all the models

ttm-1024-96-r1/config.json

@@ -0,0 +1,47 @@
+{
+  "adaptive_patching_levels": 3,
+  "architectures": [
+    "TinyTimeMixerForPrediction"
+  ],
+  "context_length": 1024,
+  "d_model": 192,
+  "d_model_scale": 3,
+  "decoder_adaptive_patching_levels": 0,
+  "decoder_d_model": 128,
+  "decoder_d_model_scale": 2,
+  "decoder_mode": "common_channel",
+  "decoder_num_layers": 2,
+  "decoder_raw_residual": false,
+  "dropout": 0.2,
+  "expansion_factor": 2,
+  "frequency_token_vocab_size": 5,
+  "gated_attn": true,
+  "head_dropout": 0.2,
+  "init_processing": true,
+  "init_std": 0.02,
+  "loss": "mse",
+  "mode": "common_channel",
+  "model_type": "tinytimemixer",
+  "norm_eps": 1e-05,
+  "norm_mlp": "LayerNorm",
+  "num_input_channels": 1,
+  "num_layers": 2,
+  "num_patches": 16,
+  "patch_last": true,
+  "patch_length": 64,
+  "patch_stride": 64,
+  "positional_encoding_type": "sincos",
+  "post_init": false,
+  "prediction_channel_indices": null,
+  "prediction_filter_length": null,
+  "prediction_length": 96,
+  "resolution_prefix_tuning": false,
+  "scaling": "std",
+  "self_attn": false,
+  "self_attn_heads": 1,
+  "stride_ratio": 1,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_decoder": true,
+  "use_positional_encoding": false
+}

ttm-1024-96-r1/generation_config.json

@@ -0,0 +1,4 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.37.2"
+}

ttm-1024-96-r1/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52b46320096e115cbac8b4cf42c0cd37d71def0ac9ddcfcbec6621e3d99b0f20
+size 3804872

ttm-1024-96-r1/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e639e25398173ec543de7ee513b880a92c37f68f9018d1dbd251549d14b65204
+size 4792

ttm-1024-96-r2/README.md

@@ -0,0 +1,3 @@
+---
+license: apache-2.0
+---

ttm-1024-96-r2/config.json

@@ -0,0 +1,59 @@
+{
+  "adaptive_patching_levels": 3,
+  "architectures": [
+    "TinyTimeMixerForPrediction"
+  ],
+  "categorical_vocab_size_list": null,
+  "context_length": 1024,
+  "d_model": 384,
+  "d_model_scale": 3,
+  "decoder_adaptive_patching_levels": 0,
+  "decoder_d_model": 256,
+  "decoder_d_model_scale": 2,
+  "decoder_mode": "common_channel",
+  "decoder_num_layers": 2,
+  "decoder_raw_residual": false,
+  "distribution_output": "student_t",
+  "dropout": 0.4,
+  "enable_forecast_channel_mixing": false,
+  "exogenous_channel_indices": null,
+  "expansion_factor": 2,
+  "fcm_context_length": 1,
+  "fcm_gated_attn": true,
+  "fcm_mix_layers": 3,
+  "fcm_prepend_past": true,
+  "fcm_use_mixer": true,
+  "frequency_token_vocab_size": 8,
+  "gated_attn": true,
+  "head_dropout": 0.4,
+  "init_embed": "pytorch",
+  "init_linear": "pytorch",
+  "init_processing": true,
+  "init_std": 0.02,
+  "loss": "mse",
+  "mode": "common_channel",
+  "model_type": "tinytimemixer",
+  "norm_eps": 1e-05,
+  "norm_mlp": "LayerNorm",
+  "num_input_channels": 1,
+  "num_layers": 2,
+  "num_parallel_samples": 100,
+  "num_patches": 8,
+  "patch_last": true,
+  "patch_length": 128,
+  "patch_stride": 128,
+  "positional_encoding_type": "sincos",
+  "post_init": false,
+  "prediction_channel_indices": null,
+  "prediction_filter_length": null,
+  "prediction_length": 96,
+  "resolution_prefix_tuning": false,
+  "scaling": "std",
+  "self_attn": false,
+  "self_attn_heads": 1,
+  "stride_ratio": 1,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_decoder": true,
+  "use_positional_encoding": false
+}

ttm-1024-96-r2/generation_config.json

@@ -0,0 +1,4 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.37.2"
+}

ttm-1024-96-r2/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5af4b472920903f243b289e08aec727e80e3755d51dc7c3627a7f422f9190140
+size 11879480

