Upload folder using huggingface_hub
Browse files- config.json +296 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1308 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,296 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-large-960h-lv60-self",
|
| 3 |
+
"activation_dropout": 0.1,
|
| 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": true,
|
| 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": true,
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_dropout": 0.0,
|
| 52 |
+
"feat_extract_norm": "layer",
|
| 53 |
+
"feat_proj_dropout": 0.1,
|
| 54 |
+
"feat_quantizer_dropout": 0.0,
|
| 55 |
+
"final_dropout": 0.1,
|
| 56 |
+
"gradient_checkpointing": false,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_dropout_prob": 0.1,
|
| 60 |
+
"hidden_size": 1024,
|
| 61 |
+
"id2label": {
|
| 62 |
+
"0": "calendar_set",
|
| 63 |
+
"1": "wavs/audio_volume_up",
|
| 64 |
+
"10": "qa_stock",
|
| 65 |
+
"11": "play_music",
|
| 66 |
+
"12": "recommendation_events",
|
| 67 |
+
"13": "qa_definition",
|
| 68 |
+
"14": "alarm_remove",
|
| 69 |
+
"15": "play_podcasts",
|
| 70 |
+
"16": "social_query",
|
| 71 |
+
"17": "email_addcontact",
|
| 72 |
+
"18": "news_query",
|
| 73 |
+
"19": "calendar_query",
|
| 74 |
+
"2": "iot_hue_lightup",
|
| 75 |
+
"20": "music_likeness",
|
| 76 |
+
"21": "general_quirky",
|
| 77 |
+
"22": "qa_factoid",
|
| 78 |
+
"23": "takeaway_order",
|
| 79 |
+
"24": "play_wavs/audiobook",
|
| 80 |
+
"25": "iot_cleaning",
|
| 81 |
+
"26": "general_greet",
|
| 82 |
+
"27": "transport_query",
|
| 83 |
+
"28": "transport_taxi",
|
| 84 |
+
"29": "email_sendemail",
|
| 85 |
+
"3": "weather_query",
|
| 86 |
+
"30": "general_joke",
|
| 87 |
+
"31": "qa_maths",
|
| 88 |
+
"32": "social_post",
|
| 89 |
+
"33": "transport_ticket",
|
| 90 |
+
"34": "cooking_recipe",
|
| 91 |
+
"35": "music_settings",
|
| 92 |
+
"36": "calendar_remove",
|
| 93 |
+
"37": "iot_wemo_on",
|
| 94 |
+
"38": "iot_hue_lightchange",
|
| 95 |
+
"39": "play_radio",
|
| 96 |
+
"4": "iot_coffee",
|
| 97 |
+
"40": "email_querycontact",
|
| 98 |
+
"41": "transport_traffic",
|
| 99 |
+
"42": "qa_currency",
|
| 100 |
+
"43": "datetime_query",
|
| 101 |
+
"44": "iot_hue_lightoff",
|
| 102 |
+
"45": "takeaway_query",
|
| 103 |
+
"46": "lists_createoradd",
|
| 104 |
+
"47": "music_query",
|
| 105 |
+
"48": "recommendation_locations",
|
| 106 |
+
"49": "lists_query",
|
| 107 |
+
"5": "wavs/audio_volume_mute",
|
| 108 |
+
"50": "recommendation_movies",
|
| 109 |
+
"51": "iot_wemo_off",
|
| 110 |
+
"52": "iot_hue_lighton",
|
| 111 |
+
"53": "sendemail",
|
| 112 |
+
"54": "wavs/audio_volume_down",
|
| 113 |
+
"55": "play_game",
|
| 114 |
+
"56": "music",
|
| 115 |
+
"57": "datetime_convert",
|
| 116 |
+
"58": "iot_hue_lightdim",
|
| 117 |
+
"59": "query",
|
| 118 |
+
"6": "lists_remove",
|
| 119 |
+
"60": "createoradd",
|
| 120 |
+
"61": "music_dislikeness",
|
| 121 |
+
