Training in progress, step 2544

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:04369c6e2112f3462af4f06ced86d9ed2cbcf1bcec738f9218de55fde5f2a1b8
+oid sha256:911ad2e3d0d2ebd06f39bfde975f4ec1fed24558e2f4f205bc9e42c412daaceb
 size 268290900

run-0/checkpoint-1500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4f0bf19f903d94a77d6fc39c81beef687529037ca75b455f113c38ac700023bf
+oid sha256:c8441570272d35ed36e9f18f6bd0d16527a9e2d42710eb1d51026269bacf9cb1
 size 268290900

run-0/checkpoint-1500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:194907f3ae8742e7fc5889cc61ddc1e5363615ad3990f7d776865212e0ac157f
+oid sha256:0d2330c283faf1ef89e513293da27a0d9c98088e492878025ef9dd6a349d9fd2
 size 536643898

run-0/checkpoint-1500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:693ec081eb0fa0af7d13e289edb194450dd19fdada23346e2af4292ea228535f
+oid sha256:eb7b4645e59a5c5cb66f63e8bcb5f006c535fd0ea63db9ae152dd586fd465b28
 size 1064

run-0/checkpoint-1500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f723f6fd73eb11c5c4d9d54d74d86750bd77c3dcce4973543d3debbd02258e20
+oid sha256:a795e38fc17d23ea7e491b0b97849dc60c777dbce157fcfdc837d87b44f51896
 size 5176

run-0/checkpoint-2000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ae54fbb8e47d41dffd5b41d2dfb55c4acea8da5d4685675cfe8991143a3ba916
+oid sha256:9872b439699dbc1d99eb062dc66705f88905f7a575b870b84c01a510f613ddeb
 size 268290900

run-0/checkpoint-2000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4e809fba9c2258238643b58736855d30f8f1b5422c045a02dcf37605ecf079c1
+oid sha256:f3bccc0dac79f644d34cb7a7615417a1ac91d2536eed7d00ebf5d0ab36d44f0f
 size 536643898

run-0/checkpoint-2000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:67d1a8a9215f18c672cf798b531337b9e0636922d8c4373e76080f8fec19ceb2
+oid sha256:2bf6d4c2c13eda2c2dfacdbfc8684055c5603a92b6c0c386d7c757b108195d12
 size 1064

run-0/checkpoint-2000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f723f6fd73eb11c5c4d9d54d74d86750bd77c3dcce4973543d3debbd02258e20
+oid sha256:a795e38fc17d23ea7e491b0b97849dc60c777dbce157fcfdc837d87b44f51896
 size 5176

run-0/checkpoint-2500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37141eb81c320a95907118f3844d6ff0e5ca494a0131ed566d10d41260aff8c2
+oid sha256:13e1427b12cea77de5548b32cd6b78135c0f5818b2b8494d97fd918f717bc16a
 size 268290900

run-0/checkpoint-2500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:de6c8d6cf573fa25320d04b662fe1f3c10658244b1801522f5b6b1dd785644bb
+oid sha256:864723fe4b0523145bbbd5adb4f48699e6ed52fe6e7ad9ab7f0a3f1fae6f1d55
 size 536643898

