Add ESM2 UniRef short sequences (≤50 AA) with train/validation splits

README.md

@@ -0,0 +1,29 @@
+---
+dataset_info:
+  config_name: nvidia_esm2_uniref_pretraining_data_leq50aa
+  features:
+  - name: ur50_id
+    dtype: string
+  - name: ur90_id
+    dtype: string
+  - name: sequence
+    dtype: string
+  - name: length
+    dtype: int32
+  splits:
+  - name: train
+    num_bytes: 877
+    num_examples: 10
+  - name: validation
+    num_bytes: 852
+    num_examples: 10
+  download_size: 5735
+  dataset_size: 1729
+configs:
+- config_name: nvidia_esm2_uniref_pretraining_data_leq50aa
+  data_files:
+  - split: train
+    path: nvidia_esm2_uniref_pretraining_data_leq50aa/train-*
+  - split: validation
+    path: nvidia_esm2_uniref_pretraining_data_leq50aa/validation-*
+---

nvidia_esm2_uniref_pretraining_data_leq50aa/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:068e0a32b5ef8475d142d4e1690eee6d12d0fe083dca051eff34058b4bdd2075
+size 2885

nvidia_esm2_uniref_pretraining_data_leq50aa/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2cb2d5c524b39367c60a9c625f06964365a6094c932326db1eea59ef4b46b1de
+size 2850