"62": "podcasts",
|
| 122 |
+
"63": "joke",
|
| 123 |
+
"64": "set",
|
| 124 |
+
"65": "hue_lightup",
|
| 125 |
+
"66": "factoid",
|
| 126 |
+
"67": "wavs/audio_volume_other",
|
| 127 |
+
"68": "hue_lightoff",
|
| 128 |
+
"69": "quirky",
|
| 129 |
+
"7": "email_query",
|
| 130 |
+
"70": "querycontact",
|
| 131 |
+
"71": "radio",
|
| 132 |
+
"72": "addcontact",
|
| 133 |
+
"73": "greet",
|
| 134 |
+
"74": "ticket",
|
| 135 |
+
"75": "traffic",
|
| 136 |
+
"76": "cooking_query",
|
| 137 |
+
"77": "remove",
|
| 138 |
+
"78": "currency",
|
| 139 |
+
"79": "coffee",
|
| 140 |
+
"8": "alarm_set",
|
| 141 |
+
"80": "game",
|
| 142 |
+
"81": "wemo_on",
|
| 143 |
+
"82": "definition",
|
| 144 |
+
"83": "events",
|
| 145 |
+
"84": "post",
|
| 146 |
+
"85": "hue_lightdim",
|
| 147 |
+
"86": "convert",
|
| 148 |
+
"87": "wemo_off",
|
| 149 |
+
"88": "cleaning",
|
| 150 |
+
"89": "settings",
|
| 151 |
+
"9": "alarm_query",
|
| 152 |
+
"90": "volume_other"
|
| 153 |
+
},
|
| 154 |
+
"initializer_range": 0.02,
|
| 155 |
+
"intermediate_size": 4096,
|
| 156 |
+
"label2id": {
|
| 157 |
+
"addcontact": "72",
|
| 158 |
+
"alarm_query": "9",
|
| 159 |
+
"alarm_remove": "14",
|
| 160 |
+
"alarm_set": "8",
|
| 161 |
+
"calendar_query": "19",
|
| 162 |
+
"calendar_remove": "36",
|
| 163 |
+
"calendar_set": "0",
|
| 164 |
+
"cleaning": "88",
|
| 165 |
+
"coffee": "79",
|
| 166 |
+
"convert": "86",
|
| 167 |
+
"cooking_query": "76",
|
| 168 |
+
"cooking_recipe": "34",
|
| 169 |
+
"createoradd": "60",
|
| 170 |
+
"currency": "78",
|
| 171 |
+
"datetime_convert": "57",
|
| 172 |
+
"datetime_query": "43",
|
| 173 |
+
"definition": "82",
|
| 174 |
+
"email_addcontact": "17",
|
| 175 |
+
"email_query": "7",
|
| 176 |
+
"email_querycontact": "40",
|
| 177 |
+
"email_sendemail": "29",
|
| 178 |
+
"events": "83",
|
| 179 |
+
"factoid": "66",
|
| 180 |
+
"game": "80",
|
| 181 |
+
"general_greet": "26",
|
| 182 |
+
"general_joke": "30",
|
| 183 |
+
"general_quirky": "21",
|
| 184 |
+
"greet": "73",
|
| 185 |
+
"hue_lightdim": "85",
|
| 186 |
+
"hue_lightoff": "68",
|
| 187 |
+
"hue_lightup": "65",
|
| 188 |
+
"iot_cleaning": "25",
|
| 189 |
+
"iot_coffee": "4",
|
| 190 |
+
"iot_hue_lightchange": "38",
|
| 191 |
+
"iot_hue_lightdim": "58",
|
| 192 |
+
"iot_hue_lightoff": "44",
|
| 193 |
+
"iot_hue_lighton": "52",
|
| 194 |
+
"iot_hue_lightup": "2",
|
| 195 |
+
"iot_wemo_off": "51",
|
| 196 |
+
"iot_wemo_on": "37",
|
| 197 |
+
"joke": "63",
|
| 198 |
+
"lists_createoradd": "46",
|
| 199 |
+
"lists_query": "49",
|
| 200 |
+
"lists_remove": "6",
|
| 201 |
+
"music": "56",
|
| 202 |
+
"music_dislikeness": "61",
|
| 203 |
+
"music_likeness": "20",
|
| 204 |
+
"music_query": "47",
|
| 205 |
+
"music_settings": "35",
|
| 206 |
+
"news_query": "18",
|
| 207 |
+
"play_game": "55",
|
| 208 |
+
"play_music": "11",
|
| 209 |
+
"play_podcasts": "15",
|
| 210 |
+
"play_radio": "39",
|
| 211 |
+
