Qwen90M step 835000; complete resumable checkpoint

complete.json

@@ -1,18 +1,18 @@
 {
-  "step": 830000,
+  "step": 835000,
   "files": {
     "config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7",
     "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4",
-    "model.safetensors": "148b9cd485ba37f4e9cce26a66ff0cccc598fbe722764f410999681d873c39a5",
+    "model.safetensors": "434bc5e66416d00f38daf4d1c20bd1d7d620b0fc06f97108edac543c845dc140",
     "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03",
     "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300",
     "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263",
     "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6",
-    "trainer.pt": "f6ee31a07eec8634f2a4579290acbcefb3cab5d7972f2153a10c3041f0c32cb5",
+    "trainer.pt": "8fc02097e06a5d8dde4c92534118b36500d860341974727dbdf1235db028d4da",
     "training_config.json": "4e8f8b2fc7193aabdbca12082e7c7cc3a0ed080bd5ff16e060cc02e10ffc53dd",
     "performance.json": "eb45b2d9ecb051b6b7e592c18f2a0bfebf4278147b783ba131f7a9a4453a9ec7",
-    "shard_scheduler.json": "d229c2adbc90c3e90d75f2ded34df13e9ded5f7d2aa7cf78783ee6b128d17bdd",
-    "coverage.json": "537e5ac8f9bc339cc5920f3da24a226dfbd2e65ffea7c0b2d60b4a270d3159d5",
+    "shard_scheduler.json": "15ec84a118722aa0da30b1eedddde0fa6f0d20effd89f99c89d67b1b955f9b9e",
+    "coverage.json": "2268abb420311a79a47ad9499884d787f6b2b1069f26033253f7830b5909fc05",
     "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"
   },
   "signature": "052aedaef69b550457da29474e17df28fa635c99594a7b9c0881f3ac3827000e"

coverage.json

@@ -1,107 +1,107 @@
 {
-  "step": 830000,
+  "step": 835000,
   "stats": {
-    "climbmix/blocks": 29880000,
-    "climbmix/targets": 49950066776,
-    "climbmix/input_tokens": 49953231760,
-    "ultra_qa/blocks": 11205000,
-    "ultra_qa/targets": 18279978645,
-    "ultra_qa/input_tokens": 18280013833,
-    "cortex/blocks": 11205000,
-    "cortex/targets": 426071773,
-    "cortex/input_tokens": 17811149534,
-    "ultra_style/blocks": 11205000,
-    "ultra_style/targets": 17722842692,
-    "ultra_style/input_tokens": 17723772365,
-    "openmath/blocks": 7470000,
-    "openmath/targets": 11542438616,
-    "openmath/input_tokens": 11546284486,
-    "rewrite6/blocks": 3735000,
-    "rewrite6/targets": 6122985883,
-    "rewrite6/input_tokens": 6129699061
+    "climbmix/blocks": 30060000,
+    "climbmix/targets": 50250362065,
+    "climbmix/input_tokens": 50253546093,
+    "ultra_qa/blocks": 11272500,
+    "ultra_qa/targets": 18390084306,
+    "ultra_qa/input_tokens": 18390119687,
+    "cortex/blocks": 11272500,
+    "cortex/targets": 428641870,
+    "cortex/input_tokens": 17918505142,
+    "ultra_style/blocks": 11272500,
+    "ultra_style/targets": 17829475019,
+    "ultra_style/input_tokens": 17830410407,
+    "openmath/blocks": 7515000,
+    "openmath/targets": 11612035358,
+    "openmath/input_tokens": 11615904568,
+    "rewrite6/blocks": 3757500,
+    "rewrite6/targets": 6159911854,
+    "rewrite6/input_tokens": 6166665342
   },
   "sources": {
     "climbmix": {
-      "page": 4092,
-      "row": 49,
+      "page": 4116,
+      "row": 3220,
       "start": {
-        "file": 198,
+        "file": 199,
         "reader": {
           "format": "parquet-row-v1",
-          "row_group": 7,
+          "row_group": 20,
           "row": 1024
         }
       },
       "epoch": 0,
-      "offset": 2048,
-      "documents": 16760881
+      "offset": 0,
+      "documents": 16862356
