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, 16:00 UTC
Add minimum version for `huggingface_hub` to enable Xet downloads (#15873)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
Apr 2, 2025 at 10:03 UTC
51826d51fa6ef36963ddd79e99dc77c7660ffbf5
3 files changed
+6
-1
requirements/common.txt
+1
@@ -7,6 +7,7 @@ tqdm
7
blake3
8
py-cpuinfo
9
transformers >= 4.50.3
10
+huggingface-hub[hf-xet] >= 0.30.0 # Required for Xet downloads.
11
tokenizers >= 0.19.1 # Required for Llama 3.
12
protobuf # Required by LlamaTokenizer.
13
fastapi[standard] >= 0.115.0 # Required by FastAPI's form models in the OpenAI API server's audio transcriptions endpoint.
requirements/test.in
+1
@@ -31,6 +31,7 @@ opencv-python-headless >= 4.11.0 # required for video test
31
datamodel_code_generator # required for minicpm3 test
32
lm-eval[api]==0.4.8 # required for model evaluation test
33
transformers==4.50.3
34
+huggingface-hub[hf-xet]>=0.30.0 # Required for Xet downloads.
35
# quantization
36
bitsandbytes>=0.45.3
37
buildkite-test-collector==0.1.9
requirements/test.txt
+4
-1
@@ -152,14 +152,17 @@ genson==1.3.0
152
# via datamodel-code-generator
153
h11==0.14.0
154
# via httpcore
155
+hf-xet==0.1.4
156
+ # via huggingface-hub
157
hiredis==3.0.0
158
# via tensorizer
159
httpcore==1.0.6
160
# via httpx
161
httpx==0.27.2
162
# via -r requirements/test.in
161
-huggingface-hub==0.26.2
163
+huggingface-hub==0.30.1
164
# via
165
+ # -r requirements/test.in
166
# accelerate
167
# datasets
168
# evaluate