"play_wavs/audiobook": "24",
|
| 212 |
+
"podcasts": "62",
|
| 213 |
+
"post": "84",
|
| 214 |
+
"qa_currency": "42",
|
| 215 |
+
"qa_definition": "13",
|
| 216 |
+
"qa_factoid": "22",
|
| 217 |
+
"qa_maths": "31",
|
| 218 |
+
"qa_stock": "10",
|
| 219 |
+
"query": "59",
|
| 220 |
+
"querycontact": "70",
|
| 221 |
+
"quirky": "69",
|
| 222 |
+
"radio": "71",
|
| 223 |
+
"recommendation_events": "12",
|
| 224 |
+
"recommendation_locations": "48",
|
| 225 |
+
"recommendation_movies": "50",
|
| 226 |
+
"remove": "77",
|
| 227 |
+
"sendemail": "53",
|
| 228 |
+
"set": "64",
|
| 229 |
+
"settings": "89",
|
| 230 |
+
"social_post": "32",
|
| 231 |
+
"social_query": "16",
|
| 232 |
+
"takeaway_order": "23",
|
| 233 |
+
"takeaway_query": "45",
|
| 234 |
+
"ticket": "74",
|
| 235 |
+
"traffic": "75",
|
| 236 |
+
"transport_query": "27",
|
| 237 |
+
"transport_taxi": "28",
|
| 238 |
+
"transport_ticket": "33",
|
| 239 |
+
"transport_traffic": "41",
|
| 240 |
+
"volume_other": "90",
|
| 241 |
+
"wavs/audio_volume_down": "54",
|
| 242 |
+
"wavs/audio_volume_mute": "5",
|
| 243 |
+
"wavs/audio_volume_other": "67",
|
| 244 |
+
"wavs/audio_volume_up": "1",
|
| 245 |
+
"weather_query": "3",
|
| 246 |
+
"wemo_off": "87",
|
| 247 |
+
"wemo_on": "81"
|
| 248 |
+
},
|
| 249 |
+
"layer_norm_eps": 1e-05,
|
| 250 |
+
"layerdrop": 0.1,
|
| 251 |
+
"mask_feature_length": 10,
|
| 252 |
+
"mask_feature_min_masks": 0,
|
| 253 |
+
"mask_feature_prob": 0.0,
|
| 254 |
+
"mask_time_length": 10,
|
| 255 |
+
"mask_time_min_masks": 2,
|
| 256 |
+
"mask_time_prob": 0.05,
|
| 257 |
+
"model_type": "wav2vec2",
|
| 258 |
+
"num_adapter_layers": 3,
|
| 259 |
+
"num_attention_heads": 16,
|
| 260 |
+
"num_codevector_groups": 2,
|
| 261 |
+
"num_codevectors_per_group": 320,
|
| 262 |
+
"num_conv_pos_embedding_groups": 16,
|
| 263 |
+
"num_conv_pos_embeddings": 128,
|
| 264 |
+
"num_feat_extract_layers": 7,
|
| 265 |
+
"num_hidden_layers": 24,
|
| 266 |
+
"num_negatives": 100,
|
| 267 |
+
"output_hidden_size": 1024,
|
| 268 |
+
"pad_token_id": 0,
|
| 269 |
+
"proj_codevector_dim": 256,
|
| 270 |
+
"tdnn_dilation": [
|
| 271 |
+
1,
|
| 272 |
+
2,
|
| 273 |
+
3,
|
| 274 |
+
1,
|
| 275 |
+
1
|
| 276 |
+
],
|
| 277 |
+
"tdnn_dim": [
|
| 278 |
+
512,
|
| 279 |
+
512,
|
| 280 |
+
512,
|
| 281 |
+
512,
|
| 282 |
+
1500
|
| 283 |
+
],
|
| 284 |
+
"tdnn_kernel": [
|
| 285 |
+
5,
|
| 286 |
+
3,
|
| 287 |
+
3,
|
| 288 |
+
1,
|
| 289 |
+
1
|
| 290 |
+
],
|
| 291 |
+
"torch_dtype": "float32",
|
| 292 |
+
"transformers_version": "4.45.2",
|
| 293 |
+
"use_weighted_layer_sum": false,
|
| 294 |
+
"vocab_size": 32,
|
| 295 |
+
"xvector_output_dim": 512
|
| 296 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c489d880e03d06dc00b5272bac80748e203e42508bda2332d59f1e3233119d38
|
| 3 |
+
size 1262950804
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e0cabc3951b5d157753d58f27b7a7de2d0dc27a03c2c999d6240a522e19e6073
|
