Training in progress, step 2862

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a5ead783c8cd394f3b7cff5bb9210de7d7720379c213b3f0ee696e1fc5a9d7e7
+oid sha256:76cadb93027e5ec338e6724a0b24825a1fc09abf2eef793105f9414585309730
 size 268290900

run-2/checkpoint-2500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0d1660491774a1f9d0cb694570aa3990f735a7dd9bed466f984bafab6467a23e
+oid sha256:6623ae75fa42bc651bb78d6a329587e2a40ed8c59df585f178a98ba6e2df18ff
 size 268290900

run-2/checkpoint-2500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4a951e9b9b4cadebbb0b6d1ffb0443200eb5604256b6bc66f189d0db13e26bc
+oid sha256:3aeb58ef529b4a5188f21c262ab45c9233aa2fcbad4984461dde21f2136039cf
 size 536643898

run-2/checkpoint-2500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2b997ea9328158f28ff50e49c476931e703b821360bc8325e0c4d100e032c865
+oid sha256:b7c943a77651a184c6653dd4706c14c9a27d52136e81942a2b361c74acb9f665
 size 1064

run-2/checkpoint-2500/trainer_state.json

@@ -10,107 +10,107 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5738709677419355,
-      "eval_loss": 0.1914977878332138,
-      "eval_runtime": 5.3865,
-      "eval_samples_per_second": 575.517,
-      "eval_steps_per_second": 12.067,
+      "eval_accuracy": 0.5948387096774194,
+      "eval_loss": 0.2074165940284729,
+      "eval_runtime": 5.1592,
+      "eval_samples_per_second": 600.867,
+      "eval_steps_per_second": 12.599,
       "step": 318
     },
     {
       "epoch": 1.5723270440251573,
-      "grad_norm": 0.5165097713470459,
-      "learning_rate": 1.606918238993711e-05,
-      "loss": 0.3072,
+      "grad_norm": 0.538582980632782,
+      "learning_rate": 1.650593990216632e-05,
+      "loss": 0.3336,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8141935483870968,
-      "eval_loss": 0.09490247070789337,
-      "eval_runtime": 5.6092,
-      "eval_samples_per_second": 552.66,
-      "eval_steps_per_second": 11.588,
+      "eval_accuracy": 0.8258064516129032,
+      "eval_loss": 0.09868818521499634,
+      "eval_runtime": 5.3751,
+      "eval_samples_per_second": 576.732,
+      "eval_steps_per_second": 12.093,
       "step": 636
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8741935483870967,
-      "eval_loss": 0.06484736502170563,
-      "eval_runtime": 6.0767,
-      "eval_samples_per_second": 510.146,
-      "eval_steps_per_second": 10.697,
+      "eval_accuracy": 0.8829032258064516,
+      "eval_loss": 0.06500214338302612,
+      "eval_runtime": 5.0949,
+      "eval_samples_per_second": 608.447,
+      "eval_steps_per_second": 12.758,
       "step": 954
     },
     {
       "epoch": 3.1446540880503147,
-      "grad_norm": 0.43419399857521057,
-      "learning_rate": 1.2138364779874214e-05,
-      "loss": 0.1097,
+      "grad_norm": 0.4604446291923523,
+      "learning_rate": 1.3011879804332637e-05,
+      "loss": 0.1147,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8958064516129032,
-      "eval_loss": 0.05110108479857445,
