Upload with huggingface_hub
Browse files- config.json +298 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +917 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,298 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"adapter_attn_dim": null,
|
| 5 |
+
"adapter_kernel_size": 3,
|
| 6 |
+
"adapter_stride": 2,
|
| 7 |
+
"add_adapter": false,
|
| 8 |
+
"apply_spec_augment": true,
|
| 9 |
+
"architectures": [
|
| 10 |
+
"Wav2Vec2ForSequenceClassification"
|
| 11 |
+
],
|
| 12 |
+
"attention_dropout": 0.1,
|
| 13 |
+
"bos_token_id": 1,
|
| 14 |
+
"classifier_proj_size": 256,
|
| 15 |
+
"codevector_dim": 256,
|
| 16 |
+
"contrastive_logits_temperature": 0.1,
|
| 17 |
+
"conv_bias": false,
|
| 18 |
+
"conv_dim": [
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512,
|
| 25 |
+
512
|
| 26 |
+
],
|
| 27 |
+
"conv_kernel": [
|
| 28 |
+
10,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
3,
|
| 33 |
+
2,
|
| 34 |
+
2
|
| 35 |
+
],
|
| 36 |
+
"conv_stride": [
|
| 37 |
+
5,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2,
|
| 43 |
+
2
|
| 44 |
+
],
|
| 45 |
+
"ctc_loss_reduction": "sum",
|
| 46 |
+
"ctc_zero_infinity": false,
|
| 47 |
+
"diversity_loss_weight": 0.1,
|
| 48 |
+
"do_stable_layer_norm": false,
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_norm": "group",
|
| 52 |
+
"feat_proj_dropout": 0.1,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"freeze_feat_extract_train": true,
|
| 56 |
+
"hidden_act": "gelu",
|
| 57 |
+
"hidden_dropout": 0.1,
|
| 58 |
+
"hidden_size": 768,
|
| 59 |
+
"id2label": {
|
| 60 |
+
"0": "calendar_set",
|
| 61 |
+
"1": "wavs/audio_volume_mute",
|
| 62 |
+
"10": "iot_coffee",
|
| 63 |
+
"11": "calendar_remove",
|
| 64 |
+
"12": "email_sendemail",
|
| 65 |
+
"13": "general_joke",
|
| 66 |
+
"14": "lists_query",
|
| 67 |
+
"15": "iot_cleaning",
|
| 68 |
+
"16": "social_query",
|
| 69 |
+
"17": "cooking_recipe",
|
| 70 |
+
"18": "play_game",
|
| 71 |
+
"19": "weather_query",
|
| 72 |
+
"2": "qa_stock",
|
| 73 |
+
"20": "iot_hue_lightoff",
|
| 74 |
+
"21": "qa_factoid",
|
| 75 |
+
"22": "play_wavs/audiobook",
|
| 76 |
+
"23": "news_query",
|
| 77 |
+
"24": "qa_maths",
|
| 78 |
+
"25": "email_query",
|
| 79 |
+
"26": "recommendation_movies",
|
| 80 |
+
"27": "transport_traffic",
|
| 81 |
+
"28": "takeaway_order",
|
| 82 |
+
"29": "wavs/audio_volume_other",
|
| 83 |
+
"3": "social_post",
|
| 84 |
+
"30": "datetime_convert",
|
| 85 |
+
"31": "music",
|
| 86 |
+
"32": "recommendation_locations",
|
| 87 |
+
"33": "recommendation_events",
|
| 88 |
+
"34": "qa_currency",
|
| 89 |
+
"35": "lists_createoradd",
|
| 90 |
+
"36": "datetime_query",
|
| 91 |
+
"37": "transport_ticket",
|
| 92 |
+
"38": "takeaway_query",
|
| 93 |
+
"39": "general_greet",
|
| 94 |
+
"4": "play_radio",
|
| 95 |
+
"40": "qa_definition",
|
| 96 |
+
"41": "play_podcasts",
|
| 97 |
+
"42": "transport_taxi",
|
| 98 |
+
"43": "alarm_remove",
|
| 99 |
+
"44": "iot_hue_lightchange",
|
| 100 |
+
"45": "email_querycontact",
|
| 101 |
+
"46": "iot_hue_lightdim",
|
| 102 |
+
"47": "alarm_set",
|
| 103 |
+
"48": "iot_hue_lightup",
|
| 104 |
+
"49": "transport_query",
|
| 105 |
+
"5": "calendar_query",
|
| 106 |
+
"50": "iot_wemo_on",
|
| 107 |
+
"51": "music_likeness",
|
| 108 |
+
"52": "alarm_query",
|
| 109 |
+
"53": "music_dislikeness",
|
| 110 |
+
"54": "lists_remove",
|
| 111 |
+
"55": "iot_hue_lighton",
|
| 112 |
+
"56": "wavs/audio_volume_down",
|
| 113 |
+
"57": "factoid",
|
| 114 |
+
"58": "iot_wemo_off",
|
| 115 |
+
"59": "query",
|
| 116 |
+
"6": "music_query",
|
| 117 |
+
"60": "set",
|
| 118 |
+
"61": "quirky",
|
| 119 |
+
"62": "email_addcontact",
|
| 120 |
+
"63": "music_settings",
|
| 121 |
+
"64": "joke",
|
| 122 |
+
"65": "podcasts",
|
| 123 |
+
"66": "game",
|
| 124 |
+
"67": "coffee",
|
| 125 |
+
"68": "radio",
|
| 126 |
+
"69": "post",
|
| 127 |
+
"7": "general_quirky",
|
| 128 |
+
"70": "convert",
|
| 129 |
+
"71": "remove",
|
| 130 |
+
"72": "greet",
|
| 131 |
+
"73": "cooking_query",
|
| 132 |
+
"74": "sendemail",
|
| 133 |
+
"75": "traffic",
|
| 134 |
+
"76": "hue_lightup",
|
| 135 |
+
"77": "hue_lightoff",
|
| 136 |
+
"78": "currency",
|
| 137 |
+
"79": "wemo_off",
|
| 138 |
+
"8": "play_music",
|
| 139 |
+
"80": "hue_lightdim",
|
| 140 |
+
"81": "createoradd",
|
| 141 |
+
"82": "ticket",
|
| 142 |
+
"83": "volume_other",
|
| 143 |
+
"84": "cleaning",
|
| 144 |
+
"85": "querycontact",
|
| 145 |
+
"86": "wemo_on",
|
| 146 |
+
"87": "addcontact",
|
| 147 |
+
"9": "wavs/audio_volume_up"
|
| 148 |
+
},
|
| 149 |
+
"initializer_range": 0.02,
|
| 150 |
+
"intermediate_size": 3072,
|
| 151 |
+
"label2id": {
|
| 152 |
+
"addcontact": "87",
|
| 153 |
+
"alarm_query": "52",
|
| 154 |
+
"alarm_remove": "43",
|
| 155 |
+
"alarm_set": "47",
|
| 156 |
+
"calendar_query": "5",
|
| 157 |
+
"calendar_remove": "11",
|
| 158 |
+
"calendar_set": "0",
|
| 159 |
+
"cleaning": "84",
|
| 160 |
+
"coffee": "67",
|
| 161 |
+
"convert": "70",
|
| 162 |
+
"cooking_query": "73",
|
| 163 |
+
"cooking_recipe": "17",
|
| 164 |
+
"createoradd": "81",
|
| 165 |
+
"currency": "78",
|
| 166 |
+
"datetime_convert": "30",
|
| 167 |
+
"datetime_query": "36",
|
| 168 |
+
"email_addcontact": "62",
|
| 169 |
+
"email_query": "25",
|
