Training in progress, step 1000, checkpoint

last-checkpoint/config.json

@@ -8,184 +8,184 @@
   "hidden_dropout_prob": 0.1,
   "hidden_size": 768,
   "id2label": {
-    "0": "Water Supply",
-    "1": "Sexual Orientation and Gender Identity",
-    "2": "Investment Climate",
-    "3": "Environment",
-    "4": "Water Resources Management",
-    "5": "Labor Markets",
-    "6": "Governance",
-    "7": "Competition Policy",
-    "8": "Long-Term Finance",
-    "9": "Food Security Update",
-    "10": "Social Protection",
-    "11": "Inequality and Shared Prosperity",
-    "12": "Infectious diseases and Vaccines",
-    "13": "Jobs and Development",
-    "14": "Energy",
-    "15": "Education and Technology",
-    "16": "Debt Relief",
-    "17": "Financial Sector",
-    "18": "Measuring Poverty",
-    "19": "Education",
-    "20": "Water",
-    "21": "Procurement for Development",
-    "22": "Livestock and Sustainability",
-    "23": "Disaster Risk Management",
-    "24": "Financial Inclusion",
-    "25": "Trade Facilitation and Logistics",
-    "26": "Financial Integrity",
-    "27": "Health",
-    "28": "Regional Integration",
-    "29": "Forests and Landscapes",
-    "30": "Jobs & Development",
-    "31": "Payment Systems",
-    "32": "Agribusiness and Value Chains",
-    "33": "Disability Inclusion",
-    "34": "Marine Plastic Pollution",
-    "35": "Global Financing Facility for Women, Children and Adolescents",
-    "36": "Skills Development",
-    "37": "Sustainable Infrastructure Finance",
-    "38": "Land",
-    "39": "Sustainable Communities",
-    "40": "Higher Education",
-    "41": "Climate Change",
-    "42": "Small and Medium Enterprises Finance",
-    "43": "Teachers",
-    "44": "Migration",
-    "45": "Debt",
-    "46": "Urban Development",
-    "47": "Mining Investment and Governance Review",
-    "48": "Innovation and Entrepreneurship",
-    "49": "Natural Capital",
-    "50": "Transport",
-    "51": "Gender",
-    "52": "Safety Nets and Cash Transfers",
-    "53": "Inclusive Cities",
-    "54": "Extractive Industries",
-    "55": "Global Value Chains",
-    "56": "Agriculture and Food",
-    "57": "Universal Health Coverage",
-    "58": "Sanitation",
-    "59": "Gas Flaring Reduction",
-    "60": "COVID-19 Hub",
-    "61": "Girls' Education",
-    "62": "Macroeconomics",
-    "63": "Trade",
-    "64": "Pandemic Preparedness and COVID-19",
-    "65": "Infrastructure",
-    "66": "Oceans, Fisheries, and Coastal Economies",
-    "67": "Social Sustainability and Inclusion",
-    "68": "Community-Driven Development",
-    "69": "Credit Infrastructure",
-    "70": "Pollution",
-    "71": "Pensions",
-    "72": "Financial Stability",
-    "73": "One Health",
-    "74": "Indigenous Peoples",
-    "75": "Digital Development",
-    "76": "Climate-Smart Agriculture",
-    "77": "Food System Jobs",
-    "78": "Early Childhood Development",
-    "79": "Biodiversity",
-    "80": "Poverty",
-    "81": "Fragility, Conflict, and Violence",
-    "82": "Competitiveness",
-    "83": "Social Inclusion",
-    "84": "Nutrition",
-    "85": "Taxes and Government Revenue",
-    "86": "Water in Agriculture"
+    "0": "Global Financing Facility for Women, Children and Adolescents",
+    "1": "Competitiveness",
+    "2": "Sustainable Communities",
+    "3": "Education",
+    "4": "Natural Capital",
+    "5": "Regional Integration",
+    "6": "Jobs and Development",
+    "7": "Innovation and Entrepreneurship",
+    "8": "Inclusive Cities",
+    "9": "Jobs & Development",
+    "10": "Transport",
+    "11": "Sustainable Infrastructure Finance",
+    "12": "Pandemic Preparedness and COVID-19",
+    "13": "Taxes and Government Revenue",
