benchmark eba8e0f953ffead6

.gitattributes

@@ -58,3 +58,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+probe_subset.i64 filter=lfs diff=lfs merge=lfs -text
+train_families.json filter=lfs diff=lfs merge=lfs -text
+train_offsets.i64 filter=lfs diff=lfs merge=lfs -text
+train_three_di.u8 filter=lfs diff=lfs merge=lfs -text
+train_tokens.u8 filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,60 @@
+---
+license: cc-by-4.0
+pretty_name: PLM speedrun, AFDB v6
+tags:
+- protein
+- masked-language-modeling
+- benchmark
+---
+
+# PLM speedrun, AFDB v6
+
+The fixed data of the SpeedrunningPLMs ten-minute protein masked language model benchmark
+([code](https://github.com/Synthyra/SpeedrunningPLMs)): AlphaFold DB proteins from the family-disjoint split of the
+Alignment Atlas (`Synthyra/alignment_atlas_families_afdb_v6` at `83737fa33e72`), tokenized once in the ESM2 alphabet, with the
+held-out splits' 15 percent masks materialized so every model is scored on the same hidden residues. No held-out
+protein has a training relative at 30 percent identity over 80 percent coverage by sequence (MMseqs2) or by structure
+(Foldseek). Benchmark id `eba8e0f953ffead6`: the SHA-256 of this manifest, carried by every run's result.
+
+## Objective
+
+A model reads a protein (CLS, residues, EOS) with 15 percent of the standard residues replaced by MASK and nothing
+else, and predicts the hidden residues; the loss is the cross entropy over the 20 standard residues on the masked
+positions. `val_select` (one protein per validation family, truncated to 1,024 tokens) selects weights in a
+run; validation and test are scored at full length (up to 2,048 tokens). The train stream is one
+uint8 token stream of 1,754,034 proteins (392,592,221 tokens, 333,143 families) in a seeded
+shuffled order, with document offsets and the Foldseek 3Di state of every residue beside it; the 3Di states are never
+an input, only the target of a reported structure probe on a fixed 20,000-protein subset.
+
+## Splits
+
+| Split | Proteins | Families | Residues | Scored (masked) residues | Width |
+|---|---|---|---|---|---|
+| test | 25,778 | 4,807 | 6,098,489 | 915,208 | 1,927 |
+| val_select | 4,751 | 4,751 | 1,066,372 | 156,231 | 1,024 |
+| validation | 25,459 | 4,751 | 5,919,918 | 889,145 | 1,978 |
+
+## Files
+
+| File | Bytes |
+|---|---|
+| `probe_subset.i64` | 160,000 |
+| `test.pt` | 149,531,512 |
+| `train_families.json` | 23,938,324 |
+| `train_offsets.i64` | 14,032,280 |
+| `train_three_di.u8` | 392,592,221 |
+| `train_tokens.u8` | 392,592,221 |
+| `val_select.pt` | 14,766,458 |
+| `validation.pt` | 151,576,442 |
+
+Every file's SHA-256 is in `manifest.json`; the loaders verify them.
+
+`train_tokens.u8` and `train_three_di.u8` are flat uint8 streams, `train_offsets.i64` the int64 start of every
+protein, `probe_subset.i64` the probe subset's indices, `train_families.json` the family of every training protein;
+`val_select.pt`, `validation.pt` and `test.pt` hold padded `input_ids`, `masked`, `three_di`, `family` and `lengths`
+tensors and the list of `families`. `speedrunning_plms.benchmark.load_train` and `load_held_out` read them and verify the digests.
+
+## Source and terms
+
+AlphaFold DB sequences (CC-BY-4.0) and Foldseek 3Di states, through the Alignment Atlas family split; the source
+dataset's taxonomy is not audited here (see the workspace record).

