Skip to content

Add univa to models #1526

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Jun 23, 2025
Merged

Add univa to models #1526

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
86ba583
add univa
merveenoyan Jun 12, 2025
230ef8f
add processor
merveenoyan Jun 12, 2025
f363509
Update packages/tasks/src/model-libraries-snippets.ts
merveenoyan Jun 20, 2025
5be783a
Update packages/tasks/src/model-libraries.ts
merveenoyan Jun 20, 2025
06e5c74
address Lucain's comment
merveenoyan Jun 20, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 11 additions & 0 deletions packages/tasks/src/model-libraries-snippets.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1408,6 +1408,17 @@ model = SwarmFormerModel.from_pretrained("${model.id}")
`,
];

export const univa = (model: ModelData): string[] => [
`from univa.models.qwen2p5vl.modeling_univa_qwen2p5vl import UnivaQwen2p5VLForConditionalGeneration
model = UnivaQwen2p5VLForConditionalGeneration.from_pretrained(
"${model.id}",
torch_dtype=torch.bfloat16,
attn_implementation="flash_attention_2",
).to("cuda")
processor = AutoProcessor.from_pretrained("${model.id}")
`,
];

const mlx_unknown = (model: ModelData): string[] => [
`# Download the model from the Hub
pip install huggingface_hub[hf_xet]
Expand Down
8 changes: 8 additions & 0 deletions packages/tasks/src/model-libraries.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1027,6 +1027,14 @@ export const MODEL_LIBRARIES_UI_ELEMENTS = {
countDownloads: `path_extension:"pt"`,
snippets: snippets.ultralytics,
},
"univa": {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

as mentioned by @Vaibhavs10 let's get the model listed under https://huggingface.co/models?other=uniworld (or https://huggingface.co/models?other=univa) before shipping this PR

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

before this PR @NielsRogge has opened this PR https://huggingface.co/LanguageBind/UniWorld-V1/discussions/3/files they haven't responded, pinged them again

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

before this PR @NielsRogge has opened this PR https://huggingface.co/LanguageBind/UniWorld-V1/discussions/3/files they haven't responded, pinged them again

Done! Thanks!

prettyLabel: "univa",
repoName: "univa",
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
"univa": {
prettyLabel: "univa",
repoName: "univa",
"uniworld": {
prettyLabel: "UniWorld",
repoName: "UniWorld",

Shouldn't the library be named UniWorld instead to follow the repo naming convention? I feel it would be more natural that way

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the library itself was called univa, i.e. import univa, the model is loaded from there. if the authors choose to train new models loaded with this code it would be more unified imo

repoUrl: "https://github.com/PKU-YuanGroup/UniWorld-V1",
snippets: snippets.univa,
filter: true,
countDownloads: `path_extension:"safetensors"`,
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks like the repo already have a config.json file so would be good to keep the counting on that file (if loaded at each .from_pretrained). This is the default rule if we don't set one.

Problem with counting path_extension:"safetensors" is that it's non-canonical + will count 17 downloads each time a user downloads the repo (given the 17 shards). (taken from files section)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

fixed this, thanks for the heads up!

},
"uni-3dar": {
prettyLabel: "Uni-3DAR",
repoName: "Uni-3DAR",
Expand Down
Loading