+    "14": "Labor Markets",
+    "15": "Measuring Poverty",
+    "16": "Food Security Update",
+    "17": "Disability Inclusion",
+    "18": "Agriculture and Food",
+    "19": "Trade Facilitation and Logistics",
+    "20": "Social Protection",
+    "21": "One Health",
+    "22": "Biodiversity",
+    "23": "Social Inclusion",
+    "24": "Credit Infrastructure",
+    "25": "Water Supply",
+    "26": "Early Childhood Development",
+    "27": "Food System Jobs",
+    "28": "Migration",
+    "29": "Indigenous Peoples",
+    "30": "Universal Health Coverage",
+    "31": "Financial Sector",
+    "32": "Procurement for Development",
+    "33": "Inequality and Shared Prosperity",
+    "34": "COVID-19 Hub",
+    "35": "Poverty",
+    "36": "Financial Stability",
+    "37": "Digital Development",
+    "38": "Long-Term Finance",
+    "39": "Gas Flaring Reduction",
+    "40": "Mining Investment and Governance Review",
+    "41": "Small and Medium Enterprises Finance",
+    "42": "Infrastructure",
+    "43": "Health",
+    "44": "Sexual Orientation and Gender Identity",
+    "45": "Nutrition",
+    "46": "Financial Inclusion",
+    "47": "Fragility, Conflict, and Violence",
+    "48": "Debt Relief",
+    "49": "Disaster Risk Management",
+    "50": "Water in Agriculture",
+    "51": "Livestock and Sustainability",
+    "52": "Global Value Chains",
+    "53": "Competition Policy",
+    "54": "Pollution",
+    "55": "Urban Development",
+    "56": "Gender",
+    "57": "Safety Nets and Cash Transfers",
+    "58": "Forests and Landscapes",
+    "59": "Water Resources Management",
+    "60": "Extractive Industries",
+    "61": "Social Sustainability and Inclusion",
+    "62": "Energy",
+    "63": "Girls' Education",
+    "64": "Environment",
+    "65": "Marine Plastic Pollution",
+    "66": "Education and Technology",
+    "67": "Financial Integrity",
+    "68": "Oceans, Fisheries, and Coastal Economies",
+    "69": "Sanitation",
+    "70": "Land",
+    "71": "Higher Education",
+    "72": "Teachers",
+    "73": "Investment Climate",
+    "74": "Debt",
+    "75": "Climate Change",
+    "76": "Trade",
+    "77": "Skills Development",
+    "78": "Agribusiness and Value Chains",
+    "79": "Climate-Smart Agriculture",
+    "80": "Pensions",
+    "81": "Infectious diseases and Vaccines",
+    "82": "Payment Systems",
+    "83": "Community-Driven Development",
+    "84": "Water",
+    "85": "Governance",
+    "86": "Macroeconomics"
   },
   "initializer_range": 0.02,
   "intermediate_size": 3072,
   "label2id": {
-    "Agribusiness and Value Chains": 32,
-    "Agriculture and Food": 56,
-    "Biodiversity": 79,
-    "COVID-19 Hub": 60,
-    "Climate Change": 41,
-    "Climate-Smart Agriculture": 76,
-    "Community-Driven Development": 68,
-    "Competition Policy": 7,
-    "Competitiveness": 82,
-    "Credit Infrastructure": 69,
-    "Debt": 45,
-    "Debt Relief": 16,
-    "Digital Development": 75,
-    "Disability Inclusion": 33,
-    "Disaster Risk Management": 23,
-    "Early Childhood Development": 78,
-    "Education": 19,
-    "Education and Technology": 15,
-    "Energy": 14,
-    "Environment": 3,
-    "Extractive Industries": 54,
-    "Financial Inclusion": 24,
-    "Financial Integrity": 26,
-    "Financial Sector": 17,
-    "Financial Stability": 72,
-    "Food Security Update": 9,
-    "Food System Jobs": 77,
-    "Forests and Landscapes": 29,
-    "Fragility, Conflict, and Violence": 81,
-    "Gas Flaring Reduction": 59,
-    "Gender": 51,
-    "Girls' Education": 61,
-    "Global Financing Facility for Women, Children and Adolescents": 35,
-    "Global Value Chains": 55,
-    "Governance": 6,