ttm-1024-96-r2/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4280fa8a9513a77547ec686aee3c38f6cbe088849148b78e22c9d0655ff6ef53
+size 4856

ttm-1536-96-r2/README.md

@@ -0,0 +1,3 @@
+---
+license: apache-2.0
+---

ttm-1536-96-r2/config.json

@@ -0,0 +1,59 @@
+{
+  "adaptive_patching_levels": 3,
+  "architectures": [
+    "TinyTimeMixerForPrediction"
+  ],
+  "categorical_vocab_size_list": null,
+  "context_length": 1536,
+  "d_model": 384,
+  "d_model_scale": 3,
+  "decoder_adaptive_patching_levels": 0,
+  "decoder_d_model": 256,
+  "decoder_d_model_scale": 2,
+  "decoder_mode": "common_channel",
+  "decoder_num_layers": 2,
+  "decoder_raw_residual": false,
+  "distribution_output": "student_t",
+  "dropout": 0.4,
+  "enable_forecast_channel_mixing": false,
+  "exogenous_channel_indices": null,
+  "expansion_factor": 2,
+  "fcm_context_length": 1,
+  "fcm_gated_attn": true,
+  "fcm_mix_layers": 3,
+  "fcm_prepend_past": true,
+  "fcm_use_mixer": true,
+  "frequency_token_vocab_size": 8,
+  "gated_attn": true,
+  "head_dropout": 0.4,
+  "init_embed": "pytorch",
+  "init_linear": "pytorch",
+  "init_processing": true,
+  "init_std": 0.02,
+  "loss": "mse",
+  "mode": "common_channel",
+  "model_type": "tinytimemixer",
+  "norm_eps": 1e-05,
+  "norm_mlp": "LayerNorm",
+  "num_input_channels": 1,
+  "num_layers": 2,
+  "num_parallel_samples": 100,
+  "num_patches": 12,
+  "patch_last": true,
+  "patch_length": 128,
+  "patch_stride": 128,
+  "positional_encoding_type": "sincos",
+  "post_init": false,
+  "prediction_channel_indices": null,
+  "prediction_filter_length": null,
+  "prediction_length": 96,
+  "resolution_prefix_tuning": false,
+  "scaling": "std",
+  "self_attn": false,
+  "self_attn_heads": 1,
+  "stride_ratio": 1,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_decoder": true,
+  "use_positional_encoding": false
+}

ttm-1536-96-r2/generation_config.json

@@ -0,0 +1,4 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.37.2"
+}

ttm-1536-96-r2/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:559ef888f0fbb79ec54a2dbf7108ac3d13bec5a4da147de63cf11e5deb4260d1
+size 12344144

ttm-1536-96-r2/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a4adf648f23a2b1d9939f7365e65f6dd41b38a09e18fde6479e5883c373717e
+size 4920

