Qwen90M step 690000; complete resumable checkpoint

complete.json

@@ -1,19 +1,19 @@
 {
-  "step": 685000,
+  "step": 690000,
   "files": {
     "config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7",
     "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4",
-    "model.safetensors": "2f198abaec427b39945f67e2dc4ae368c1e1f3bc749571b34df3aa6e1fabdbbc",
+    "model.safetensors": "060f0081aaffd4b21592f326ed75b02afab281b3ffe2de294ff5041c4fd8f70f",
     "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03",
     "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300",
     "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263",
     "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6",
-    "trainer.pt": "fef7bf8813882524f5ca30eee990ca62de8fcdd9a4aac08f179ed7a0bfb2f7cb",
-    "training_config.json": "13fe2f3f4e97181994bd5fcebc103bd2c14b5ea2ce9da3213db6d9bf51f1066f",
-    "performance.json": "5949473856ba60c6ea6e09958b880f6c16bf1db77e55259c9de9f88123633edf",
-    "shard_scheduler.json": "fa63f6790e325f60a69dde36bb42495987d8d2fe00e2528f2132581755f7dea1",
-    "coverage.json": "327c66438413b6fcd5c9a5d61ad54b0268663e1e6f30a1e6729f0395214b9ec3",
+    "trainer.pt": "24ae65d790f92deee8ff8c64623716e72d101b0c319d3796bed4c28c51aec73b",
+    "training_config.json": "d8df67626674cd77078628cbede1a700b576ca242c6f22a365ea49ba7ebdeea5",
+    "performance.json": "96c7d3993f53da33aeba321717c341456e5878e35efa676b6ffe17d014497e29",
+    "shard_scheduler.json": "dc02b0a99d28c9305b60760d51bc8f62064c31b5e9e3882cfe674a2fb80005a8",
+    "coverage.json": "fd32998195f03fdfef6945cb5182b881b8e2c1599651d3d35fa9078908d2cd99",
     "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"
   },
-  "signature": "33a459803c8cffa8929a90def998b4ad2c7c52fd1245365a9ee946b5552d845a"
+  "signature": "1dee8333ad826e9fb3d197c33535515cdb9de560b94f624cae15e044b464f0ac"
 }