-    "Health": 27,
-    "Higher Education": 40,
-    "Inclusive Cities": 53,
-    "Indigenous Peoples": 74,
-    "Inequality and Shared Prosperity": 11,
-    "Infectious diseases and Vaccines": 12,
-    "Infrastructure": 65,
-    "Innovation and Entrepreneurship": 48,
-    "Investment Climate": 2,
-    "Jobs & Development": 30,
-    "Jobs and Development": 13,
-    "Labor Markets": 5,
-    "Land": 38,
-    "Livestock and Sustainability": 22,
-    "Long-Term Finance": 8,
-    "Macroeconomics": 62,
-    "Marine Plastic Pollution": 34,
-    "Measuring Poverty": 18,
-    "Migration": 44,
-    "Mining Investment and Governance Review": 47,
-    "Natural Capital": 49,
-    "Nutrition": 84,
-    "Oceans, Fisheries, and Coastal Economies": 66,
-    "One Health": 73,
-    "Pandemic Preparedness and COVID-19": 64,
-    "Payment Systems": 31,
-    "Pensions": 71,
-    "Pollution": 70,
-    "Poverty": 80,
-    "Procurement for Development": 21,
-    "Regional Integration": 28,
-    "Safety Nets and Cash Transfers": 52,
-    "Sanitation": 58,
-    "Sexual Orientation and Gender Identity": 1,
-    "Skills Development": 36,
-    "Small and Medium Enterprises Finance": 42,
-    "Social Inclusion": 83,
-    "Social Protection": 10,
-    "Social Sustainability and Inclusion": 67,
-    "Sustainable Communities": 39,
-    "Sustainable Infrastructure Finance": 37,
-    "Taxes and Government Revenue": 85,
-    "Teachers": 43,
-    "Trade": 63,
-    "Trade Facilitation and Logistics": 25,
-    "Transport": 50,
-    "Universal Health Coverage": 57,
-    "Urban Development": 46,
-    "Water": 20,
-    "Water Resources Management": 4,
-    "Water Supply": 0,
-    "Water in Agriculture": 86
+    "Agribusiness and Value Chains": 78,
+    "Agriculture and Food": 18,
+    "Biodiversity": 22,
+    "COVID-19 Hub": 34,
+    "Climate Change": 75,
+    "Climate-Smart Agriculture": 79,
+    "Community-Driven Development": 83,
+    "Competition Policy": 53,
+    "Competitiveness": 1,
+    "Credit Infrastructure": 24,
+    "Debt": 74,
+    "Debt Relief": 48,
+    "Digital Development": 37,
+    "Disability Inclusion": 17,
+    "Disaster Risk Management": 49,
+    "Early Childhood Development": 26,
+    "Education": 3,
+    "Education and Technology": 66,
+    "Energy": 62,
+    "Environment": 64,
+    "Extractive Industries": 60,
+    "Financial Inclusion": 46,
+    "Financial Integrity": 67,
+    "Financial Sector": 31,
+    "Financial Stability": 36,
+    "Food Security Update": 16,
+    "Food System Jobs": 27,
+    "Forests and Landscapes": 58,
+    "Fragility, Conflict, and Violence": 47,
+    "Gas Flaring Reduction": 39,
+    "Gender": 56,
+    "Girls' Education": 63,
+    "Global Financing Facility for Women, Children and Adolescents": 0,
+    "Global Value Chains": 52,
+    "Governance": 85,
+    "Health": 43,
+    "Higher Education": 71,
+    "Inclusive Cities": 8,
+    "Indigenous Peoples": 29,
+    "Inequality and Shared Prosperity": 33,
+    "Infectious diseases and Vaccines": 81,
+    "Infrastructure": 42,
+    "Innovation and Entrepreneurship": 7,
+    "Investment Climate": 73,
+    "Jobs & Development": 9,
+    "Jobs and Development": 6,
+    "Labor Markets": 14,
+    "Land": 70,
+    "Livestock and Sustainability": 51,
+    "Long-Term Finance": 38,
+    "Macroeconomics": 86,
+    "Marine Plastic Pollution": 65,
+    "Measuring Poverty": 15,
+    "Migration": 28,
+    "Mining Investment and Governance Review": 40,
+    "Natural Capital": 4,
+    "Nutrition": 45,
+    "Oceans, Fisheries, and Coastal Economies": 68,
+    "One Health": 21,
+    "Pandemic Preparedness and COVID-19": 12,
+    "Payment Systems": 82,