run-0/checkpoint-2500/scheduler.pt

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

run-0/checkpoint-2500/trainer_state.json

@@ -10,107 +10,107 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5861290322580646,
-      "eval_loss": 0.19721022248268127,
-      "eval_runtime": 6.85,
-      "eval_samples_per_second": 452.554,
-      "eval_steps_per_second": 9.489,
+      "eval_accuracy": 0.6758064516129032,
+      "eval_loss": 0.40983492136001587,
+      "eval_runtime": 5.8092,
+      "eval_samples_per_second": 533.639,
+      "eval_steps_per_second": 11.189,
       "step": 318
     },
     {
       "epoch": 1.5723270440251573,
-      "grad_norm": 0.5261461734771729,
-      "learning_rate": 1.650593990216632e-05,
-      "loss": 0.3175,
+      "grad_norm": 0.86518794298172,
+      "learning_rate": 1.606918238993711e-05,
+      "loss": 0.645,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8193548387096774,
-      "eval_loss": 0.09563933312892914,
-      "eval_runtime": 5.9266,
-      "eval_samples_per_second": 523.068,
-      "eval_steps_per_second": 10.968,
+      "eval_accuracy": 0.8412903225806452,
+      "eval_loss": 0.142293781042099,
+      "eval_runtime": 6.1991,
+      "eval_samples_per_second": 500.073,
+      "eval_steps_per_second": 10.485,
       "step": 636
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8774193548387097,
-      "eval_loss": 0.06404151022434235,
-      "eval_runtime": 5.6772,
-      "eval_samples_per_second": 546.04,
-      "eval_steps_per_second": 11.449,
+      "eval_accuracy": 0.8938709677419355,
+      "eval_loss": 0.07486626505851746,
+      "eval_runtime": 5.5967,
+      "eval_samples_per_second": 553.895,
+      "eval_steps_per_second": 11.614,
       "step": 954
     },
     {
       "epoch": 3.1446540880503147,
-      "grad_norm": 0.44202694296836853,
-      "learning_rate": 1.3011879804332637e-05,
-      "loss": 0.1109,
+      "grad_norm": 0.7068584561347961,
+      "learning_rate": 1.2138364779874214e-05,
+      "loss": 0.1671,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8996774193548387,
-      "eval_loss": 0.04958628490567207,
-      "eval_runtime": 5.7185,
-      "eval_samples_per_second": 542.104,
-      "eval_steps_per_second": 11.367,
+      "eval_accuracy": 0.9141935483870968,
+      "eval_loss": 0.053671687841415405,
+      "eval_runtime": 5.7734,
+      "eval_samples_per_second": 536.949,
+      "eval_steps_per_second": 11.259,
       "step": 1272
     },
     {
       "epoch": 4.716981132075472,
-      "grad_norm": 0.3023432493209839,
-      "learning_rate": 9.517819706498952e-06,
-      "loss": 0.071,
+      "grad_norm": 0.464497447013855,
+      "learning_rate": 8.207547169811321e-06,
+      "loss": 0.0849,
       "step": 1500
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.9077419354838709,
-      "eval_loss": 0.04074610769748688,
-      "eval_runtime": 5.8617,
-      "eval_samples_per_second": 528.859,
-      "eval_steps_per_second": 11.089,
+      "eval_accuracy": 0.9238709677419354,
+      "eval_loss": 0.04483605548739433,
+      "eval_runtime": 6.0581,
+      "eval_samples_per_second": 511.712,
+      "eval_steps_per_second": 10.729,
       "step": 1590
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.9103225806451613,
-      "eval_loss": 0.035789892077445984,
-      "eval_runtime": 5.6338,
-      "eval_samples_per_second": 550.247,
-      "eval_steps_per_second": 11.537,
+      "eval_accuracy": 0.9261290322580645,
+      "eval_loss": 0.04038877785205841,
+      "eval_runtime": 5.6704,
+      "eval_samples_per_second": 546.695,
+      "eval_steps_per_second": 11.463,
       "step": 1908
     },
     {
       "epoch": 6.289308176100629,
-      "grad_norm": 0.2652527689933777,
-      "learning_rate": 6.02375960866527e-06,
-      "loss": 0.0559,
+      "grad_norm": 0.36644139885902405,
+      "learning_rate": 4.276729559748428e-06,
+      "loss": 0.0648,
       "step": 2000
     },
     {
       "epoch": 7.0,
-      "eval_accuracy": 0.9190322580645162,
-      "eval_loss": 0.03312822803854942,
-      "eval_runtime": 5.9997,
-      "eval_samples_per_second": 516.69,
-      "eval_steps_per_second": 10.834,
+      "eval_accuracy": 0.9312903225806451,
+      "eval_loss": 0.038099613040685654,
+      "eval_runtime": 5.8179,
+      "eval_samples_per_second": 532.835,
+      "eval_steps_per_second": 11.172,
       "step": 2226
     },
     {
       "epoch": 7.861635220125786,
-      "grad_norm": 0.27172261476516724,
-      "learning_rate": 2.5296995108315863e-06,
-      "loss": 0.0491,
+      "grad_norm": 0.39690807461738586,
+      "learning_rate": 3.459119496855346e-07,
+      "loss": 0.0579,
       "step": 2500
     }
   ],
   "logging_steps": 500,
-  "max_steps": 2862,
+  "max_steps": 2544,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 9,
+  "num_train_epochs": 8,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -124,12 +124,12 @@
       "attributes": {}
     }
   },
-  "total_flos": 651155886807636.0,
+  "total_flos": 698672024112468.0,
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.7760026697362031,
-    "num_train_epochs": 9,
-    "temperature": 12
+    "alpha": 0.3417365166940647,
+    "num_train_epochs": 8,
+    "temperature": 2
   }
 }

run-0/checkpoint-2500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f723f6fd73eb11c5c4d9d54d74d86750bd77c3dcce4973543d3debbd02258e20
+oid sha256:a795e38fc17d23ea7e491b0b97849dc60c777dbce157fcfdc837d87b44f51896
 size 5176

