hndc commited on
Commit
6da1dbc
1 Parent(s): 7e3d9f5

Training in progress, step 2500

Browse files
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0f9dabcdddabb3935e507a3e58ad176dc78f06793f6bb9f6905d3a7b80abe651
3
  size 268290900
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ab3a110e6d712c68628a3e4063442b8b8e554daf23e53e67eace5ba1727b245
3
  size 268290900
run-6/checkpoint-2500/config.json ADDED
@@ -0,0 +1,331 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "distilbert-base-uncased",
3
+ "activation": "gelu",
4
+ "architectures": [
5
+ "DistilBertForSequenceClassification"
6
+ ],
7
+ "attention_dropout": 0.1,
8
+ "dim": 768,
9
+ "dropout": 0.1,
10
+ "hidden_dim": 3072,
11
+ "id2label": {
12
+ "0": "restaurant_reviews",
13
+ "1": "nutrition_info",
14
+ "2": "account_blocked",
15
+ "3": "oil_change_how",
16
+ "4": "time",
17
+ "5": "weather",
18
+ "6": "redeem_rewards",
19
+ "7": "interest_rate",
20
+ "8": "gas_type",
21
+ "9": "accept_reservations",
22
+ "10": "smart_home",
23
+ "11": "user_name",
24
+ "12": "report_lost_card",
25
+ "13": "repeat",
26
+ "14": "whisper_mode",
27
+ "15": "what_are_your_hobbies",
28
+ "16": "order",
29
+ "17": "jump_start",
30
+ "18": "schedule_meeting",
31
+ "19": "meeting_schedule",
32
+ "20": "freeze_account",
33
+ "21": "what_song",
34
+ "22": "meaning_of_life",
35
+ "23": "restaurant_reservation",
36
+ "24": "traffic",
37
+ "25": "make_call",
38
+ "26": "text",
39
+ "27": "bill_balance",
40
+ "28": "improve_credit_score",
41
+ "29": "change_language",
42
+ "30": "no",
43
+ "31": "measurement_conversion",
44
+ "32": "timer",
45
+ "33": "flip_coin",
46
+ "34": "do_you_have_pets",
47
+ "35": "balance",
48
+ "36": "tell_joke",
49
+ "37": "last_maintenance",
50
+ "38": "exchange_rate",
51
+ "39": "uber",
52
+ "40": "car_rental",
53
+ "41": "credit_limit",
54
+ "42": "oos",
55
+ "43": "shopping_list",
56
+ "44": "expiration_date",
57
+ "45": "routing",
58
+ "46": "meal_suggestion",
59
+ "47": "tire_change",
60
+ "48": "todo_list",
61
+ "49": "card_declined",
62
+ "50": "rewards_balance",
63
+ "51": "change_accent",
64
+ "52": "vaccines",
65
+ "53": "reminder_update",
66
+ "54": "food_last",
67
+ "55": "change_ai_name",
68
+ "56": "bill_due",
69
+ "57": "who_do_you_work_for",
70
+ "58": "share_location",
71
+ "59": "international_visa",
72
+ "60": "calendar",
73
+ "61": "translate",
74
+ "62": "carry_on",
75
+ "63": "book_flight",
76
+ "64": "insurance_change",
77
+ "65": "todo_list_update",
78
+ "66": "timezone",
79
+ "67": "cancel_reservation",
80
+ "68": "transactions",
81
+ "69": "credit_score",
82
+ "70": "report_fraud",
83
+ "71": "spending_history",
84
+ "72": "directions",
85
+ "73": "spelling",
86
+ "74": "insurance",
87
+ "75": "what_is_your_name",
88
+ "76": "reminder",
89
+ "77": "where_are_you_from",
90
+ "78": "distance",
91
+ "79": "payday",
92
+ "80": "flight_status",
93
+ "81": "find_phone",
94
+ "82": "greeting",
95
+ "83": "alarm",
96
+ "84": "order_status",
97
+ "85": "confirm_reservation",
98
+ "86": "cook_time",
99
+ "87": "damaged_card",
100
+ "88": "reset_settings",
101
+ "89": "pin_change",
102
+ "90": "replacement_card_duration",
103
+ "91": "new_card",
104
+ "92": "roll_dice",
105
+ "93": "income",
106
+ "94": "taxes",
107
+ "95": "date",
108
+ "96": "who_made_you",
109
+ "97": "pto_request",
110
+ "98": "tire_pressure",
111
+ "99": "how_old_are_you",
112
+ "100": "rollover_401k",
113
+ "101": "pto_request_status",
114
+ "102": "how_busy",
115
+ "103": "application_status",
116
+ "104": "recipe",
117
+ "105": "calendar_update",
118
+ "106": "play_music",