+    "Pensions": 80,
+    "Pollution": 54,
+    "Poverty": 35,
+    "Procurement for Development": 32,
+    "Regional Integration": 5,
+    "Safety Nets and Cash Transfers": 57,
+    "Sanitation": 69,
+    "Sexual Orientation and Gender Identity": 44,
+    "Skills Development": 77,
+    "Small and Medium Enterprises Finance": 41,
+    "Social Inclusion": 23,
+    "Social Protection": 20,
+    "Social Sustainability and Inclusion": 61,
+    "Sustainable Communities": 2,
+    "Sustainable Infrastructure Finance": 11,
+    "Taxes and Government Revenue": 13,
+    "Teachers": 72,
+    "Trade": 76,
+    "Trade Facilitation and Logistics": 19,
+    "Transport": 10,
+    "Universal Health Coverage": 30,
+    "Urban Development": 55,
+    "Water": 84,
+    "Water Resources Management": 59,
+    "Water Supply": 25,
+    "Water in Agriculture": 50
   },
   "layer_norm_eps": 1e-07,
   "max_position_embeddings": 512,

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5c7c809bfb9e0c09a991cbc751afbd7d153af1ea8be3d21c60d821e480d58397
+oid sha256:56824edd4761257a415044881a1993322b91ee41909153e8164b07d4cdb8e299
 size 567860028

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:54c71b342dcaf811b4743acd66cc3b8dab74d01a77947a9e074165554a191a3f
+oid sha256:65d705b23d30e8d777f2c745aee65fc860e7de16b51a91182b36a495553f35f3
 size 1135783354

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4c6930021eee2ce84b130095d62c41e90293ccc2830acaf7c168280128912a1f
+oid sha256:34649568145f465443beb487f2295ab3dad9e5f49f758646dd823029413e18fe
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:abe099e1b5028bec45c9b921f52c62bf9191fdd92a0c19da1f44a14388797be1
+oid sha256:8e136870ce715682b288c29fb4d05a0aec61f1a74b2cc393d7fb9e66e4965261
 size 1064

last-checkpoint/trainer_state.json

@@ -1,24 +1,24 @@
 {
-  "best_metric": 0.03578707203269005,
-  "best_model_checkpoint": "doc-topic-model_eval-00_train-01/checkpoint-20000",
-  "epoch": 9.861932938856016,
+  "best_metric": 0.08824141323566437,
+  "best_model_checkpoint": "doc-topic-model_eval-00_train-01/checkpoint-1000",
+  "epoch": 0.4930966469428008,
   "eval_steps": 1000,
-  "global_step": 20000,
+  "global_step": 1000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.2465483234714004,
-      "grad_norm": 0.32120612263679504,
-      "learning_rate": 1.95069033530572e-05,
-      "loss": 0.1666,
+      "grad_norm": 0.32877257466316223,
+      "learning_rate": 1.9950690335305722e-05,
+      "loss": 0.166,
       "step": 500
     },
     {
       "epoch": 0.4930966469428008,
-      "grad_norm": 0.33755889534950256,
-      "learning_rate": 1.90138067061144e-05,
+      "grad_norm": 0.3466370701789856,
+      "learning_rate": 1.9901380670611442e-05,
       "loss": 0.0936,
       "step": 1000
     },
@@ -26,515 +26,30 @@
       "epoch": 0.4930966469428008,
       "eval_accuracy": 0.9814660487265615,
       "eval_f1": 0.0,
-      "eval_loss": 0.08871060609817505,
+      "eval_loss": 0.08824141323566437,
       "eval_precision": 0.0,
       "eval_recall": 0.0,
-      "eval_runtime": 26.4203,
-      "eval_samples_per_second": 306.961,
-      "eval_steps_per_second": 19.19,
+      "eval_runtime": 11.7844,
+      "eval_samples_per_second": 688.2,
+      "eval_steps_per_second": 2.715,
       "step": 1000
-    },
-    {
-      "epoch": 0.7396449704142012,
-      "grad_norm": 0.38695186376571655,
-      "learning_rate": 1.85207100591716e-05,
-      "loss": 0.0861,
-      "step": 1500
-    },
-    {
-      "epoch": 0.9861932938856016,
-      "grad_norm": 0.39813894033432007,
-      "learning_rate": 1.80276134122288e-05,
-      "loss": 0.0774,
-      "step": 2000
-    },
-    {
-      "epoch": 0.9861932938856016,