run-0/checkpoint-2544/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": "4.42.4",
+  "vocab_size": 30522
+}

run-0/checkpoint-2544/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:911ad2e3d0d2ebd06f39bfde975f4ec1fed24558e2f4f205bc9e42c412daaceb
+size 268290900

run-0/checkpoint-2544/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8354f2f99d5bbbe17e6c97d24264769b920eaeb2462df7ccb3cba1bacb145f8b
+size 536643898

run-0/checkpoint-2544/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6a7287bbe104afc32e0b289aa227410a1b5d10bc0c28d69081c1e0bfce4e57e
+size 14244

run-0/checkpoint-2544/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:214eac02be0209c50c951afc980c6dcbdad93d966fbfca385bff547659925bf8
+size 1064

run-0/checkpoint-2544/special_tokens_map.json

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

run-0/checkpoint-2544/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-0/checkpoint-2544/trainer_state.json

@@ -0,0 +1,135 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 8.0,
+  "eval_steps": 500,
+  "global_step": 2544,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.6758064516129032,
+      "eval_loss": 0.40983492136001587,
+      "eval_runtime": 5.8092,
+      "eval_samples_per_second": 533.639,
+      "eval_steps_per_second": 11.189,
+      "step": 318
+    },
+    {
+      "epoch": 1.5723270440251573,
+      "grad_norm": 0.86518794298172,
+      "learning_rate": 1.606918238993711e-05,
+      "loss": 0.645,
+      "step": 500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8412903225806452,
+      "eval_loss": 0.142293781042099,
+      "eval_runtime": 6.1991,
+      "eval_samples_per_second": 500.073,
+      "eval_steps_per_second": 10.485,
+      "step": 636
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8938709677419355,
+      "eval_loss": 0.07486626505851746,
+      "eval_runtime": 5.5967,
+      "eval_samples_per_second": 553.895,
+      "eval_steps_per_second": 11.614,
+      "step": 954
+    },
+    {
+      "epoch": 3.1446540880503147,
+      "grad_norm": 0.7068584561347961,
+      "learning_rate": 1.2138364779874214e-05,
+      "loss": 0.1671,
+      "step": 1000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9141935483870968,
+      "eval_loss": 0.053671687841415405,
+      "eval_runtime": 5.7734,
+      "eval_samples_per_second": 536.949,
+      "eval_steps_per_second": 11.259,
+      "step": 1272
+    },
+    {
+      "epoch": 4.716981132075472,
+      "grad_norm": 0.464497447013855,
+      "learning_rate": 8.207547169811321e-06,
+      "loss": 0.0849,
+      "step": 1500
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9238709677419354,
+      "eval_loss": 0.04483605548739433,
+      "eval_runtime": 6.0581,
+      "eval_samples_per_second": 511.712,
+      "eval_steps_per_second": 10.729,
+      "step": 1590
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9261290322580645,
+      "eval_loss": 0.04038877785205841,
+      "eval_runtime": 5.6704,
+      "eval_samples_per_second": 546.695,
+      "eval_steps_per_second": 11.463,
+      "step": 1908
+    },
+    {
+      "epoch": 6.289308176100629,
+      "grad_norm": 0.36644139885902405,
+      "learning_rate": 4.276729559748428e-06,
+      "loss": 0.0648,
+      "step": 2000
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9312903225806451,
+      "eval_loss": 0.038099613040685654,
+      "eval_runtime": 5.8179,
+      "eval_samples_per_second": 532.835,
+      "eval_steps_per_second": 11.172,
+      "step": 2226
+    },
+    {
+      "epoch": 7.861635220125786,
+      "grad_norm": 0.39690807461738586,
+      "learning_rate": 3.459119496855346e-07,
+      "loss": 0.0579,
+      "step": 2500
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 2544,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 8,
+  "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": 698672024112468.0,
+  "train_batch_size": 48,
+  "trial_name": null,
+  "trial_params": {
+    "alpha": 0.3417365166940647,
+    "num_train_epochs": 8,
+    "temperature": 2
+  }
+}

run-0/checkpoint-2544/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a795e38fc17d23ea7e491b0b97849dc60c777dbce157fcfdc837d87b44f51896
+size 5176

runs/Aug10_17-36-17_bc748cae6928/events.out.tfevents.1723317271.bc748cae6928.590.9

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c129deb002bc1610182c4d240fbc924510a6ed7c677ff900e4509b9297102713
-size 29156
+oid sha256:32dd61c6ed8211d54c4d55148530043bc4bdace6e9e0a20ba580931aeee24549
+size 30044