coverage.json

@@ -1,107 +1,107 @@
 {
-  "step": 685000,
+  "step": 690000,
   "stats": {
-    "climbmix/blocks": 24660000,
-    "climbmix/targets": 41224942465,
-    "climbmix/input_tokens": 41227554983,
-    "ultra_qa/blocks": 9247500,
-    "ultra_qa/targets": 15086681728,
-    "ultra_qa/input_tokens": 15086710887,
-    "cortex/blocks": 9247500,
-    "cortex/targets": 351611655,
-    "cortex/input_tokens": 14699417911,
-    "ultra_style/blocks": 9247500,
-    "ultra_style/targets": 14626945782,
-    "ultra_style/input_tokens": 14627713723,
-    "openmath/blocks": 6165000,
-    "openmath/targets": 9526167364,
-    "openmath/input_tokens": 9529340892,
-    "rewrite6/blocks": 3082500,
-    "rewrite6/targets": 5053314593,
-    "rewrite6/input_tokens": 5058855101
+    "climbmix/blocks": 24840000,
+    "climbmix/targets": 41524942926,
+    "climbmix/input_tokens": 41527574676,
+    "ultra_qa/blocks": 9315000,
+    "ultra_qa/targets": 15196630228,
+    "ultra_qa/input_tokens": 15196659602,
+    "cortex/blocks": 9315000,
+    "cortex/targets": 354172942,
+    "cortex/input_tokens": 14806782028,
+    "ultra_style/blocks": 9315000,
+    "ultra_style/targets": 14733529377,
+    "ultra_style/input_tokens": 14734302959,
+    "openmath/blocks": 6210000,
+    "openmath/targets": 9595770485,
+    "openmath/input_tokens": 9598967555,
+    "rewrite6/blocks": 3105000,
+    "rewrite6/targets": 5090189858,
+    "rewrite6/input_tokens": 5095770661
   },
   "sources": {
     "climbmix": {
-      "page": 3377,
-      "row": 676,
+      "page": 3402,
+      "row": 485,
       "start": {
-        "file": 163,
+        "file": 164,
         "reader": {
           "format": "parquet-row-v1",
-          "row_group": 37,
+          "row_group": 54,
           "row": 1024
         }
       },
       "epoch": 0,
-      "offset": 4096,
-      "documents": 13832868
+      "offset": 12288,
+      "documents": 13935077
     },
     "rewrite6": {
-      "page": 21,
-      "row": 1829,
+      "page": 35,
+      "row": 1992,
       "start": {
         "file": 0,
         "reader": {
           "format": "jsonl-row-v1",
-          "row": 86016
+          "row": 143360
         }
       },
       "epoch": 38,
-      "offset": 0,
-      "documents": 7893197
+      "offset": 2048,
+      "documents": 7950704
     },
     "openmath": {
-      "page": 2110,
-      "row": 2518,
+      "page": 2126,
+      "row": 549,
       "start": {
         "file": 19,
         "reader": {
           "format": "parquet-row-v1",
-          "row_group": 346,
-          "row": 210
+          "row_group": 411,
+          "row": 746
         }
       },
       "epoch": 0,
       "offset": 0,
-      "documents": 8645078
+      "documents": 8708645
     },
     "ultra_qa": {
-      "page": 912,
-      "row": 2704,
+      "page": 919,
+      "row": 1367,
       "start": {
         "file": 7,
         "reader": {
           "format": "parquet-row-v1",
           "row_group": 1,
-          "row": 33616
+          "row": 62288
         }
       },
       "epoch": 0,
       "offset": 2048,
-      "documents": 3738256
+      "documents": 3765591
     },
     "ultra_style": {
-      "page": 1393,
-      "row": 671,
+      "page": 1403,
+      "row": 1688,
       "start": {
         "file": 8,
         "reader": {
           "format": "parquet-row-v1",
-          "row_group": 2,
-          "row": 56143
+          "row_group": 3,
+          "row": 11049
         }
       },
       "epoch": 0,
       "offset": 2048,
-      "documents": 5706399
+      "documents": 5748376
     },
     "cortex": {
-      "page": 179346,
-      "row": 60,
+      "page": 180655,
+      "row": 117,
       "start": null,
       "epoch": 0,
       "offset": 0,
-      "documents": 22956348
+      "documents": 23123957
     }
   }
 }

latest.json

@@ -1 +1 @@
-{"step": 685000, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "2f198abaec427b39945f67e2dc4ae368c1e1f3bc749571b34df3aa6e1fabdbbc", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "fef7bf8813882524f5ca30eee990ca62de8fcdd9a4aac08f179ed7a0bfb2f7cb", "training_config.json": "13fe2f3f4e97181994bd5fcebc103bd2c14b5ea2ce9da3213db6d9bf51f1066f", "performance.json": "5949473856ba60c6ea6e09958b880f6c16bf1db77e55259c9de9f88123633edf", "shard_scheduler.json": "fa63f6790e325f60a69dde36bb42495987d8d2fe00e2528f2132581755f7dea1", "coverage.json": "327c66438413b6fcd5c9a5d61ad54b0268663e1e6f30a1e6729f0395214b9ec3", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "33a459803c8cffa8929a90def998b4ad2c7c52fd1245365a9ee946b5552d845a", "folder": "step-000000685000"}
+{"step": 690000, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "060f0081aaffd4b21592f326ed75b02afab281b3ffe2de294ff5041c4fd8f70f", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "24ae65d790f92deee8ff8c64623716e72d101b0c319d3796bed4c28c51aec73b", "training_config.json": "d8df67626674cd77078628cbede1a700b576ca242c6f22a365ea49ba7ebdeea5", "performance.json": "96c7d3993f53da33aeba321717c341456e5878e35efa676b6ffe17d014497e29", "shard_scheduler.json": "dc02b0a99d28c9305b60760d51bc8f62064c31b5e9e3882cfe674a2fb80005a8", "coverage.json": "fd32998195f03fdfef6945cb5182b881b8e2c1599651d3d35fa9078908d2cd99", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "1dee8333ad826e9fb3d197c33535515cdb9de560b94f624cae15e044b464f0ac", "folder": "step-000000690000"}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f198abaec427b39945f67e2dc4ae368c1e1f3bc749571b34df3aa6e1fabdbbc
+oid sha256:060f0081aaffd4b21592f326ed75b02afab281b3ffe2de294ff5041c4fd8f70f
 size 360622744