-      "eval_accuracy": 0.9814660487265615,
-      "eval_f1": 0.0,
-      "eval_loss": 0.06950555741786957,
-      "eval_precision": 0.0,
-      "eval_recall": 0.0,
-      "eval_runtime": 25.7659,
-      "eval_samples_per_second": 314.757,
-      "eval_steps_per_second": 19.677,
-      "step": 2000
-    },
-    {
-      "epoch": 1.232741617357002,
-      "grad_norm": 0.452305406332016,
-      "learning_rate": 1.7534516765285997e-05,
-      "loss": 0.0685,
-      "step": 2500
-    },
-    {
-      "epoch": 1.4792899408284024,
-      "grad_norm": 0.46256884932518005,
-      "learning_rate": 1.70414201183432e-05,
-      "loss": 0.0622,
-      "step": 3000
-    },
-    {
-      "epoch": 1.4792899408284024,
-      "eval_accuracy": 0.9821973723372592,
-      "eval_f1": 0.09950534088465122,
-      "eval_loss": 0.05685883387923241,
-      "eval_precision": 0.7958715596330275,
-      "eval_recall": 0.05307027605719966,
-      "eval_runtime": 26.3616,
-      "eval_samples_per_second": 307.644,
-      "eval_steps_per_second": 19.232,
-      "step": 3000
-    },
-    {
-      "epoch": 1.725838264299803,
-      "grad_norm": 0.35877692699432373,
-      "learning_rate": 1.6548323471400396e-05,
-      "loss": 0.0579,
-      "step": 3500
-    },
-    {
-      "epoch": 1.972386587771203,
-      "grad_norm": 0.43726620078086853,
-      "learning_rate": 1.6055226824457594e-05,
-      "loss": 0.0529,
-      "step": 4000
-    },
-    {
-      "epoch": 1.972386587771203,
-      "eval_accuracy": 0.9842609521379877,
-      "eval_f1": 0.3101198981176616,
-      "eval_loss": 0.050210438668727875,
-      "eval_precision": 0.8264900662251655,
-      "eval_recall": 0.1908694654737325,
-      "eval_runtime": 26.3196,
-      "eval_samples_per_second": 308.135,
-      "eval_steps_per_second": 19.263,
-      "step": 4000
-    },
-    {
-      "epoch": 2.2189349112426036,
-      "grad_norm": 0.2717958092689514,
-      "learning_rate": 1.5562130177514792e-05,
-      "loss": 0.0495,
-      "step": 4500
-    },
-    {
-      "epoch": 2.465483234714004,
-      "grad_norm": 0.4179055690765381,
-      "learning_rate": 1.5069033530571993e-05,
-      "loss": 0.0475,
-      "step": 5000
-    },
-    {
-      "epoch": 2.465483234714004,
-      "eval_accuracy": 0.9854528962399195,
-      "eval_f1": 0.4501232186863817,
-      "eval_loss": 0.04653949663043022,
-      "eval_precision": 0.7516550366791913,
-      "eval_recall": 0.32125105146440314,
-      "eval_runtime": 26.3415,
-      "eval_samples_per_second": 307.879,
-      "eval_steps_per_second": 19.247,
-      "step": 5000
-    },
-    {
-      "epoch": 2.712031558185404,
-      "grad_norm": 0.3518264591693878,
-      "learning_rate": 1.4575936883629191e-05,
-      "loss": 0.0457,
-      "step": 5500
-    },
-    {
-      "epoch": 2.9585798816568047,
-      "grad_norm": 0.4679907560348511,
-      "learning_rate": 1.4082840236686392e-05,
-      "loss": 0.0444,
-      "step": 6000
-    },
-    {
-      "epoch": 2.9585798816568047,
-      "eval_accuracy": 0.9860665844636252,
-      "eval_f1": 0.4852610084297608,
-      "eval_loss": 0.043688371777534485,
-      "eval_precision": 0.7695117901029558,
-      "eval_recall": 0.3543626213963447,
-      "eval_runtime": 26.4572,
-      "eval_samples_per_second": 306.533,
-      "eval_steps_per_second": 19.163,
-      "step": 6000
-    },
-    {
-      "epoch": 3.2051282051282053,
-      "grad_norm": 0.21664711833000183,
-      "learning_rate": 1.3589743589743592e-05,
-      "loss": 0.0422,
-      "step": 6500
-    },
-    {
-      "epoch": 3.4516765285996054,
-      "grad_norm": 0.33109939098358154,
-      "learning_rate": 1.309664694280079e-05,
