Upload UrbanFM datasets

.gitattributes

@@ -58,3 +58,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+data_analysis/benchmark_st_vis.pdf filter=lfs diff=lfs merge=lfs -text
+data_analysis/benchmark_t_vis.pdf filter=lfs diff=lfs merge=lfs -text

DATASET_CARD.md

@@ -0,0 +1,35 @@
+---
+license: mit
+task_categories:
+  - time-series-forecasting
+language:
+  - en
+tags:
+  - spatio-temporal
+  - urban-computing
+  - foundation-model
+size_categories:
+  - 1B<n<10B
+---
+
+# UrbanFM Datasets
+
+Datasets for [UrbanFM: Scaling Urban Spatio-Temporal Foundation Models](https://github.com/Onedean/UrbanFM).
+
+## Structure
+
+- `eval_datasets/` — evaluation datasets (traffic flow, speed, occupancy, etc.)
+- `full_pretrain_datasets/` — full-scale pre-training datasets
+- `pretrain_datasets/` — pre-training datasets
+- `clean.ipynb`, `spatial_index_generation.py`, `data_index_process.sh` — data processing utilities
+
+## Download
+
+```bash
+pip install huggingface_hub
+huggingface-cli download Onedean/UrbanFM-datasets --repo-type dataset --local-dir ./datasets
+```
+
+## Citation
+
+If you use these datasets, please cite the UrbanFM paper and star the [GitHub repository](https://github.com/Onedean/UrbanFM).

README.md

@@ -0,0 +1,32 @@
+# UrbanFM Datasets
+
+The full dataset (~10 GB) is hosted on Hugging Face:
+
+**https://huggingface.co/datasets/Onedean/UrbanFM-datasets**
+
+## Download
+
+Install the Hugging Face CLI:
+
+```bash
+pip install huggingface_hub
+```
+
+Download the entire datasets folder into the project root:
+
+```bash
+huggingface-cli download Onedean/UrbanFM-datasets --repo-type dataset --local-dir ./datasets
+```
+
+Or use the provided script from the project root:
+
+```bash
+bash scripts/download_datasets.sh
+```
+
+## Contents
+
+- `eval_datasets/` — evaluation datasets
+- `full_pretrain_datasets/` — full pre-training datasets
+- `pretrain_datasets/` — pre-training datasets
+- `clean.ipynb`, `spatial_index_generation.py`, `data_index_process.sh` — data processing scripts