     },
     "rewrite6": {
-      "page": 28,
-      "row": 636,
+      "page": 42,
+      "row": 790,
       "start": {
         "file": 0,
         "reader": {
           "format": "jsonl-row-v1",
-          "row": 114688
+          "row": 172032
         }
       },
       "epoch": 46,
       "offset": 0,
-      "documents": 9563908
+      "documents": 9621406
     },
     "openmath": {
-      "page": 2557,
-      "row": 2242,
+      "page": 2572,
+      "row": 4072,
       "start": {
-        "file": 23,
+        "file": 24,
         "reader": {
           "format": "parquet-row-v1",
-          "row_group": 430,
-          "row": 522
+          "row_group": 55,
+          "row": 313
         }
       },
       "epoch": 0,
       "offset": 0,
-      "documents": 10475714
+      "documents": 10538984
     },
     "ultra_qa": {
-      "page": 1105,
-      "row": 3374,
+      "page": 1112,
+      "row": 1914,
       "start": {
         "file": 8,
         "reader": {
           "format": "parquet-row-v1",
-          "row_group": 5,
-          "row": 46357
+          "row_group": 6,
+          "row": 10507
         }
       },
       "epoch": 0,
       "offset": 2048,
-      "documents": 4529454
+      "documents": 4556666
     },
     "ultra_style": {
-      "page": 1687,
-      "row": 3474,
+      "page": 1698,
+      "row": 306,
       "start": {
         "file": 10,
         "reader": {
           "format": "parquet-row-v1",
-          "row_group": 0,
-          "row": 60537
+          "row_group": 1,
+          "row": 20264
         }
       },
       "epoch": 0,
-      "offset": 2048,
-      "documents": 6913426
+      "offset": 0,
+      "documents": 6955314
     },
     "cortex": {
-      "page": 217307,
-      "row": 113,
+      "page": 218617,
+      "row": 105,
       "start": null,
       "epoch": 0,
       "offset": 0,
-      "documents": 27815409
+      "documents": 27983081
     }
   }
 }

latest.json

@@ -1 +1 @@
-{"step": 830000, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "148b9cd485ba37f4e9cce26a66ff0cccc598fbe722764f410999681d873c39a5", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "f6ee31a07eec8634f2a4579290acbcefb3cab5d7972f2153a10c3041f0c32cb5", "training_config.json": "4e8f8b2fc7193aabdbca12082e7c7cc3a0ed080bd5ff16e060cc02e10ffc53dd", "performance.json": "eb45b2d9ecb051b6b7e592c18f2a0bfebf4278147b783ba131f7a9a4453a9ec7", "shard_scheduler.json": "d229c2adbc90c3e90d75f2ded34df13e9ded5f7d2aa7cf78783ee6b128d17bdd", "coverage.json": "537e5ac8f9bc339cc5920f3da24a226dfbd2e65ffea7c0b2d60b4a270d3159d5", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "052aedaef69b550457da29474e17df28fa635c99594a7b9c0881f3ac3827000e", "folder": "step-000000830000"}
+{"step": 835000, "files": {"config.json": "cbc72137da3dae65085a2a8c39ad7634d16bf92d785ca00d283ee82fe32b37e7", "generation_config.json": "54ed0d19f6226888c91157964ede8a892bd5ae175bfad85336152b3d9b6cd7c4", "model.safetensors": "434bc5e66416d00f38daf4d1c20bd1d7d620b0fc06f97108edac543c845dc140", "tokenization_byte.py": "279b9990f77b1eaedc7c1184fe13d60273a713ead856200f1fb9ae323c7b1e03", "tokenizer_config.json": "0ecbd4fd39e35125800d806af981ba0cacd8a7ba5f02c9e729c4ecc54af6d300", "special_tokens_map.json": "20d4b94d0e657c3f44d9123522c106e7315b6c1a1b6d259478f17bd2918b1263", "byte_vocab.json": "57a1cd2e70dd5b7cc6fd2eaa12ebfcd173c1620fa5d45c2c6282511c29b6a5f6", "trainer.pt": "8fc02097e06a5d8dde4c92534118b36500d860341974727dbdf1235db028d4da", "training_config.json": "4e8f8b2fc7193aabdbca