| 3 |
+
size 2526152656
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e570cd5f9d55b7177de3ebb2d7cae79bc4c203b31bae473d44da24a40539c0aa
|
| 3 |
+
size 14308
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:417c32b40ce266eae8ab814d0af12409a7770b12ed2c5afd1b1ffbf9fa27b42f
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1308 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8005753739930955,
|
| 3 |
+
"best_model_checkpoint": "results/facebook/wav2vec2-large-960h-lv60-self/42/_retain/checkpoint-30000",
|
| 4 |
+
"epoch": 75.80543272267846,
|
| 5 |
+
"eval_steps": 400,
|
| 6 |
+
"global_step": 30000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.010739102969046,
|
| 13 |
+
"grad_norm": 3.2389800548553467,
|
| 14 |
+
"learning_rate": 6.666666666666667e-05,
|
| 15 |
+
"loss": 4.0919,
|
| 16 |
+
"step": 400
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.010739102969046,
|
| 20 |
+
"eval_accuracy": 0.10586881472957423,
|
| 21 |
+
"eval_f1_macro": 0.008128718856806105,
|
| 22 |
+
"eval_loss": 3.68546199798584,
|
| 23 |
+
"eval_runtime": 133.4265,
|
| 24 |
+
"eval_samples_per_second": 65.129,
|
| 25 |
+
"eval_steps_per_second": 2.039,
|
| 26 |
+
"step": 400
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 2.021478205938092,
|
| 30 |
+
"grad_norm": 5.276159286499023,
|
| 31 |
+
"learning_rate": 0.00013333333333333334,
|
| 32 |
+
"loss": 2.9391,
|
| 33 |
+
"step": 800
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 2.021478205938092,
|
| 37 |
+
"eval_accuracy": 0.5268124280782509,
|
| 38 |
+
"eval_f1_macro": 0.2773414885221941,
|
| 39 |
+
"eval_loss": 1.907711386680603,
|
| 40 |
+
"eval_runtime": 132.9453,
|
| 41 |
+
"eval_samples_per_second": 65.365,
|
| 42 |
+
"eval_steps_per_second": 2.046,
|
| 43 |
+
"step": 800
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.0322173089071383,
|
| 47 |
+
"grad_norm": 5.944188117980957,
|
| 48 |
+
"learning_rate": 0.0002,
|
| 49 |
+
"loss": 1.583,
|
| 50 |
+
"step": 1200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.0322173089071383,
|
| 54 |
+
"eval_accuracy": 0.6894131185270426,
|
| 55 |
+
"eval_f1_macro": 0.48707209156248815,
|
| 56 |
+
"eval_loss": 1.2798452377319336,
|
| 57 |
+
"eval_runtime": 133.055,
|
| 58 |
+
"eval_samples_per_second": 65.311,
|
| 59 |
+
"eval_steps_per_second": 2.044,
|
| 60 |
+
"step": 1200
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 4.042956411876184,
|
| 64 |
+
"grad_norm": 6.609740257263184,
|
| 65 |
+
"learning_rate": 0.0002666666666666667,
|
| 66 |
+
"loss": 1.0089,
|
| 67 |
+
"step": 1600
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 4.042956411876184,
|
| 71 |
+
"eval_accuracy": 0.7447640966628308,
|
| 72 |
+
"eval_f1_macro": 0.5630866427455141,
|
| 73 |
+
"eval_loss": 1.1743698120117188,
|
| 74 |
+
"eval_runtime": 132.7655,
|
| 75 |
+
"eval_samples_per_second": 65.454,
|
| 76 |
+
"eval_steps_per_second": 2.049,
|
| 77 |