clean.ipynb

@@ -0,0 +1,904 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import pandas as pd\n",
+    "from tqdm import tqdm"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 67,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 40it [00:01, 23.31it/s]\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Path to the root folder containing subdirectories\n",
+    "input_root_folder = \"/data/weichen/st_datasets/world_st_traffic/processed_data_outlier\"\n",
+    "output_root_folder = \"/data/weichen/st_datasets/world_st_traffic/pretrain_datasets\"\n",
+    "\n",
+    "detectors = pd.read_csv(\"detectors_public.csv\")\n",
+    "\n",
+    "# Initialize variables to store the total number of rows (T) and details\n",
+    "total_rows = 0\n",
+    "file_shapes = []\n",
+    "\n",
+    "def process_column_name(col):\n",
+    "    try:\n",
+    "        # 尝试将列名转换为浮点数，再转换为整数，然后转换为字符串\n",
+    "        col_int = int(float(col))\n",
+    "        # 如果转换后的整数长度为 7，则在前面加上 '0'\n",
+    "        if len(str(col_int)) == 7:\n",
+    "            return '0' + str(col_int)\n",
+    "        else:\n",
+    "            return str(col_int)  # 返回原始列名或转换后的列名\n",
+    "    except ValueError:\n",
+    "        return col  # 如果转换失败，返回原始列名\n",
+    "\n",
+    "# Traverse through the root folder to process CSV files\n",
+    "for subdir, _, files in tqdm(os.walk(input_root_folder), desc=\"Processing files\"):\n",
+    "    for file_name in files:\n",
+    "        if file_name.endswith('.csv'):\n",
+    "            \n",
+    "            dataset_name = file_name.split('.')[0]\n",
+    "            city = dataset_name.split('_')[0]\n",
+    "            type = dataset_name.split('_')[1]\n",
+    "            \n",
+    "            if city != \"losangeles\" and city != \"stuttgart\":\n",
+    "                continue\n",
+    "            \n",
+    "            if city == \"losangeles\":\n",
+    "                city = \"losanageles\"\n",
+    "            \n",
+    "            # 读取 CSV 文件\n",
+    "            file_path = os.path.join(subdir, file_name)\n",
+    "            df = pd.read_csv(file_path)\n",
+    "            \n",
+    "            # 先将 time 列转换为 datetime 类型\n",
+    "            df['time'] = pd.to_datetime(df['time'])\n",
+    "            \n",
+    "            # 将 time 列设置为索引\n",
+    "            df.set_index('time', inplace=True)\n",
+    "\n",
+    "            # 使用 resample 将数据采样到5分钟的频率，这里取均值 （下采样）\n",
+    "            df_5min = df.resample('5min').mean()\n",
+    "            \n",
+    "            # 将缺失值进行线性插值 （上采样） \n",
+    "            df_5min = df_5min.interpolate(method='linear')\n",
+    "            \n",
+    "            if not os.path.exists(f\"{output_root_folder}/{dataset_name}\"):\n",
+    "                os.makedirs(f\"{output_root_folder}/{dataset_name}\")\n",
+    "            \n",
+    "            if city == \"stuttgart\":\n",
+    "                df_5min.columns = df_5min.columns.map(process_column_name)\n",
+    "            \n",
+    "            # 保存文件\n",
+    "            df_5min.to_pickle(f'{output_root_folder}/{dataset_name}/{dataset_name}_temporal.pkl')\n",
+    "            \n",
+    "            detector_ids = list(df_5min.columns)\n",
+    "            \n",
+    "            detectors_city = detectors[detectors[\"citycode\"] == city].copy()\n",
+    "\n",
+    "            selected_detectors = detectors_city[detectors_city[\"detid\"].isin(detector_ids)].copy()\n",
+    "\n",
+    "            selected_detectors = (selected_detectors.drop_duplicates(subset=\"detid\").set_index(\"detid\").reindex(detector_ids).reset_index())[[\"detid\", \"lat\", \"long\"]].copy()\n",
+    "\n",
+    "            selected_detectors.rename(columns={\"detid\": \"ID\", \"lat\": \"Latitude\", \"long\": \"Longitude\"}, inplace=True)\n",
+    "\n",
+    "            selected_detectors.to_pickle(f'{output_root_folder}/{dataset_name}/{dataset_name}_spatial.pkl')\n",
+    "            "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### new process"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 0it [00:00, ?it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in utrecht flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 2it [00:01,  1.40it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in bolton occ\n",
+      "Total NaNs in df_processed: 0 in bolton speed\n",
+      "Total NaNs in df_processed: 0 in bolton flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 3it [00:02,  1.00it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in toronto flow\n",
+      "Total NaNs in df_processed: 0 in toronto occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 4it [00:08,  2.64s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in paris flow\n",
+      "Total NaNs in df_processed: 0 in paris occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 5it [00:38, 12.32s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in wolfsburg flow\n",