12082e7c7cc3a0ed080bd5ff16e060cc02e10ffc53dd", "performance.json": "eb45b2d9ecb051b6b7e592c18f2a0bfebf4278147b783ba131f7a9a4453a9ec7", "shard_scheduler.json": "15ec84a118722aa0da30b1eedddde0fa6f0d20effd89f99c89d67b1b955f9b9e", "coverage.json": "2268abb420311a79a47ad9499884d787f6b2b1069f26033253f7830b5909fc05", "README.md": "9ce0f9efacbc46177a7e818b6aed274d84e7145ff944b5c7d6ecd5d33b9d4419"}, "signature": "052aedaef69b550457da29474e17df28fa635c99594a7b9c0881f3ac3827000e", "folder": "step-000000835000"}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:148b9cd485ba37f4e9cce26a66ff0cccc598fbe722764f410999681d873c39a5
+oid sha256:434bc5e66416d00f38daf4d1c20bd1d7d620b0fc06f97108edac543c845dc140
 size 360622744

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": 77014.49798515823,
-      "required_pages_per_second_upper_estimate": 0.005328041668868387
+      "required_token_positions_per_second": 76407.9111220439,
+      "required_pages_per_second_upper_estimate": 0.0052860765822027025
     },
     {
       "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": 9626.812248144779,
-      "required_pages_per_second_upper_estimate": 0.0040305452575985145
+      "required_token_positions_per_second": 9550.988890255487,
+      "required_pages_per_second_upper_estimate": 0.003998799601022032
     },
     {
       "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": 19253.624496289558,
-      "required_pages_per_second_upper_estimate": 0.00401002223035429
+      "required_token_positions_per_second": 19101.977780510973,
+      "required_pages_per_second_upper_estimate": 0.00397843821865044
     },
     {
       "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": 28880.436744434337,
-      "required_pages_per_second_upper_estimate": 0.0017313380849552794
+      "required_token_positions_per_second": 28652.966670766462,
+      "required_pages_per_second_upper_estimate": 0.001717701601365582
     },
     {
       "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": 28880.436744434337,
-      "required_pages_per_second_upper_estimate": 0.002245326925263646
+      "required_token_positions_per_second": 28652.966670766462,
+      "required_pages_per_second_upper_estimate": 0.0022276421275710828
     },
     {
       "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": 23,
-      "required_token_positions_per_second": 28880.436744434337,
-      "required_pages_per_second_upper_estimate": 0.3704472331606103
+      "required_token_positions_per_second": 28652.966670766462,
+      "required_pages_per_second_upper_estimate": 0.3675294912939349
     }
   ],
   "gpu_measurement": "Selected forward/backward estimate; replaced by live successful update timings.",
@@ -158,408 +158,408 @@
     "prefetch_batches": 63,
     "prefetch_capacity_limit": 512,
     "batch_memory_estimate_bytes": 720750,
-    "producer_batch_seconds": 0.10640840249212147,
-    "producer_p95_seconds": 0.2958205579998321,
-    "gpu_update_seconds": 0.9573262428355817,
-    "producer_batches_per_second": 9.397754092530809,
-    "gpu_batches_per_second": 1.0445759817865428,
-    "producer_headroom": 8.99671661649523,
+    "producer_batch_seconds": 0.10879093923438177,
+    "producer_p95_seconds": 0.32612680799502414,
+    "gpu_update_seconds": 0.964926261133309,
+    "producer_batches_per_second": 9.191941967203505,
+    "gpu_batches_per_second": 1.0363486209044583,
+    "producer_headroom": 8.86954619496803,
     "sustainable": true,
-    "buffer_seconds": 60.31155329864164,