ttm-r1/README.md

@@ -0,0 +1,213 @@
+---
+license: apache-2.0
+pipeline_tag: time-series-forecasting
+tags:
+- time series
+- forecasting
+- pretrained models
+- foundation models
+- time series foundation models
+- time-series
+---
+
+# Granite-TimeSeries-TTM-R1 Model Card
+
+<p align="center" width="100%">
+<img src="ttm_image.webp" width="600">
+</p>
+
+TinyTimeMixers (TTMs) are compact pre-trained models for Multivariate Time-Series Forecasting, open-sourced by IBM Research. 
+**With less than 1 Million parameters, TTM (accepted in NeurIPS 24) introduces the notion of the first-ever “tiny” pre-trained models for Time-Series Forecasting.** 
+
+
+TTM outperforms several popular benchmarks demanding billions of parameters in zero-shot and few-shot forecasting. TTMs are lightweight 
+forecasters, pre-trained on publicly available time series data with various augmentations. TTM provides state-of-the-art zero-shot forecasts and can easily be 
+fine-tuned for multi-variate forecasts with just 5% of the training data to be competitive.  Refer to our [paper](https://arxiv.org/pdf/2401.03955.pdf) for more details.
+
+
+**The current open-source version supports point forecasting use-cases specifically ranging from minutely to hourly resolutions 
+(Ex. 10 min, 15 min, 1 hour.).**
+
+**Note that zeroshot, fine-tuning and inference tasks using TTM can easily be executed in 1 GPU machine or in laptops too!!**
+
+
+**New updates:** TTM-R1 comprises TTM variants pre-trained on 250M public training samples. We have another set of TTM models released recently under TTM-R2 trained on a much larger pretraining
+dataset (~700M samples) which can be accessed from [here](https://huggingface.co/ibm-granite/granite-timeseries-ttm-r2). In general, TTM-R2 models perform better than 
+TTM-R1 models as they are trained on larger pretraining dataset. However, the choice of R1 vs R2 depends on your target data distribution. Hence requesting users to 
+try both R1 and R2 variants and pick the best for your data.
+
+
+
+## Model Description
+
+TTM falls under the category of “focused pre-trained models”, wherein each pre-trained TTM is tailored for a particular forecasting 
+setting (governed by the context length and forecast length). Instead of building one massive model supporting all forecasting settings, 
+we opt for the approach of constructing smaller pre-trained models, each focusing on a specific forecasting setting, thereby 
+yielding more accurate results. Furthermore, this approach ensures that our models remain extremely small and exceptionally fast, 
+facilitating easy deployment without demanding a ton of resources. 
+
+Hence, in this model card, we plan to release several pre-trained 
+TTMs that can cater to many common forecasting settings in practice. Additionally, we have released our source code along with 
+our pretraining scripts that users can utilize to pretrain models on their own. Pretraining TTMs is very easy and fast, taking 
+only 3-6 hours using 6 A100 GPUs, as opposed to several days or weeks in traditional approaches.
+
+Each pre-trained model will be released in a different branch name in this model card. Kindly access the required model using our 
+getting started [notebook](https://github.com/IBM/tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb) mentioning the branch name.
+
+
+## Model Releases (along with the branch name where the models are stored):
+
+
+- **512-96:** Given the last 512 time-points (i.e. context length), this model can forecast up to next 96 time-points (i.e. forecast length)
+  in future. This model is targeted towards a forecasting setting of context length 512 and forecast length 96 and
+  recommended for hourly and minutely resolutions (Ex. 10 min, 15 min, 1 hour, etc). This model refers to the TTM-Q variant used in the paper. (branch name: main) [[Benchmark Scripts]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/tinytimemixer/ttm-r1_benchmarking_512_96.ipynb)
+
+- **1024-96:** Given the last 1024 time-points (i.e. context length), this model can forecast up to next 96 time-points (i.e. forecast length)
+  in future. This model is targeted towards a long forecasting setting of context length 1024 and forecast length 96 and
+  recommended for hourly and minutely resolutions (Ex. 10 min, 15 min, 1 hour, etc). (branch name: 1024-96-v1) [[Benchmark Scripts]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/tinytimemixer/ttm-r1_benchmarking_1024_96.ipynb)
+
+
+    
+
+## Model Capabilities with example scripts
+
+The below model scripts can be used for any of the above TTM models. Please update the HF model URL and branch name in the `from_pretrained` call appropriately to pick the model of your choice.
+
+- Getting Started [[colab]](https://colab.research.google.com/github/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb) 
+- Zeroshot Multivariate Forecasting [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb)
+- Finetuned Multivariate Forecasting:
+  - Channel-Independent Finetuning [[Example 1]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb) [[Example 2]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/tinytimemixer/ttm_m4_hourly.ipynb)
+  - Channel-Mix Finetuning [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/tutorial/ttm_channel_mix_finetuning.ipynb)
+- **New Releases (extended features released on October 2024)**
+  - Finetuning and Forecasting with Exogenous/Control Variables [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/tutorial/ttm_with_exog_tutorial.ipynb)
+  - Finetuning and Forecasting with static categorical features [Example: To be added soon]
+  - Rolling Forecasts - Extend forecast lengths beyond 96 via rolling capability [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_rolling_prediction_getting_started.ipynb)