-      "loss": 0.0394,
-      "step": 7000
-    },
-    {
-      "epoch": 3.4516765285996054,
-      "eval_accuracy": 0.9865711410632538,
-      "eval_f1": 0.5374664388577007,
-      "eval_loss": 0.04136768355965614,
-      "eval_precision": 0.7431155507559395,
-      "eval_recall": 0.4209681119522826,
-      "eval_runtime": 26.4486,
-      "eval_samples_per_second": 306.632,
-      "eval_steps_per_second": 19.169,
-      "step": 7000
-    },
-    {
-      "epoch": 3.698224852071006,
-      "grad_norm": 0.3816561996936798,
-      "learning_rate": 1.2603550295857989e-05,
-      "loss": 0.0403,
-      "step": 7500
-    },
-    {
-      "epoch": 3.9447731755424065,
-      "grad_norm": 0.44973593950271606,
-      "learning_rate": 1.2110453648915189e-05,
-      "loss": 0.039,
-      "step": 8000
-    },
-    {
-      "epoch": 3.9447731755424065,
-      "eval_accuracy": 0.9868616862961861,
-      "eval_f1": 0.5516540917005224,
-      "eval_loss": 0.040279507637023926,
-      "eval_precision": 0.750493485984998,
-      "eval_recall": 0.43610919935765086,
-      "eval_runtime": 27.0585,
-      "eval_samples_per_second": 299.721,
-      "eval_steps_per_second": 18.737,
-      "step": 8000
-    },
-    {
-      "epoch": 4.191321499013807,
-      "grad_norm": 0.43002060055732727,
-      "learning_rate": 1.1617357001972386e-05,
-      "loss": 0.0363,
-      "step": 8500
-    },
-    {
-      "epoch": 4.437869822485207,
-      "grad_norm": 0.3201405704021454,
-      "learning_rate": 1.1124260355029586e-05,
-      "loss": 0.0356,
-      "step": 9000
-    },
-    {
-      "epoch": 4.437869822485207,
-      "eval_accuracy": 0.9870360134359454,
-      "eval_f1": 0.5708655876143561,
-      "eval_loss": 0.03906617686152458,
-      "eval_precision": 0.7385287691187181,
-      "eval_recall": 0.465244322092223,
-      "eval_runtime": 26.4636,
-      "eval_samples_per_second": 306.458,
-      "eval_steps_per_second": 19.158,
-      "step": 9000
-    },
-    {
-      "epoch": 4.684418145956608,
-      "grad_norm": 0.47237107157707214,
-      "learning_rate": 1.0631163708086787e-05,
-      "loss": 0.0359,
-      "step": 9500
-    },
-    {
-      "epoch": 4.930966469428008,
-      "grad_norm": 0.5127544403076172,
-      "learning_rate": 1.0138067061143987e-05,
-      "loss": 0.0346,
-      "step": 10000
-    },
-    {
-      "epoch": 4.930966469428008,
-      "eval_accuracy": 0.9873010473801324,
-      "eval_f1": 0.5763593380614657,
-      "eval_loss": 0.03822428733110428,
-      "eval_precision": 0.7549857549857549,
-      "eval_recall": 0.46608549361474344,
-      "eval_runtime": 26.8486,
-      "eval_samples_per_second": 302.064,
-      "eval_steps_per_second": 18.884,
-      "step": 10000
-    },
-    {
-      "epoch": 5.177514792899408,
-      "grad_norm": 0.2535999119281769,
-      "learning_rate": 9.644970414201184e-06,
-      "loss": 0.0321,
-      "step": 10500
-    },
-    {
-      "epoch": 5.424063116370808,
-      "grad_norm": 0.3756001591682434,
-      "learning_rate": 9.151873767258384e-06,
-      "loss": 0.0317,
-      "step": 11000
-    },
-    {
-      "epoch": 5.424063116370808,
-      "eval_accuracy": 0.9874201000609436,
-      "eval_f1": 0.5807670508218401,
-      "eval_loss": 0.03766867518424988,
-      "eval_precision": 0.7594811612106238,
-      "eval_recall": 0.47013841095052383,
-      "eval_runtime": 27.4901,
-      "eval_samples_per_second": 295.015,
-      "eval_steps_per_second": 18.443,
-      "step": 11000
-    },
-    {
-      "epoch": 5.670611439842209,
-      "grad_norm": 0.3415575325489044,
-      "learning_rate": 8.658777120315582e-06,
-      "loss": 0.0325,
-      "step": 11500