-      "eval_runtime": 5.4249,
-      "eval_samples_per_second": 571.437,
-      "eval_steps_per_second": 11.982,
+      "eval_accuracy": 0.8996774193548387,
+      "eval_loss": 0.049765754491090775,
+      "eval_runtime": 5.2676,
+      "eval_samples_per_second": 588.5,
+      "eval_steps_per_second": 12.34,
       "step": 1272
     },
     {
       "epoch": 4.716981132075472,
-      "grad_norm": 0.2985864579677582,
-      "learning_rate": 8.207547169811321e-06,
-      "loss": 0.072,
+      "grad_norm": 0.31128990650177,
+      "learning_rate": 9.517819706498952e-06,
+      "loss": 0.0724,
       "step": 1500
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.9048387096774193,
-      "eval_loss": 0.04280891269445419,
-      "eval_runtime": 5.5512,
-      "eval_samples_per_second": 558.437,
-      "eval_steps_per_second": 11.709,
+      "eval_accuracy": 0.91,
+      "eval_loss": 0.04065703973174095,
+      "eval_runtime": 5.2744,
+      "eval_samples_per_second": 587.744,
+      "eval_steps_per_second": 12.324,
       "step": 1590
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.9074193548387097,
-      "eval_loss": 0.038190603256225586,
-      "eval_runtime": 5.4902,
-      "eval_samples_per_second": 564.646,
-      "eval_steps_per_second": 11.839,
+      "eval_accuracy": 0.912258064516129,
+      "eval_loss": 0.03559587895870209,
+      "eval_runtime": 5.1021,
+      "eval_samples_per_second": 607.594,
+      "eval_steps_per_second": 12.74,
       "step": 1908
     },
     {
       "epoch": 6.289308176100629,
-      "grad_norm": 0.2691451609134674,
-      "learning_rate": 4.276729559748428e-06,
-      "loss": 0.0579,
+      "grad_norm": 0.269597589969635,
+      "learning_rate": 6.02375960866527e-06,
+      "loss": 0.0567,
       "step": 2000
     },
     {
       "epoch": 7.0,
-      "eval_accuracy": 0.9116129032258065,
-      "eval_loss": 0.03593328222632408,
-      "eval_runtime": 5.5022,
-      "eval_samples_per_second": 563.415,
-      "eval_steps_per_second": 11.814,
+      "eval_accuracy": 0.9203225806451613,
+      "eval_loss": 0.032886430621147156,
+      "eval_runtime": 5.44,
+      "eval_samples_per_second": 569.85,
+      "eval_steps_per_second": 11.948,
       "step": 2226
     },
     {
       "epoch": 7.861635220125786,
-      "grad_norm": 0.265292763710022,
-      "learning_rate": 3.459119496855346e-07,
-      "loss": 0.0522,
+      "grad_norm": 0.28186649084091187,
+      "learning_rate": 2.5296995108315863e-06,
+      "loss": 0.0496,
       "step": 2500
     }
   ],
   "logging_steps": 500,
-  "max_steps": 2544,
+  "max_steps": 2862,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 8,
+  "num_train_epochs": 9,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -128,8 +128,8 @@
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.5962321311067933,
-    "num_train_epochs": 8,
-    "temperature": 20
+    "alpha": 0.186668286700706,
+    "num_train_epochs": 9,
+    "temperature": 8
   }
 }

