Update image_bands to nested list format, make public

README.md

@@ -34,7 +34,10 @@ dataset_info:
     - name: label_name
       dtype: string
     - name: image_bands
-      dtype: binary
+      list:
+        list:
+          list:
+            dtype: float32
 ---
 
 # Galaxy10 AION-1 Benchmark
@@ -81,11 +84,11 @@ These are the exact indices used in the AION-1 paper. The test set produces the
 
 ### Reconstructing `image_bands`
 
-The `image_bands` column stores raw float32 flux values as bytes. To reconstruct:
+The `image_bands` column stores float32 flux values as a nested list (4 bands x 96 x 96 pixels). To reconstruct:
 
 ```python
 import numpy as np
-cutout = np.frombuffer(row["image_bands"], dtype=np.float32).reshape(4, 96, 96)
+cutout = np.array(row["image_bands"], dtype=np.float32)  # (4, 96, 96)
 # cutout[0] = g-band, cutout[1] = r-band, cutout[2] = i-band, cutout[3] = z-band
 ```
 
@@ -102,7 +105,7 @@ from aion.model import AION
 codec_manager = CodecManager(device="cuda")
 model = AION.from_pretrained("polymathic-ai/aion-base").to("cuda").eval()
 
-cutout = np.frombuffer(row["image_bands"], dtype=np.float32).reshape(4, 96, 96)
+cutout = np.array(row["image_bands"], dtype=np.float32)  # (4, 96, 96)
 image_flux = torch.tensor(cutout).unsqueeze(0).to("cuda")
 image = LegacySurveyImage(flux=image_flux, bands=["DES-G", "DES-R", "DES-I", "DES-Z"])
 tokens = codec_manager.encode(image)

data/test-00000-of-00002.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0175190b3b2b073ce26532e975a3ddb777a8efe09db2777fb407016efff21993
-size 117153124
+oid sha256:c20137463d746fe87bfb10ab7fe37300b7937a4e637b98c8738125c41f1c1901
+size 117825385

data/test-00001-of-00002.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f68559ba8cfb2f9693717f3008b5e9419265ae78e60e7599e906881560a349e3
-size 117425679
+oid sha256:3b87c03d6edfa6c9d900951285d0a0a862eeaa75cc818cbf644a104b14aa0d19
+size 118095541

data/train-00000-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36e9fc41e181e0d13fc448841234692e648c8505df1b473067338e003fd31655
+size 191764220

data/train-00001-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4fbad13c64ce93e19b6ae43ae2427aa7bc9eec7283425846442d56f1782142a9
+size 191167418

data/train-00002-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9723f89b3f32c9a6375b36854bc6f60f1d4f9d5917b315a73667b6489d189be
+size 190863581

data/train-00003-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb6ca59c5e552671df2c067e28e48870792edd84d121bfb8cd9c140847a7c7c
+size 192166852

data/train-00004-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9009b753c04053ce3667fdcf97863506f5fb72abcee266ede841d61930d4017f
+size 193150724

data/train-00005-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abbe1e483941ba1303012901287f19602773d9fef0db684ba769cc8a3f32528b
+size 191659413

data/train-00006-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bc3fc46aff61e317520f87d93addf0d145c8ec53411e1414f5c2eb957a2c747
+size 190292852

data/train-00007-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:deeeeb6a97218c2ac3a9375d53e02a7ae2796d868f7d826553d0dd2907dae738
+size 191045548

data/train-00008-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5321c14200ccdef597e86723f7d00c0db0a56f9fc94536bf8e438a670f6747c
+size 192296752

data/train-00009-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c0843f232621e4bdcb2b4fa12e0346c7a3639cb4530c604d2891d4a3f9856e6
+size 191201342

data/train-00010-of-00011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdb33d753307f85576729c479b7b65110b5f1b7bafcb0c99ff3e2bd453106f6b
+size 189853759