| 170 |
+
"email_querycontact": "45",
|
| 171 |
+
"email_sendemail": "12",
|
| 172 |
+
"factoid": "57",
|
| 173 |
+
"game": "66",
|
| 174 |
+
"general_greet": "39",
|
| 175 |
+
"general_joke": "13",
|
| 176 |
+
"general_quirky": "7",
|
| 177 |
+
"greet": "72",
|
| 178 |
+
"hue_lightdim": "80",
|
| 179 |
+
"hue_lightoff": "77",
|
| 180 |
+
"hue_lightup": "76",
|
| 181 |
+
"iot_cleaning": "15",
|
| 182 |
+
"iot_coffee": "10",
|
| 183 |
+
"iot_hue_lightchange": "44",
|
| 184 |
+
"iot_hue_lightdim": "46",
|
| 185 |
+
"iot_hue_lightoff": "20",
|
| 186 |
+
"iot_hue_lighton": "55",
|
| 187 |
+
"iot_hue_lightup": "48",
|
| 188 |
+
"iot_wemo_off": "58",
|
| 189 |
+
"iot_wemo_on": "50",
|
| 190 |
+
"joke": "64",
|
| 191 |
+
"lists_createoradd": "35",
|
| 192 |
+
"lists_query": "14",
|
| 193 |
+
"lists_remove": "54",
|
| 194 |
+
"music": "31",
|
| 195 |
+
"music_dislikeness": "53",
|
| 196 |
+
"music_likeness": "51",
|
| 197 |
+
"music_query": "6",
|
| 198 |
+
"music_settings": "63",
|
| 199 |
+
"news_query": "23",
|
| 200 |
+
"play_game": "18",
|
| 201 |
+
"play_music": "8",
|
| 202 |
+
"play_podcasts": "41",
|
| 203 |
+
"play_radio": "4",
|
| 204 |
+
"play_wavs/audiobook": "22",
|
| 205 |
+
"podcasts": "65",
|
| 206 |
+
"post": "69",
|
| 207 |
+
"qa_currency": "34",
|
| 208 |
+
"qa_definition": "40",
|
| 209 |
+
"qa_factoid": "21",
|
| 210 |
+
"qa_maths": "24",
|
| 211 |
+
"qa_stock": "2",
|
| 212 |
+
"query": "59",
|
| 213 |
+
"querycontact": "85",
|
| 214 |
+
"quirky": "61",
|
| 215 |
+
"radio": "68",
|
| 216 |
+
"recommendation_events": "33",
|
| 217 |
+
"recommendation_locations": "32",
|
| 218 |
+
"recommendation_movies": "26",
|
| 219 |
+
"remove": "71",
|
| 220 |
+
"sendemail": "74",
|
| 221 |
+
"set": "60",
|
| 222 |
+
"social_post": "3",
|
| 223 |
+
"social_query": "16",
|
| 224 |
+
"takeaway_order": "28",
|
| 225 |
+
"takeaway_query": "38",
|
| 226 |
+
"ticket": "82",
|
| 227 |
+
"traffic": "75",
|
| 228 |
+
"transport_query": "49",
|
| 229 |
+
"transport_taxi": "42",
|
| 230 |
+
"transport_ticket": "37",
|
| 231 |
+
"transport_traffic": "27",
|
| 232 |
+
"volume_other": "83",
|
| 233 |
+
"wavs/audio_volume_down": "56",
|
| 234 |
+
"wavs/audio_volume_mute": "1",
|
| 235 |
+
"wavs/audio_volume_other": "29",
|
| 236 |
+
"wavs/audio_volume_up": "9",
|
| 237 |
+
"weather_query": "19",
|
| 238 |
+
"wemo_off": "79",
|
| 239 |
+
"wemo_on": "86"
|
| 240 |
+
},
|
| 241 |
+
"layer_norm_eps": 1e-05,
|
| 242 |
+
"layerdrop": 0.0,
|
| 243 |
+
"mask_channel_length": 10,
|
| 244 |
+
"mask_channel_min_space": 1,
|
| 245 |
+
"mask_channel_other": 0.0,
|
| 246 |
+
"mask_channel_prob": 0.0,
|
| 247 |
+
"mask_channel_selection": "static",
|
| 248 |
+
"mask_feature_length": 10,
|
| 249 |
+
"mask_feature_min_masks": 0,
|
| 250 |
+
"mask_feature_prob": 0.0,
|
| 251 |
+
"mask_time_length": 10,
|
| 252 |
+
"mask_time_min_masks": 2,
|
| 253 |
+
"mask_time_min_space": 1,
|
| 254 |
+
"mask_time_other": 0.0,
|
| 255 |
+
"mask_time_prob": 0.05,
|
| 256 |
+
"mask_time_selection": "static",
|
| 257 |
+
"model_type": "wav2vec2",
|
| 258 |
+
"no_mask_channel_overlap": false,
|
| 259 |
+
"no_mask_time_overlap": false,
|
| 260 |
+
"num_adapter_layers": 3,
|
| 261 |
+
"num_attention_heads": 12,
|
| 262 |
+
"num_codevector_groups": 2,
|
| 263 |
+
"num_codevectors_per_group": 320,
|
| 264 |
+
"num_conv_pos_embedding_groups": 16,
|
| 265 |
+
"num_conv_pos_embeddings": 128,
|
| 266 |
+
"num_feat_extract_layers": 7,
|
| 267 |
+
"num_hidden_layers": 12,
|
| 268 |
+
"num_negatives": 100,
|
| 269 |
+
"output_hidden_size": 768,
|
| 270 |
+
"pad_token_id": 0,
|
| 271 |
+
"proj_codevector_dim": 256,
|
| 272 |
+
"tdnn_dilation": [
|
| 273 |
+
1,
|
| 274 |
+
2,
|
| 275 |
+
3,
|
| 276 |
+
1,
|
| 277 |
+
1
|
| 278 |
+
],
|
| 279 |
+
"tdnn_dim": [
|
| 280 |
+
512,
|
| 281 |
+
512,
|
| 282 |
+
512,
|
| 283 |
+
512,
|
| 284 |
+
1500
|
| 285 |
+
],
|
| 286 |
+
"tdnn_kernel": [
|
| 287 |
+
5,
|
| 288 |
+
3,
|
| 289 |
+
3,
|
| 290 |
+
1,
|
| 291 |
+
1
|
| 292 |
+
],
|
| 293 |
+
"torch_dtype": "float32",
|
| 294 |
+
"transformers_version": "4.45.2",
|
| 295 |
+
"use_weighted_layer_sum": false,
|
| 296 |
+
"vocab_size": 32,
|
| 297 |
+
"xvector_output_dim": 512
|
| 298 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2eb595d9bffebf49d469d7ba67f390dd4a3bbbafe03327475cefbfcbfb2fbf9e
|
| 3 |
+
size 378390784
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7378f65dc440ef3323f4064ec7cdf0a00ec197d73ef4cb485cb02f4e166988c
|
| 3 |
+
size 756909370
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8d42fbb35be881a336af3b263d82ee7d97c2eca75108898734086b9ca86cfe08
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b1d6302a7353dd67847dce608afaaff547c670ef8ef4c52a5a5ac9772ebc0155
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,917 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8215622660344397,
|
| 3 |
+
"best_model_checkpoint": "results_unlearning/facebook/wav2vec2-base/42/checkpoint-26000",
|
| 4 |
+
"epoch": 57.30027548209367,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 26000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.1019283746556474,
|
| 13 |
+
"grad_norm": 3.954470634460449,
|
| 14 |
+
"learning_rate": 8.333333333333333e-05,
|
| 15 |
+
