Add precomputed metadata for efficient training

README.md

@@ -10,18 +10,42 @@ dataset_info:
       sequence: float64
   - name: valid_mask
     sequence: bool
+  - name: metadata
+    struct:
+    - name: contact_density
+      dtype: float64
+    - name: has_valid_structure
+      dtype: bool
+    - name: invalid_residues
+      dtype: int64
+    - name: long_range_contacts
+      dtype: int64
+    - name: medium_range_contacts
+      dtype: int64
+    - name: seq_length
+      dtype: int64
+    - name: short_range_contacts
+      dtype: int64
+    - name: token_count
+      dtype: int64
+    - name: tokenization_ratio
+      dtype: float64
+    - name: total_contacts
+      dtype: int64
+    - name: valid_residues
+      dtype: int64
   splits:
   - name: train
-    num_bytes: 167543183
+    num_bytes: 169570266
     num_examples: 25299
   - name: validation
-    num_bytes: 1333050
+    num_bytes: 1350998
     num_examples: 224
   - name: test
-    num_bytes: 353260
+    num_bytes: 356465
     num_examples: 40
-  download_size: 96000751
-  dataset_size: 169229493
+  download_size: 96822838
+  dataset_size: 171277729
 configs:
 - config_name: default
   data_files:

data/test-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a359079d2cc081561b3c0078472639ccfc5e9f3fab7c014de59a2b9c2e02e1a
-size 249682
+oid sha256:70e4ac87fa2244a7956711f95cc10a782be9b4815f8566b94c0c860d40c8883b
+size 256673

data/train-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b35f0e35c6dc8f8cd12076619a35f3c7ef9dbda321fa10d356975e1927d309b
-size 94929947
+oid sha256:5100759304a7f7ee8884f01453f1fcdcc95cdbd3b9ddcad17ef2b9f7ec10a42f
+size 95731674

data/validation-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6889ca472706bdf0f5020f4e6196fc18e39d0724c2ba2bed5c40563fa082b2a8
-size 821122
+oid sha256:b7d95fc3582f3ad1c5ce92f62c73d7f24c5477a013ed1eff0db1c8dfb2e2663f
+size 834491