Training in progress, step 3498

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:317126ec6d8a3f2c90d7cec7cc1f69d8da7db62fe758f658e0c4a624eed85808
+oid sha256:84858a93d4119f469c5e2546173e1c8adceac2a2161aaabae542daec8ad0627e
 size 268290900

run-2/checkpoint-3498/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-3498/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84858a93d4119f469c5e2546173e1c8adceac2a2161aaabae542daec8ad0627e
+size 268290900

run-2/checkpoint-3498/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc4380034b48d1e03ea3e55fb61a07a49e45895968a418fb2c398ec98a121fe4
+size 536643898

run-2/checkpoint-3498/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6d9fe861d36f55d22d38a1a6527e0b0186e397a4999235f510693643125ee91
+size 14244

run-2/checkpoint-3498/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15a2f41f6c8f936b738823806a6a529f14ce9fae9c7ead30edc4833526538d8b
+size 1064

run-2/checkpoint-3498/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-3498/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-3498/trainer_state.json

@@ -0,0 +1,209 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 11.0,
+  "eval_steps": 500,
+  "global_step": 3498,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.9968553459119497,
+      "grad_norm": 0.429341197013855,
+      "learning_rate": 0.00013794880138427262,
+      "loss": 0.3323,
+      "step": 317
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8664516129032258,
+      "eval_loss": 0.07744947820901871,
+      "eval_runtime": 2.7326,
+      "eval_samples_per_second": 1134.466,
+      "eval_steps_per_second": 23.787,
+      "step": 318
+    },
+    {
+      "epoch": 1.9937106918238994,
+      "grad_norm": 0.2267828732728958,
+      "learning_rate": 0.00014653290645786366,
+      "loss": 0.0632,
+      "step": 634
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9290322580645162,
+      "eval_loss": 0.034380555152893066,
+      "eval_runtime": 2.6579,
+      "eval_samples_per_second": 1166.326,
+      "eval_steps_per_second": 24.455,
+      "step": 636
+    },
+    {
+      "epoch": 2.990566037735849,
+      "grad_norm": 0.17225514352321625,
+      "learning_rate": 0.00013031400584782778,
+      "loss": 0.0332,
+      "step": 951
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.937741935483871,
+      "eval_loss": 0.025218434631824493,
+      "eval_runtime": 2.6227,
+      "eval_samples_per_second": 1181.985,
+      "eval_steps_per_second": 24.784,
+      "step": 954
+    },
+    {
+      "epoch": 3.9874213836477987,
+      "grad_norm": 0.10644444078207016,
+      "learning_rate": 0.0001140951052377919,
+      "loss": 0.0254,
+      "step": 1268
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9448387096774193,
+      "eval_loss": 0.02074083872139454,
+      "eval_runtime": 2.6543,
+      "eval_samples_per_second": 1167.909,
+      "eval_steps_per_second": 24.488,
+      "step": 1272
+    },
+    {
+      "epoch": 4.984276729559748,
+      "grad_norm": 0.08869536221027374,
+      "learning_rate": 9.787620462775599e-05,
+      "loss": 0.0223,
+      "step": 1585
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9412903225806452,
+      "eval_loss": 0.020231781527400017,
+      "eval_runtime": 2.6557,
+      "eval_samples_per_second": 1167.32,
+      "eval_steps_per_second": 24.476,
+      "step": 1590
+    },
+    {
+      "epoch": 5.981132075471698,
+      "grad_norm": 0.08641496300697327,
+      "learning_rate": 8.16573040177201e-05,
+      "loss": 0.0201,
+      "step": 1902
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9441935483870968,
+      "eval_loss": 0.01878916658461094,
+      "eval_runtime": 2.635,
+      "eval_samples_per_second": 1176.481,
+      "eval_steps_per_second": 24.668,
+      "step": 1908
+    },
+    {
+      "epoch": 6.977987421383648,
+      "grad_norm": 0.07254982739686966,
+      "learning_rate": 6.543840340768423e-05,
+      "loss": 0.0183,
+      "step": 2219
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9467741935483871,
+      "eval_loss": 0.017883770167827606,
+      "eval_runtime": 2.651,
+      "eval_samples_per_second": 1169.38,
+      "eval_steps_per_second": 24.519,
+      "step": 2226
+    },
+    {
+      "epoch": 7.9748427672955975,
+      "grad_norm": 0.0713103711605072,
+      "learning_rate": 4.9219502797648335e-05,
+      "loss": 0.0171,
+      "step": 2536
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9493548387096774,
+      "eval_loss": 0.017295770347118378,
+      "eval_runtime": 2.6543,
+      "eval_samples_per_second": 1167.916,
+      "eval_steps_per_second": 24.489,
+      "step": 2544
+    },
+    {
+      "epoch": 8.971698113207546,
+      "grad_norm": 0.0737316906452179,
+      "learning_rate": 3.3000602187612454e-05,
+      "loss": 0.0162,
+      "step": 2853
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9480645161290323,
+      "eval_loss": 0.01655430532991886,
+      "eval_runtime": 2.6551,
+      "eval_samples_per_second": 1167.55,
+      "eval_steps_per_second": 24.481,
+      "step": 2862
+    },
+    {
+      "epoch": 9.968553459119496,
+      "grad_norm": 0.06824547797441483,
+      "learning_rate": 1.6781701577576563e-05,
+      "loss": 0.0155,
+      "step": 3170
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.947741935483871,
+      "eval_loss": 0.016286808997392654,
+      "eval_runtime": 2.6473,
+      "eval_samples_per_second": 1170.992,
+      "eval_steps_per_second": 24.553,
+      "step": 3180
+    },
+    {
+      "epoch": 10.965408805031446,
+      "grad_norm": 0.058784905821084976,
+      "learning_rate": 5.628009675406775e-07,
+      "loss": 0.0149,
+      "step": 3487
+    }
+  ],
+  "logging_steps": 317,
+  "max_steps": 3498,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 11,
+  "save_steps": 1000000000.0,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 907183422801936.0,
+  "train_batch_size": 48,
+  "trial_name": null,
+  "trial_params": {
+    "alpha": 0.5336013283534605,
+    "fp16": true,
+    "learning_rate": 0.00016014245712748368,
+    "lr_scheduler": "cosine",
+    "num_train_epochs": 11,
+    "temperature": 9,
+    "warmup_steps": 368,
+    "weight_decay": 0.06096677657018104
+  }
+}

run-2/checkpoint-3498/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2566b0f411ad51fb0202ffe64a8d3a8e1fc1c209fd2b66172dc9e60005165aed
+size 5240

runs/Oct02_17-16-46_01f54d6a8452/events.out.tfevents.1727891833.01f54d6a8452.2040.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:561aa4cd98b6250cff7b08e84c54a8c65b561b32082658594d40ef86f94bfa5b
+size 18715

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc8932ec57c460304ed4be82a2209d7cbf66b9151826791db21403076d84aeb5
+oid sha256:2566b0f411ad51fb0202ffe64a8d3a8e1fc1c209fd2b66172dc9e60005165aed
 size 5240