+  - Helper scripts for optimal Learning Rate suggestions for Finetuning [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/tutorial/ttm_with_exog_tutorial.ipynb)
+ 
+## Benchmarks
+
+TTM outperforms popular benchmarks such as TimesFM, Moirai, Chronos, Lag-Llama, Moment, GPT4TS, TimeLLM, LLMTime in zero/fewshot forecasting while reducing computational requirements significantly. 
+Moreover, TTMs are lightweight and can be executed even on CPU-only machines, enhancing usability and fostering wider
+adoption in resource-constrained environments. For more details, refer to our [paper](https://arxiv.org/pdf/2401.03955.pdf) TTM-Q referred in the paper maps to the `512-96` model
+uploaded in the main branch. For other variants (TTM-B, TTM-E and TTM-A) please refer [here](https://huggingface.co/ibm-granite/granite-timeseries-ttm-r2). For more details, refer to the paper.
+
+## Recommended Use
+1. Users have to externally standard scale their data independently for every channel before feeding it to the model (Refer to [TSP](https://github.com/IBM/tsfm/blob/main/tsfm_public/toolkit/time_series_preprocessor.py), our data processing utility for data scaling.)
+2. The current open-source version supports only minutely and hourly resolutions(Ex. 10 min, 15 min, 1 hour.). Other lower resolutions (say weekly, or monthly) are currently not supported in this version, as the model needs a minimum context length of 512 or 1024.
+3. Enabling any upsampling or prepending zeros to virtually increase the context length for shorter-length datasets is not recommended and will
+   impact the model performance. 
+
+
+## Model Details
+
+For more details on TTM architecture and benchmarks, refer to our [paper](https://arxiv.org/pdf/2401.03955.pdf).
+
+TTM-1 currently supports 2 modes:
+
+ - **Zeroshot forecasting**: Directly apply the pre-trained model on your target data to get an initial forecast (with no training).
+
+ - **Finetuned forecasting**: Finetune the pre-trained model with a subset of your target data to further improve the forecast.
+
+**Since, TTM models are extremely small and fast, it is practically very easy to finetune the model with your available target data in few minutes 
+to get more accurate forecasts.**
+
+The current release supports multivariate forecasting via both channel independence and channel-mixing approaches. 
+Decoder Channel-Mixing can be enabled during fine-tuning for capturing strong channel-correlation patterns across 
+time-series variates, a critical capability lacking in existing counterparts.
+
+In addition, TTM also supports exogenous infusion and categorical data infusion.
+
+
+### Model Sources
+
+- **Repository:** https://github.com/ibm-granite/granite-tsfm/tree/main/tsfm_public/models/tinytimemixer
+- **Paper:** https://arxiv.org/pdf/2401.03955.pdf
+
+
+### Blogs and articles on TTM:
+-  Refer to our [wiki](https://github.com/ibm-granite/granite-tsfm/wiki)
+
+  
+## Uses
+
+```
+# Load Model from HF Model Hub mentioning the branch name in revision field
+
+model = TinyTimeMixerForPrediction.from_pretrained(
+                "https://huggingface.co/ibm/TTM", revision="main"
+            ) 
+
+# Do zeroshot
+zeroshot_trainer = Trainer(
+        model=model,
+        args=zeroshot_forecast_args,
+        )
+    )
+
+zeroshot_output = zeroshot_trainer.evaluate(dset_test)
+
+
+# Freeze backbone and enable few-shot or finetuning:
+
+# freeze backbone
+for param in model.backbone.parameters():
+  param.requires_grad = False
+
+finetune_forecast_trainer = Trainer(
+        model=model,
+        args=finetune_forecast_args,
+        train_dataset=dset_train,
+        eval_dataset=dset_val,
+        callbacks=[early_stopping_callback, tracking_callback],
+        optimizers=(optimizer, scheduler),
+    )
+finetune_forecast_trainer.train()
+fewshot_output = finetune_forecast_trainer.evaluate(dset_test)
+
+```
+
+
+## Training Data
+
+The original r1 TTM models were trained on a collection of datasets from the Monash Time Series Forecasting repository. The datasets used include:
+ - Australian Electricity Demand: https://zenodo.org/records/4659727 
+ - Australian Weather: https://zenodo.org/records/4654822 
+ - Bitcoin dataset: https://zenodo.org/records/5122101 
+ - KDD Cup 2018 dataset: https://zenodo.org/records/4656756 
+ - London Smart Meters: https://zenodo.org/records/4656091 
+ - Saugeen River Flow: https://zenodo.org/records/4656058
+ - Solar Power: https://zenodo.org/records/4656027 
+ - Sunspots: https://zenodo.org/records/4654722
+ - Solar: https://zenodo.org/records/4656144 
+ - US Births: https://zenodo.org/records/4656049 
+ - Wind Farms Production data: https://zenodo.org/records/4654858 
+ - Wind Power: https://zenodo.org/records/4656032
+
+
+## Citation
+Kindly cite the following paper, if you intend to use our model or its associated architectures/approaches in your 
+work
+
+**BibTeX:**
+
+```
+@inproceedings{ekambaram2024tinytimemixersttms,
+      title={Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series},
+      author={Vijay Ekambaram and Arindam Jati and Pankaj Dayama and Sumanta Mukherjee and Nam H. Nguyen and Wesley M. Gifford and Chandra Reddy and Jayant Kalagnanam},
+      booktitle={Advances in Neural Information Processing Systems (NeurIPS 2024)},
+      year={2024},
+}
+```
+
+## Model Card Authors
+
+Vijay Ekambaram, Arindam Jati, Pankaj Dayama, Wesley M. Gifford, Sumanta Mukherjee, Chandra Reddy and Jayant Kalagnanam
+
+
+## IBM Public Repository Disclosure: 
+
+All content in this repository including code has been provided by IBM under the associated 
+open source software license and IBM is under no obligation to provide enhancements, 
+updates, or support. IBM developers produced this code as an 
+open source project (not as an IBM product), and IBM makes no assertions as to 
+the level of quality nor security, and will not be maintaining this code going forward.