+      "Total NaNs in df_processed: 0 in wolfsburg occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 6it [00:40,  8.81s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in birmingham speed\n",
+      "Total NaNs in df_processed: 0 in birmingham flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 7it [00:41,  6.31s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in innsbruck flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 8it [00:42,  4.62s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in santander occ\n",
+      "Total NaNs in df_processed: 0 in santander flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 9it [00:44,  3.70s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in melbourne flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 10it [00:47,  3.75s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in augsburg flow\n",
+      "Total NaNs in df_processed: 0 in augsburg occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 11it [00:49,  3.09s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in stuttgart occ\n",
+      "Total NaNs in df_processed: 0 in stuttgart flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 12it [00:51,  2.61s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in strasbourg occ\n",
+      "Total NaNs in df_processed: 0 in strasbourg flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 13it [00:53,  2.50s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in toulouse occ\n",
+      "Total NaNs in df_processed: 0 in toulouse flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 14it [00:55,  2.53s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in graz flow\n",
+      "Total NaNs in df_processed: 0 in graz occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 15it [00:58,  2.46s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in cagliari occ\n",
+      "Total NaNs in df_processed: 0 in cagliari flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 16it [01:02,  2.93s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in constance occ\n",
+      "Total NaNs in df_processed: 0 in constance flow\n",
+      "Total NaNs in df_processed: 0 in constance speed\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 17it [01:03,  2.57s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in kassel flow\n",
+      "Total NaNs in df_processed: 0 in kassel occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 18it [01:05,  2.18s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in essen occ\n",
+      "Total NaNs in df_processed: 0 in essen speed\n",
+      "Total NaNs in df_processed: 0 in essen flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 19it [01:08,  2.50s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in taipeh flow\n",
+      "Total NaNs in df_processed: 0 in taipeh occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 20it [01:11,  2.70s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in bremen occ\n",
+      "Total NaNs in df_processed: 0 in bremen flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 21it [01:16,  3.24s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in groningen occ\n",
+      "Total NaNs in df_processed: 0 in groningen speed\n",
+      "Total NaNs in df_processed: 0 in groningen flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 22it [01:17,  2.58s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in darmstadt occ\n",
+      "Total NaNs in df_processed: 0 in darmstadt flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 23it [01:21,  3.10s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in losanageles flow\n",
+      "Total NaNs in df_processed: 0 in losanageles occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 24it [01:24,  3.14s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in basel occ\n",
+      "Total NaNs in df_processed: 0 in basel flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 25it [01:25,  2.56s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in torino flow\n",
+      "Total NaNs in df_processed: 0 in torino speed\n",
+      "Total NaNs in df_processed: 0 in torino occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 26it [01:30,  3.26s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in london flow\n",
+      "Total NaNs in df_processed: 0 in london occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 27it [02:13, 15.20s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in bern flow\n",
+      "Total NaNs in df_processed: 0 in bern occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 28it [02:17, 11.60s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in rotterdam occ\n",
+      "Total NaNs in df_processed: 0 in rotterdam speed\n",
+      "Total NaNs in df_processed: 0 in rotterdam flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 29it [02:18,  8.61s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in marseille occ\n",
+      "Total NaNs in df_processed: 0 in marseille flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 30it [02:22,  7.15s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in vilnius occ\n",