+
"step": 1600
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 5.053695514845231,
|
| 81 |
+
"grad_norm": 9.530195236206055,
|
| 82 |
+
"learning_rate": 0.0003333333333333333,
|
| 83 |
+
"loss": 0.7348,
|
| 84 |
+
"step": 2000
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 5.053695514845231,
|
| 88 |
+
"eval_accuracy": 0.7604142692750288,
|
| 89 |
+
"eval_f1_macro": 0.5961285021365654,
|
| 90 |
+
"eval_loss": 1.1527246236801147,
|
| 91 |
+
"eval_runtime": 127.8747,
|
| 92 |
+
"eval_samples_per_second": 67.957,
|
| 93 |
+
"eval_steps_per_second": 2.127,
|
| 94 |
+
"step": 2000
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 6.0644346178142765,
|
| 98 |
+
"grad_norm": 6.680343151092529,
|
| 99 |
+
"learning_rate": 0.0004,
|
| 100 |
+
"loss": 0.5957,
|
| 101 |
+
"step": 2400
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 6.0644346178142765,
|
| 105 |
+
"eval_accuracy": 0.7676639815880322,
|
| 106 |
+
"eval_f1_macro": 0.6054951189790404,
|
| 107 |
+
"eval_loss": 1.198480248451233,
|
| 108 |
+
"eval_runtime": 128.28,
|
| 109 |
+
"eval_samples_per_second": 67.742,
|
| 110 |
+
"eval_steps_per_second": 2.12,
|
| 111 |
+
"step": 2400
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 7.075173720783323,
|
| 115 |
+
"grad_norm": 3.774092435836792,
|
| 116 |
+
"learning_rate": 0.00046666666666666666,
|
| 117 |
+
"loss": 0.521,
|
| 118 |
+
"step": 2800
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 7.075173720783323,
|
| 122 |
+
"eval_accuracy": 0.7630609896432681,
|
| 123 |
+
"eval_f1_macro": 0.5903658522565237,
|
| 124 |
+
"eval_loss": 1.1921718120574951,
|
| 125 |
+
"eval_runtime": 128.2033,
|
| 126 |
+
"eval_samples_per_second": 67.783,
|
| 127 |
+
"eval_steps_per_second": 2.122,
|
| 128 |
+
"step": 2800
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 8.085912823752368,
|
| 132 |
+
"grad_norm": 3.719675302505493,
|
| 133 |
+
"learning_rate": 0.0004962962962962963,
|
| 134 |
+
"loss": 0.4667,
|
| 135 |
+
"step": 3200
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 8.085912823752368,
|
| 139 |
+
"eval_accuracy": 0.7619102416570771,
|
| 140 |
+
"eval_f1_macro": 0.6061718024259425,
|
| 141 |
+
"eval_loss": 1.2508888244628906,
|
| 142 |
+
"eval_runtime": 109.5839,
|
| 143 |
+
"eval_samples_per_second": 79.3,
|
| 144 |
+
"eval_steps_per_second": 2.482,
|
| 145 |
+
"step": 3200
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 9.096651926721416,
|
| 149 |
+
"grad_norm": 3.703678607940674,
|
| 150 |
+
"learning_rate": 0.0004888888888888889,
|
| 151 |
+
"loss": 0.3861,
|
| 152 |
+
"step": 3600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 9.096651926721416,
|
| 156 |
+
"eval_accuracy": 0.7640966628308401,
|
| 157 |
+
"eval_f1_macro": 0.5910106640214171,
|
| 158 |
+
"eval_loss": 1.2851234674453735,
|
| 159 |
+
"eval_runtime": 109.2588,
|
| 160 |
+
"eval_samples_per_second": 79.536,
|
| 161 |
+
"eval_steps_per_second": 2.49,