+    "buffer_seconds": 60.79035445139847,
     "buffer_drain_seconds": null,
-    "shard_target_mib": 16,
-    "upload_pages_per_commit": 83,
+    "shard_target_mib": 32,
+    "upload_pages_per_commit": 165,
     "max_pending_pages": 4096,
-    "generation_bytes_per_second": 311300.6360700732,
-    "upload_bytes_per_second": 3156742.7663847003,
+    "generation_bytes_per_second": 666855.3792614836,
+    "upload_bytes_per_second": 3927796.2312513823,
     "upload_sustainable": true,
     "logical_page_rows": 4096,
     "logical_cortex_page_rows": 128,
-    "estimated_full_shards_per_hour": 66.79786979271552,
+    "estimated_full_shards_per_hour": 71.54582039539042,
     "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": 1.913415155999246,
-      "bytes": 2704493,
-      "pages": 37,
+      "seconds": 1.5374675419989217,
+      "bytes": 6254376,
+      "pages": 36,
       "shards": 2
     },
     {
-      "seconds": 1.6119605509993562,
-      "bytes": 6041654,
-      "pages": 28,
-      "shards": 2
+      "seconds": 1.5695442620017275,
+      "bytes": 11600439,
+      "pages": 29,
+      "shards": 3
     },
     {
-      "seconds": 1.6828175019982154,
-      "bytes": 10664698,
-      "pages": 40,
+      "seconds": 1.7552413789999264,
+      "bytes": 6467958,
+      "pages": 38,
       "shards": 3
     },
     {
-      "seconds": 1.565426788998593,
-      "bytes": 2415576,
-      "pages": 30,
+      "seconds": 1.6743759019991558,
+      "bytes": 6275115,
+      "pages": 37,
       "shards": 2
     },
     {
-      "seconds": 1.5388405060002697,
-      "bytes": 6065532,
-      "pages": 36,
-      "shards": 2
+      "seconds": 2.272818253000878,
+      "bytes": 7589463,
+      "pages": 39,
+      "shards": 3
     },
     {
-      "seconds": 1.3961634170009347,
-      "bytes": 973137,
+      "seconds": 1.5132595840004797,
+      "bytes": 5701329,
       "pages": 28,
-      "shards": 1
+      "shards": 2
     },
     {
-      "seconds": 2.0744263130000036,
-      "bytes": 17948904,
-      "pages": 44,
-      "shards": 5
+      "seconds": 1.5889989900024375,
+      "bytes": 6430899,
+      "pages": 37,
+      "shards": 2
     },
     {
-      "seconds": 1.631489851999504,
-      "bytes": 5794503,
-      "pages": 28,
-      "shards": 2
+      "seconds": 1.6179743659995438,
+      "bytes": 8234542,
+      "pages": 31,
+      "shards": 3
     },
     {
-      "seconds": 1.5992371809988981,
-      "bytes": 2589704,
+      "seconds": 1.6133322549976583,
+      "bytes": 6059251,
       "pages": 37,
       "shards": 2
     },
     {
-      "seconds": 1.4273169570005848,
-      "bytes": 6151606,
-      "pages": 36,
+      "seconds": 1.6795359120005742,
+      "bytes": 5921876,
+      "pages": 24,
       "shards": 2
     },
     {
-      "seconds": 1.4220550049976737,
-      "bytes": 4624242,
-      "pages": 30,
+      "seconds": 1.7577662710027653,
+      "bytes": 6866333,
+      "pages": 46,
+      "shards": 3
+    },
+    {
+      "seconds": 1.93685000899859,
+      "bytes": 6023833,
+      "pages": 33,
       "shards": 2
     },
     {
-      "seconds": 2.047509969001112,
-      "bytes": 7597965,
-      "pages": 31,
+      "seconds": 1.8011142590003146,
+      "bytes": 7499768,
+      "pages": 34,
       "shards": 3
     },
     {
-      "seconds": 1.836660891000065,
-      "bytes": 12233872,
-      "pages": 41,
+      "seconds": 1.6402012090002245,
+      "bytes": 10865294,
+      "pages": 22,
       "shards": 3
     },
     {
-      "seconds": 1.7772512390001793,
-      "bytes": 2617112,
