Improve model card metadata and description

Hi! I'm Niels from the Hugging Face team.

I've noticed this model is part of the impressive Toto 2.0 family. I'm opening this PR to further improve the model card metadata and documentation:
- Added `library_name: pytorch` to enable automated code snippet features.
- Added `datasets: Datadog/BOOM` to the metadata to link the model with its evaluation dataset.
- Added the `u-mup` tag to highlight the scaling technology used.
- Included the full list of authors in the model description.
- Ensured the paper and GitHub repository are clearly linked.

The performance metrics already present in the `model-index` are preserved.

README.md

@@ -1,4 +1,9 @@
 ---
+license: apache-2.0
+pipeline_tag: time-series-forecasting
+library_name: pytorch
+datasets:
+- Datadog/BOOM
 tags:
 - time-series-forecasting
 - foundation-models
@@ -9,64 +14,65 @@ tags:
 - observability
 - safetensors
 - pytorch_model_hub_mixin
-license: apache-2.0
-pipeline_tag: time-series-forecasting
+- u-mup
 thumbnail: https://web-assets.dd-static.net/42588/1778691695-toto-2-hero.png
 model-index:
 - name: Toto-2.0-313m
   results:
-    - task:
-        type: time-series-forecasting
-      dataset:
-        name: BOOM
-        type: BOOM
-      metrics:
-        - name: CRPS
-          type: CRPS
-          value: 0.351
-        - name: MASE
-          type: MASE
-          value: 0.585
-      source:
-        name: BOOM 💥 Observability Time-Series Forecasting Leaderboard
-        url: https://huggingface.co/spaces/Datadog/BOOM
-    - task:
-        type: time-series-forecasting
-      dataset:
-        name: GIFT-Eval
-        type: GIFT-Eval
-      metrics:
-        - name: CRPS
-          type: CRPS
-          value: 0.481
-        - name: MASE
-          type: MASE
-          value: 0.703
-      source:
-        name: GIFT-Eval Time Series Forecasting Leaderboard
-        url: https://huggingface.co/spaces/Salesforce/GIFT-Eval
-    - task:
-        type: time-series-forecasting
-      dataset:
-        name: TIME
-        type: TIME
-      metrics:
-        - name: CRPS
-          type: CRPS
-          value: 0.535
-        - name: MASE
-          type: MASE
-          value: 0.642
-      source:
-        name: TIME Benchmark Leaderboard
-        url: https://huggingface.co/spaces/Real-TSF/TIME-leaderboard
+  - task:
+      type: time-series-forecasting
+    dataset:
+      name: BOOM
+      type: BOOM
+    metrics:
+    - type: CRPS
+      value: 0.351
+      name: CRPS
+    - type: MASE
+      value: 0.585
+      name: MASE
+    source:
+      url: https://huggingface.co/spaces/Datadog/BOOM
+      name: BOOM 💥 Observability Time-Series Forecasting Leaderboard
+  - task:
+      type: time-series-forecasting
+    dataset:
+      name: GIFT-Eval
+      type: GIFT-Eval
+    metrics:
+    - type: CRPS
+      value: 0.481
+      name: CRPS
+    - type: MASE
+      value: 0.703
+      name: MASE
+    source:
+      url: https://huggingface.co/spaces/Salesforce/GIFT-Eval
+      name: GIFT-Eval Time Series Forecasting Leaderboard
+  - task:
+      type: time-series-forecasting
+    dataset:
+      name: TIME
+      type: TIME
+    metrics:
+    - type: CRPS
+      value: 0.535
+      name: CRPS
+    - type: MASE
+      value: 0.642
+      name: MASE
+    source:
+      url: https://huggingface.co/spaces/Real-TSF/TIME-leaderboard
+      name: TIME Benchmark Leaderboard
 ---
 