+      "Total NaNs in df_processed: 0 in vilnius flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 31it [02:23,  5.20s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in munich occ\n",
+      "Total NaNs in df_processed: 0 in munich flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 32it [02:24,  3.99s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in zurich flow\n",
+      "Total NaNs in df_processed: 0 in zurich occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 33it [02:28,  4.02s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in madrid flow\n",
+      "Total NaNs in df_processed: 0 in madrid occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 34it [02:36,  5.31s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in luzern flow\n",
+      "Total NaNs in df_processed: 0 in luzern occ\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 35it [03:10, 13.78s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in speyer occ\n",
+      "Total NaNs in df_processed: 0 in speyer flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 36it [03:12, 10.30s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in hamburg occ\n",
+      "Total NaNs in df_processed: 0 in hamburg flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 37it [03:34, 13.70s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in frankfurt occ\n",
+      "Total NaNs in df_processed: 0 in frankfurt flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 38it [03:34,  9.80s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in manchester occ\n",
+      "Total NaNs in df_processed: 0 in manchester speed\n",
+      "Total NaNs in df_processed: 0 in manchester flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 39it [03:38,  7.93s/it]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total NaNs in df_processed: 0 in bordeaux occ\n",
+      "Total NaNs in df_processed: 0 in bordeaux flow\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Processing files: 40it [03:40,  5.51s/it]\n",
+      "Processing files: 40it [03:40,  5.51s/it]\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "from tqdm import tqdm\n",
+    "import seaborn as sns\n",
+    "\n",
+    "# Path to the root folder containing subdirectories\n",
+    "input_root_folder = \"/data/weichen/st_datasets/world_st_traffic/processed_data_outlier\"\n",
+    "output_root_folder = \"/data/weichen/st_datasets/world_st_traffic/pretrain_datasets\"\n",
+    "\n",
+    "detectors = pd.read_csv(\"detectors_public.csv\")\n",
+    "\n",
+    "def process_column_name(col):\n",
+    "    try:\n",
+    "        col_int = int(float(col))\n",
+    "        if len(str(col_int)) == 7:\n",
+    "            return '0' + str(col_int)\n",
+    "        else:\n",
+    "            return str(col_int)\n",
+    "    except ValueError:\n",
+    "        return col\n",
+    "\n",
+    "# Function to split data into valid segments based on missing time gaps\n",
+    "def split_segments(df, min_gap):\n",
+    "    gaps = (df.index[1:] - df.index[:-1]).total_seconds() / 60  # Calculate gaps in minutes\n",
+    "    break_points = [-1] + (gaps > min_gap).nonzero()[0].tolist() + [len(df)-1]\n",
+    "    segments = [df.iloc[break_points[i]+1:break_points[i + 1]+1] for i in range(len(break_points) - 1)]\n",
+    "    return [seg for seg in segments if not seg.empty]\n",
+    "\n",
+    "# Minimum gap to consider as a missing time block\n",
+    "min_gap_minutes = 60  # Adjust based on your specific dataset\n",
+    "\n",
+    "\n",
+    "# Function to plot mean and standard deviation with discontinuities\n",
+    "def plot_with_discontinuities(means, std_devs, output_path, title):\n",
+    "    plt.figure(figsize=(20, 6))\n",
+    "    segments = split_segments(pd.DataFrame({'mean': means, 'std': std_devs}), min_gap_minutes)\n",
+    "    \n",
+    "    for segment in segments:\n",
+    "        mean_segment = segment['mean']\n",
+    "        std_segment = segment['std']\n",
+    "        plt.plot(mean_segment.index, mean_segment, label='Mean', color='red')\n",
+    "        plt.fill_between(mean_segment.index, mean_segment - std_segment, mean_segment + std_segment, color='orange', alpha=0.5, label='Standard Deviation')\n",
+    "    \n",
+    "    plt.title(title)\n",
+    "    plt.xlabel('Date')\n",
+    "    plt.ylabel('Values')\n",
+    "    # plt.legend()\n",
+    "    plt.grid(True)\n",
+    "    plt.tight_layout()\n",
+    "    plt.savefig(output_path)\n",
+    "    plt.close()\n",
+    "\n",
+    "\n",
+    "# Traverse through the root folder to process CSV files\n",
+    "for subdir, _, files in tqdm(os.walk(input_root_folder), desc=\"Processing files\"):\n",
+    "    for file_name in files:\n",
+    "        if file_name.endswith('.csv'):\n",
+    "            dataset_name = file_name.split('.')[0]\n",
+    "            city = dataset_name.split('_')[0]\n",
+    "            type = dataset_name.split('_')[1]\n",
+    "            \n",
+    "            if city == \"losangeles\":\n",
