docs: restore overwritten README

README.md

@@ -1,5 +1,5 @@
 ---
-license: cc-by-nc-nd-4.0
+license: cc-by-nc-4.0
 tags:
   - time series
   - forecasting
@@ -15,24 +15,14 @@ library_name: yug
 
 # Yug
 
-**Yug** is a 271.8M-parameter time series foundation model developed by
-Birla AI Labs for **zero-shot probabilistic forecasting**. A single checkpoint
-forecasts a previously unseen series without task-specific training or
-per-series tuning: given a historical context and a forecast horizon, it
-returns a predictive distribution summarised by quantile levels. It supports
-**univariate**, **multivariate**, and **covariate-informed** forecasting
-within a single architecture, and decodes a whole block of future points per
-forward pass rather than one step at a time.
+**Yug** is a 271.8M-parameter time series foundation model developed by Birla AI Labs for **zero-shot probabilistic forecasting**. A single checkpoint forecasts a previously unseen series without task-specific training or per-series tuning: given a historical context and a forecast horizon, it returns a predictive distribution summarised by quantile levels. It supports **univariate** and **covariate-informed** forecasting within a single architecture.
 
-<!-- TODO(release): one or two sentences on the architecture family / what it's
-     inspired by, once that's confirmed for public description (mirrors the
-     "Inspired by the T5 encoder" line in comparable model cards). -->
 
 ## Links
 
-- 💻 [GitHub](https://github.com/birla-ai-labs/model-release)
-- 📓 [Quickstart notebook](https://github.com/birla-ai-labs/model-release/blob/main/notebooks/quickstart.ipynb)
-- 🐛 [Issues and questions](https://github.com/birla-ai-labs/model-release/issues)
+- 💻 [GitHub](https://github.com/birla-ai-labs/yug)
+- 📓 [Quickstart notebook](https://github.com/birla-ai-labs/yug/blob/main/notebooks/quickstart.ipynb)
+- 🐛 [Issues and questions](https://github.com/birla-ai-labs/yug/issues)
 <!-- TODO(release): add a technical report / paper link here once public. -->
 
 ## Overview
@@ -40,15 +30,11 @@ forward pass rather than one step at a time.
 | Capability | Yug |
 |---|---|
 | Univariate forecasting | ✅ |
-| Cross-learning across items | ✅ |
-| Multivariate forecasting | ✅ |
 | Covariates | ✅ |
 | Probabilistic (quantile) output | ✅ |
-| Max context length | 8192 |
+| Max context length | 2048 |
 | Quantile levels | 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 |
 
-Quantile levels are fixed at training time: this checkpoint cannot emit a
-level it was not trained on.
 
 ## Model architecture
 
@@ -95,31 +81,32 @@ pipeline = YugPipeline.from_pretrained(
     "birlaailabs/yug", device_map="cuda"
 )
 
-# Synthetic daily signal: trend + weekly/annual seasonality + noise
-HORIZON = 48
-rng = np.random.default_rng(0)
-t = np.arange(1000)
+# Synthetic signal: trend + 64-step season + 12-step season + noise
+CONTEXT, HORIZON, PERIOD = 2000, 256, 64
+rng = np.random.default_rng([20240902, 315595964, 0])
+t = np.arange(CONTEXT + HORIZON)
+trend = np.linspace(0.0, rng.uniform(50, 150), len(t))
+phase1, amp1 = rng.uniform(0, 6), rng.uniform(10, 30)
+phase2, amp2 = rng.uniform(0, 6), rng.uniform(2, 8)
 signal = (
-    100.0                                    # level
-    + 0.05 * t                               # trend
-    + 12.0 * np.sin(2 * np.pi * t / 7)      # weekly cycle
-    + 25.0 * np.sin(2 * np.pi * t / 365)    # annual cycle
-    + rng.normal(0, 3.0, len(t))             # noise
+    trend                                                 # trend
+    + amp1 * np.sin(2 * np.pi * t / PERIOD + phase1)      # 64-step season
+    + amp2 * np.sin(2 * np.pi * t / 12 + phase2)          # 12-step season
+    + rng.normal(0, 1.0, len(t))                          # noise
 ).astype(np.float32)
 
-context, truth = signal[:-HORIZON], signal[-HORIZON:]
+context, truth = signal[:CONTEXT], signal[CONTEXT:]
 
 forecast = pipeline.predict(
     context,
     prediction_length=HORIZON,
     freq="D",            # a model input, not metadata: pandas offset alias
-    num_samples=100,
     seed=0,              # makes the call reproducible
 )
 
-print(forecast.median)                # (48,)      point forecast
-print(forecast.quantile(0.9))         # (48,)
-print(forecast.interval())            # {"lower": (48,), "upper": (48,)}
+print(forecast.median)                # (256,)     point forecast
+print(forecast.quantile(0.9))         # (256,)
+print(forecast.interval())            # {"lower": (256,), "upper": (256,)}
 print(forecast.to_dataframe().head())
 ```
 
@@ -168,23 +155,21 @@ If you find Yug useful for your research, please consider citing it:
 
 ```bibtex
 @software{yug_2026,
-  title  = {Yug: a foundation model for probabilistic time-series forecasting},
-  author = {Aaditya Jain*, Debdeep Sanyal*, Aaryan Nagpal, Dhruv Kumar, Murari Mandal, Saurabh Deshpande},
+  title  = {Yug: a foundation model for zero-shot probabilistic time-series forecasting},
+  author = {Aaditya Jain* and Debdeep Sanyal* and Aaryan Nagpal and Dhruv Kumar and Murari Mandal and Saurabh Deshpande},
   year   = {2026},
-  url    = {https://github.com/birla-ai-labs/model-release},
+  organization = {Birla AI Labs},
+  url    = {https://github.com/birla-ai-labs/yug}
 }
 ```
-
 *Equal contribution.
 
+## Acknowledgements
+
+We thank Karthik Sridhar, Deepak Lenka, Akash Kokare, Manya Pandey, and Varya Srivastava, as well as the extended Birla AI Labs team for their support.
+
 ## License
 
-Model weights in this repository are licensed under
-[CC-BY-NC-ND 4.0](https://creativecommons.org/licenses/by-nc-nd/4.0/) —
-noncommercial use only, no redistribution of modified/fine-tuned versions,
-attribution required. The inference code and model architecture on
-[GitHub](https://github.com/birla-ai-labs/model-release) are licensed
-separately under
-[PolyForm Noncommercial 1.0.0](https://github.com/birla-ai-labs/model-release/blob/main/LICENSE).
+Model weights : [CC-BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/)
 
-<!-- Internal checkpoint provenance: aaditya_trainings_V19, epoch 1, step 12,000, loss 0.1270 -->
+Code : licensed separately — see [GitHub repository](https://github.com/birla-ai-labs/yug) for its license.