Remove dataset data files (keep Croissant + datasheet)

README.md

@@ -1,32 +0,0 @@
----
-license: apache-2.0
-configs:
-- config_name: metadata
-  default: true
-  data_files:
-  - split: test
-    path: beans_next_metadata.parquet
-- config_name: audio
-  data_files:
-  - split: test
-    path: beans_next_audio.parquet
----
-
-# beans-next
-
-**beans-next** ships two root Parquet files:
-
-- **`beans_next_metadata.parquet`** — one row per evaluation sample (columns include ``subset``, ``tier``, ``sample_id``, and ``audio_id`` or ``audio_ids`` / ``query_audio_id`` for tier 4).
-- **`beans_next_audio.parquet`** — one row per **unique** audio blob (deduplicated by SHA-256); columns include ``audio_id``, ``sha256``, and ``audio_bytes`` (canonical WAV).
-
-Tier 3 can be added later by regenerating both Parquets and re-uploading.
-
-## Loading with ``datasets``
-
-```python
-from datasets import load_dataset
-
-meta = load_dataset("EarthSpeciesProject/beans-next", "metadata", split="test")
-audio = load_dataset("EarthSpeciesProject/beans-next", "audio", split="test")
-# Join on meta['audio_id'] etc. to audio['audio_id']
-```

beans_next_audio.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0bb36bc6822386e637f30ef90c5f68a763fe1f47dee5d50e4cec6b7a03f9db5b
-size 763363342

beans_next_metadata.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0cda777adc6eea46f6b806453d508e1acdd5b1aad725e2c407914be644f28bb9
-size 126170