-    },
-    {
-      "epoch": 5.9171597633136095,
-      "grad_norm": 0.37248894572257996,
-      "learning_rate": 8.165680473372781e-06,
-      "loss": 0.0308,
-      "step": 12000
-    },
-    {
-      "epoch": 5.9171597633136095,
-      "eval_accuracy": 0.9873846677154641,
-      "eval_f1": 0.5882788288079929,
-      "eval_loss": 0.0372396819293499,
-      "eval_precision": 0.7444392413954577,
-      "eval_recall": 0.4862736101552344,
-      "eval_runtime": 26.4199,
-      "eval_samples_per_second": 306.965,
-      "eval_steps_per_second": 19.19,
-      "step": 12000
-    },
-    {
-      "epoch": 6.16370808678501,
-      "grad_norm": 0.47570303082466125,
-      "learning_rate": 7.67258382642998e-06,
-      "loss": 0.0292,
-      "step": 12500
-    },
-    {
-      "epoch": 6.410256410256411,
-      "grad_norm": 0.4004118740558624,
-      "learning_rate": 7.17948717948718e-06,
-      "loss": 0.0288,
-      "step": 13000
-    },
-    {
-      "epoch": 6.410256410256411,
-      "eval_accuracy": 0.9875915926130646,
-      "eval_f1": 0.5967110415035239,
-      "eval_loss": 0.036796875298023224,
-      "eval_precision": 0.7503475440222428,
-      "eval_recall": 0.4952970864877265,
-      "eval_runtime": 26.3632,
-      "eval_samples_per_second": 307.626,
-      "eval_steps_per_second": 19.231,
-      "step": 13000
-    },
-    {
-      "epoch": 6.65680473372781,
-      "grad_norm": 0.32691922783851624,
-      "learning_rate": 6.686390532544379e-06,
-      "loss": 0.0279,
-      "step": 13500
-    },
-    {
-      "epoch": 6.903353057199211,
-      "grad_norm": 0.3402620553970337,
-      "learning_rate": 6.193293885601579e-06,
-      "loss": 0.0284,
-      "step": 14000
-    },
-    {
-      "epoch": 6.903353057199211,
-      "eval_accuracy": 0.9876780475360347,
-      "eval_f1": 0.6049259292920113,
-      "eval_loss": 0.036554474383592606,
-      "eval_precision": 0.745436219061485,
-      "eval_recall": 0.5089852412632867,
-      "eval_runtime": 26.4132,
-      "eval_samples_per_second": 307.043,
-      "eval_steps_per_second": 19.195,
-      "step": 14000
-    },
-    {
-      "epoch": 7.149901380670611,
-      "grad_norm": 0.33896222710609436,
-      "learning_rate": 5.700197238658778e-06,
-      "loss": 0.0272,
-      "step": 14500
-    },
-    {
-      "epoch": 7.396449704142012,
-      "grad_norm": 0.5163460373878479,
-      "learning_rate": 5.207100591715976e-06,
-      "loss": 0.026,
-      "step": 15000
-    },
-    {
-      "epoch": 7.396449704142012,
-      "eval_accuracy": 0.9877928483353884,
-      "eval_f1": 0.6092991608074394,
-      "eval_loss": 0.0364510752260685,
-      "eval_precision": 0.7488849241748439,
-      "eval_recall": 0.5135734495679437,
-      "eval_runtime": 26.576,
-      "eval_samples_per_second": 305.162,
-      "eval_steps_per_second": 19.077,
-      "step": 15000
-    },
-    {
-      "epoch": 7.642998027613412,
-      "grad_norm": 0.29480230808258057,
-      "learning_rate": 4.714003944773176e-06,
-      "loss": 0.0262,
-      "step": 15500
-    },
-    {
-      "epoch": 7.889546351084813,
-      "grad_norm": 0.37596189975738525,
-      "learning_rate": 4.220907297830375e-06,
-      "loss": 0.0268,
-      "step": 16000
-    },
-    {
-      "epoch": 7.889546351084813,
-      "eval_accuracy": 0.9879019799594654,
-      "eval_f1": 0.6212282570110046,
-      "eval_loss": 0.036032263189554214,
-      "eval_precision": 0.7400359446030236,
-      "eval_recall": 0.5352909688766536,
-      "eval_runtime": 26.3511,
-      "eval_samples_per_second": 307.767,
-      "eval_steps_per_second": 19.24,
-      "step": 16000
-    },
-    {
-      "epoch": 8.136094674556213,