"loss": 3.9037,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.1019283746556474,
|
| 20 |
+
"eval_accuracy": 0.1856750686298977,
|
| 21 |
+
"eval_f1_macro": 0.0338842756461234,
|
| 22 |
+
"eval_loss": 3.2653889656066895,
|
| 23 |
+
"eval_runtime": 201.4102,
|
| 24 |
+
"eval_samples_per_second": 39.789,
|
| 25 |
+
"eval_steps_per_second": 1.246,
|
| 26 |
+
"step": 500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 2.203856749311295,
|
| 30 |
+
"grad_norm": 7.994827747344971,
|
| 31 |
+
"learning_rate": 0.00016666666666666666,
|
| 32 |
+
"loss": 2.4697,
|
| 33 |
+
"step": 1000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 2.203856749311295,
|
| 37 |
+
"eval_accuracy": 0.5384327427002745,
|
| 38 |
+
"eval_f1_macro": 0.2623597839382389,
|
| 39 |
+
"eval_loss": 1.8842540979385376,
|
| 40 |
+
"eval_runtime": 47.0501,
|
| 41 |
+
"eval_samples_per_second": 170.329,
|
| 42 |
+
"eval_steps_per_second": 5.335,
|
| 43 |
+
"step": 1000
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.3057851239669422,
|
| 47 |
+
"grad_norm": 8.498858451843262,
|
| 48 |
+
"learning_rate": 0.00025,
|
| 49 |
+
"loss": 1.6285,
|
| 50 |
+
"step": 1500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.3057851239669422,
|
| 54 |
+
"eval_accuracy": 0.6313950586473671,
|
| 55 |
+
"eval_f1_macro": 0.4048605043298036,
|
| 56 |
+
"eval_loss": 1.509529709815979,
|
| 57 |
+
"eval_runtime": 34.1201,
|
| 58 |
+
"eval_samples_per_second": 234.876,
|
| 59 |
+
"eval_steps_per_second": 7.356,
|
| 60 |
+
"step": 1500
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 4.40771349862259,
|
| 64 |
+
"grad_norm": 9.255996704101562,
|
| 65 |
+
"learning_rate": 0.0003333333333333333,
|
| 66 |
+
"loss": 1.4076,
|
| 67 |
+
"step": 2000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 4.40771349862259,
|
| 71 |
+
"eval_accuracy": 0.6400049912652858,
|
| 72 |
+
"eval_f1_macro": 0.42658784153796886,
|
| 73 |
+
"eval_loss": 1.5477856397628784,
|
| 74 |
+
"eval_runtime": 34.313,
|
| 75 |
+
"eval_samples_per_second": 233.556,
|
| 76 |
+
"eval_steps_per_second": 7.315,
|
| 77 |
+
"step": 2000
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 5.509641873278237,
|
| 81 |
+
"grad_norm": 8.901646614074707,
|
| 82 |
+
"learning_rate": 0.0004166666666666667,
|
| 83 |
+
"loss": 1.3687,
|
| 84 |
+
"step": 2500
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 5.509641873278237,
|
| 88 |
+
"eval_accuracy": 0.6643374095333167,
|
| 89 |
+
"eval_f1_macro": 0.4590986266426292,
|
| 90 |
+
"eval_loss": 1.4123471975326538,
|
| 91 |
+
"eval_runtime": 34.1574,
|
| 92 |
+
"eval_samples_per_second": 234.62,
|
| 93 |
+
"eval_steps_per_second": 7.348,
|
| 94 |
+
"step": 2500
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 6.6115702479338845,
|
| 98 |
+
"grad_norm": 6.906811714172363,
|
| 99 |
+
"learning_rate": 0.0005,
|
| 100 |
+
"loss": 1.3757,
|
| 101 |
+
"step": 3000
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 6.6115702479338845,
|
| 105 |
+
"eval_accuracy": 0.6093087097579236,
|
| 106 |
+
"eval_f1_macro": 0.42633143981000005,
|
| 107 |
+
"eval_loss": 1.642005205154419,
|
| 108 |
+
"eval_runtime": 34.1921,
|
| 109 |
+
"eval_samples_per_second": 234.381,
|
| 110 |
+
"eval_steps_per_second": 7.341,
|
| 111 |
+
"step": 3000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 7.7134986225895315,
|
| 115 |
+
"grad_norm": 9.686969757080078,
|
| 116 |
+
"learning_rate": 0.0004907407407407408,
|
| 117 |
+
"loss": 1.3816,
|
| 118 |
+
"step": 3500
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 7.7134986225895315,
|
| 122 |
+
"eval_accuracy": 0.6442475667581732,
|
| 123 |
+
"eval_f1_macro": 0.4625270466241396,
|
| 124 |
+
"eval_loss": 1.520119309425354,
|
| 125 |
+
"eval_runtime": 176.8139,
|
| 126 |
+
"eval_samples_per_second": 45.324,
|
| 127 |
+
"eval_steps_per_second": 1.42,
|
| 128 |
+
"step": 3500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 8.81542699724518,
|
| 132 |
+
"grad_norm": 9.070294380187988,
|
| 133 |
+
"learning_rate": 0.00048148148148148144,
|
| 134 |
+
"loss": 1.2373,
|
| 135 |
+
"step": 4000
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 8.81542699724518,
|
| 139 |
+
"eval_accuracy": 0.6640878462690292,
|
| 140 |
+
"eval_f1_macro": 0.4936672291314334,
|
| 141 |
+
"eval_loss": 1.4355494976043701,
|
| 142 |
+
"eval_runtime": 34.0662,
|
| 143 |
+
"eval_samples_per_second": 235.248,
|
| 144 |
+
"eval_steps_per_second": 7.368,
|
| 145 |
+
"step": 4000
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 9.917355371900827,
|
| 149 |
+
"grad_norm": 10.048285484313965,
|
| 150 |
+
"learning_rate": 0.00047222222222222224,
|
| 151 |
+
"loss": 1.1039,
|
| 152 |
+
"step": 4500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 9.917355371900827,
|
| 156 |
+
"eval_accuracy": 0.6771899176441227,
|
| 157 |
+
"eval_f1_macro": 0.48714340971176784,
|
| 158 |
+
"eval_loss": 1.3826853036880493,
|
| 159 |
+
"eval_runtime": 33.9169,
|
| 160 |
+
"eval_samples_per_second": 236.284,
|
| 161 |
+
"eval_steps_per_second": 7.4,
|
| 162 |
+
"step": 4500
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 11.019283746556473,
|
| 166 |
+
"grad_norm": 6.188383102416992,
|
| 167 |
+
"learning_rate": 0.000462962962962963,
|
| 168 |
+
"loss": 1.037,
|
| 169 |
+