+    "                city = \"losanageles\"\n",
+    "\n",
+    "            file_path = os.path.join(subdir, file_name)\n",
+    "            df = pd.read_csv(file_path)\n",
+    "\n",
+    "            # Convert 'time' column to datetime and set as index\n",
+    "            df['time'] = pd.to_datetime(df['time'])\n",
+    "            df.set_index('time', inplace=True)\n",
+    "\n",
+    "            # Determine original sampling rate\n",
+    "            original_sampling_rate = (df.index[1] - df.index[0]).total_seconds() / 60\n",
+    "\n",
+    "            # Split data into valid segments\n",
+    "            segments = split_segments(df, min_gap_minutes)\n",
+    "\n",
+    "            # Process each segment\n",
+    "            processed_segments = []\n",
+    "            for segment in segments:                \n",
+    "                resampled_segment = segment.resample('5min').mean()\n",
+    "                \n",
+    "                # First fill NaN at the beginning and end using forward and backward fill\n",
+    "                resampled_segment = resampled_segment.ffill().bfill()\n",
+    "                \n",
+    "                interpolated_segment = resampled_segment.interpolate(method='linear')\n",
+    "                processed_segments.append(interpolated_segment)\n",
+    "\n",
+    "            # Concatenate processed segments\n",
+    "            df_processed = pd.concat(processed_segments)\n",
+    "            \n",
+    "            if type == 'occ':  # Multiply flow values by 10 if type is 'occ'\n",
+    "                df_processed *= 100\n",
+    "            \n",
+    "            output_dir = f\"{output_root_folder}/{dataset_name}\"\n",
+    "            if not os.path.exists(output_dir):\n",
+    "                os.makedirs(output_dir)\n",
+    "\n",
+    "            if city == \"stuttgart\":\n",
+    "                df_processed.columns = df_processed.columns.map(process_column_name)\n",
+    "\n",
+    "            # Save temporal data\n",
+    "            df_processed.to_pickle(f'{output_dir}/{dataset_name}_temporal.pkl')\n",
+    "            \n",
+    "            # Count NaNs in df_processed\n",
+    "            total_nans = df_processed.isna().sum().sum()\n",
+    "\n",
+    "            # Log or print the NaN statistics\n",
+    "            print(f'Total NaNs in df_processed: {total_nans} in {city} {type}')\n",
+    "            \n",
+    "\n",
+    "            detector_ids = list(df_processed.columns)\n",
+    "            detectors_city = detectors[detectors[\"citycode\"] == city].copy()\n",
+    "            selected_detectors = detectors_city[detectors_city[\"detid\"].isin(detector_ids)].copy()\n",
+    "            selected_detectors = (\n",
+    "                selected_detectors.drop_duplicates(subset=\"detid\")\n",
+    "                .set_index(\"detid\")\n",
+    "                .reindex(detector_ids)\n",
+    "                .reset_index()\n",
+    "            )[[\"detid\", \"lat\", \"long\"]].copy()\n",
+    "            selected_detectors.rename(columns={\"detid\": \"ID\", \"lat\": \"Latitude\", \"long\": \"Longitude\"}, inplace=True)\n",
+    "\n",
+    "            # Save spatial data\n",
+    "            selected_detectors.to_pickle(f'{output_dir}/{dataset_name}_spatial.pkl')\n",
+    "            \n",
+    "            \n",
+    "            # 绘制热力图\n",
+    "            value_min = df_processed.min().min()\n",
+    "            value_max = df_processed.max().max()\n",
+    "\n",
+    "            plt.figure(figsize=(12, 6))\n",
+    "            sns.heatmap(df_processed.T, cbar=True, cmap=\"viridis\", xticklabels=False, yticklabels=False, vmin=value_min, vmax=value_max)\n",
+    "            plt.title(f\"Heatmap for {city}\")\n",
+    "            plt.savefig(f'{output_dir}/{dataset_name}_heatmap.png')\n",
+    "            plt.close()\n",
+    "            \n",
+    "\n",
+    "            # 计算每个时间点的平均值和标准差\n",
+    "            means = df_processed.mean(axis=1)\n",
+    "            std_devs = df_processed.std(axis=1)\n",
+    "\n",
+    "            # # 绘制平均值和标准差图\n",
+    "            # plt.figure(figsize=(20, 6))\n",
+    "            # plt.plot(means.index, means, label='Mean', color='red')\n",
+    "            # plt.fill_between(means.index, means - std_devs, means + std_devs, color='orange', alpha=0.5, label='Standard Deviation')\n",
+    "            # plt.title(f'{city}')\n",
+    "            # plt.xlabel('Date')\n",
+    "            # plt.ylabel('Values')\n",
+    "            # plt.legend()\n",
+    "            # plt.grid(True)\n",
+    "\n",
+    "            # # 保存平均值图\n",
+    "            # plt.savefig(f'{output_dir}/{dataset_name}_vis.png')\n",
+    "            # plt.close()\n",
+    "            \n",
+    "            # 绘制平均值和标准差图\n",
+    "            plot_with_discontinuities(means, std_devs, f'{output_dir}/{dataset_name}_vis.png', title=f'{city}')"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "geoai",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