run-2/checkpoint-2500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:605b5083623cb90e46bc6aa64e7e8f2cfc283cd343d0688245d2c254a0200467
+oid sha256:e09a3826ab4ffc76c9981d2ccde475f4c95a27c3ed7ffae8225ce76a3cd34eeb
 size 5240

run-2/checkpoint-2862/config.json

@@ -0,0 +1,331 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "id2label": {
+    "0": "restaurant_reviews",
+    "1": "nutrition_info",
+    "2": "account_blocked",
+    "3": "oil_change_how",
+    "4": "time",
+    "5": "weather",
+    "6": "redeem_rewards",
+    "7": "interest_rate",
+    "8": "gas_type",
+    "9": "accept_reservations",
+    "10": "smart_home",
+    "11": "user_name",
+    "12": "report_lost_card",
+    "13": "repeat",
+    "14": "whisper_mode",
+    "15": "what_are_your_hobbies",
+    "16": "order",
+    "17": "jump_start",
+    "18": "schedule_meeting",
+    "19": "meeting_schedule",
+    "20": "freeze_account",
+    "21": "what_song",
+    "22": "meaning_of_life",
+    "23": "restaurant_reservation",
+    "24": "traffic",
+    "25": "make_call",
+    "26": "text",
+    "27": "bill_balance",
+    "28": "improve_credit_score",
+    "29": "change_language",
+    "30": "no",
+    "31": "measurement_conversion",
+    "32": "timer",
+    "33": "flip_coin",
+    "34": "do_you_have_pets",
+    "35": "balance",
+    "36": "tell_joke",
+    "37": "last_maintenance",
+    "38": "exchange_rate",
+    "39": "uber",
+    "40": "car_rental",
+    "41": "credit_limit",
+    "42": "oos",
+    "43": "shopping_list",
+    "44": "expiration_date",
+    "45": "routing",
+    "46": "meal_suggestion",
+    "47": "tire_change",
+    "48": "todo_list",
+    "49": "card_declined",
+    "50": "rewards_balance",
+    "51": "change_accent",
+    "52": "vaccines",
+    "53": "reminder_update",
+    "54": "food_last",
+    "55": "change_ai_name",
+    "56": "bill_due",
+    "57": "who_do_you_work_for",
+    "58": "share_location",
+    "59": "international_visa",
+    "60": "calendar",
+    "61": "translate",
+    "62": "carry_on",
+    "63": "book_flight",
+    "64": "insurance_change",
+    "65": "todo_list_update",
+    "66": "timezone",
+    "67": "cancel_reservation",
+    "68": "transactions",
+    "69": "credit_score",
+    "70": "report_fraud",
+    "71": "spending_history",
+    "72": "directions",
+    "73": "spelling",
+    "74": "insurance",
+    "75": "what_is_your_name",
+    "76": "reminder",
+    "77": "where_are_you_from",
+    "78": "distance",
+    "79": "payday",
+    "80": "flight_status",
+    "81": "find_phone",
+    "82": "greeting",
+    "83": "alarm",
+    "84": "order_status",
+    "85": "confirm_reservation",
+    "86": "cook_time",
+    "87": "damaged_card",
+    "88": "reset_settings",
+    "89": "pin_change",
+    "90": "replacement_card_duration",
+    "91": "new_card",
+    "92": "roll_dice",
+    "93": "income",
+    "94": "taxes",
+    "95": "date",
+    "96": "who_made_you",
+    "97": "pto_request",
+    "98": "tire_pressure",
+    "99": "how_old_are_you",
+    "100": "rollover_401k",
+    "101": "pto_request_status",
+    "102": "how_busy",
+    "103": "application_status",
+    "104": "recipe",
+    "105": "calendar_update",
+    "106": "play_music",
+    "107": "yes",
+    "108": "direct_deposit",
+    "109": "credit_limit_change",
+    "110": "gas",
+    "111": "pay_bill",
+    "112": "ingredients_list",
+    "113": "lost_luggage",
+    "114": "goodbye",
+    "115": "what_can_i_ask_you",
+    "116": "book_hotel",
+    "117": "are_you_a_bot",
+    "118": "next_song",
+    "119": "change_speed",
+    "120": "plug_type",
+    "121": "maybe",
+    "122": "w2",
+    "123": "oil_change_when",
+    "124": "thank_you",
+    "125": "shopping_list_update",
+    "126": "pto_balance",
+    "127": "order_checks",
+    "128": "travel_alert",
+    "129": "fun_fact",
+    "130": "sync_device",
+    "131": "schedule_maintenance",
+    "132": "apr",
+    "133": "transfer",
+    "134": "ingredient_substitution",
+    "135": "calories",
+    "136": "current_location",
+    "137": "international_fees",
+    "138": "calculator",
+    "139": "definition",
+    "140": "next_holiday",
+    "141": "update_playlist",