"step": 5000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 11.019283746556473,
|
| 173 |
+
"eval_accuracy": 0.6820564012977289,
|
| 174 |
+
"eval_f1_macro": 0.5233785564818548,
|
| 175 |
+
"eval_loss": 1.4125980138778687,
|
| 176 |
+
"eval_runtime": 33.9515,
|
| 177 |
+
"eval_samples_per_second": 236.042,
|
| 178 |
+
"eval_steps_per_second": 7.393,
|
| 179 |
+
"step": 5000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 12.121212121212121,
|
| 183 |
+
"grad_norm": 5.904200077056885,
|
| 184 |
+
"learning_rate": 0.0004537037037037037,
|
| 185 |
+
"loss": 0.9419,
|
| 186 |
+
"step": 5500
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 12.121212121212121,
|
| 190 |
+
"eval_accuracy": 0.7108809583229349,
|
| 191 |
+
"eval_f1_macro": 0.5375488759123677,
|
| 192 |
+
"eval_loss": 1.2807058095932007,
|
| 193 |
+
"eval_runtime": 34.1787,
|
| 194 |
+
"eval_samples_per_second": 234.473,
|
| 195 |
+
"eval_steps_per_second": 7.344,
|
| 196 |
+
"step": 5500
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 13.223140495867769,
|
| 200 |
+
"grad_norm": 6.450353145599365,
|
| 201 |
+
"learning_rate": 0.0004444444444444444,
|
| 202 |
+
"loss": 0.855,
|
| 203 |
+
"step": 6000
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 13.223140495867769,
|
| 207 |
+
"eval_accuracy": 0.7106313950586474,
|
| 208 |
+
"eval_f1_macro": 0.551305442830446,
|
| 209 |
+
"eval_loss": 1.3112666606903076,
|
| 210 |
+
"eval_runtime": 50.9709,
|
| 211 |
+
"eval_samples_per_second": 157.227,
|
| 212 |
+
"eval_steps_per_second": 4.924,
|
| 213 |
+
"step": 6000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 14.325068870523417,
|
| 217 |
+
"grad_norm": 8.54566764831543,
|
| 218 |
+
"learning_rate": 0.0004351851851851852,
|
| 219 |
+
"loss": 0.7855,
|
| 220 |
+
"step": 6500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 14.325068870523417,
|
| 224 |
+
"eval_accuracy": 0.7096331420014974,
|
| 225 |
+
"eval_f1_macro": 0.5583041388143256,
|
| 226 |
+
"eval_loss": 1.2415180206298828,
|
| 227 |
+
"eval_runtime": 138.5717,
|
| 228 |
+
"eval_samples_per_second": 57.833,
|
| 229 |
+
"eval_steps_per_second": 1.811,
|
| 230 |
+
"step": 6500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 15.426997245179063,
|
| 234 |
+
"grad_norm": 6.042281627655029,
|
| 235 |
+
"learning_rate": 0.00042592592592592595,
|
| 236 |
+
"loss": 0.7192,
|
| 237 |
+
"step": 7000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 15.426997245179063,
|
| 241 |
+
"eval_accuracy": 0.7216121786872972,
|
| 242 |
+
"eval_f1_macro": 0.5653564656146333,
|
| 243 |
+
"eval_loss": 1.2567986249923706,
|
| 244 |
+
"eval_runtime": 34.1278,
|
| 245 |
+
"eval_samples_per_second": 234.823,
|
| 246 |
+
"eval_steps_per_second": 7.355,
|
| 247 |
+
"step": 7000
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 16.52892561983471,
|
| 251 |
+
"grad_norm": 7.013518333435059,
|
| 252 |
+
"learning_rate": 0.0004166666666666667,
|
| 253 |
+
"loss": 0.6721,
|
| 254 |
+
"step": 7500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 16.52892561983471,
|
| 258 |
+
"eval_accuracy": 0.719241327676566,
|
| 259 |
+
"eval_f1_macro": 0.5674881280347366,
|
| 260 |
+
"eval_loss": 1.2789088487625122,
|
| 261 |
+
"eval_runtime": 34.173,
|
| 262 |
+
"eval_samples_per_second": 234.512,
|
| 263 |
+
"eval_steps_per_second": 7.345,
|
| 264 |
+
"step": 7500
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 17.63085399449036,
|
| 268 |
+
"grad_norm": 5.327983379364014,
|
| 269 |
+
"learning_rate": 0.0004074074074074074,
|
| 270 |
+
"loss": 0.618,
|
| 271 |
+
"step": 8000
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 17.63085399449036,
|
| 275 |
+
"eval_accuracy": 0.7237334664337409,
|
| 276 |
+
"eval_f1_macro": 0.5611898965872011,
|
| 277 |
+
"eval_loss": 1.2278856039047241,
|
| 278 |
+
"eval_runtime": 34.0889,
|
| 279 |
+
"eval_samples_per_second": 235.091,
|
| 280 |
+
"eval_steps_per_second": 7.363,
|
| 281 |
+
"step": 8000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 18.732782369146005,
|
| 285 |
+
"grad_norm": 5.324774265289307,
|
| 286 |
+
"learning_rate": 0.0003981481481481481,
|
| 287 |
+
"loss": 0.5673,
|
| 288 |
+
"step": 8500
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 18.732782369146005,
|
| 292 |
+
"eval_accuracy": 0.7378337908659846,
|
| 293 |
+
"eval_f1_macro": 0.5708025777794506,
|
| 294 |
+
"eval_loss": 1.2691614627838135,
|
| 295 |
+
"eval_runtime": 34.0618,
|
| 296 |
+
"eval_samples_per_second": 235.278,
|
| 297 |
+
"eval_steps_per_second": 7.369,
|
| 298 |
+
"step": 8500
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 19.834710743801654,
|
| 302 |
+
"grad_norm": 7.3289384841918945,
|
| 303 |
+
"learning_rate": 0.0003888888888888889,
|
| 304 |
+
"loss": 0.523,
|
| 305 |
+
"step": 9000
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 19.834710743801654,
|
| 309 |
+
"eval_accuracy": 0.736585974544547,
|
| 310 |
+
"eval_f1_macro": 0.583453506873929,
|
| 311 |
+
"eval_loss": 1.2533445358276367,
|
| 312 |
+
"eval_runtime": 34.1698,
|
| 313 |
+
"eval_samples_per_second": 234.534,
|
| 314 |
+
"eval_steps_per_second": 7.346,
|
| 315 |
+
"step": 9000
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 20.9366391184573,
|
| 319 |
+
"grad_norm": 4.2654643058776855,