-      "grad_norm": 0.1833580881357193,
-      "learning_rate": 3.7278106508875745e-06,
-      "loss": 0.025,
-      "step": 16500
-    },
-    {
-      "epoch": 8.382642998027613,
-      "grad_norm": 0.4305359125137329,
-      "learning_rate": 3.234714003944773e-06,
-      "loss": 0.0248,
-      "step": 17000
-    },
-    {
-      "epoch": 8.382642998027613,
-      "eval_accuracy": 0.9878098558612186,
-      "eval_f1": 0.6178861788617886,
-      "eval_loss": 0.03607296571135521,
-      "eval_precision": 0.7372773536895675,
-      "eval_recall": 0.53177334250975,
-      "eval_runtime": 26.3987,
-      "eval_samples_per_second": 307.212,
-      "eval_steps_per_second": 19.205,
-      "step": 17000
-    },
-    {
-      "epoch": 8.629191321499015,
-      "grad_norm": 0.4935995936393738,
-      "learning_rate": 2.7416173570019726e-06,
-      "loss": 0.0243,
-      "step": 17500
-    },
-    {
-      "epoch": 8.875739644970414,
-      "grad_norm": 0.5695288181304932,
-      "learning_rate": 2.2485207100591717e-06,
-      "loss": 0.0247,
-      "step": 18000
-    },
-    {
-      "epoch": 8.875739644970414,
-      "eval_accuracy": 0.9878693822016242,
-      "eval_f1": 0.6174748603351955,
-      "eval_loss": 0.035977210849523544,
-      "eval_precision": 0.7429554742955474,
-      "eval_recall": 0.5282557161428462,
-      "eval_runtime": 26.442,
-      "eval_samples_per_second": 306.709,
-      "eval_steps_per_second": 19.174,
-      "step": 18000
-    },
-    {
-      "epoch": 9.122287968441814,
-      "grad_norm": 0.3517369031906128,
-      "learning_rate": 1.755424063116371e-06,
-      "loss": 0.0238,
-      "step": 18500
-    },
-    {
-      "epoch": 9.368836291913215,
-      "grad_norm": 0.5258517265319824,
-      "learning_rate": 1.2623274161735703e-06,
-      "loss": 0.0236,
-      "step": 19000
-    },
-    {
-      "epoch": 9.368836291913215,
-      "eval_accuracy": 0.9880196153464574,
-      "eval_f1": 0.6215357062905753,
-      "eval_loss": 0.035947080701589584,
-      "eval_precision": 0.7497299632750054,
-      "eval_recall": 0.5307792307104076,
-      "eval_runtime": 26.2793,
-      "eval_samples_per_second": 308.608,
-      "eval_steps_per_second": 19.293,
-      "step": 19000
-    },
-    {
-      "epoch": 9.615384615384615,
-      "grad_norm": 0.4185117185115814,
-      "learning_rate": 7.692307692307694e-07,
-      "loss": 0.0233,
-      "step": 19500
-    },
-    {
-      "epoch": 9.861932938856016,
-      "grad_norm": 0.34625717997550964,
-      "learning_rate": 2.7613412228796843e-07,
-      "loss": 0.0229,
-      "step": 20000
-    },
-    {
-      "epoch": 9.861932938856016,
-      "eval_accuracy": 0.987977096531882,
-      "eval_f1": 0.6214467401490472,
-      "eval_loss": 0.03578707203269005,
-      "eval_precision": 0.7461423060437206,
-      "eval_recall": 0.5324615737554484,
-      "eval_runtime": 27.6359,
-      "eval_samples_per_second": 293.458,
-      "eval_steps_per_second": 18.346,
-      "step": 20000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 20280,
+  "max_steps": 202800,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 100,
   "save_steps": 1000,
   "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 5,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
     "TrainerControl": {
       "args": {
         "should_epoch_stop": false,
@@ -546,7 +61,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 452312037970056.0,
+  "total_flos": 22728529355328.0,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:71f491aa3c6c9aaa72371007033b32fdf24de876c9ab31ea4ee1caa9108217ee
+oid sha256:c3238ded44964259c0f3714028dd1f6d78a9114d80b100ba672efc9a29ee50e3
 size 5240