-      "pages": 37,
+      "seconds": 1.7843240430011065,
+      "bytes": 6397585,
+      "pages": 42,
       "shards": 2
     },
     {
-      "seconds": 1.5108418660020106,
-      "bytes": 11046966,
-      "pages": 31,
+      "seconds": 1.4411041519997525,
+      "bytes": 2727332,
+      "pages": 38,
       "shards": 3
     },
     {
-      "seconds": 1.4716807509976206,
-      "bytes": 1094934,
-      "pages": 36,
+      "seconds": 1.913498541998706,
+      "bytes": 6177731,
+      "pages": 28,
+      "shards": 2
+    },
+    {
+      "seconds": 2.387397704002069,
+      "bytes": 11768966,
+      "pages": 40,
+      "shards": 4
+    },
+    {
+      "seconds": 1.479657814001257,
+      "bytes": 836473,
+      "pages": 27,
       "shards": 1
     },
     {
-      "seconds": 1.608142429999134,
-      "bytes": 7266564,
-      "pages": 31,
+      "seconds": 1.9911568649986293,
+      "bytes": 11961950,
+      "pages": 42,
       "shards": 3
     },
     {
-      "seconds": 1.734035388999473,
-      "bytes": 6027691,
+      "seconds": 1.958790611999575,
+      "bytes": 2439502,
       "pages": 29,
       "shards": 2
     },
     {
-      "seconds": 1.5976459840021562,
-      "bytes": 11304870,
-      "pages": 44,
+      "seconds": 1.8466758049980854,
+      "bytes": 10680095,
+      "pages": 40,
       "shards": 3
     },
     {
-      "seconds": 1.606354307001311,
-      "bytes": 7330991,
-      "pages": 32,
-      "shards": 3
+      "seconds": 1.7479712950007524,
+      "bytes": 6009903,
+      "pages": 28,
+      "shards": 2
     },
     {
-      "seconds": 1.5559647560003214,
-      "bytes": 6263907,
+      "seconds": 2.2005035470028815,
+      "bytes": 2644353,
       "pages": 37,
       "shards": 2
     },
     {
-      "seconds": 1.8874185050008236,
-      "bytes": 2469051,
-      "pages": 30,
+      "seconds": 1.5131899659972987,
+      "bytes": 5828306,
+      "pages": 28,
       "shards": 2
     },
     {
-      "seconds": 24.49332847400001,
-      "bytes": 9957694,
-      "pages": 37,
-      "shards": 3
+      "seconds": 1.8049468739991426,
+      "bytes": 4697774,
+      "pages": 26,
+      "shards": 2
     },
     {
-      "seconds": 1.9510848309982975,
-      "bytes": 11992256,
-      "pages": 40,
-      "shards": 3
+      "seconds": 2.496297884001251,
+      "bytes": 13693480,
+      "pages": 42,
+      "shards": 4
     },
     {
-      "seconds": 1.68195884799934,
-      "bytes": 2652603,
+      "seconds": 1.8265930059969833,
+      "bytes": 5855746,
       "pages": 36,
-      "shards": 3
-    },
-    {
-      "seconds": 1.9732827289990382,
-      "bytes": 10394345,
-      "pages": 35,
-      "shards": 3
-    },
-    {
-      "seconds": 1.4147246300017287,
-      "bytes": 2402371,
-      "pages": 27,
       "shards": 2
     },
     {
-      "seconds": 2.1062088899998344,
-      "bytes": 6113756,
-      "pages": 37,
+      "seconds": 1.3942693540011533,
+      "bytes": 2663289,
+      "pages": 34,
       "shards": 2
     },
     {
-      "seconds": 1.8720440979996056,
-      "bytes": 6184562,
+      "seconds": 1.5531271910003852,
+      "bytes": 10487346,
       "pages": 37,
-      "shards": 2
+      "shards": 3
     },
     {
-      "seconds": 1.9332000800022797,
-      "bytes": 8278775,
+      "seconds": 1.5084796250012005,
+      "bytes": 1063958,
       "pages": 32,
-      "shards": 3
+      "shards": 1
     },
     {
-      "seconds": 2.060914223002328,
-      "bytes": 10411136,
-      "pages": 38,
+      "seconds": 1.9720517980022123,
