Clean metadata and add pipeline tag
Browse filesThis PR updates the model card metadata by:
- Removing the `arxiv:2609.33437` tag from the YAML metadata; the arXiv ID remains linked in the Markdown content.
- Adding `pipeline_tag: image-classification` so the model is categorized correctly on the Hub.
No `library_name` is added because the model is loaded with custom SMAT code rather than a standard Hugging Face library loader. The model card body is unchanged.
README.md
CHANGED
|
@@ -1,16 +1,17 @@
|
|
| 1 |
---
|
| 2 |
-
license: mit
|
| 3 |
base_model: openai/clip-vit-large-patch14
|
|
|
|
|
|
|
| 4 |
tags:
|
| 5 |
- smat
|
| 6 |
- model-merging
|
| 7 |
- vision
|
| 8 |
-
- arxiv:2609.33437
|
| 9 |
---
|
|
|
|
| 10 |
# CLIP ViT large-patch14 路 FT 路 RESISC45
|
| 11 |
|
| 12 |
A FT expert from Table 2 of **SMAT: Simple and Efficient Merge-Aware Training** (seed 42). SMAT trains experts with model merging in mind.
|
| 13 |
|
| 14 |
[Paper](https://arxiv.org/abs/2609.33437) 路 [GitHub & usage](https://github.com/egangu/smat/blob/main/docs/HUGGINGFACE.md)
|
| 15 |
|
| 16 |
-
`encoder.pt` is the original FP32 vision-encoder state dictionary; load it with the SMAT code and the matching OpenAI CLIP base model. Dataset terms apply separately.
|
|
|
|
| 1 |
---
|
|
|
|
| 2 |
base_model: openai/clip-vit-large-patch14
|
| 3 |
+
license: mit
|
| 4 |
+
pipeline_tag: image-classification
|
| 5 |
tags:
|
| 6 |
- smat
|
| 7 |
- model-merging
|
| 8 |
- vision
|
|
|
|
| 9 |
---
|
| 10 |
+
|
| 11 |
# CLIP ViT large-patch14 路 FT 路 RESISC45
|
| 12 |
|
| 13 |
A FT expert from Table 2 of **SMAT: Simple and Efficient Merge-Aware Training** (seed 42). SMAT trains experts with model merging in mind.
|
| 14 |
|
| 15 |
[Paper](https://arxiv.org/abs/2609.33437) 路 [GitHub & usage](https://github.com/egangu/smat/blob/main/docs/HUGGINGFACE.md)
|
| 16 |
|
| 17 |
+
`encoder.pt` is the original FP32 vision-encoder state dictionary; load it with the SMAT code and the matching OpenAI CLIP base model. Dataset terms apply separately.
|