ttm-r1/config.json

@@ -0,0 +1,46 @@
+{
+  "adaptive_patching_levels": 3,
+  "architectures": [
+    "TinyTimeMixerForPrediction"
+  ],
+  "context_length": 512,
+  "d_model": 192,
+  "decoder_adaptive_patching_levels": 0,
+  "decoder_d_model": 128,
+  "decoder_mode": "common_channel",
+  "decoder_num_layers": 2,
+  "decoder_raw_residual": false,
+  "distribution_output": "student_t",
+  "dropout": 0.2,
+  "expansion_factor": 2,
+  "frequency_token_vocab_size": 5,
+  "gated_attn": true,
+  "head_dropout": 0.2,
+  "init_processing": true,
+  "init_std": 0.02,
+  "loss": "mse",
+  "mode": "common_channel",
+  "model_type": "tinytimemixer",
+  "norm_eps": 1e-05,
+  "norm_mlp": "LayerNorm",
+  "num_input_channels": 1,
+  "num_layers": 2,
+  "num_parallel_samples": 100,
+  "num_patches": 8,
+  "patch_last": true,
+  "patch_length": 64,
+  "patch_stride": 64,
+  "positional_encoding_type": "sincos",
+  "post_init": false,
+  "prediction_channel_indices": null,
+  "prediction_filter_length": null,
+  "prediction_length": 96,
+  "resolution_prefix_tuning": false,
+  "scaling": "std",
+  "self_attn": false,
+  "self_attn_heads": 1,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_decoder": true,
+  "use_positional_encoding": false
+}

ttm-r1/generation_config.json

@@ -0,0 +1,4 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.37.2"
+}

ttm-r1/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30f8d9be619518e18c4c15c6ee504e491b9624b48e251661cd5f967bce097731
+size 3240592

