biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[VLM] Fully dynamic prompt replacement in merged input processor (#11199)

Browse Source 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2024-12-15 01:52:18 +08:00
committed by GitHub
parent 9c3dadd1c9
commit 93abf23a64
12 changed files with 565 additions and 506 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/plugins/vllm_add_dummy_model/vllm_add_dummy_model/my_llava.py
View File

@@ -3,14 +3,14 @@ from typing import Optional
import torch
from vllm.model_executor.models.llava import (LlavaForConditionalGeneration,
LlavaProcessor,
LlavaMultiModalProcessor,
get_max_llava_image_tokens)
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.multimodal import MULTIMODAL_REGISTRY
@MULTIMODAL_REGISTRY.register_max_image_tokens(get_max_llava_image_tokens)
@MULTIMODAL_REGISTRY.register_processor(LlavaProcessor)
@MULTIMODAL_REGISTRY.register_processor(LlavaMultiModalProcessor)
class MyLlava(LlavaForConditionalGeneration):
def compute_logits(