119
+ "107": "yes",
120
+ "108": "direct_deposit",
121
+ "109": "credit_limit_change",
122
+ "110": "gas",
123
+ "111": "pay_bill",
124
+ "112": "ingredients_list",
125
+ "113": "lost_luggage",
126
+ "114": "goodbye",
127
+ "115": "what_can_i_ask_you",
128
+ "116": "book_hotel",
129
+ "117": "are_you_a_bot",
130
+ "118": "next_song",
131
+ "119": "change_speed",
132
+ "120": "plug_type",
133
+ "121": "maybe",
134
+ "122": "w2",
135
+ "123": "oil_change_when",
136
+ "124": "thank_you",
137
+ "125": "shopping_list_update",
138
+ "126": "pto_balance",
139
+ "127": "order_checks",
140
+ "128": "travel_alert",
141
+ "129": "fun_fact",
142
+ "130": "sync_device",
143
+ "131": "schedule_maintenance",
144
+ "132": "apr",
145
+ "133": "transfer",
146
+ "134": "ingredient_substitution",
147
+ "135": "calories",
148
+ "136": "current_location",
149
+ "137": "international_fees",
150
+ "138": "calculator",
151
+ "139": "definition",
152
+ "140": "next_holiday",
153
+ "141": "update_playlist",
154
+ "142": "mpg",
155
+ "143": "min_payment",
156
+ "144": "change_user_name",
157
+ "145": "restaurant_suggestion",
158
+ "146": "travel_notification",
159
+ "147": "cancel",
160
+ "148": "pto_used",
161
+ "149": "travel_suggestion",
162
+ "150": "change_volume"
163
+ },
164
+ "initializer_range": 0.02,
165
+ "label2id": {
166
+ "accept_reservations": 9,
167
+ "account_blocked": 2,
168
+ "alarm": 83,
169
+ "application_status": 103,
170
+ "apr": 132,
171
+ "are_you_a_bot": 117,
172
+ "balance": 35,
173
+ "bill_balance": 27,
174
+ "bill_due": 56,
175
+ "book_flight": 63,
176
+ "book_hotel": 116,
177
+ "calculator": 138,
178
+ "calendar": 60,
179
+ "calendar_update": 105,
180
+ "calories": 135,
181
+ "cancel": 147,
182
+ "cancel_reservation": 67,
183
+ "car_rental": 40,
184
+ "card_declined": 49,
185
+ "carry_on": 62,
186
+ "change_accent": 51,
187
+ "change_ai_name": 55,
188
+ "change_language": 29,
189
+ "change_speed": 119,
190
+ "change_user_name": 144,
191
+ "change_volume": 150,
192
+ "confirm_reservation": 85,
193
+ "cook_time": 86,
194
+ "credit_limit": 41,
195
+ "credit_limit_change": 109,
196
+ "credit_score": 69,
197
+ "current_location": 136,
198
+ "damaged_card": 87,
199
+ "date": 95,
200
+ "definition": 139,
201
+ "direct_deposit": 108,
202
+ "directions": 72,
203
+ "distance": 78,
204
+ "do_you_have_pets": 34,
205
+ "exchange_rate": 38,
206
+ "expiration_date": 44,
207
+ "find_phone": 81,
208
+ "flight_status": 80,
209
+ "flip_coin": 33,
210
+ "food_last": 54,
211
+ "freeze_account": 20,
212
+ "fun_fact": 129,
213
+ "gas": 110,
214
+ "gas_type": 8,
215
+ "goodbye": 114,
216
+ "greeting": 82,
217
+ "how_busy": 102,
218
+ "how_old_are_you": 99,
219
+ "improve_credit_score": 28,
220
+ "income": 93,
221
+ "ingredient_substitution": 134,
222
+ "ingredients_list": 112,
223
+ "insurance": 74,
224
+ "insurance_change": 64,
225
+ "interest_rate": 7,
226
+ "international_fees": 137,
227
+ "international_visa": 59,
228
+ "jump_start": 17,
229
+ "last_maintenance": 37,
230
+ "lost_luggage": 113,
231
+ "make_call": 25,
232
+ "maybe": 121,
233
+ "meal_suggestion": 46,
234
+ "meaning_of_life": 22,
235
+ "measurement_conversion": 31,
236
+ "meeting_schedule": 19,
237
+ "min_payment": 143,
238
+ "mpg": 142,
239
+ "new_card": 91,
240
+ "next_holiday": 140,
241
+ "next_song": 118,
242
+ "no": 30,
243
+ "nutrition_info": 1,
244
+ "oil_change_how": 3,
245
+ "oil_change_when": 123,
246
+ "oos": 42,
247
+ "order": 16,
248
+ "order_checks": 127,
249
+ "order_status": 84,
250
+ "pay_bill": 111,
251
+ "payday": 79,
252
+ "pin_change": 89,
253
+ "play_music": 106,