|
| 320 |
+
"learning_rate": 0.00037962962962962966,
|
| 321 |
+
"loss": 0.4761,
|
| 322 |
+
"step": 9500
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 20.9366391184573,
|
| 326 |
+
"eval_accuracy": 0.7508110806089343,
|
| 327 |
+
"eval_f1_macro": 0.5780908306225785,
|
| 328 |
+
"eval_loss": 1.2073005437850952,
|
| 329 |
+
"eval_runtime": 34.1144,
|
| 330 |
+
"eval_samples_per_second": 234.915,
|
| 331 |
+
"eval_steps_per_second": 7.358,
|
| 332 |
+
"step": 9500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 22.038567493112946,
|
| 336 |
+
"grad_norm": 7.886093616485596,
|
| 337 |
+
"learning_rate": 0.00037037037037037035,
|
| 338 |
+
"loss": 0.4481,
|
| 339 |
+
"step": 10000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 22.038567493112946,
|
| 343 |
+
"eval_accuracy": 0.7504367357125031,
|
| 344 |
+
"eval_f1_macro": 0.5902163691095943,
|
| 345 |
+
"eval_loss": 1.2306609153747559,
|
| 346 |
+
"eval_runtime": 34.1357,
|
| 347 |
+
"eval_samples_per_second": 234.769,
|
| 348 |
+
"eval_steps_per_second": 7.353,
|
| 349 |
+
"step": 10000
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 23.140495867768596,
|
| 353 |
+
"grad_norm": 5.739956855773926,
|
| 354 |
+
"learning_rate": 0.0003611111111111111,
|
| 355 |
+
"loss": 0.4137,
|
| 356 |
+
"step": 10500
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 23.140495867768596,
|
| 360 |
+
"eval_accuracy": 0.7481906663339156,
|
| 361 |
+
"eval_f1_macro": 0.6000675997156686,
|
| 362 |
+
"eval_loss": 1.2535877227783203,
|
| 363 |
+
"eval_runtime": 34.1657,
|
| 364 |
+
"eval_samples_per_second": 234.563,
|
| 365 |
+
"eval_steps_per_second": 7.347,
|
| 366 |
+
"step": 10500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 24.242424242424242,
|
| 370 |
+
"grad_norm": 6.5599589347839355,
|
| 371 |
+
"learning_rate": 0.0003518518518518519,
|
| 372 |
+
"loss": 0.3771,
|
| 373 |
+
"step": 11000
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 24.242424242424242,
|
| 377 |
+
"eval_accuracy": 0.7523084601946594,
|
| 378 |
+
"eval_f1_macro": 0.6003714574908554,
|
| 379 |
+
"eval_loss": 1.1841113567352295,
|
| 380 |
+
"eval_runtime": 33.9543,
|
| 381 |
+
"eval_samples_per_second": 236.023,
|
| 382 |
+
"eval_steps_per_second": 7.392,
|
| 383 |
+
"step": 11000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 25.34435261707989,
|
| 387 |
+
"grad_norm": 4.988476753234863,
|
| 388 |
+
"learning_rate": 0.00034259259259259263,
|
| 389 |
+
"loss": 0.3456,
|
| 390 |
+
"step": 11500
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 25.34435261707989,
|
| 394 |
+
"eval_accuracy": 0.7511854255053656,
|
| 395 |
+
"eval_f1_macro": 0.613830898261218,
|
| 396 |
+
"eval_loss": 1.2578365802764893,
|
| 397 |
+
"eval_runtime": 34.0485,
|
| 398 |
+
"eval_samples_per_second": 235.37,
|
| 399 |
+
"eval_steps_per_second": 7.372,
|
| 400 |
+
"step": 11500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 26.446280991735538,
|
| 404 |
+
"grad_norm": 10.72720718383789,
|
| 405 |
+
"learning_rate": 0.0003333333333333333,
|
| 406 |
+
"loss": 0.3206,
|
| 407 |
+
"step": 12000
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 26.446280991735538,
|
| 411 |
+
"eval_accuracy": 0.7565510356875468,
|
| 412 |
+
"eval_f1_macro": 0.6008526427475953,
|
| 413 |
+
"eval_loss": 1.2595282793045044,
|
| 414 |
+
"eval_runtime": 34.0387,
|
| 415 |
+
"eval_samples_per_second": 235.438,
|
| 416 |
+
"eval_steps_per_second": 7.374,
|
| 417 |
+
"step": 12000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 27.548209366391184,
|
| 421 |
+
"grad_norm": 5.9753098487854,
|
| 422 |
+
"learning_rate": 0.00032407407407407406,
|
| 423 |
+
"loss": 0.3083,
|
| 424 |
+
"step": 12500
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 27.548209366391184,
|
| 428 |
+
"eval_accuracy": 0.7525580234589468,
|
| 429 |
+
"eval_f1_macro": 0.6225560622894577,
|
| 430 |
+
"eval_loss": 1.2758995294570923,
|
| 431 |
+
"eval_runtime": 33.9587,
|
| 432 |
+
"eval_samples_per_second": 235.992,
|
| 433 |
+
"eval_steps_per_second": 7.391,
|
| 434 |
+
"step": 12500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 28.650137741046834,
|
| 438 |
+
"grad_norm": 7.985760688781738,
|
| 439 |
+
"learning_rate": 0.0003148148148148148,
|
| 440 |
+
"loss": 0.286,
|
| 441 |
+
"step": 13000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 28.650137741046834,
|
| 445 |
+
"eval_accuracy": 0.7631644621911655,
|
| 446 |
+
"eval_f1_macro": 0.615134929599282,
|
| 447 |
+
"eval_loss": 1.2176945209503174,
|
| 448 |
+
"eval_runtime": 34.0075,
|
| 449 |
+
"eval_samples_per_second": 235.654,
|
| 450 |
+
"eval_steps_per_second": 7.381,
|
| 451 |
+
"step": 13000
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 29.75206611570248,
|
| 455 |
+
"grad_norm": 3.7680389881134033,
|
| 456 |
+
"learning_rate": 0.0003055555555555556,
|
| 457 |
+
"loss": 0.2678,
|
| 458 |
+
"step": 13500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 29.75206611570248,
|
| 462 |
+
"eval_accuracy": 0.7686548540054904,
|
| 463 |
+
"eval_f1_macro": 0.6317793794263403,
|
| 464 |
+
"eval_loss": 1.2493242025375366,
|
| 465 |
+
"eval_runtime": 34.052,
|
| 466 |
+
"eval_samples_per_second": 235.346,
|
| 467 |
+
"eval_steps_per_second": 7.371,
|
| 468 |
+
"step": 13500
