[model] MT-LSTM
Reply
Category Archives: CC0
[model] AE-MT-LSTM
[model] AE-WIFI-LSTM
[model] UWB Model
[dataset] Self-Play Game Records of Go for Professional Level I
CGI圍棋自我對弈棋譜 #307 - #403
- AlphaZero 訓練時所產生的自我對弈棋譜，共計 1,000,791 譜
- 在與計畫主持人合作前提下，有條件免費使用
- 資料集說明：
- 本資料集中包含多個 .sgf 檔案，每個 .sgf 檔案皆以 gzip 格式壓縮，存取前需要使用 gzip 指令解壓縮，如 gzip -d 307.sgf.gz
- 每個 .sgf 檔案中包含多場遊戲的紀錄 (episode)；每場遊戲的紀錄以一行表示，每行的格式為 ID LEN SGF
- ID 代表此場遊戲的序號
- LEN 代表此場遊戲中包含的棋步數量
- SGF 為此場遊戲的詳細記錄，內部包含多個 SGF 資料欄位，依出現順序說明如下：
| 欄位 | 說明 |
|---|---|
| FF | SGF 格式版號，出現於檔案開頭，目前使用 FF[4] |
| CA | 文件編碼，目前使用 CA[UTF-8] |
| DT | 產生此 SGF 紀錄的時間，如 DT[2022/07/08] |
| EV | 此盤棋譜在不同貼目下的勝率，從前至後分別為貼目-8.5至貼目22.5的勝率 |
| KM | 貼目 (Komi)，如 KM[7.5] 表示貼目為 7.5 目 |
| PB | 黑方的玩家名稱，例如 PB[CGI] |
| PW | 白方的玩家名稱，例如 PB[CGI] |
| RE | 該局對弈結果，第一個字母表示勝利者，例如 RE[B+2.5] 表示黑贏 2.5 目、RE[W+R] 表示黑認輸、RE[0] 表示平手 |
| SZ | 棋盤大小，例如 SZ[19] 表示 19 路棋盤 |
| B | 黑方一步棋，以棋盤左上角為基準，以小寫英文字母 a, b, ... 表示橫軸與縱軸位置，例如 ;B[pd] 表示黑子下在棋盤右上角星位 |
| W | 白方一步棋，格式同上 |
| C | 註解；於本資料集，用於紀錄該棋步的程式搜尋結果資訊，格式為 "棋步:搜尋次數,棋譜:搜尋次數,...,*其餘相關資訊"，例如 "A1:5,B1:6,...,*..." 表示A1搜尋5次、B1搜尋6次，以此類推 |
[dataset] iVS-Dataset 台灣地區之自駕車視角圖資
iVS-Dataset Statement
The dataset annotates the four commonly seen object types into four different categories of objects namely:
1.scooter
2.vehicle
3.pedestrian
4.bicycle
The iVS dataset consists of 89,002 annotated training images and 6,400 annotated testing images of resolution 1920x1080
目前本團隊已經公開89002張訓練資料與6400測試資料台灣
地區交通場域之行人、機車、自行車與汽車之標記資料庫，
免費提供學術界與產業界人士使用，進一步加速台灣地區深
度學習ADAS技術開發時程。
標註資訊包括物件類別與物件位置框。
iVS-ODSEG-Dataset Statement
We provide 36,000 annotated 1280x720 training images, which part of them are revised from the original BDD100K semantic segmentation dataset with additional object bounding box and classes labels and part of them are our custom semantic segmentation dataset with additional object bounding box and classes labels.
1.background
2.main_lane
3.alter_lane
4.double_line
5.dashed_line
6.single_line
[dataset] UWB dataset
資料集內含使用UWB超寬頻技術的三角定位軌跡，以及真實軌跡座標。
| Column Name | Description |
|---|---|
| D1 | The Distance between UWB base station 1 and receiver. The coordinate of base station 1: (0, 0) |
| D2 | The Distance between UWB base station 2 and receiver. The coordinate of base station 2: (300, 200) |
| D3 | The Distance between UWB base station 3 and receiver. The coordinate of base station 3: (0, 400) |
| rX | According to D1, D2, D3, the receiver uses the x-axis coordinate calculated by the triangulation method |
| rY | According to D1, D2, D3, the receiver uses the y-axis coordinate calculated by the triangulation method |
| gX | Receiver real x-axis coordinate |
| gY | Receiver real y-axis coordinate |
[dataset] smartphone wifi RSSI、GPS and multi-sensor fingerprinting data (1203_NYCU)
The dataset includes wifi RSSI data, GPS data,IMU sensor data and the position information collected by smartphone. These data are for training boundary localization AI models.
[dataset] smartphone wifi RSSI、GPS and multi-sensor fingerprinting data (0508_UIA)
The dataset includes wifi RSSI data, GPS data,IMU sensor data and the position information collected by smartphone. These data are for training boundary localization AI models.
[dataset] Self-Play Game Records of Connect6 for Advanced Level I
Self-Play Game Records of Connect6 for Advanced Level I
-
A dataset of game records of Connect6 programs for advanced level, generated by AlphaZero self-play learning, including 1,024,027 moves in total.
- SGF #1000 ~ #1024, including 1,024,027 moves in total (for advanced level).
-
Description:
- Each *.sgf file contains multiple games; each game is saved as a single line with data fields as described below.
| Field | Description |
|---|---|
| GM | The game type, e.g., GM[7] for this dataset |
| FF | The SGF format version, appears at the first of a SGF file, e.g., FF[4] for this dataset |
| B | A move of black, encoding with characters A, B, ..., e.g., ;B[PD] means a move of black at Q16 |
| W | A move of white, the format is the same as that of black |