[VLM] Add TP support for Phi-4-MM (#14453)

Signed-off-by: Isotr0py <2037008807@qq.com>
2025-03-08 21:57:14 +08:00
parent cb8bdfade2
commit 03fe18ae0f
4 changed files with 50 additions and 295 deletions
--- a/examples/offline_inference/audio_language.py
+++ b/examples/offline_inference/audio_language.py
@@ -77,6 +77,7 @@ def run_phi4mm(questions: str, audio_count: int):
        enable_lora=True,
        max_lora_rank=320,
        lora_extra_vocab_size=0,
+        limit_mm_per_prompt={"audio": audio_count},
    )
    lora_request = LoRARequest("speech", 1, speech_lora_path)
    # To maintain code compatibility in this script, we add LoRA here.