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 30.853994490358126,
|
| 472 |
+
"grad_norm": 2.3487305641174316,
|
| 473 |
+
"learning_rate": 0.0002962962962962963,
|
| 474 |
+
"loss": 0.2488,
|
| 475 |
+
"step": 14000
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 30.853994490358126,
|
| 479 |
+
"eval_accuracy": 0.7685300723733467,
|
| 480 |
+
"eval_f1_macro": 0.6379453782996546,
|
| 481 |
+
"eval_loss": 1.2121572494506836,
|
| 482 |
+
"eval_runtime": 33.8764,
|
| 483 |
+
"eval_samples_per_second": 236.566,
|
| 484 |
+
"eval_steps_per_second": 7.409,
|
| 485 |
+
"step": 14000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 31.955922865013775,
|
| 489 |
+
"grad_norm": 5.986231803894043,
|
| 490 |
+
"learning_rate": 0.00028703703703703703,
|
| 491 |
+
"loss": 0.2324,
|
| 492 |
+
"step": 14500
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 31.955922865013775,
|
| 496 |
+
"eval_accuracy": 0.7707761417519341,
|
| 497 |
+
"eval_f1_macro": 0.6291965084001815,
|
| 498 |
+
"eval_loss": 1.2486802339553833,
|
| 499 |
+
"eval_runtime": 33.8585,
|
| 500 |
+
"eval_samples_per_second": 236.691,
|
| 501 |
+
"eval_steps_per_second": 7.413,
|
| 502 |
+
"step": 14500
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 33.05785123966942,
|
| 506 |
+
"grad_norm": 4.138526439666748,
|
| 507 |
+
"learning_rate": 0.0002777777777777778,
|
| 508 |
+
"loss": 0.212,
|
| 509 |
+
"step": 15000
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 33.05785123966942,
|
| 513 |
+
"eval_accuracy": 0.7655353132018967,
|
| 514 |
+
"eval_f1_macro": 0.6182777113261247,
|
| 515 |
+
"eval_loss": 1.2596852779388428,
|
| 516 |
+
"eval_runtime": 33.8607,
|
| 517 |
+
"eval_samples_per_second": 236.675,
|
| 518 |
+
"eval_steps_per_second": 7.413,
|
| 519 |
+
"step": 15000
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 34.15977961432507,
|
| 523 |
+
"grad_norm": 4.644781112670898,
|
| 524 |
+
"learning_rate": 0.0002685185185185186,
|
| 525 |
+
"loss": 0.2003,
|
| 526 |
+
"step": 15500
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 34.15977961432507,
|
| 530 |
+
"eval_accuracy": 0.7730222111305216,
|
| 531 |
+
"eval_f1_macro": 0.6195180591624613,
|
| 532 |
+
"eval_loss": 1.2190001010894775,
|
| 533 |
+
"eval_runtime": 33.8524,
|
| 534 |
+
"eval_samples_per_second": 236.734,
|
| 535 |
+
"eval_steps_per_second": 7.415,
|
| 536 |
+
"step": 15500
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 35.26170798898072,
|
| 540 |
+
"grad_norm": 2.8191895484924316,
|
| 541 |
+
"learning_rate": 0.00025925925925925926,
|
| 542 |
+
"loss": 0.1822,
|
| 543 |
+
"step": 16000
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 35.26170798898072,
|
| 547 |
+
"eval_accuracy": 0.7781382580484153,
|
| 548 |
+
"eval_f1_macro": 0.6369508969483741,
|
| 549 |
+
"eval_loss": 1.2363426685333252,
|
| 550 |
+
"eval_runtime": 40.925,
|
| 551 |
+
"eval_samples_per_second": 195.822,
|
| 552 |
+
"eval_steps_per_second": 6.133,
|
| 553 |
+
"step": 16000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 36.36363636363637,
|
| 557 |
+
"grad_norm": 4.859241962432861,
|
| 558 |
+
"learning_rate": 0.00025,
|
| 559 |
+
"loss": 0.1743,
|
| 560 |
+
"step": 16500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 36.36363636363637,
|
| 564 |
+
"eval_accuracy": 0.777763913151984,
|
| 565 |
+
"eval_f1_macro": 0.6468738399385165,
|
| 566 |
+
"eval_loss": 1.2521047592163086,
|
| 567 |
+
"eval_runtime": 40.9791,
|
| 568 |
+
"eval_samples_per_second": 195.563,
|
| 569 |
+
"eval_steps_per_second": 6.125,
|
| 570 |
+
"step": 16500
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 37.46556473829201,
|
| 574 |
+
"grad_norm": 3.027918577194214,
|
| 575 |
+
"learning_rate": 0.00024074074074074072,
|
| 576 |
+
"loss": 0.1604,
|
| 577 |
+
"step": 17000
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 37.46556473829201,
|
| 581 |
+
"eval_accuracy": 0.7700274519590716,
|
| 582 |
+
"eval_f1_macro": 0.6320426231926091,
|
| 583 |
+
"eval_loss": 1.2722282409667969,
|
| 584 |
+
"eval_runtime": 41.0083,
|
| 585 |
+
"eval_samples_per_second": 195.424,
|
| 586 |
+
"eval_steps_per_second": 6.121,
|
| 587 |
+
"step": 17000
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 38.56749311294766,
|
| 591 |
+
"grad_norm": 5.677507400512695,
|
| 592 |
+
"learning_rate": 0.0002314814814814815,
|
| 593 |
+
"loss": 0.1523,
|
| 594 |
+
"step": 17500
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 38.56749311294766,
|
| 598 |
+
"eval_accuracy": 0.7745195907162465,
|
| 599 |
+
"eval_f1_macro": 0.6419239322758532,
|
| 600 |
+
"eval_loss": 1.2623705863952637,
|
| 601 |
+
"eval_runtime": 40.8771,
|
| 602 |
+
"eval_samples_per_second": 196.051,
|
| 603 |
+
"eval_steps_per_second": 6.14,
|
| 604 |
+
"step": 17500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 39.66942148760331,
|
| 608 |
+
"grad_norm": 5.041018009185791,
|
| 609 |
+
"learning_rate": 0.0002222222222222222,
|
| 610 |
+
"loss": 0.1384,
|
| 611 |
+
"step": 18000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 39.66942148760331,
|
| 615 |
+
"eval_accuracy": 0.7866234090341901,
|
| 616 |
+
"eval_f1_macro": 0.6659870204162003,
|
| 617 |
+
"eval_loss": 1.189522624015808,
|
| 618 |
+
"eval_runtime": 40.9611,
|
| 619 |