254
+ "plug_type": 120,
255
+ "pto_balance": 126,
256
+ "pto_request": 97,
257
+ "pto_request_status": 101,
258
+ "pto_used": 148,
259
+ "recipe": 104,
260
+ "redeem_rewards": 6,
261
+ "reminder": 76,
262
+ "reminder_update": 53,
263
+ "repeat": 13,
264
+ "replacement_card_duration": 90,
265
+ "report_fraud": 70,
266
+ "report_lost_card": 12,
267
+ "reset_settings": 88,
268
+ "restaurant_reservation": 23,
269
+ "restaurant_reviews": 0,
270
+ "restaurant_suggestion": 145,
271
+ "rewards_balance": 50,
272
+ "roll_dice": 92,
273
+ "rollover_401k": 100,
274
+ "routing": 45,
275
+ "schedule_maintenance": 131,
276
+ "schedule_meeting": 18,
277
+ "share_location": 58,
278
+ "shopping_list": 43,
279
+ "shopping_list_update": 125,
280
+ "smart_home": 10,
281
+ "spelling": 73,
282
+ "spending_history": 71,
283
+ "sync_device": 130,
284
+ "taxes": 94,
285
+ "tell_joke": 36,
286
+ "text": 26,
287
+ "thank_you": 124,
288
+ "time": 4,
289
+ "timer": 32,
290
+ "timezone": 66,
291
+ "tire_change": 47,
292
+ "tire_pressure": 98,
293
+ "todo_list": 48,
294
+ "todo_list_update": 65,
295
+ "traffic": 24,
296
+ "transactions": 68,
297
+ "transfer": 133,
298
+ "translate": 61,
299
+ "travel_alert": 128,
300
+ "travel_notification": 146,
301
+ "travel_suggestion": 149,
302
+ "uber": 39,
303
+ "update_playlist": 141,
304
+ "user_name": 11,
305
+ "vaccines": 52,
306
+ "w2": 122,
307
+ "weather": 5,
308
+ "what_are_your_hobbies": 15,
309
+ "what_can_i_ask_you": 115,
310
+ "what_is_your_name": 75,
311
+ "what_song": 21,
312
+ "where_are_you_from": 77,
313
+ "whisper_mode": 14,
314
+ "who_do_you_work_for": 57,
315
+ "who_made_you": 96,
316
+ "yes": 107
317
+ },
318
+ "max_position_embeddings": 512,
319
+ "model_type": "distilbert",
320
+ "n_heads": 12,
321
+ "n_layers": 6,
322
+ "pad_token_id": 0,
323
+ "problem_type": "single_label_classification",
324
+ "qa_dropout": 0.1,
325
+ "seq_classif_dropout": 0.2,
326
+ "sinusoidal_pos_embds": false,
327
+ "tie_weights_": true,
328
+ "torch_dtype": "float32",
329
+ "transformers_version": "4.37.0",
330
+ "vocab_size": 30522
331
+ }
run-6/checkpoint-2500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ab3a110e6d712c68628a3e4063442b8b8e554daf23e53e67eace5ba1727b245
3
+ size 268290900
run-6/checkpoint-2500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c57438b8950b4b28138f3146c6d0d955283915bd7981c91fb2618e706b860178
3
+ size 536643898
run-6/checkpoint-2500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb87084ca039ffe048289dee8c9f62dcbc7a8cf00837c72835a7ebfe2c345795
3
+ size 14244
run-6/checkpoint-2500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d116f7be47ce88249a782ce7aafaf5b673a05bfa0ebdf9f14ab4c327c94c6b04
3
+ size 1064
run-6/checkpoint-2500/special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
run-6/checkpoint-2500/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
run-6/checkpoint-2500/tokenizer_config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "[PAD]",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "100": {
12
+ "content": "[UNK]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "101": {
20
+ "content": "[CLS]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "102": {
28
+ "content": "[SEP]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "103": {
36
+ "content": "[MASK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "clean_up_tokenization_spaces": true,
45
+ "cls_token": "[CLS]",
46
+ "do_lower_case": true,
47
+ "mask_token": "[MASK]",
48
+ "model_max_length": 512,
49
+ "pad_token": "[PAD]",
50
+ "sep_token": "[SEP]",
51
+ "strip_accents": null,
52