performance.json

@@ -49,11 +49,11 @@
   "selected": {
     "microbatch": 90,
     "mode": "compiled",
-    "microbatch_seconds": 0.9403511730000105,
+    "microbatch_seconds": 0.9388940140000273,
     "peak_gb": 89.37790203094482,
     "fits_budget": true,
-    "estimated_update_seconds": 0.9403511730000105,
-    "input_tokens_per_second": 196011.87863887308
+    "estimated_update_seconds": 0.9388940140000273,
+    "input_tokens_per_second": 196316.08813302612
   },
   "measurements": [
     {

shard_scheduler.json

@@ -22,8 +22,8 @@
       "tokens_per_second": 598315.063954628,
       "note": "Bounded fresh sample; full-page latency is an extrapolation.",
       "estimated_pages_for_buffer": 1,
-      "required_token_positions_per_second": 75042.20393901474,
-      "required_pages_per_second_upper_estimate": 0.005191593790403495
+      "required_token_positions_per_second": 75790.38877363384,
+      "required_pages_per_second_upper_estimate": 0.005243354953290438
     },
     {
       "source": "rewrite6",
@@ -36,8 +36,8 @@
       "tokens_per_second": 8533.957213384612,
       "note": "Bounded fresh sample; full-page latency is an extrapolation.",
       "estimated_pages_for_buffer": 1,
-      "required_token_positions_per_second": 9380.275492376843,
-      "required_pages_per_second_upper_estimate": 0.00392732546623137
+      "required_token_positions_per_second": 9473.79859670423,
+      "required_pages_per_second_upper_estimate": 0.0039664816370287475
     },
     {
       "source": "openmath",
@@ -50,8 +50,8 @@
       "tokens_per_second": 380144.35761751566,
       "note": "Bounded fresh sample; full-page latency is an extrapolation.",
       "estimated_pages_for_buffer": 1,
-      "required_token_positions_per_second": 18760.550984753685,
-      "required_pages_per_second_upper_estimate": 0.003907328021124296
+      "required_token_positions_per_second": 18947.59719340846,
+      "required_pages_per_second_upper_estimate": 0.00394628481364685
     },
     {
       "source": "ultra_qa",
@@ -64,8 +64,8 @@
       "tokens_per_second": 43742.16556775435,
       "note": "Bounded fresh sample; full-page latency is an extrapolation.",
       "estimated_pages_for_buffer": 1,
-      "required_token_positions_per_second": 28140.826477130526,
-      "required_pages_per_second_upper_estimate": 0.0016869995787499043
+      "required_token_positions_per_second": 28421.395790112692,
+      "required_pages_per_second_upper_estimate": 0.0017038192806586498
     },
     {
       "source": "ultra_style",
@@ -78,8 +78,8 @@
       "tokens_per_second": 22018.86276523721,
       "note": "Bounded fresh sample; full-page latency is an extrapolation.",
       "estimated_pages_for_buffer": 1,
-      "required_token_positions_per_second": 28140.826477130526,
-      "required_pages_per_second_upper_estimate": 0.0021878254801825318
+      "required_token_positions_per_second": 28421.395790112692,
+      "required_pages_per_second_upper_estimate": 0.0022096385101729103
     },
     {
       "source": "cortex",
@@ -92,8 +92,8 @@
       "tokens_per_second": 910315.184299418,
       "note": "Bounded fresh sample; full-page latency is an extrapolation.",
       "estimated_pages_for_buffer": 22,
-      "required_token_positions_per_second": 28140.826477130526,
-      "required_pages_per_second_upper_estimate": 0.3609603067832702
+      "required_token_positions_per_second": 28421.395790112692,