+    "142": "mpg",
+    "143": "min_payment",
+    "144": "change_user_name",
+    "145": "restaurant_suggestion",
+    "146": "travel_notification",
+    "147": "cancel",
+    "148": "pto_used",
+    "149": "travel_suggestion",
+    "150": "change_volume"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "accept_reservations": 9,
+    "account_blocked": 2,
+    "alarm": 83,
+    "application_status": 103,
+    "apr": 132,
+    "are_you_a_bot": 117,
+    "balance": 35,
+    "bill_balance": 27,
+    "bill_due": 56,
+    "book_flight": 63,
+    "book_hotel": 116,
+    "calculator": 138,
+    "calendar": 60,
+    "calendar_update": 105,
+    "calories": 135,
+    "cancel": 147,
+    "cancel_reservation": 67,
+    "car_rental": 40,
+    "card_declined": 49,
+    "carry_on": 62,
+    "change_accent": 51,
+    "change_ai_name": 55,
+    "change_language": 29,
+    "change_speed": 119,
+    "change_user_name": 144,
+    "change_volume": 150,
+    "confirm_reservation": 85,
+    "cook_time": 86,
+    "credit_limit": 41,
+    "credit_limit_change": 109,
+    "credit_score": 69,
+    "current_location": 136,
+    "damaged_card": 87,
+    "date": 95,
+    "definition": 139,
+    "direct_deposit": 108,
+    "directions": 72,
+    "distance": 78,
+    "do_you_have_pets": 34,
+    "exchange_rate": 38,
+    "expiration_date": 44,
+    "find_phone": 81,
+    "flight_status": 80,
+    "flip_coin": 33,
+    "food_last": 54,
+    "freeze_account": 20,
+    "fun_fact": 129,
+    "gas": 110,
+    "gas_type": 8,
+    "goodbye": 114,
+    "greeting": 82,
+    "how_busy": 102,
+    "how_old_are_you": 99,
+    "improve_credit_score": 28,
+    "income": 93,
+    "ingredient_substitution": 134,
+    "ingredients_list": 112,
+    "insurance": 74,
+    "insurance_change": 64,
+    "interest_rate": 7,
+    "international_fees": 137,
+    "international_visa": 59,
+    "jump_start": 17,
+    "last_maintenance": 37,
+    "lost_luggage": 113,
+    "make_call": 25,
+    "maybe": 121,
+    "meal_suggestion": 46,
+    "meaning_of_life": 22,
+    "measurement_conversion": 31,
+    "meeting_schedule": 19,
+    "min_payment": 143,
+    "mpg": 142,
+    "new_card": 91,
+    "next_holiday": 140,
+    "next_song": 118,
+    "no": 30,
+    "nutrition_info": 1,
+    "oil_change_how": 3,
+    "oil_change_when": 123,
+    "oos": 42,
+    "order": 16,
+    "order_checks": 127,
+    "order_status": 84,
+    "pay_bill": 111,
+    "payday": 79,
+    "pin_change": 89,
+    "play_music": 106,
+    "plug_type": 120,
+    "pto_balance": 126,
+    "pto_request": 97,
+    "pto_request_status": 101,
+    "pto_used": 148,
+    "recipe": 104,
+    "redeem_rewards": 6,
+    "reminder": 76,
+    "reminder_update": 53,
+    "repeat": 13,
+    "replacement_card_duration": 90,
+    "report_fraud": 70,
+    "report_lost_card": 12,
+    "reset_settings": 88,
+    "restaurant_reservation": 23,
+    "restaurant_reviews": 0,
+    "restaurant_suggestion": 145,
+    "rewards_balance": 50,
+    "roll_dice": 92,
+    "rollover_401k": 100,
+    "routing": 45,
+    "schedule_maintenance": 131,
+    "schedule_meeting": 18,
+    "share_location": 58,
+    "shopping_list": 43,
+    "shopping_list_update": 125,
+    "smart_home": 10,
+    "spelling": 73,
+    "spending_history": 71,
+    "sync_device": 130,
+    "taxes": 94,
+    "tell_joke": 36,
+    "text": 26,
+    "thank_you": 124,
+    "time": 4,
+    "timer": 32,
+    "timezone": 66,
+    "tire_change": 47,
+    "tire_pressure": 98,
+    "todo_list": 48,
+    "todo_list_update": 65,
+    "traffic": 24,
+    "transactions": 68,
+    "transfer": 133,
+    "translate": 61,
+    "travel_alert": 128,
+    "travel_notification": 146,
+    "travel_suggestion": 149,
+    "uber": 39,
+    "update_playlist": 141,
+    "user_name": 11,
+    "vaccines": 52,
+    "w2": 122,
+    "weather": 5,
+    "what_are_your_hobbies": 15,
+    "what_can_i_ask_you": 115,
+    "what_is_your_name": 75,
+    "what_song": 21,
+    "where_are_you_from": 77,
+    "whisper_mode": 14,
+    "who_do_you_work_for": 57,
+    "who_made_you": 96,
+    "yes": 107
+  },
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.44.2",
+  "vocab_size": 30522
+}