+
"eval_samples_per_second": 195.649,
|
| 620 |
+
"eval_steps_per_second": 6.128,
|
| 621 |
+
"step": 18000
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 40.77134986225895,
|
| 625 |
+
"grad_norm": 3.659827947616577,
|
| 626 |
+
"learning_rate": 0.00021296296296296298,
|
| 627 |
+
"loss": 0.135,
|
| 628 |
+
"step": 18500
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 40.77134986225895,
|
| 632 |
+
"eval_accuracy": 0.7850012478163214,
|
| 633 |
+
"eval_f1_macro": 0.6720105896795865,
|
| 634 |
+
"eval_loss": 1.1996605396270752,
|
| 635 |
+
"eval_runtime": 40.9502,
|
| 636 |
+
"eval_samples_per_second": 195.701,
|
| 637 |
+
"eval_steps_per_second": 6.129,
|
| 638 |
+
"step": 18500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 41.8732782369146,
|
| 642 |
+
"grad_norm": 5.367224216461182,
|
| 643 |
+
"learning_rate": 0.0002037037037037037,
|
| 644 |
+
"loss": 0.1239,
|
| 645 |
+
"step": 19000
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 41.8732782369146,
|
| 649 |
+
"eval_accuracy": 0.7908659845270776,
|
| 650 |
+
"eval_f1_macro": 0.6804665897748318,
|
| 651 |
+
"eval_loss": 1.2242608070373535,
|
| 652 |
+
"eval_runtime": 40.9731,
|
| 653 |
+
"eval_samples_per_second": 195.592,
|
| 654 |
+
"eval_steps_per_second": 6.126,
|
| 655 |
+
"step": 19000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 42.97520661157025,
|
| 659 |
+
"grad_norm": 3.64471173286438,
|
| 660 |
+
"learning_rate": 0.00019444444444444446,
|
| 661 |
+
"loss": 0.1131,
|
| 662 |
+
"step": 19500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 42.97520661157025,
|
| 666 |
+
"eval_accuracy": 0.7947342151235338,
|
| 667 |
+
"eval_f1_macro": 0.653675258584708,
|
| 668 |
+
"eval_loss": 1.1854939460754395,
|
| 669 |
+
"eval_runtime": 40.6521,
|
| 670 |
+
"eval_samples_per_second": 197.136,
|
| 671 |
+
"eval_steps_per_second": 6.174,
|
| 672 |
+
"step": 19500
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 44.07713498622589,
|
| 676 |
+
"grad_norm": 7.800271987915039,
|
| 677 |
+
"learning_rate": 0.00018518518518518518,
|
| 678 |
+
"loss": 0.1076,
|
| 679 |
+
"step": 20000
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 44.07713498622589,
|
| 683 |
+
"eval_accuracy": 0.7879960069877714,
|
| 684 |
+
"eval_f1_macro": 0.6313626574356537,
|
| 685 |
+
"eval_loss": 1.2797794342041016,
|
| 686 |
+
"eval_runtime": 40.7712,
|
| 687 |
+
"eval_samples_per_second": 196.56,
|
| 688 |
+
"eval_steps_per_second": 6.156,
|
| 689 |
+
"step": 20000
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 45.17906336088154,
|
| 693 |
+
"grad_norm": 3.2701990604400635,
|
| 694 |
+
"learning_rate": 0.00017592592592592595,
|
| 695 |
+
"loss": 0.0999,
|
| 696 |
+
"step": 20500
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 45.17906336088154,
|
| 700 |
+
"eval_accuracy": 0.7922385824806588,
|
| 701 |
+
"eval_f1_macro": 0.6725445304133565,
|
| 702 |
+
"eval_loss": 1.2082223892211914,
|
| 703 |
+
"eval_runtime": 39.2309,
|
| 704 |
+
"eval_samples_per_second": 204.278,
|
| 705 |
+
"eval_steps_per_second": 6.398,
|
| 706 |
+
"step": 20500
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 46.28099173553719,
|
| 710 |
+
"grad_norm": 3.959247350692749,
|
| 711 |
+
"learning_rate": 0.00016666666666666666,
|
| 712 |
+
"loss": 0.0898,
|
| 713 |
+
"step": 21000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 46.28099173553719,
|
| 717 |
+
"eval_accuracy": 0.80059895183429,
|
| 718 |
+
"eval_f1_macro": 0.6765711205444297,
|
| 719 |
+
"eval_loss": 1.1848937273025513,
|
| 720 |
+
"eval_runtime": 40.6064,
|
| 721 |
+
"eval_samples_per_second": 197.358,
|
| 722 |
+
"eval_steps_per_second": 6.181,
|
| 723 |
+
"step": 21000
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 47.382920110192835,
|
| 727 |
+
"grad_norm": 3.7939414978027344,
|
| 728 |
+
"learning_rate": 0.0001574074074074074,
|
| 729 |
+
"loss": 0.0861,
|
| 730 |
+
"step": 21500
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 47.382920110192835,
|
| 734 |
+
"eval_accuracy": 0.7979785375592713,
|
| 735 |
+
"eval_f1_macro": 0.6842554850331637,
|
| 736 |
+
"eval_loss": 1.177991509437561,
|
| 737 |
+
"eval_runtime": 40.9664,
|
| 738 |
+
"eval_samples_per_second": 195.624,
|
| 739 |
+
"eval_steps_per_second": 6.127,
|
| 740 |
+
"step": 21500
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 48.484848484848484,
|
| 744 |
+
"grad_norm": 2.4407143592834473,
|
| 745 |
+
"learning_rate": 0.00014814814814814815,
|
| 746 |
+
"loss": 0.0743,
|
| 747 |
+
"step": 22000
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 48.484848484848484,
|
| 751 |
+
"eval_accuracy": 0.8023458946843025,
|
| 752 |
+
"eval_f1_macro": 0.672574066052015,
|
| 753 |
+
"eval_loss": 1.23497474193573,
|
| 754 |
+
"eval_runtime": 40.9537,
|
| 755 |
+
"eval_samples_per_second": 195.684,
|
| 756 |
+
"eval_steps_per_second": 6.129,
|
| 757 |
+
"step": 22000
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 49.586776859504134,
|
| 761 |
+
"grad_norm": 1.4111738204956055,
|
| 762 |
+
"learning_rate": 0.0001388888888888889,
|
| 763 |
+
"loss": 0.0725,
|
| 764 |
+
"step": 22500
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 49.586776859504134,
|
| 768 |
+
"eval_accuracy": 0.8037184926378838,
|
| 769 |
+