+      "required_pages_per_second_upper_estimate": 0.36455914867834804
     }
   ],
   "gpu_measurement": "Selected forward/backward estimate; replaced by live successful update timings.",
@@ -158,409 +158,61 @@
     "prefetch_batches": 62,
     "prefetch_capacity_limit": 512,
     "batch_memory_estimate_bytes": 720750,
-    "producer_batch_seconds": 0.3391172111871015,
-    "producer_p95_seconds": 0.355980565000209,
-    "gpu_update_seconds": 0.9824871356379303,
-    "producer_batches_per_second": 2.948832931538438,
-    "gpu_batches_per_second": 1.0178250317249178,
-    "producer_headroom": 2.897190420382001,
+    "producer_batch_seconds": 0.15217620494523487,
+    "producer_p95_seconds": 0.1721363199994812,
+    "gpu_update_seconds": 0.9727882544606855,
+    "producer_batches_per_second": 6.571329600182103,
+    "gpu_batches_per_second": 1.0279729380104417,
+    "producer_headroom": 6.392512251246982,
     "sustainable": true,
-    "buffer_seconds": 60.91420240955168,
+    "buffer_seconds": 60.3128717765625,
     "buffer_drain_seconds": null,
     "shard_target_mib": 16,
-    "upload_pages_per_commit": 83,
+    "upload_pages_per_commit": 79,
     "max_pending_pages": 4096,
-    "generation_bytes_per_second": 385732.1530428105,
-    "upload_bytes_per_second": 4949797.310566265,
+    "generation_bytes_per_second": 373719.53067025304,
+    "upload_bytes_per_second": 4658638.935595296,
     "upload_sustainable": true,
     "logical_page_rows": 4096,
     "logical_cortex_page_rows": 128,
-    "estimated_full_shards_per_hour": 82.769140658028,
+    "estimated_full_shards_per_hour": 80.19151153641408,
     "selection_policy": "Smallest shard within 90% of measured local peak, enlarged to amortize 30 seconds of measured page production; bounded to 256 MiB."
   },
   "upload_measurements": [
     {
-      "seconds": 1.6456713849984226,
-      "bytes": 6071875,
-      "pages": 38,
-      "shards": 2
-    },
-    {
-      "seconds": 1.52896361799867,
-      "bytes": 13388189,
-      "pages": 35,
-      "shards": 4
-    },
-    {
-      "seconds": 1.3202310080014286,
-      "bytes": 5863529,
-      "pages": 31,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5651336049995734,
-      "bytes": 6044120,
-      "pages": 37,
-      "shards": 2
-    },
-    {
-      "seconds": 1.1207633359990723,
-      "bytes": 2404167,
-      "pages": 29,
-      "shards": 2
-    },
-    {
-      "seconds": 1.4597645020003256,
-      "bytes": 6256363,
-      "pages": 37,
-      "shards": 2
-    },
-    {
-      "seconds": 1.2800013509986456,
-      "bytes": 667279,
-      "pages": 18,
+      "seconds": 0.9768797280003128,
+      "bytes": 23252,
+      "pages": 1,
       "shards": 1
     },
     {
-      "seconds": 1.4021912890020758,
-      "bytes": 11567150,
-      "pages": 32,
-      "shards": 4
-    },
-    {
-      "seconds": 1.280043183000089,
-      "bytes": 12206708,
-      "pages": 48,
-      "shards": 3
-    },
-    {
-      "seconds": 0.8664777250014595,
-      "bytes": 2384518,
-      "pages": 29,
-      "shards": 2
-    },
-    {
-      "seconds": 1.027337190000253,
-      "bytes": 5987324,
-      "pages": 24,
-      "shards": 2
-    },
-    {
-      "seconds": 1.2340771259987378,
-      "bytes": 10655993,
-      "pages": 48,
-      "shards": 3
-    },
-    {
-      "seconds": 1.3521600980020594,
-      "bytes": 2290704,
-      "pages": 29,
-      "shards": 2
-    },
-    {
-      "seconds": 1.766581726995355,