data_analysis/HA.ipynb

@@ -0,0 +1,236 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Load Data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pickle\n",
+    "import torch\n",
+    "import numpy as np"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def masked_mse(preds, labels, null_val):\n",
+    "    if torch.isnan(null_val):\n",
+    "        mask = ~torch.isnan(labels)\n",
+    "    else:\n",
+    "        mask = (labels != null_val)\n",
+    "    mask = mask.float()\n",
+    "    mask /= torch.mean((mask))\n",
+    "    mask = torch.where(torch.isnan(mask), torch.zeros_like(mask), mask)\n",
+    "    loss = (preds - labels)**2\n",
+    "    loss = loss * mask\n",
+    "    loss = torch.where(torch.isnan(loss), torch.zeros_like(loss), loss)\n",
+    "    return torch.mean(loss)\n",
+    "\n",
+    "\n",
+    "def masked_rmse(preds, labels, null_val):\n",
+    "    return torch.sqrt(masked_mse(preds=preds, labels=labels, null_val=null_val))\n",
+    "\n",
+    "\n",
+    "def masked_mae(preds, labels, null_val):\n",
+    "    if torch.isnan(null_val):\n",
+    "        mask = ~torch.isnan(labels)\n",
+    "    else:\n",
+    "        mask = (labels != null_val)\n",
+    "    mask = mask.float()\n",
+    "    mask /= torch.mean((mask))\n",
+    "    mask = torch.where(torch.isnan(mask), torch.zeros_like(mask), mask)\n",
+    "    loss = torch.abs(preds - labels)\n",
+    "    loss = loss * mask\n",
+    "    loss = torch.where(torch.isnan(loss), torch.zeros_like(loss), loss)\n",
+    "    return torch.mean(loss)\n",
+    "\n",
+    "\n",
+    "def masked_mape(preds, labels, null_val):\n",
+    "    if torch.isnan(null_val):\n",
+    "        mask = ~torch.isnan(labels)\n",
+    "    else:\n",
+    "        mask = (labels != null_val)\n",
+    "    mask = mask.float()\n",
+    "    mask /= torch.mean((mask))\n",
+    "    mask = torch.where(torch.isnan(mask), torch.zeros_like(mask), mask)\n",
+    "    loss = torch.abs(preds - labels) / labels\n",
+    "    loss = loss * mask\n",
+    "    loss = torch.where(torch.isnan(loss), torch.zeros_like(loss), loss)\n",
+    "    return torch.mean(loss)\n",
+    "\n",
+    "\n",
+    "def compute_all_metrics(preds, labels, null_val):\n",
+    "    mae = masked_mae(preds, labels, null_val).item()\n",
+    "    mape = masked_mape(preds, labels, null_val).item()\n",
+    "    rmse = masked_rmse(preds, labels, null_val).item()\n",
+    "    return mae, mape, rmse"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "datasets_list = ['pems03_flow', 'pems04_flow', 'pems07_flow', 'pems08_flow', 'occpairs_occupancy', 'occhamburg_occupancy', 'pemsbay_speed', 'metrla_speed', 'trafficsh_speed', 'bikenyc_inflow', 'taxinyc_inflow', 'tdrive_inflow']\n",
+    "folder_path = \"/data/weichen/ST-Library/datasets/eval_datasets\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for few_shot_ratio in [0.1, 1.0]:\n",
+    "    print('-'*150)\n",
+    "    if few_shot_ratio == 0.1:\n",
+    "        print(f\"\\t\\t\\t\\t\\t\\t eval_shot: Few-shot\")\n",
+    "    else:\n",
+    "        print(f\"\\t\\t\\t\\t\\t\\t eval_shot: Full-shot\")\n",
+    "    print('-'*150)\n",
+    "    for dataset in datasets_list:\n",
+    "        \n",
+    "        print('*'*30)\n",
+    "        print(f\"Dataset: {dataset}\")\n",
+    "        print('*'*30)\n",
+    "        \n",
+    "        for num_steps in [12, 24]:\n",
+    "        \n",
+    "            with open(f\"{folder_path}/{dataset}/{dataset}_temporal.pkl\", 'rb') as f:\n",
+    "                df = pickle.load(f)\n",
+    "                raw_temporal_data = torch.tensor(df.values).unsqueeze(-1)\n",
+    "                f.close()\n",
+    "\n",
+    "            T, N, _ = raw_temporal_data.shape\n",
+    "\n",
+    "            # Add time-based features if specified\n",
+    "            feature_list = [raw_temporal_data]\n",
+    "\n",
+    "            # add_time_of_day\n",
+    "            time_ind = (df.index.values - df.index.values.astype('datetime64[D]')) / np.timedelta64(1, 'D') * 288\n",
+    "            time_of_day = np.tile(time_ind, [1, N, 1]).transpose((2, 1, 0))\n",
+    "            feature_list.append(torch.tensor(time_of_day, dtype=torch.float32))\n",
+    "\n",
+    "            # add_day_of_week:\n",
+    "            dow = df.index.dayofweek\n",
+    "            dow_tiled = np.tile(dow, [1, N, 1]).transpose((2, 1, 0))\n",
+    "            day_of_week = dow_tiled / 7 * 7\n",
+    "            feature_list.append(torch.tensor(day_of_week, dtype=torch.float32))\n",
+    "\n",
+    "            temporal_data = torch.cat(feature_list, dim=-1).numpy()  # Concatenate features along the channel dimension\n",