run-2/checkpoint-2862/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76cadb93027e5ec338e6724a0b24825a1fc09abf2eef793105f9414585309730
+size 268290900

run-2/checkpoint-2862/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8667c798442f1b02b7b39c7231255f4e609b3a1cb62d4ee309ff3b2f187a645
+size 536643898

run-2/checkpoint-2862/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb284f354e17b6a985f5b0a1ec258af1589c2ddbdcd17b3670ab73d9e89918e1
+size 14244

run-2/checkpoint-2862/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a738824491ff8b261b6db2a842e15bf80355d04a07892a16df2c4715b2ee556
+size 1064

run-2/checkpoint-2862/special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

run-2/checkpoint-2862/tokenizer_config.json

@@ -0,0 +1,55 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

run-2/checkpoint-2862/trainer_state.json

@@ -0,0 +1,144 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 9.0,
+  "eval_steps": 500,
+  "global_step": 2862,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.5948387096774194,
+      "eval_loss": 0.2074165940284729,
+      "eval_runtime": 5.1592,
+      "eval_samples_per_second": 600.867,
+      "eval_steps_per_second": 12.599,
+      "step": 318
+    },
+    {
+      "epoch": 1.5723270440251573,
+      "grad_norm": 0.538582980632782,
+      "learning_rate": 1.650593990216632e-05,
+      "loss": 0.3336,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8258064516129032,
+      "eval_loss": 0.09868818521499634,
+      "eval_runtime": 5.3751,
+      "eval_samples_per_second": 576.732,
+      "eval_steps_per_second": 12.093,
+      "step": 636
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8829032258064516,
+      "eval_loss": 0.06500214338302612,
+      "eval_runtime": 5.0949,
+      "eval_samples_per_second": 608.447,
+      "eval_steps_per_second": 12.758,
+      "step": 954
+    },
+    {
+      "epoch": 3.1446540880503147,
+      "grad_norm": 0.4604446291923523,
+      "learning_rate": 1.3011879804332637e-05,
+      "loss": 0.1147,
+      "step": 1000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8996774193548387,
+      "eval_loss": 0.049765754491090775,
+      "eval_runtime": 5.2676,
+      "eval_samples_per_second": 588.5,
+      "eval_steps_per_second": 12.34,
+      "step": 1272
+    },
+    {
+      "epoch": 4.716981132075472,
+      "grad_norm": 0.31128990650177,
+      "learning_rate": 9.517819706498952e-06,
+      "loss": 0.0724,
+      "step": 1500
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.91,
+      "eval_loss": 0.04065703973174095,
+      "eval_runtime": 5.2744,
+      "eval_samples_per_second": 587.744,
+      "eval_steps_per_second": 12.324,
+      "step": 1590
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.912258064516129,
+      "eval_loss": 0.03559587895870209,
+      "eval_runtime": 5.1021,
+      "eval_samples_per_second": 607.594,
+      "eval_steps_per_second": 12.74,
+      "step": 1908
+    },
+    {
+      "epoch": 6.289308176100629,
+      "grad_norm": 0.269597589969635,
+      "learning_rate": 6.02375960866527e-06,
+      "loss": 0.0567,
+      "step": 2000
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9203225806451613,
+      "eval_loss": 0.032886430621147156,
+      "eval_runtime": 5.44,
+      "eval_samples_per_second": 569.85,
+      "eval_steps_per_second": 11.948,
+      "step": 2226
+    },
+    {
+      "epoch": 7.861635220125786,
+      "grad_norm": 0.28186649084091187,
+      "learning_rate": 2.5296995108315863e-06,
+      "loss": 0.0496,
+      "step": 2500
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9251612903225807,
+      "eval_loss": 0.03102920949459076,
+      "eval_runtime": 5.1362,
+      "eval_samples_per_second": 603.565,
+      "eval_steps_per_second": 12.655,
+      "step": 2544
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 2862,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 9,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 651155886807636.0,
+  "train_batch_size": 48,
+  "trial_name": null,
+  "trial_params": {
+    "alpha": 0.186668286700706,
+    "num_train_epochs": 9,
+    "temperature": 8
+  }
+}

run-2/checkpoint-2862/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e09a3826ab4ffc76c9981d2ccde475f4c95a27c3ed7ffae8225ce76a3cd34eeb
+size 5240

runs/Oct11_14-59-09_984d9c1bc9e1/events.out.tfevents.1728659967.984d9c1bc9e1.1239.3

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9b17341e5a06c4aec47da48b8c6de627b5668536e6c1b500d5592dafe02385de
-size 15546
+oid sha256:3811fb78909962c92f90f74639e572a48c46429d4c5ee3f956f5d629085de5ef
+size 16757