"eval_f1_macro": 0.687093106679298,
|
| 770 |
+
"eval_loss": 1.1898678541183472,
|
| 771 |
+
"eval_runtime": 41.112,
|
| 772 |
+
"eval_samples_per_second": 194.931,
|
| 773 |
+
"eval_steps_per_second": 6.105,
|
| 774 |
+
"step": 22500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 50.68870523415978,
|
| 778 |
+
"grad_norm": 2.446298122406006,
|
| 779 |
+
"learning_rate": 0.00012962962962962963,
|
| 780 |
+
"loss": 0.063,
|
| 781 |
+
"step": 23000
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 50.68870523415978,
|
| 785 |
+
"eval_accuracy": 0.7988520089842776,
|
| 786 |
+
"eval_f1_macro": 0.6845006699400662,
|
| 787 |
+
"eval_loss": 1.2227065563201904,
|
| 788 |
+
"eval_runtime": 41.0323,
|
| 789 |
+
"eval_samples_per_second": 195.31,
|
| 790 |
+
"eval_steps_per_second": 6.117,
|
| 791 |
+
"step": 23000
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 51.790633608815426,
|
| 795 |
+
"grad_norm": 3.328188419342041,
|
| 796 |
+
"learning_rate": 0.00012037037037037036,
|
| 797 |
+
"loss": 0.061,
|
| 798 |
+
"step": 23500
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 51.790633608815426,
|
| 802 |
+
"eval_accuracy": 0.8082106313950587,
|
| 803 |
+
"eval_f1_macro": 0.6785356281203615,
|
| 804 |
+
"eval_loss": 1.1609222888946533,
|
| 805 |
+
"eval_runtime": 42.2613,
|
| 806 |
+
"eval_samples_per_second": 189.63,
|
| 807 |
+
"eval_steps_per_second": 5.939,
|
| 808 |
+
"step": 23500
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 52.892561983471076,
|
| 812 |
+
"grad_norm": 2.284496545791626,
|
| 813 |
+
"learning_rate": 0.0001111111111111111,
|
| 814 |
+
"loss": 0.056,
|
| 815 |
+
"step": 24000
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 52.892561983471076,
|
| 819 |
+
"eval_accuracy": 0.8119540803593711,
|
| 820 |
+
"eval_f1_macro": 0.6977082191883057,
|
| 821 |
+
"eval_loss": 1.160672664642334,
|
| 822 |
+
"eval_runtime": 40.9648,
|
| 823 |
+
"eval_samples_per_second": 195.632,
|
| 824 |
+
"eval_steps_per_second": 6.127,
|
| 825 |
+
"step": 24000
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 53.99449035812672,
|
| 829 |
+
"grad_norm": 3.7205893993377686,
|
| 830 |
+
"learning_rate": 0.00010185185185185185,
|
| 831 |
+
"loss": 0.051,
|
| 832 |
+
"step": 24500
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 53.99449035812672,
|
| 836 |
+
"eval_accuracy": 0.8083354130272024,
|
| 837 |
+
"eval_f1_macro": 0.6629286097479332,
|
| 838 |
+
"eval_loss": 1.224273920059204,
|
| 839 |
+
"eval_runtime": 40.9557,
|
| 840 |
+
"eval_samples_per_second": 195.675,
|
| 841 |
+
"eval_steps_per_second": 6.129,
|
| 842 |
+
"step": 24500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 55.09641873278237,
|
| 846 |
+
"grad_norm": 0.219478040933609,
|
| 847 |
+
"learning_rate": 9.259259259259259e-05,
|
| 848 |
+
"loss": 0.0447,
|
| 849 |
+
"step": 25000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 55.09641873278237,
|
| 853 |
+
"eval_accuracy": 0.8098327926129274,
|
| 854 |
+
"eval_f1_macro": 0.666826551929984,
|
| 855 |
+
"eval_loss": 1.199051856994629,
|
| 856 |
+
"eval_runtime": 41.0107,
|
| 857 |
+
"eval_samples_per_second": 195.412,
|
| 858 |
+
"eval_steps_per_second": 6.12,
|
| 859 |
+
"step": 25000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 56.19834710743802,
|
| 863 |
+
"grad_norm": 3.8260011672973633,
|
| 864 |
+
"learning_rate": 8.333333333333333e-05,
|
| 865 |
+
"loss": 0.0409,
|
| 866 |
+
"step": 25500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 56.19834710743802,
|
| 870 |
+
"eval_accuracy": 0.8124532068879461,
|
| 871 |
+
"eval_f1_macro": 0.699449604989361,
|
| 872 |
+
"eval_loss": 1.2268598079681396,
|
| 873 |
+
"eval_runtime": 40.8996,
|
| 874 |
+
"eval_samples_per_second": 195.943,
|
| 875 |
+
"eval_steps_per_second": 6.137,
|
| 876 |
+
"step": 25500
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 57.30027548209367,
|
| 880 |
+
"grad_norm": 0.5541914105415344,
|
| 881 |
+
"learning_rate": 7.407407407407407e-05,
|
| 882 |
+
"loss": 0.039,
|
| 883 |
+
"step": 26000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 57.30027548209367,
|
| 887 |
+
"eval_accuracy": 0.8215622660344397,
|
| 888 |
+
"eval_f1_macro": 0.6928701132826408,
|
| 889 |
+
"eval_loss": 1.1886591911315918,
|
| 890 |
+
"eval_runtime": 40.5313,
|
| 891 |
+
"eval_samples_per_second": 197.724,
|
| 892 |
+
"eval_steps_per_second": 6.193,
|
| 893 |
+
"step": 26000
|
| 894 |
+
}
|
| 895 |
+
],
|
| 896 |
+
"logging_steps": 500,
|
| 897 |
+
"max_steps": 30000,
|
| 898 |
+
"num_input_tokens_seen": 0,
|
| 899 |
+
"num_train_epochs": 67,
|
| 900 |
+
"save_steps": 1000,
|
| 901 |
+
"stateful_callbacks": {
|
| 902 |
+
"TrainerControl": {
|
| 903 |
+
"args": {
|
| 904 |
+
"should_epoch_stop": false,
|
| 905 |
+
"should_evaluate": false,
|
| 906 |
+
"should_log": false,
|
| 907 |
+
"should_save": true,
|
| 908 |
+
"should_training_stop": false
|
| 909 |
+
},
|
| 910 |
+
"attributes": {}
|
| 911 |
+
}
|
| 912 |
+
},
|
| 913 |
+
"total_flos": 1.5103646535813897e+20,
|
| 914 |
+
"train_batch_size": 32,
|
| 915 |
+
"trial_name": null,
|
| 916 |
+
"trial_params": null
|
| 917 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e1e7e8d962fba8418949e1365bf6d9bd403c8fe8015481536a1cd9d10372b363
|
| 3 |
+
size 5240
|