+    "\n",
+    "            # train_rate = few_shot_ratio if few_shot_ratio <= train_val_test_rate[0] else train_val_test_rate[0]\n",
+    "            train_rate = 0.6 * few_shot_ratio\n",
+    "            valid_rate = 0.2\n",
+    "            test_rate = 0.2\n",
+    "\n",
+    "            train = temporal_data[:int(T * train_rate)]\n",
+    "            test = temporal_data[int(T * (1 - test_rate)):]\n",
+    "            \n",
+    "            # Create a dictionary to store lists of observations for each (tod, dow) pair\n",
+    "            history = {}\n",
+    "            # Iterate through the historical data to populate the history dictionary\n",
+    "            for i in range(train.shape[0]-num_steps-1):\n",
+    "                \n",
+    "                key = (train[i, 0, 1], train[i, 0, 2])  # (tod, dow) tuple\n",
+    "                if key not in history:\n",
+    "                    history[key] = []\n",
+    "                    \n",
+    "                # Collect observations for the next num_steps steps\n",
+    "                future_values = train[i+1:i+num_steps+1,:,0]\n",
+    "                \n",
+    "                history[key].append(future_values)  # Collect and store future values\n",
+    "            \n",
+    "            history_averages = {}\n",
+    "\n",
+    "            # Calculate the average values for each (tod, dow) pair\n",
+    "            for k, v in history.items():\n",
+    "                stacked_array = np.stack(v, axis=-1)\n",
+    "                averages = np.mean(stacked_array, axis=-1)\n",
+    "                history_averages[k] = averages\n",
+    "            \n",
+    "            \n",
+    "            preds = []\n",
+    "            labels = []\n",
+    "\n",
+    "            # Retrieve labels and predictions\n",
+    "            for i in range(test.shape[0]-num_steps-1):\n",
+    "\n",
+    "                key = (test[i, 0, 1], test[i, 0, 2])  # (tod, dow) tuple\n",
+    "                \n",
+    "                labels.append(test[i+1:i+num_steps+1,:,0])\n",
+    "                \n",
+    "                try:\n",
+    "                    preds.append(history_averages[key])\n",
+    "                except KeyError:\n",
+    "                    # If the (tod, dow) pair is not present in the history dictionary, predict last point\n",
+    "                    preds.append(torch.tensor(test[i,:,0]).expand(num_steps, -1))\n",
+    "                \n",
+    "            labels = np.stack(labels, axis=0)\n",
+    "            preds = np.stack(preds, axis=0)\n",
+    "            \n",
+    "            labels = torch.Tensor(labels)\n",
+    "            preds = torch.Tensor(preds)\n",
+    "\n",
+    "            # handle the precision issue when performing inverse transform to label\n",
+    "            mask_value = torch.tensor(0)\n",
+    "\n",
+    "            test_mae = []\n",
+    "            test_mape = []\n",
+    "            test_rmse = []\n",
+    "\n",
+    "            # Calculate metrics\n",
+    "            for i in range(12):\n",
+    "                res = compute_all_metrics(preds[:,i,:], labels[:,i,:], mask_value)\n",
+    "                test_mae.append(res[0])\n",
+    "                test_mape.append(res[1] * 100)\n",
+    "                test_rmse.append(res[2])\n",
+    "\n",
+    "            mae_mean = np.mean(test_mae)\n",
+    "            mae_std = 0\n",
+    "            rmse_mean = np.mean(test_rmse)\n",
+    "            rmse_std = 0\n",
+    "            mape_mean = np.mean(test_mape)\n",
+    "            mape_std = 0\n",
+    "            \n",
+    "            if num_steps == 12:\n",
+    "                print('HA - Short Forecasting' + \"\\t\\t MAE:\" + \"& $\" + f\"{mae_mean:.2f}\" + \"\\\\textcolor{gray}{\\\\text{\\scriptsize±\" + f'{mae_std:.2f}'+\"}}$\" + \"\\t\\t RMSE:\" + \"& $\" + f\"{rmse_mean:.2f}\" + \"\\\\textcolor{gray}{\\\\text{\\scriptsize±\" + f'{rmse_std:.2f}'+\"}}$\" + \"\\t\\t MAPE:\" + \"& $\" + f\"{mape_mean:.2f}\" + \"\\\\textcolor{gray}{\\\\text{\\scriptsize±\" + f'{mape_std:.2f}'+\"}}$\")\n",
+    "            else:\n",
+    "                print('HA - Long Forecasting' + \"\\t\\t MAE:\" + \"& $\" + f\"{mae_mean:.2f}\" + \"\\\\textcolor{gray}{\\\\text{\\scriptsize±\" + f'{mae_std:.2f}'+\"}}$\" + \"\\t\\t RMSE:\" + \"& $\" + f\"{rmse_mean:.2f}\" + \"\\\\textcolor{gray}{\\\\text{\\scriptsize±\" + f'{rmse_std:.2f}'+\"}}$\" + \"\\t\\t MAPE:\" + \"& $\" + f\"{mape_mean:.2f}\" + \"\\\\textcolor{gray}{\\\\text{\\scriptsize±\" + f'{mape_std:.2f}'+\"}}$\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.13"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

