[Bugfix][Model] Add base class for vision-language models #4809
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
DarkLight1337
changed the title
ClassVar to indicate vision modelsClassVar to indicate vision models and improve error handling when incorrect image_feature_size is passed
DarkLight1337
changed the title
ClassVar to indicate vision models and improve error handling when incorrect image_feature_size is passed
DarkLight1337
changed the title
ywang96
reviewed
May 15, 2024
ywang96
reviewed
May 15, 2024
rkooo567
reviewed
May 16, 2024
| @@ -172,7 +174,7 @@ def forward( | |||
| image_features = image_input | |||
| vision_embeddings = self.multi_modal_projector(image_features) | |||
| inputs_embeds = self.language_model.get_input_embeddings(input_ids) | |||
| _merge_vision_embeddings( | |||
| inputs_embeds = _merge_vision_embeddings( | |||
There was a problem hiding this comment.
No, it is just to make explicit the fact that inputs_embeds is modified.
|
@simon-mo The |
rkooo567
approved these changes
May 17, 2024
|
@DarkLight1337 @ywang96 @rkooo567 Is this PR ready for merge? |
Yes. |
robertgshaw2-neuralmagic
pushed a commit
to neuralmagic/nm-vllm
that referenced
this pull request
May 19, 2024
dtrifiro
pushed a commit
to dtrifiro/vllm
that referenced
this pull request
May 21, 2024
Temirulan
pushed a commit
to Temirulan/vllm-whisper
that referenced
this pull request
Sep 6, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds a base class
VLMBaseto avoid importingLlavaForConditionalGenerationinvllm/model_executor/model_loader/loader.py, thus solving #4807.Along the way, I have also ported the improved error handling logic regarding
image_feature_sizefor LLaVA model.FIX #4807