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, 05:30 UTC
[Bugfix] Set the minimum python version for gpt-oss (#26392)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Jee Jee Li committed
Oct 9, 2025 at 11:35 UTC
0c52d6ef81d1fd45effd5e212374d34510446dd7
3 files changed
+39
-10
requirements/common.txt
-1
@@ -49,4 +49,3 @@ pybase64 # fast base64 implementation
49
cbor2 # Required for cross-language serialization of hashable objects
50
setproctitle # Used to set process names for better debugging and monitoring
51
openai-harmony >= 0.0.3 # Required for gpt-oss
52
-gpt-oss >= 0.0.7
requirements/test.in
+1
@@ -55,3 +55,4 @@ fastsafetensors>=0.1.10
55
pydantic>=2.10 # 2.9 leads to error on python 3.10
56
decord==0.6.0
57
terratorch @ git+https://github.com/IBM/terratorch.git@1.1.rc3 # required for PrithviMAE test
58
+gpt-oss >= 0.0.7; python_version > '3.11'
\ No newline at end of file
requirements/test.txt
+38
-9
@@ -10,18 +10,19 @@ aenum==3.1.16
10
# via lightly
11
affine==2.4.0
12
# via rasterio
13
-aiohappyeyeballs==2.4.3
13
+aiohappyeyeballs==2.6.1
14
# via aiohttp
15
-aiohttp==3.10.11
15
+aiohttp==3.13.0
16
# via
17
# aiohttp-cors
18
# datasets
19
# fsspec
20
+ # gpt-oss
21
# lm-eval
22
# ray
23
aiohttp-cors==0.8.1
24
# via ray
24
-aiosignal==1.3.1
25
+aiosignal==1.4.0
26
# via aiohttp
27
albucore==0.0.16
28
# via terratorch
@@ -103,6 +104,8 @@ chardet==5.2.0
104
# via mbstrdecoder
105
charset-normalizer==3.4.0
106
# via requests
107
+chz==0.3.0
108
+ # via gpt-oss
109
click==8.1.7
110
# via
111
# black
@@ -173,7 +176,9 @@ distlib==0.3.9
176
dnspython==2.7.0
177
# via email-validator
178
docker==7.1.0
176
- # via mlflow
179
+ # via
180
+ # gpt-oss
181
+ # mlflow
182
docopt==0.6.2
183
# via num2words
184
docstring-parser==0.17.0
@@ -199,7 +204,9 @@ eval-type-backport==0.2.2
204
evaluate==0.4.3
205
# via lm-eval
206
fastapi==0.116.1
202
- # via mlflow-skinny
207
+ # via
208
+ # gpt-oss
209
+ # mlflow-skinny
210
fastparquet==2024.11.0
211
# via genai-perf
212
fastrlock==0.8.2
@@ -274,6 +281,8 @@ google-resumable-media==2.7.2
281
# via google-cloud-storage
282
googleapis-common-protos==1.70.0
283
# via google-api-core
284
+gpt-oss==0.0.8
285
+ # via -r requirements/test.in
286
graphene==3.4.3
287
# via mlflow
288
graphql-core==3.2.6
@@ -301,6 +310,8 @@ hf-xet==1.1.7
310
# via huggingface-hub
311
hiredis==3.0.0
312
# via tensorizer
313
+html2text==2025.4.15
314
+ # via gpt-oss
315
httpcore==1.0.6
316
# via httpx
317
httpx==0.27.2
@@ -435,6 +446,7 @@ lm-eval @ git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b772215
446
lxml==5.3.0
447
# via
448
# blobfile
449
+ # gpt-oss
450
# sacrebleu
451
mako==1.3.10
452
# via alembic
@@ -604,6 +616,8 @@ omegaconf==2.3.0
616
# lightning
617
open-clip-torch==2.32.0
618
# via -r requirements/test.in
619
+openai-harmony==0.0.4
620
+ # via gpt-oss
621
opencensus==0.11.4
622
# via ray
623
opencensus-context==0.1.3
@@ -724,7 +738,9 @@ prometheus-client==0.22.0
738
# opentelemetry-exporter-prometheus
739
# ray
740
propcache==0.2.0
727
- # via yarl
741
+ # via
742
+ # aiohttp
743
+ # yarl
744
proto-plus==1.26.1
745
# via google-api-core
746
protobuf==5.28.3
@@ -773,10 +789,12 @@ pydantic==2.11.7
789
# albumentations
790
# datamodel-code-generator
791
# fastapi
792
+ # gpt-oss
793
# lightly
794
# mistral-common
795
# mlflow-skinny
796
# mteb
797
+ # openai-harmony
798
# pydantic-extra-types
799
# ray
800
pydantic-core==2.33.2
@@ -907,6 +925,7 @@ requests==2.32.3
925
# evaluate
926
# google-api-core
927
# google-cloud-storage
928
+ # gpt-oss
929
# huggingface-hub
930
# lightly
931
# lm-eval
@@ -1052,6 +1071,8 @@ starlette-testclient==0.4.1
1071
# via schemathesis
1072
statsmodels==0.14.4
1073
# via genai-perf
1074
+structlog==25.4.0
1075
+ # via gpt-oss
1076
sympy==1.13.3
1077
# via
1078
# einx
@@ -1064,14 +1085,17 @@ tblib==3.1.0
1085
# via -r requirements/test.in
1086
tcolorpy==0.1.6
1087
# via pytablewriter
1067
-tenacity==9.0.0
1088
+tenacity==9.1.2
1089
# via
1090
+ # gpt-oss
1091
# lm-eval
1092
# plotly
1093
tensorboardx==2.6.4
1094
# via lightning
1095
tensorizer==2.10.1
1096
# via -r requirements/test.in
1097
+termcolor==3.1.0
1098
+ # via gpt-oss
1099
terratorch @ git+https://github.com/IBM/terratorch.git@07184fcf91a1324f831ff521dd238d97fe350e3e
1100
# via -r requirements/test.in
1101
threadpoolctl==3.5.0
@@ -1080,8 +1104,9 @@ tifffile==2025.3.30
1104
# via
1105
# scikit-image
1106
# terratorch
1083
-tiktoken==0.7.0
1107
+tiktoken==0.12.0
1108
# via
1109
+ # gpt-oss
1110
# lm-eval
1111
# mistral-common
1112
timm==1.0.17
@@ -1201,8 +1226,10 @@ typeshed-client==2.8.2
1226
# via jsonargparse
1227
typing-extensions==4.12.2
1228
# via
1229
+ # aiosignal
1230
# albumentations
1231
# alembic
1232
+ # chz
1233
# fastapi
1234
# graphene
1235
# huggingface-hub
@@ -1242,7 +1269,9 @@ urllib3==2.2.3
1269
# responses
1270
# tritonclient
1271
uvicorn==0.35.0
1245
- # via mlflow-skinny
1272
+ # via
1273
+ # gpt-oss
1274
+ # mlflow-skinny
1275
vector-quantize-pytorch==1.21.2
1276
# via -r requirements/test.in
1277
virtualenv==20.31.2