+      "bytes": 7314841,
+      "pages": 27,
       "shards": 3
     },
     {
-      "seconds": 1.9575544710023678,
-      "bytes": 7322575,
-      "pages": 31,
+      "seconds": 1.793492888999026,
+      "bytes": 11753421,
+      "pages": 38,
       "shards": 3
     },
     {
-      "seconds": 1.413499946000229,
-      "bytes": 1206249,
-      "pages": 35,
-      "shards": 1
-    },
-    {
-      "seconds": 2.568047996999667,
-      "bytes": 10001629,
-      "pages": 35,
+      "seconds": 1.5141362069989555,
+      "bytes": 6622032,
+      "pages": 34,
       "shards": 3
     },
     {
-      "seconds": 1.5792092789997696,
-      "bytes": 2645048,
-      "pages": 36,
+      "seconds": 1.5094903019999037,
+      "bytes": 6206747,
+      "pages": 38,
       "shards": 2
     },
     {
-      "seconds": 2.121729968999716,
-      "bytes": 11759509,
-      "pages": 29,
-      "shards": 3
+      "seconds": 1.7292475489994104,
+      "bytes": 6419421,
+      "pages": 37,
+      "shards": 2
     },
     {
-      "seconds": 1.7052124439978797,
-      "bytes": 6177071,
-      "pages": 37,
+      "seconds": 1.6914732460027153,
+      "bytes": 2366813,
+      "pages": 27,
       "shards": 2
     },
     {
-      "seconds": 2.175897373002954,
-      "bytes": 7190309,
-      "pages": 36,
+      "seconds": 1.8487111699978414,
+      "bytes": 10763047,
+      "pages": 31,
       "shards": 3
     },
     {
-      "seconds": 1.8149941150004452,
-      "bytes": 6150813,
-      "pages": 37,
-      "shards": 2
+      "seconds": 1.605599512000481,
+      "bytes": 8445757,
+      "pages": 39,
+      "shards": 3
     },
     {
-      "seconds": 1.5219117590022506,
-      "bytes": 2396663,
+      "seconds": 1.6673135730015929,
+      "bytes": 6027005,
       "pages": 28,
       "shards": 2
     },
     {
-      "seconds": 1.5374675419989217,
-      "bytes": 6254376,
+      "seconds": 1.4991452859976562,
+      "bytes": 1121499,
       "pages": 36,
-      "shards": 2
-    },
-    {
-      "seconds": 1.5695442620017275,
-      "bytes": 11600439,
-      "pages": 29,
-      "shards": 3
+      "shards": 1
     },
     {
-      "seconds": 1.7552413789999264,
-      "bytes": 6467958,
-      "pages": 38,
-      "shards": 3
+      "seconds": 2.2660662760026753,
+      "bytes": 11276737,
+      "pages": 36,
+      "shards": 4
     },
     {
-      "seconds": 1.6743759019991558,
-      "bytes": 6275115,
-      "pages": 37,
+      "seconds": 1.8010207819970674,
+      "bytes": 5858104,
+      "pages": 32,
       "shards": 2
     },
     {
-      "seconds": 2.272818253000878,
-      "bytes": 7589463,
-      "pages": 39,
-      "shards": 3
+      "seconds": 1.3410764899963397,
+      "bytes": 1019830,
+      "pages": 31,
+      "shards": 1
     },
     {
-      "seconds": 1.5132595840004797,
-      "bytes": 5701329,
-      "pages": 28,
-      "shards": 2
+      "seconds": 1.8211984559966368,
+      "bytes": 13136869,
+      "pages": 32,
+      "shards": 4
     },
     {
-      "seconds": 1.5889989900024375,
-      "bytes": 6430899,
-      "pages": 37,
+      "seconds": 1.4420257410019985,
+      "bytes": 5442386,
+      "pages": 33,
       "shards": 2
     },
     {
-      "seconds": 1.6179743659995438,
-      "bytes": 8234542,
-      "pages": 31,
+      "seconds": 2.6930293979967246,
+      "bytes": 5991867,
+      "pages": 29,
       "shards": 3
     },
     {
-      "seconds": 1.6133322549976583,
-      "bytes": 6059251,
-      "pages": 37,
-      "shards": 2
+      "seconds": 1.7729905640007928,
