github/InvokeAI
1
1
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI.git synced 2026-04-23 03:00:31 -04:00
Code Issues Packages Projects Releases Wiki Activity

Make LLaVA Onevision node work with 0 images, and other minor improvements.

Browse Source
This commit is contained in:
Ryan Dick
2025-02-26 20:00:30 +00:00
committed by psychedelicious
parent e9714fe476
commit 41de112932
2 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
invokeai/app/invocations/llava_onevision_vllm.py
View File

@@ -14,7 +14,7 @@ from invokeai.backend.util.devices import TorchDevice
class LlavaOnevisionVllmInvocation(BaseInvocation):
"""Run a LLaVA OneVision VLLM model."""
image: list[ImageField] | ImageField | None = InputField(description="Input image.")
images: list[ImageField] | ImageField | None = InputField(default=None, description="Input image.")
prompt: str = InputField(
default="",
description="Input text prompt.",
@@ -27,10 +27,10 @@ class LlavaOnevisionVllmInvocation(BaseInvocation):
# )
def _get_images(self, context: InvocationContext) -> list[Image]:
if self.image is None:
if self.images is None:
return []
image_fields = self.image if isinstance(self.image, list) else [self.image]
image_fields = self.images if isinstance(self.images, list) else [self.images]
return [context.images.get_pil(image_field.image_name, "RGB") for image_field in image_fields]
@torch.no_grad()

4
invokeai/backend/llava_onevision_model.py
View File

@@ -38,8 +38,8 @@ class LlavaOnevisionModel(RawModel):
conversation = [{"role": "user", "content": content}]
prompt = self._processor.apply_chat_template(conversation, add_generation_prompt=True)
inputs = self._processor(images=images, text=prompt, return_tensors="pt").to(device=device, dtype=dtype)
output = self._vllm_model.generate(**inputs, max_new_tokens=200, do_sample=False)
inputs = self._processor(images=images or None, text=prompt, return_tensors="pt").to(device=device, dtype=dtype)
output = self._vllm_model.generate(**inputs, max_new_tokens=400, do_sample=False)
output_str: str = self._processor.decode(output[0][2:], skip_special_tokens=True)
# The output_str will include the prompt, so we extract the response.
response = output_str.split("assistant\n", 1)[1].strip()