ttm-r2/README.md

@@ -0,0 +1,227 @@
+---
+license: apache-2.0
+pipeline_tag: time-series-forecasting
+tags:
+- time series
+- forecasting
+- pretrained models
+- foundation models
+- time series foundation models
+- time-series
+---
+
+# Granite-TimeSeries-TTM-R2 Model Card
+
+<p align="center" width="100%">
+<img src="ttm_image.webp" width="600">
+</p>
+
+TinyTimeMixers (TTMs) are compact pre-trained models for Multivariate Time-Series Forecasting, open-sourced by IBM Research. 
+**With model sizes starting from 1M params, TTM (accepted in NeurIPS 24) introduces the notion of the first-ever “tiny” pre-trained models for Time-Series Forecasting.** 
+
+
+TTM outperforms several popular benchmarks demanding billions of parameters in zero-shot and few-shot forecasting. TTMs are lightweight 
+forecasters, pre-trained on publicly available time series data with various augmentations. TTM provides state-of-the-art zero-shot forecasts and can easily be 
+fine-tuned for multi-variate forecasts with just 5% of the training data to be competitive.  Refer to our [paper](https://arxiv.org/pdf/2401.03955.pdf) for more details.
+
+
+**The current open-source version supports point forecasting use-cases specifically ranging from minutely to hourly resolutions 
+(Ex. 10 min, 15 min, 1 hour.).**
+
+**Note that zeroshot, fine-tuning and inference tasks using TTM can easily be executed in 1 GPU machine or in laptops too!!**
+
+
+**TTM-R2 comprises TTM variants pre-trained on larger pretraining datasets (~700M samples).** We have another set of TTM models released under `TTM-R1` trained on ~250M samples 
+which can be accessed from [here](https://huggingface.co/ibm-granite/granite-timeseries-ttm-r1) In general, `TTM-R2` models perform better than `TTM-R1` models as they are 
+trained on larger pretraining dataset. However, the choice of R1 vs R2 depends on your target data distribution. Hence requesting users to try both
+R1 and R2 variants and pick the best for your data.
+
+
+
+## Model Description
+
+TTM falls under the category of “focused pre-trained models”, wherein each pre-trained TTM is tailored for a particular forecasting 
+setting (governed by the context length and forecast length). Instead of building one massive model supporting all forecasting settings, 
+we opt for the approach of constructing smaller pre-trained models, each focusing on a specific forecasting setting, thereby 
+yielding more accurate results. Furthermore, this approach ensures that our models remain extremely small and exceptionally fast, 
+facilitating easy deployment without demanding a ton of resources. 
+
+Hence, in this model card, we plan to release several pre-trained 
+TTMs that can cater to many common forecasting settings in practice. Additionally, we have released our source code along with 
+our pretraining scripts that users can utilize to pretrain models on their own. Pretraining TTMs is very easy and fast, taking 
+3-6 hours for R1 versions and 12-24 hours for R2 versions using 6 A100 GPUs, as opposed to several days or weeks in traditional approaches.
+
+Each pre-trained model will be released in a different branch name in this model card. Kindly access the required model using our 
+getting started [notebook](https://github.com/IBM/tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb) mentioning the branch name.
+
+
+## Model Releases (along with the branch name where the models are stored):
+
+
+- **512-96-r2**: Given the last 512 time-points (i.e. context length), this model can forecast up to next 96 time-points (i.e. forecast length)
+    in future. This model is pre-trained with a larger pretraining dataset for improved accuracy. Recommended for hourly and minutely
+    resolutions (Ex. 10 min, 15 min, 1 hour, etc).   (branch name: 512-96-r2) [[Benchmarks]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/tinytimemixer/ttm-r2_benchmarking_512_96.ipynb)
+  
+- **1024-96-r2**: Given the last 1024 time-points (i.e. context length), this model can forecast up to next 96 time-points (i.e. forecast length)
+    in future. This model is pre-trained with a larger pretraining dataset for improved accuracy. Recommended for hourly and minutely
+    resolutions (Ex. 10 min, 15 min, 1 hour, etc).   (branch name: 1024-96-r2) [[Benchmarks]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/tinytimemixer/ttm-r2_benchmarking_1024_96.ipynb)
+ 
+
+- **1536-96-r2**: Given the last 1536 time-points (i.e. context length), this model can forecast up to next 96 time-points (i.e. forecast length)
+    in future. This model is pre-trained with a larger pretraining dataset for improved accuracy. Recommended for hourly and minutely
+    resolutions (Ex. 10 min, 15 min, 1 hour, etc).   (branch name: 1536-96-r2) [[Benchmarks]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/tinytimemixer/ttm-r2_benchmarking_1536_96.ipynb)
+
+    
+
+## Model Capabilities with example scripts
+
+The below model scripts can be used for any of the above TTM models. Please update the HF model URL and branch name in the `from_pretrained` call appropriately to pick the model of your choice.
+
+- Getting Started [[colab]](https://colab.research.google.com/github/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb) 
+- Zeroshot Multivariate Forecasting [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb)
+- Finetuned Multivariate Forecasting:
+  - Channel-Independent Finetuning [[Example 1]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_getting_started.ipynb) [[Example 2]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/tinytimemixer/ttm_m4_hourly.ipynb)
+  - Channel-Mix Finetuning [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/tutorial/ttm_channel_mix_finetuning.ipynb)
+- **New Releases (extended features released on October 2024)**
+  - Finetuning and Forecasting with Exogenous/Control Variables [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/tutorial/ttm_with_exog_tutorial.ipynb)
+  - Finetuning and Forecasting with static categorical features [Example: To be added soon]
+  - Rolling Forecasts - Extend forecast lengths beyond 96 via rolling capability [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/hfdemo/ttm_rolling_prediction_getting_started.ipynb)
+  - Helper scripts for optimal Learning Rate suggestions for Finetuning [[Example]](https://github.com/ibm-granite/granite-tsfm/blob/main/notebooks/tutorial/ttm_with_exog_tutorial.ipynb)
