fork 0

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, 22:00 UTC

[CI/Build] Enable mypy typing for remaining folders (#6268)

Benjamin Muskalla committed Jul 10, 2024 at 16:15 UTC b422d4961a3052c5b4bcfc3747a1ad55acfe7eb8
3 files changed +21 -20
.github/workflows/mypy.yaml
+10 -8
@@ -32,20 +32,22 @@ jobs:
32 pip install types-setuptools
33 - name: Mypy
34 run: |
35 + mypy tests --config-file pyproject.toml
36 + mypy vllm/*.py --config-file pyproject.toml
37 mypy vllm/attention --config-file pyproject.toml
38 mypy vllm/core --config-file pyproject.toml
39 mypy vllm/distributed --config-file pyproject.toml
40 + mypy vllm/engine --config-file pyproject.toml
41 mypy vllm/entrypoints --config-file pyproject.toml
42 mypy vllm/executor --config-file pyproject.toml
43 + mypy vllm/inputs --config-file pyproject.toml
44 + mypy vllm/logging --config-file pyproject.toml
45 + mypy vllm/lora --config-file pyproject.toml
46 + mypy vllm/model_executor --config-file pyproject.toml
47 mypy vllm/multimodal --config-file pyproject.toml
41 - mypy vllm/usage --config-file pyproject.toml
42 - mypy vllm/*.py --config-file pyproject.toml
48 + mypy vllm/platforms --config-file pyproject.toml
49 + mypy vllm/spec_decode --config-file pyproject.toml
50 mypy vllm/transformers_utils --config-file pyproject.toml
44 - mypy vllm/engine --config-file pyproject.toml
51 + mypy vllm/usage --config-file pyproject.toml
52 mypy vllm/worker --config-file pyproject.toml
46 - mypy vllm/spec_decode --config-file pyproject.toml
47 - mypy vllm/model_executor --config-file pyproject.toml
48 - mypy vllm/lora --config-file pyproject.toml
49 - mypy vllm/logging --config-file pyproject.toml
50 - mypy tests --config-file pyproject.toml
53
format.sh
+9 -9
@@ -96,23 +96,23 @@ echo 'vLLM yapf: Done'
96
97 # Run mypy
98 echo 'vLLM mypy:'
99 +mypy tests --config-file pyproject.toml
100 +mypy vllm/*.py --config-file pyproject.toml
101 mypy vllm/attention --config-file pyproject.toml
102 mypy vllm/core --config-file pyproject.toml
103 mypy vllm/distributed --config-file pyproject.toml
104 +mypy vllm/engine --config-file pyproject.toml
105 mypy vllm/entrypoints --config-file pyproject.toml
106 mypy vllm/executor --config-file pyproject.toml
107 +mypy vllm/logging --config-file pyproject.toml
108 +mypy vllm/lora --config-file pyproject.toml
109 +mypy vllm/model_executor --config-file pyproject.toml
110 mypy vllm/multimodal --config-file pyproject.toml
105 -mypy vllm/usage --config-file pyproject.toml
106 -mypy vllm/*.py --config-file pyproject.toml
111 +mypy vllm/prompt_adapter --config-file pyproject.toml
112 +mypy vllm/spec_decode --config-file pyproject.toml
113 mypy vllm/transformers_utils --config-file pyproject.toml
108 -mypy vllm/engine --config-file pyproject.toml
114 +mypy vllm/usage --config-file pyproject.toml
115 mypy vllm/worker --config-file pyproject.toml
110 -mypy vllm/spec_decode --config-file pyproject.toml
111 -mypy vllm/model_executor --config-file pyproject.toml
112 -mypy vllm/lora --config-file pyproject.toml
113 -mypy vllm/logging --config-file pyproject.toml
114 -mypy vllm/prompt_adapter --config-file pyproject.toml
115 -mypy tests --config-file pyproject.toml
116
117
118 # If git diff returns a file that is in the skip list, the file may be checked anyway:
vllm/platforms/cuda.py
+2 -3
@@ -34,11 +34,10 @@ def get_physical_device_capability(device_id: int = 0) -> Tuple[int, int]:
34 def device_id_to_physical_device_id(device_id: int) -> int:
35 if "CUDA_VISIBLE_DEVICES" in os.environ:
36 device_ids = os.environ["CUDA_VISIBLE_DEVICES"].split(",")
37 - device_ids = [int(device_id) for device_id in device_ids]
37 physical_device_id = device_ids[device_id]
38 + return int(physical_device_id)
39 else:
40 - physical_device_id = device_id
41 - return physical_device_id
40 + return device_id
41
42
43 class CudaPlatform(Platform):