-      "bytes": 12101436,
-      "pages": 41,
-      "shards": 3
-    },
-    {
-      "seconds": 1.0664826619977248,
-      "bytes": 6396203,
-      "pages": 31,
-      "shards": 3
-    },
-    {
-      "seconds": 1.070789342003991,
-      "bytes": 6236244,
-      "pages": 37,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5840121689980151,
-      "bytes": 6547762,
-      "pages": 36,
-      "shards": 2
-    },
-    {
-      "seconds": 1.2987593140060198,
-      "bytes": 2393188,
-      "pages": 30,
-      "shards": 2
-    },
-    {
-      "seconds": 1.808516741002677,
-      "bytes": 10401411,
-      "pages": 34,
-      "shards": 3
-    },
-    {
-      "seconds": 1.1967187990012462,
-      "bytes": 13416312,
-      "pages": 42,
-      "shards": 4
-    },
-    {
-      "seconds": 0.995916217994818,
-      "bytes": 1003302,
-      "pages": 27,
-      "shards": 1
-    },
-    {
-      "seconds": 0.9674175760010257,
-      "bytes": 6098431,
+      "seconds": 1.0272072289999414,
+      "bytes": 7383107,
       "pages": 38,
-      "shards": 2
-    },
-    {
-      "seconds": 1.0489661490064464,
-      "bytes": 6157279,
-      "pages": 33,
-      "shards": 3
-    },
-    {
-      "seconds": 1.4413105299972813,
-      "bytes": 6425996,
-      "pages": 38,
-      "shards": 2
-    },
-    {
-      "seconds": 1.128546699997969,
-      "bytes": 5952550,
-      "pages": 28,
       "shards": 3
     },
     {
-      "seconds": 1.3368553659965983,
-      "bytes": 12766871,
-      "pages": 38,
-      "shards": 4
-    },
-    {
-      "seconds": 1.6370693080025376,
-      "bytes": 5715454,
-      "pages": 28,
-      "shards": 2
-    },
-    {
-      "seconds": 1.2990395020024152,
-      "bytes": 2748829,
-      "pages": 38,
-      "shards": 2
-    },
-    {
-      "seconds": 1.300822404999053,
-      "bytes": 6046016,
+      "seconds": 1.1484548910002559,
+      "bytes": 5982767,
       "pages": 37,
       "shards": 2
     },
     {
-      "seconds": 1.4766436999998405,
-      "bytes": 10238885,
-      "pages": 34,
-      "shards": 3
-    },
-    {
-      "seconds": 1.4429230239984463,
-      "bytes": 2385665,
-      "pages": 28,
-      "shards": 2
-    },
-    {
-      "seconds": 1.539863748002972,
-      "bytes": 11993569,
-      "pages": 43,
-      "shards": 3
-    },
-    {
-      "seconds": 1.4481774710002355,
-      "bytes": 7324064,
-      "pages": 31,
-      "shards": 3
-    },
-    {
-      "seconds": 1.494102450000355,
-      "bytes": 5356034,
-      "pages": 33,
-      "shards": 2
-    },
-    {
-      "seconds": 1.1517664069979219,
-      "bytes": 6283260,
-      "pages": 38,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5029984030043124,
-      "bytes": 2302830,
-      "pages": 28,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5264721800049301,
-      "bytes": 6345437,
-      "pages": 38,
-      "shards": 2
-    },
-    {
-      "seconds": 1.7716346369998064,
-      "bytes": 17764299,
-      "pages": 34,
-      "shards": 5
-    },
-    {
-      "seconds": 1.4270490490016527,
-      "bytes": 1152119,
-      "pages": 35,
-      "shards": 1
-    },
-    {
-      "seconds": 1.5365510039991932,
-      "bytes": 6323603,
-      "pages": 37,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5189877529992373,
-      "bytes": 7424560,
-      "pages": 31,
-      "shards": 3
-    },
-    {
-      "seconds": 1.2933890700005577,
-      "bytes": 4945280,
