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, 03:30 UTC
Add ignore consolidated file in mistral example code (#20420)
Signed-off-by: 汪志鹏 <wangzhipeng628@gmail.com>
汪志鹏 committed
Jul 4, 2025 at 10:55 UTC
25950dca9b212743b755c8f1729a124db4f7f691
2 files changed
+2
examples/offline_inference/vision_language.py
+1
@@ -677,6 +677,7 @@ def run_mistral3(questions: list[str], modality: str) -> ModelRequestData:
677
max_num_seqs=2,
678
tensor_parallel_size=2,
679
limit_mm_per_prompt={modality: 1},
680
+ ignore_patterns=["consolidated.safetensors"],
681
)
682
683
prompts = [f"<s>[INST]{question}\n[IMG][/INST]" for question in questions]
examples/offline_inference/vision_language_multi_image.py
+1
@@ -505,6 +505,7 @@ def load_mistral3(question: str, image_urls: list[str]) -> ModelRequestData:
505
max_num_seqs=2,
506
tensor_parallel_size=2,
507
limit_mm_per_prompt={"image": len(image_urls)},
508
+ ignore_patterns=["consolidated.safetensors"],
509
)
510
511
placeholders = "[IMG]" * len(image_urls)