+      "bytes": 7619842,
+      "pages": 39,
+      "shards": 3
     },
     {
-      "seconds": 1.6795359120005742,
-      "bytes": 5921876,
-      "pages": 24,
-      "shards": 2
+      "seconds": 1.2987675089971162,
+      "bytes": 1223490,
+      "pages": 35,
+      "shards": 1
     },
     {
-      "seconds": 1.7577662710027653,
-      "bytes": 6866333,
-      "pages": 46,
-      "shards": 3
+      "seconds": 1.546362834997126,
+      "bytes": 11822887,
+      "pages": 33,
+      "shards": 4
     },
     {
-      "seconds": 1.93685000899859,
-      "bytes": 6023833,
-      "pages": 33,
+      "seconds": 1.5667999549987144,
+      "bytes": 6923893,
+      "pages": 35,
       "shards": 2
     },
     {
-      "seconds": 1.8011142590003146,
-      "bytes": 7499768,
-      "pages": 34,
-      "shards": 3
+      "seconds": 1.4826895270016394,
+      "bytes": 6163509,
+      "pages": 32,
+      "shards": 2
     },
     {
-      "seconds": 1.6402012090002245,
-      "bytes": 10865294,
-      "pages": 22,
+      "seconds": 1.7263165549957193,
+      "bytes": 7465725,
+      "pages": 39,
       "shards": 3
     },
     {
-      "seconds": 1.7843240430011065,
-      "bytes": 6397585,
-      "pages": 42,
+      "seconds": 2.012805174999812,
+      "bytes": 4595806,
+      "pages": 30,
       "shards": 2
     },
     {
-      "seconds": 1.4411041519997525,
-      "bytes": 2727332,
-      "pages": 38,
+      "seconds": 1.6076512380022905,
+      "bytes": 7510724,
+      "pages": 40,
       "shards": 3
     },
     {
-      "seconds": 1.913498541998706,
-      "bytes": 6177731,
-      "pages": 28,
+      "seconds": 1.5883049809999648,
+      "bytes": 5665128,
+      "pages": 27,
       "shards": 2
     },
     {
-      "seconds": 2.387397704002069,
-      "bytes": 11768966,
+      "seconds": 2.0745964650050155,
+      "bytes": 11457524,
+      "pages": 35,
+      "shards": 3
+    },
+    {
+      "seconds": 1.6211695970050641,
+      "bytes": 7634724,
       "pages": 40,
-      "shards": 4
+      "shards": 3
     },
     {
-      "seconds": 1.479657814001257,
-      "bytes": 836473,
-      "pages": 27,
+      "seconds": 1.4634708860030514,
+      "bytes": 987981,
+      "pages": 29,
       "shards": 1
     },
     {
-      "seconds": 1.9911568649986293,
-      "bytes": 11961950,
-      "pages": 42,
+      "seconds": 1.6518487119974452,
+      "bytes": 7534957,
+      "pages": 37,
       "shards": 3
     },
     {
-      "seconds": 1.958790611999575,
-      "bytes": 2439502,
-      "pages": 29,
-      "shards": 2
+      "seconds": 1.5817108970004483,
+      "bytes": 10450177,
+      "pages": 33,
+      "shards": 3
     },
     {
-      "seconds": 1.8466758049980854,
-      "bytes": 10680095,
-      "pages": 40,
-      "shards": 3
+      "seconds": 1.328966004999529,
+      "bytes": 1184587,
+      "pages": 36,
+      "shards": 1
     },
     {
-      "seconds": 1.7479712950007524,
-      "bytes": 6009903,
-      "pages": 28,
-      "shards": 2
+      "seconds": 2.0095185120007955,
+      "bytes": 12961722,
+      "pages": 36,
+      "shards": 4
     },
     {
-      "seconds": 2.2005035470028815,
-      "bytes": 2644353,
+      "seconds": 1.4560003849983332,
+      "bytes": 6473170,
       "pages": 37,
       "shards": 2
     }

trainer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f6ee31a07eec8634f2a4579290acbcefb3cab5d7972f2153a10c3041f0c32cb5
-size 721775359
+oid sha256:8fc02097e06a5d8dde4c92534118b36500d860341974727dbdf1235db028d4da
+size 721778239