manifest.json

@@ -0,0 +1,110 @@
+{
+  "files": {
+    "probe_subset.i64": {
+      "bytes": 160000,
+      "sha256": "5cc6afa12e994e3209c751c4070fcf612da386c589972a00465b1a9202dcf334"
+    },
+    "test.pt": {
+      "bytes": 149531512,
+      "sha256": "c3de0d7e028b84868dec49e7f72220b4b62c26ede74e5911ceb1d1093d9cefe0"
+    },
+    "train_families.json": {
+      "bytes": 23938324,
+      "sha256": "4b558fb4cfe9e80a18db4a6e987efd5bc1796b481a26753e767268e7a081fe50"
+    },
+    "train_offsets.i64": {
+      "bytes": 14032280,
+      "sha256": "04d945fe21edab918dc975a29b5f4b60309d98c427bec7035781b71d30322887"
+    },
+    "train_three_di.u8": {
+      "bytes": 392592221,
+      "sha256": "2c3f6077a1fdfcd73468d42d1cfbd6e5a84f9be8baac59988c55fbfb27451f6d"
+    },
+    "train_tokens.u8": {
+      "bytes": 392592221,
+      "sha256": "7558e3abbef2a3fc0efb1641b3b3fd2a9a2e718fb8fde9c36c122d7c6d601b57"
+    },
+    "val_select.pt": {
+      "bytes": 14766458,
+      "sha256": "4f4aa1da183d35e7dd7c2a8bbc5520b11f13b9bc6574a5ee9a909e2d8043c238"
+    },
+    "validation.pt": {
+      "bytes": 151576442,
+      "sha256": "34db70938ae0c0c92d7ea10467f0c8330f0e53a530a672269de6f74074cb9ccf"
+    }
+  },
+  "length_bins": [
+    [
+      50,
+      100
+    ],
+    [
+      100,
+      250
+    ],
+    [
+      250,
+      500
+    ],
+    [
+      500,
+      1000
+    ],
+    [
+      1000,
+      2047
+    ]
+  ],
+  "mask_seed": 20261028,
+  "max_tokens": {
+    "test": 2048,
+    "val_select": 1024,
+    "validation": 2048
+  },
+  "objective": {
+    "classes": "ACDEFGHIKLMNPQRSTVWY",
+    "cross_entropy": "collapse_free",
+    "loss": "masked residues",
+    "mask_rate": 0.15,
+    "masked": "standard residues",
+    "replacement": "mask_only"
+  },
+  "schema_version": 2,
+  "shuffle_seed": 0,
+  "source": {
+    "repo_id": "Synthyra/alignment_atlas_families_afdb_v6",
+    "revision": "83737fa33e72b74faeb5b046a45a327f660ad116"
+  },
+  "splits": {
+    "test": {
+      "families": 4807,
+      "proteins": 25778,
+      "residues": 6098489,
+      "scored_residues": 915208,
+      "width": 1927
+    },
+    "val_select": {
+      "families": 4751,
+      "proteins": 4751,
+      "residues": 1066372,
+      "scored_residues": 156231,
+      "width": 1024
+    },
+    "validation": {
+      "families": 4751,
+      "proteins": 25459,
+      "residues": 5919918,
+      "scored_residues": 889145,
+      "width": 1978
+    }
+  },
+  "tokenizer": "esm2",
+  "train": {
+    "families": 333143,
+    "probe_subset": 20000,
+    "proteins": 1754034,
+    "residues": 389084153,
+    "tokens": 392592221
+  },
+  "vocab_size": 33
+}

probe_subset.i64

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cc6afa12e994e3209c751c4070fcf612da386c589972a00465b1a9202dcf334
+size 160000

test.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3de0d7e028b84868dec49e7f72220b4b62c26ede74e5911ceb1d1093d9cefe0
+size 149531512

train_families.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b558fb4cfe9e80a18db4a6e987efd5bc1796b481a26753e767268e7a081fe50
+size 23938324

train_offsets.i64

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04d945fe21edab918dc975a29b5f4b60309d98c427bec7035781b71d30322887
+size 14032280

train_three_di.u8

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c3f6077a1fdfcd73468d42d1cfbd6e5a84f9be8baac59988c55fbfb27451f6d
+size 392592221

train_tokens.u8

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7558e3abbef2a3fc0efb1641b3b3fd2a9a2e718fb8fde9c36c122d7c6d601b57
+size 392592221

val_select.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f4aa1da183d35e7dd7c2a8bbc5520b11f13b9bc6574a5ee9a909e2d8043c238
+size 14766458

validation.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34db70938ae0c0c92d7ea10467f0c8330f0e53a530a672269de6f74074cb9ccf
+size 151576442