 # Toto-2.0-313m
 
-Toto (Time Series Optimized Transformer for [Observability](https://www.datadoghq.com/knowledge-center/observability/)) is a family of time series foundation models for multivariate forecasting developed by [Datadog](https://www.datadoghq.com/). Toto 2.0 is the current generation, featuring u-μP-scaled transformers ranging from 4m to 2.5B parameters, all trained from a single recipe. Forecast quality improves reliably with parameter count across the family.
+Toto (Time Series Optimized Transformer for [Observability](https://www.datadoghq.com/knowledge-center/observability/)) is a family of time series foundation models for multivariate forecasting developed by [Datadog](https://www.datadoghq.com/). 
+
+This model, **Toto-2.0-313m**, was presented in the paper [Toto 2.0: Time Series Forecasting Enters the Scaling Era](https://huggingface.co/papers/2605.20119) by Emaad Khwaja, Chris Lettieri, Gerald Woo, Eden Belouadah, Marc Cenac, Guillaume Jarry, Enguerrand Paquin, Xunyi Zhao, Viktoriya Zhukov, Othmane Abou-Amal, Chenghao Liu, Ameet Talwalkar, and David Asker.
 
-The family sets a new state of the art on three forecasting benchmarks: [BOOM](https://huggingface.co/spaces/Datadog/BOOM), our observability benchmark; [GIFT-Eval](https://huggingface.co/spaces/Salesforce/GIFT-Eval), the standard general-purpose benchmark; and the recent contamination-resistant [TIME](https://arxiv.org/abs/2602.12147) benchmark.
+Toto 2.0 is a generation of u-μP-scaled transformers ranging from 4m to 2.5B parameters, all trained from a single recipe. Forecast quality improves reliably with parameter count across the family. The family sets a new state of the art on three forecasting benchmarks: [BOOM](https://huggingface.co/spaces/Datadog/BOOM), our observability benchmark; [GIFT-Eval](https://huggingface.co/spaces/Salesforce/GIFT-Eval), the standard general-purpose benchmark; and the recent contamination-resistant [TIME](https://arxiv.org/abs/2602.12147) benchmark.
 
 ## 📊 Performance
 
@@ -114,7 +120,7 @@ For more examples, see the [Quick Start notebook](https://github.com/DataDog/tot
 
 ## 💾 Available Checkpoints
 
-All five Toto 2.0 sizes share the same training recipe; pick a size based on your accuracy/latency budget. Latency is forward-pass time for a 1,024-step single-pass forecast at batch size 8 on a single A100.
+All five Toto 2.0 sizes share the same training recipe. Latency is forward-pass time for a 1,024-step single-pass forecast at batch size 8 on a single A100.
 
 | Model | Params | Weights (fp32) | Latency | Recommended for |
 |:---:|:---:|:---:|---|---|
@@ -136,7 +142,7 @@ All five Toto 2.0 sizes share the same training recipe; pick a size based on you
 
 <figure>
 <img src="assets/architecture.png" alt="Overview of the Toto 2.0 architecture.">
-<figcaption>A decoder-only patched transformer whose attention layers alternate between time-axis (causal) and variate-axis (full) views of the input. Toto 2.0 adds <b>contiguous patch masking (CPM)</b> for single-pass parallel decoding, a <b>quantile output head</b> trained with pinball loss, a robust arcsinh input scaler, residual MLP patch projections, and is trained with NorMuon. See the <a href="#-additional-resources">technical report</a> for details.</figcaption>
+<figcaption>A decoder-only patched transformer whose attention layers alternate between time-axis (causal) and variate-axis (full) views of the input. Toto 2.0 adds <b>contiguous patch masking (CPM)</b> for single-pass parallel decoding, a <b>quantile output head</b> trained with pinball loss, a robust arcsinh input scaler, residual MLP patch projections, and is trained with NorMuon. See the <a href="https://arxiv.org/abs/2605.20119">technical report</a> for details.</figcaption>
 </figure>
 
 ## 🔗 Additional Resources
@@ -160,4 +166,4 @@ All five Toto 2.0 sizes share the same training recipe; pick a size based on you
       primaryClass={cs.LG},
       url={https://arxiv.org/abs/2605.20119}, 
 }
-```
+```