+ "tokenize_chinese_chars": true,
53
+ "tokenizer_class": "DistilBertTokenizer",
54
+ "unk_token": "[UNK]"
55
+ }
run-6/checkpoint-2500/trainer_state.json ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 7.861635220125786,
5
+ "eval_steps": 500,
6
+ "global_step": 2500,
7
+ "is_hyper_param_search": true,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "eval_accuracy": 0.6267741935483871,
14
+ "eval_loss": 0.19981417059898376,
15
+ "eval_runtime": 5.7482,
16
+ "eval_samples_per_second": 539.296,
17
+ "eval_steps_per_second": 11.308,
18
+ "step": 318
19
+ },
20
+ {
21
+ "epoch": 1.57,
22
+ "learning_rate": 1.685534591194969e-05,
23
+ "loss": 0.324,
24
+ "step": 500
25
+ },
26
+ {
27
+ "epoch": 2.0,
28
+ "eval_accuracy": 0.8287096774193549,
29
+ "eval_loss": 0.09645497798919678,
30
+ "eval_runtime": 5.8849,
31
+ "eval_samples_per_second": 526.769,
32
+ "eval_steps_per_second": 11.045,
33
+ "step": 636
34
+ },
35
+ {
36
+ "epoch": 3.0,
37
+ "eval_accuracy": 0.8841935483870967,
38
+ "eval_loss": 0.061595458537340164,
39
+ "eval_runtime": 5.6926,
40
+ "eval_samples_per_second": 544.564,
41
+ "eval_steps_per_second": 11.418,
42
+ "step": 954
43
+ },
44
+ {
45
+ "epoch": 3.14,
46
+ "learning_rate": 1.371069182389937e-05,
47
+ "loss": 0.1107,
48
+ "step": 1000
49
+ },
50
+ {
51
+ "epoch": 4.0,
52
+ "eval_accuracy": 0.9035483870967742,
53
+ "eval_loss": 0.046646881848573685,
54
+ "eval_runtime": 5.7741,
55
+ "eval_samples_per_second": 536.883,
56
+ "eval_steps_per_second": 11.257,
57
+ "step": 1272
58
+ },
59
+ {
60
+ "epoch": 4.72,
61
+ "learning_rate": 1.0566037735849058e-05,
62
+ "loss": 0.0692,
63
+ "step": 1500
64
+ },
65
+ {
66
+ "epoch": 5.0,
67
+ "eval_accuracy": 0.9096774193548387,
68
+ "eval_loss": 0.03824981674551964,
69
+ "eval_runtime": 5.8988,
70
+ "eval_samples_per_second": 525.532,
71
+ "eval_steps_per_second": 11.019,
72
+ "step": 1590
73
+ },
74
+ {
75
+ "epoch": 6.0,
76
+ "eval_accuracy": 0.9193548387096774,
77
+ "eval_loss": 0.033315952867269516,
78
+ "eval_runtime": 5.6959,
79
+ "eval_samples_per_second": 544.253,
80
+ "eval_steps_per_second": 11.412,
81
+ "step": 1908
82
+ },
83
+ {
84
+ "epoch": 6.29,
85
+ "learning_rate": 7.421383647798742e-06,
86
+ "loss": 0.0539,
87
+ "step": 2000
88
+ },
89
+ {
90
+ "epoch": 7.0,
91
+ "eval_accuracy": 0.9235483870967742,
92
+ "eval_loss": 0.03077365830540657,
93
+ "eval_runtime": 5.7622,
94
+ "eval_samples_per_second": 537.993,
95
+ "eval_steps_per_second": 11.28,
96
+ "step": 2226
97
+ },
98
+ {
99
+ "epoch": 7.86,
100
+ "learning_rate": 4.276729559748428e-06,
101
+ "loss": 0.0464,
102
+ "step": 2500
103
+ }
104
+ ],
105
+ "logging_steps": 500,
106
+ "max_steps": 3180,
107
+ "num_input_tokens_seen": 0,
108
+ "num_train_epochs": 10,
109
+ "save_steps": 500,
110
+ "total_flos": 708042017121348.0,
111
+ "train_batch_size": 48,
112
+ "trial_name": null,
113
+ "trial_params": {
114
+ "alpha": 0.8067006596645674,
115
+ "num_train_epochs": 10,
116
+ "temperature": 10
117
+ }
118
+ }
run-6/checkpoint-2500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57d040a4f719a987d2815ce2da25572380b6e8496ac1be556c18b875cd7deb62
3
+ size 4792
run-6/checkpoint-2500/vocab.txt ADDED
The diff for this file is too large to render. See raw diff
 
runs/Jan23_16-12-52_99e8694adbdf/events.out.tfevents.1706030977.99e8694adbdf.39534.8 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:56ae19e5ec01619174ced9b06b8ffe31861f23eede5f4f9f30f9687dc99e106b
3
- size 14784
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbadb4f6aff21c83c2a3c277a8c0c8b4eadd96c90b48b7ca9220b5add85273f0
3
+ size 15264