Add reproducible tokbench v1 model inputs

.gitattributes

@@ -71,3 +71,6 @@ gpt-oss.json filter=lfs diff=lfs merge=lfs -text
 glm-5.2.json filter=lfs diff=lfs merge=lfs -text
 gemma-4.json filter=lfs diff=lfs merge=lfs -text
 mistral-small-4.json filter=lfs diff=lfs merge=lfs -text
+models/glm-5.2/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+models/gpt-oss/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+models/nemotron-3/tokenizer.json filter=lfs diff=lfs merge=lfs -text

models/glm-5.2/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19e773648cb4e65de8660ea6365e10acca112d42a854923df93db4a6f333a82d
+size 20217442

models/gpt-oss/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0614fe83cadab421296e664e1f48f4261fa8fef6e03e63bb75c20f38e37d07d3
+size 27868174

models/manifest.json

@@ -0,0 +1,55 @@
+{
+  "corpus_revision": "54d012f11a4911b797aa6ef2dffc0fb67cbc7619",
+  "models": {
+    "deepseek-v4": {
+      "sha256": "8f9f37ca37fdc4f5fd36d5cf4d3b0e8392edb4e894fd10cc0d70b4957c8633cf",
+      "source_file": "deepseek-v4.json",
+      "source_repo": "hf-internal-testing/tokenizers-test-data",
+      "source_revision": "54d012f11a4911b797aa6ef2dffc0fb67cbc7619"
+    },
+    "glm-5.2": {
+      "sha256": "19e773648cb4e65de8660ea6365e10acca112d42a854923df93db4a6f333a82d",
+      "source_file": "glm-5.2.json",
+      "source_repo": "hf-internal-testing/tokenizers-test-data",
+      "source_revision": "54d012f11a4911b797aa6ef2dffc0fb67cbc7619"
+    },
+    "gpt-oss": {
+      "sha256": "0614fe83cadab421296e664e1f48f4261fa8fef6e03e63bb75c20f38e37d07d3",
+      "source_file": "gpt-oss.json",
+      "source_repo": "hf-internal-testing/tokenizers-test-data",
+      "source_revision": "54d012f11a4911b797aa6ef2dffc0fb67cbc7619"
+    },
+    "gpt2": {
+      "sha256": "8414cab924d8b9b33013f0d221c5862f365ee9be39c5c2bfae8a5a9e970478a6",
+      "source_file": "gpt2.json",
+      "source_repo": "hf-internal-testing/tokenizers-test-data",
+      "source_revision": "54d012f11a4911b797aa6ef2dffc0fb67cbc7619"
+    },
+    "llama-3": {
+      "sha256": "76e48799b099d43365bd24ccd8ecc5aedac831718da780552f03b0a6eb4412aa",
+      "source_file": "llama-3-tokenizer.json",
+      "source_repo": "hf-internal-testing/tokenizers-test-data",
+      "source_revision": "54d012f11a4911b797aa6ef2dffc0fb67cbc7619"
+    },
+    "minimax": {
+      "sha256": "757622126525aeeb131756849d93298070ff3f0319c455ec8c5bb0f6b1cebbe8",
+      "source_file": "tokenizer.json",
+      "source_repo": "MiniMaxAI/MiniMax-M2",
+      "source_revision": "757303d492a50514c312788b5247a4f696a4c6a3"
+    },
+    "nemotron-3": {
+      "sha256": "c6021eb6847e682f89aa52d5eb6e8c7d902a23acfc8137e25211cf84828f1592",
+      "source_file": "tokenizer.json",
+      "source_repo": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16",
+      "source_revision": "2d59de1cbd51c0adf384eb906b766d1aee0e0517"
+    },
+    "qwen2": {
+      "sha256": "f7c9b2dba4a296b1aa76c16a34b8225c0c118978400d4bb66bff0902d702f5b8",
+      "source_file": "tokenizer.json",
+      "source_repo": "Qwen/Qwen2-7B",
+      "source_revision": "453ed1575b739b5b03ce3758b23befdb0967f40e"
+    }
+  },
+  "purpose": "Exact tokenizer inputs used by the tokenizers v1 blog benchmark. All eight were verified against the retained run on eng_Latn and added_normalized_dense by token count and FNV-1a ID-stream hash.",
+  "schema_version": 1
+}

models/nemotron-3/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6021eb6847e682f89aa52d5eb6e8c7d902a23acfc8137e25211cf84828f1592
+size 17077485