data_analysis/benchmark_st_vis.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4c47c27a3770b1690d747b6ff466099109417fa873eac39af59da40e7ad1ff2
+size 838357

data_analysis/benchmark_t_vis.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd499502c6aea63045684f99e856381695553eaeef23216891e20ba150a99d78
+size 581725

data_index_process.sh

@@ -0,0 +1,12 @@
+# python spatial_index_generation.py --d 16 --type eval_datasets
+# python spatial_index_generation.py --d 32 --type eval_datasets
+# python spatial_index_generation.py --d 48 --type eval_datasets
+# python spatial_index_generation.py --d 64 --type eval_datasets
+python spatial_index_generation.py --d 128 --type eval_datasets
+
+# python spatial_index_generation.py --d 16 --type pretrain_datasets
+# python spatial_index_generation.py --d 32 --type pretrain_datasets
+# python spatial_index_generation.py --d 48 --type pretrain_datasets
+# python spatial_index_generation.py --d 64 --type pretrain_datasets
+python spatial_index_generation.py --d 128 --type pretrain_datasets
+

eval_datasets/bikenyc_inflow/bikenyc_inflow_spatial.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef3174379a18e5585099949f2b67a8860ce727f17fe1b86ff7117c34ba5ad994
+size 5572

eval_datasets/bikenyc_inflow/bikenyc_inflow_temporal.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:570da74f91d9fa3d3b2b903c825ba9f9c64b44fd15603fd0e6a1e4a8aaf2bcb9
+size 27780944

eval_datasets/bikenyc_inflow/index_128.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4315466774eed4bba7c3cefb9c6ee94e57a11f04750ce7d64e7e616e1b039a7b
+size 4973

eval_datasets/bikenyc_inflow/index_16.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:914fcbc2d4c49a5bd3194e613076e810e59b55a6fd5500da1c45b670f47e09b5
+size 4105

eval_datasets/bikenyc_inflow/index_32.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:973948b9750b1408b45d696e90ea63c8473fc38ce015652bdbc7a8714379756c
+size 4385

eval_datasets/bikenyc_inflow/index_48.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb9ff5255ba381e7826de09ef23ae89aaf45a36de0164aab6adf2419ccbf70b2
+size 4681

eval_datasets/bikenyc_inflow/index_64.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c00e99520c421fcce466c21c22f6acfb0da5bbc14f055bfe4266705302a2ae63
+size 4981

eval_datasets/bikenyc_inflow/mask_128.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:422b4ed8a2ed11a296973e98cc6de0b910b578a2df3eba5365ab331e09f8c3db
+size 536

eval_datasets/bikenyc_inflow/mask_16.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc9dbddd4321b8e28b591274ffd4a99fe590f1a1fe0bfef20dd167da30904ee2
+size 484

eval_datasets/bikenyc_inflow/mask_32.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c013af850ae19e4610913c6011bf80b40b83ff9eac9acf4fc4003bb6550da181
+size 492

eval_datasets/bikenyc_inflow/mask_48.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26c50d8ae300b489f553855922edeb78972bf0743e46b0cb0b0ed95b1b32aa19
+size 516

eval_datasets/bikenyc_inflow/mask_64.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:253fb74aa8c3ba337732607c59819b2bf522a3b5de0ba632e94c3010d2b8119a
+size 544

eval_datasets/metrla_speed/index_128.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22ca47e52be2dd7813e6518243419c73784ae96142207bc436bd822b6e7bb68a
+size 4973

eval_datasets/metrla_speed/index_16.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ebc93159fea1b6c5c7a3159822d58fdd023e8c2e2947fa4917a2a027be89874
+size 4105

eval_datasets/metrla_speed/index_32.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96cf3cb10137d9001f897b337639caab21a2cea61f1a4e9bd0f907d74de2a7ea
+size 4385

eval_datasets/metrla_speed/index_48.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ca7cc0500a42cf1e6fbb927250f6350794ba3f9738664f6ba6c708cfaaab2bb
+size 4681

eval_datasets/metrla_speed/index_64.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2aedbd0a08527d40baf4915ed80dad939417e827d53e3daa5051d6a027a5b6f1
+size 4981

eval_datasets/metrla_speed/mask_128.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b271f78191d99491e2c3e5f3368aff659f837e4b62fcc2e6d0ef5db9f85e36a8
+size 536

eval_datasets/metrla_speed/mask_16.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c9feba94b149b4ea8ca227fc0d7341d643295295b64e86ff9d578386631d14d
+size 484

eval_datasets/metrla_speed/mask_32.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cde7b7c0184ba9a91ccaae8545b1f7fd50e059af25f4eae883e9abc4e6aefc54
+size 492

eval_datasets/metrla_speed/mask_48.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60d0dbf964457b15dd712b614e36bd09e60fb47a3874751feea306e6710cf93b
+size 516

eval_datasets/metrla_speed/mask_64.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee89d83d9ed85782c5b25d39ccb21c762edf3503d1c624fa1121a234012e865e
+size 544

eval_datasets/metrla_speed/metrla_speed_spatial.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85731c8d285cd2f29bf09de171ff904302df7c275975fa1a443d22c197e03372
+size 5759

eval_datasets/metrla_speed/metrla_speed_temporal.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:265017d936e8135b625db791888a7a883e320979acceebbb54b6fa5e75ca5403
+size 57031322

eval_datasets/occhamburg_occupancy/index_128.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2dabe08547eac077143ce0724e806b3ba1f9a338657384105bf14e94d63bff7c
+size 4973