Update model card metadata and add Hugging Face paper link

This PR makes the following improvements to the model card:

- Removes the `arxiv:2610.18877` tag from the YAML metadata, because arXiv IDs should only be referenced in the Markdown content.
- Changes the `pipeline_tag` to `text-generation`, since LOTUS is a language model that outputs textual answers and rationales.
- Adds a link to the Hugging Face paper page next to the existing arXiv link.
- Keeps the existing informative content, usage examples, and evaluation results unchanged.

README.md

@@ -1,16 +1,15 @@
 ---
+base_model:
+- google/gemma-3-270m
+- google/gemma-3-1b-pt
+- Qwen/Qwen3-0.6B
 license: mit
+pipeline_tag: text-generation
 tags:
 - time-series
 - time-series-language-model
 - forecast-grounding
 - lora
-- arxiv:2610.18877
-base_model:
-- google/gemma-3-270m
-- google/gemma-3-1b-pt
-- Qwen/Qwen3-0.6B
-pipeline_tag: time-series-forecasting
 ---
 
 # LOTUS: Learning to Understand Time-series by Learning to Forecast
@@ -37,7 +36,7 @@ best average performance in every matched-backbone group with minimal trainable
 baseline by **3.36%** — and its 270M-parameter model is the top system on the TSRBench zero-shot
 reasoning benchmark ahead of backbones 25–50× larger.
 
-Paper: [arXiv:2610.18877](https://arxiv.org/abs/2610.18877) · Code: [DA2I2-SLM/Lotus](https://github.com/DA2I2-SLM/Lotus)
+Paper: [arXiv:2610.18877](https://arxiv.org/abs/2610.18877) · [Hugging Face Paper page](https://huggingface.co/papers/2610.18877) · Code: [DA2I2-SLM/Lotus](https://github.com/DA2I2-SLM/Lotus)
 · Phase-1 (pretraining) checkpoints: [hiepnh137/LangTimeFoundationModel](https://huggingface.co/hiepnh137/LangTimeFoundationModel)
 
 ## Training
@@ -228,4 +227,4 @@ split, printing accuracy next to the paper's number together with the model's ve
   journal = {arXiv preprint arXiv:2610.18877},
   year    = {2026}
 }
-```
+```