+    
+## Benchmarks
+
+TTM outperforms popular benchmarks such as TimesFM, Moirai, Chronos, Lag-Llama, Moment, GPT4TS, TimeLLM, LLMTime in zero/fewshot forecasting while reducing computational requirements significantly. 
+Moreover, TTMs are lightweight and can be executed even on CPU-only machines, enhancing usability and fostering wider
+adoption in resource-constrained environments. For more details, refer to our [paper](https://arxiv.org/pdf/2401.03955.pdf).
+- TTM-B referred in the paper maps to the `512-96-r2` model.
+- TTM-E referred in the paper maps to the `1024-96-r2` model.
+- TTM-A referred in the paper maps to the `1536-96-r2` model
+
+Please note that the Granite TTM models are pre-trained exclusively on datasets
+with clear commercial-use licenses that are approved by our legal team. As a result, the pre-training dataset used in this release differs slightly from the one used in the research
+paper, which may lead to minor variations in model performance as compared to the published results. Please refer to our paper for more details.
+
+## Recommended Use
+1. Users have to externally standard scale their data independently for every channel before feeding it to the model (Refer to [TSP](https://github.com/IBM/tsfm/blob/main/tsfm_public/toolkit/time_series_preprocessor.py), our data processing utility for data scaling.)
+2. The current open-source version supports only minutely and hourly resolutions(Ex. 10 min, 15 min, 1 hour.). Other lower resolutions (say weekly, or monthly) are currently not supported in this version, as the model needs a minimum context length of 512 or 1024.
+3. Enabling any upsampling or prepending zeros to virtually increase the context length for shorter-length datasets is not recommended and will
+   impact the model performance. 
+
+
+
+## Model Details
+
+For more details on TTM architecture and benchmarks, refer to our [paper](https://arxiv.org/pdf/2401.03955.pdf).
+
+TTM-1 currently supports 2 modes:
+
+ - **Zeroshot forecasting**: Directly apply the pre-trained model on your target data to get an initial forecast (with no training).
+
+ - **Finetuned forecasting**: Finetune the pre-trained model with a subset of your target data to further improve the forecast.
+
+**Since, TTM models are extremely small and fast, it is practically very easy to finetune the model with your available target data in few minutes 
+to get more accurate forecasts.**
+
+The current release supports multivariate forecasting via both channel independence and channel-mixing approaches. 
+Decoder Channel-Mixing can be enabled during fine-tuning for capturing strong channel-correlation patterns across 
+time-series variates, a critical capability lacking in existing counterparts.
+
+In addition, TTM also supports exogenous infusion and categorical data infusion.
+
+
+### Model Sources
+
+- **Repository:** https://github.com/ibm-granite/granite-tsfm/tree/main/tsfm_public/models/tinytimemixer
+- **Paper:** https://arxiv.org/pdf/2401.03955.pdf
+
+
+### Blogs and articles on TTM:
+-  Refer to our [wiki](https://github.com/ibm-granite/granite-tsfm/wiki)
+
+  
+## Uses
+
+```
+# Load Model from HF Model Hub mentioning the branch name in revision field
+
+model = TinyTimeMixerForPrediction.from_pretrained(
+                "https://huggingface.co/ibm/TTM", revision="main"
+            ) 
+
+# Do zeroshot
+zeroshot_trainer = Trainer(
+        model=model,
+        args=zeroshot_forecast_args,
+        )
+    )
+
+zeroshot_output = zeroshot_trainer.evaluate(dset_test)
+
+
+# Freeze backbone and enable few-shot or finetuning:
+
+# freeze backbone
+for param in model.backbone.parameters():
+  param.requires_grad = False
+
+finetune_forecast_trainer = Trainer(
+        model=model,
+        args=finetune_forecast_args,
+        train_dataset=dset_train,
+        eval_dataset=dset_val,
+        callbacks=[early_stopping_callback, tracking_callback],
+        optimizers=(optimizer, scheduler),
+    )
+finetune_forecast_trainer.train()
+fewshot_output = finetune_forecast_trainer.evaluate(dset_test)
+
+```
+
+
+## Training Data
+
+The r2 TTM models were trained on a collection of datasets as follows:
+ - Australian Electricity Demand: https://zenodo.org/records/4659727 
+ - Australian Weather: https://zenodo.org/records/4654822 
+ - Bitcoin dataset: https://zenodo.org/records/5122101 
+ - KDD Cup 2018 dataset: https://zenodo.org/records/4656756 
+ - London Smart Meters: https://zenodo.org/records/4656091 
+ - Saugeen River Flow: https://zenodo.org/records/4656058
+ - Solar Power: https://zenodo.org/records/4656027 
+ - Sunspots: https://zenodo.org/records/4654722
+ - Solar: https://zenodo.org/records/4656144 
+ - US Births: https://zenodo.org/records/4656049 
+ - Wind Farms Production data: https://zenodo.org/records/4654858 
+ - Wind Power: https://zenodo.org/records/4656032
+ - PEMSD3, PEMSD4, PEMSD7, PEMSD8, PEMS_BAY: https://drive.google.com/drive/folders/1g5v2Gq1tkOq8XO0HDCZ9nOTtRpB6-gPe
+ - LOS_LOOP: https://drive.google.com/drive/folders/1g5v2Gq1tkOq8XO0HDCZ9nOTtRpB6-gPe 
+
+
+
+## Citation
+Kindly cite the following paper, if you intend to use our model or its associated architectures/approaches in your 
+work
+
+**BibTeX:**
+
+```
+@inproceedings{ekambaram2024tinytimemixersttms,
+      title={Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series},
+      author={Vijay Ekambaram and Arindam Jati and Pankaj Dayama and Sumanta Mukherjee and Nam H. Nguyen and Wesley M. Gifford and Chandra Reddy and Jayant Kalagnanam},
+      booktitle={Advances in Neural Information Processing Systems (NeurIPS 2024)},
+      year={2024},
+}
+```
+
+## Model Card Authors
+
+Vijay Ekambaram, Arindam Jati, Pankaj Dayama, Wesley M. Gifford, Sumanta Mukherjee, Chandra Reddy and Jayant Kalagnanam
+
+
+## IBM Public Repository Disclosure: 
+
+All content in this repository including code has been provided by IBM under the associated 
+open source software license and IBM is under no obligation to provide enhancements, 
+updates, or support. IBM developers produced this code as an 
+open source project (not as an IBM product), and IBM makes no assertions as to 
+the level of quality nor security, and will not be maintaining this code going forward.