-      "pages": 34,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5587586719993851,
-      "bytes": 6311746,
-      "pages": 37,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5925084610062186,
-      "bytes": 13513629,
-      "pages": 37,
-      "shards": 4
-    },
-    {
-      "seconds": 1.9457880960035254,
-      "bytes": 1015286,
-      "pages": 33,
-      "shards": 1
-    },
-    {
-      "seconds": 1.624823366997589,
-      "bytes": 11605166,
-      "pages": 31,
-      "shards": 4
-    },
-    {
-      "seconds": 1.4567563370001153,
-      "bytes": 1138092,
-      "pages": 35,
-      "shards": 1
-    },
-    {
-      "seconds": 1.3836040919995867,
-      "bytes": 6235413,
-      "pages": 37,
-      "shards": 2
-    },
-    {
-      "seconds": 1.6395168350063614,
-      "bytes": 7620206,
-      "pages": 34,
-      "shards": 3
-    },
-    {
-      "seconds": 1.4781058270018548,
-      "bytes": 11305199,
-      "pages": 39,
-      "shards": 3
-    },
-    {
-      "seconds": 1.3714339009966352,
-      "bytes": 7352671,
-      "pages": 31,
-      "shards": 3
-    },
-    {
-      "seconds": 1.3506424060033169,
-      "bytes": 6203770,
-      "pages": 38,
-      "shards": 2
-    },
-    {
-      "seconds": 1.3150392690004082,
-      "bytes": 914685,
+      "seconds": 1.7043199660001846,
+      "bytes": 963598,
       "pages": 26,
       "shards": 1
     },
     {
-      "seconds": 1.6476991579984315,
-      "bytes": 12016198,
-      "pages": 39,
-      "shards": 4
-    },
-    {
-      "seconds": 1.4772214850017917,
-      "bytes": 1146549,
-      "pages": 36,
-      "shards": 2
-    },
-    {
-      "seconds": 1.649438596003165,
-      "bytes": 11880426,
-      "pages": 30,
-      "shards": 3
-    },
-    {
-      "seconds": 0.9699813629995333,
-      "bytes": 2462244,
-      "pages": 31,
-      "shards": 2
-    },
-    {
-      "seconds": 1.3995045109986677,
-      "bytes": 10981344,
-      "pages": 45,
-      "shards": 3
-    },
-    {
-      "seconds": 1.094142590998672,
-      "bytes": 7571723,
-      "pages": 31,
+      "seconds": 1.1770538490000035,
+      "bytes": 8003790,
+      "pages": 38,
       "shards": 3
     },
     {
-      "seconds": 1.1913003890003893,
-      "bytes": 1341787,
-      "pages": 37,
-      "shards": 1
-    },
-    {
-      "seconds": 1.4330705320026027,
-      "bytes": 5823183,
-      "pages": 26,
-      "shards": 2
-    },
-    {
-      "seconds": 1.3540045269983239,
-      "bytes": 17287207,
-      "pages": 44,
-      "shards": 5
-    },
-    {
-      "seconds": 1.2279312789978576,
-      "bytes": 1252954,
-      "pages": 35,
-      "shards": 1
-    },
-    {
-      "seconds": 1.2454847960034385,
-      "bytes": 7379042,
-      "pages": 32,
+      "seconds": 1.2935991909998847,
+      "bytes": 11779732,
+      "pages": 36,
       "shards": 3
     }
   ],
@@ -606,7 +258,7 @@
   },
   "cache_reused": true,
   "resume_probe_seconds": [
-    2.9105605440000204,
-    0.014060387999961677
+    2.7244317560000013,
+    0.015869192000025123
   ]
 }

trainer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fef7bf8813882524f5ca30eee990ca62de8fcdd9a4aac08f179ed7a0bfb2f7cb
-size 721756031
+oid sha256:24ae65d790f92deee8ff8c64623716e72d101b0c319d3796bed4c28c51aec73b
+size 721776383

training_config.json

@@ -32,7 +32,7 @@
     "openmath": 10,
     "rewrite6": 5
   },
-  "lr": 0.0001,
+  "lr": 8e-05,
   "weight_decay": 0.1,
   "lr_schedule": "auto",
   "warmup_steps": 1000,