ttm-r2/config.json

@@ -0,0 +1,59 @@
+{
+  "adaptive_patching_levels": 3,
+  "architectures": [
+    "TinyTimeMixerForPrediction"
+  ],
+  "categorical_vocab_size_list": null,
+  "context_length": 512,
+  "d_model": 192,
+  "d_model_scale": 3,
+  "decoder_adaptive_patching_levels": 0,
+  "decoder_d_model": 128,
+  "decoder_d_model_scale": 2,
+  "decoder_mode": "common_channel",
+  "decoder_num_layers": 2,
+  "decoder_raw_residual": false,
+  "distribution_output": "student_t",
+  "dropout": 0.4,
+  "enable_forecast_channel_mixing": false,
+  "exogenous_channel_indices": null,
+  "expansion_factor": 2,
+  "fcm_context_length": 1,
+  "fcm_gated_attn": true,
+  "fcm_mix_layers": 3,
+  "fcm_prepend_past": true,
+  "fcm_use_mixer": true,
+  "frequency_token_vocab_size": 5,
+  "gated_attn": true,
+  "head_dropout": 0.4,
+  "init_embed": "pytorch",
+  "init_linear": "pytorch",
+  "init_processing": true,
+  "init_std": 0.02,
+  "loss": "mse",
+  "mode": "common_channel",
+  "model_type": "tinytimemixer",
+  "norm_eps": 1e-05,
+  "norm_mlp": "LayerNorm",
+  "num_input_channels": 1,
+  "num_layers": 2,
+  "num_parallel_samples": 100,
+  "num_patches": 8,
+  "patch_last": true,
+  "patch_length": 64,
+  "patch_stride": 64,
+  "positional_encoding_type": "sincos",
+  "post_init": false,
+  "prediction_channel_indices": null,
+  "prediction_filter_length": null,
+  "prediction_length": 96,
+  "resolution_prefix_tuning": false,
+  "scaling": "std",
+  "self_attn": false,
+  "self_attn_heads": 1,
+  "stride_ratio": 1,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.2",
+  "use_decoder": true,
+  "use_positional_encoding": false
+}

ttm-r2/generation_config.json

@@ -0,0 +1,4 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.37.2"
+}

ttm-r2/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a706726a7eb01bbcb42994b7dcb3c06ea9557898dbae8d480eb04fe8ccb89710
+size 3240592

ttm-r2/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0cefcfb5a52810ab0e1103fe206c6d5b00a4c86159a4a957681031122ec23ee7
+size 4920