|
| 162 |
+
"step": 3600
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 10.107391029690461,
|
| 166 |
+
"grad_norm": 5.4869585037231445,
|
| 167 |
+
"learning_rate": 0.00048148148148148144,
|
| 168 |
+
"loss": 0.32,
|
| 169 |
+
"step": 4000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 10.107391029690461,
|
| 173 |
+
"eval_accuracy": 0.7590333716915996,
|
| 174 |
+
"eval_f1_macro": 0.5804751345832923,
|
| 175 |
+
"eval_loss": 1.4432213306427002,
|
| 176 |
+
"eval_runtime": 109.3455,
|
| 177 |
+
"eval_samples_per_second": 79.473,
|
| 178 |
+
"eval_steps_per_second": 2.488,
|
| 179 |
+
"step": 4000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 11.118130132659507,
|
| 183 |
+
"grad_norm": 2.1531548500061035,
|
| 184 |
+
"learning_rate": 0.0004740740740740741,
|
| 185 |
+
"loss": 0.2828,
|
| 186 |
+
"step": 4400
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 11.118130132659507,
|
| 190 |
+
"eval_accuracy": 0.7590333716915996,
|
| 191 |
+
"eval_f1_macro": 0.6021086310983942,
|
| 192 |
+
"eval_loss": 1.3173363208770752,
|
| 193 |
+
"eval_runtime": 109.3574,
|
| 194 |
+
"eval_samples_per_second": 79.464,
|
| 195 |
+
"eval_steps_per_second": 2.487,
|
| 196 |
+
"step": 4400
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 12.128869235628553,
|
| 200 |
+
"grad_norm": 2.9061076641082764,
|
| 201 |
+
"learning_rate": 0.00046666666666666666,
|
| 202 |
+
"loss": 0.2367,
|
| 203 |
+
"step": 4800
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 12.128869235628553,
|
| 207 |
+
"eval_accuracy": 0.7543153049482163,
|
| 208 |
+
"eval_f1_macro": 0.6092446104843484,
|
| 209 |
+
"eval_loss": 1.4384377002716064,
|
| 210 |
+
"eval_runtime": 109.3136,
|
| 211 |
+
"eval_samples_per_second": 79.496,
|
| 212 |
+
"eval_steps_per_second": 2.488,
|
| 213 |
+
"step": 4800
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 13.139608338597599,
|
| 217 |
+
"grad_norm": 2.8866333961486816,
|
| 218 |
+
"learning_rate": 0.00045925925925925925,
|
| 219 |
+
"loss": 0.2187,
|
| 220 |
+
"step": 5200
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 13.139608338597599,
|
| 224 |
+
"eval_accuracy": 0.7654775604142693,
|
| 225 |
+
"eval_f1_macro": 0.5880603922791815,
|
| 226 |
+
"eval_loss": 1.4380950927734375,
|
| 227 |
+
"eval_runtime": 109.4554,
|
| 228 |
+
"eval_samples_per_second": 79.393,
|
| 229 |
+
"eval_steps_per_second": 2.485,
|
| 230 |
+
"step": 5200
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 14.150347441566646,
|
| 234 |
+
"grad_norm": 1.7574183940887451,
|
| 235 |
+
"learning_rate": 0.00045185185185185183,
|
| 236 |
+
"loss": 0.1847,
|
| 237 |
+
"step": 5600
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 14.150347441566646,
|
| 241 |
+
"eval_accuracy": 0.7730724971231301,
|
| 242 |
+
"eval_f1_macro": 0.5690127519635726,
|
| 243 |
+
"eval_loss": 1.4231289625167847,
|
| 244 |
+
"eval_runtime": 109.3887,
|
| 245 |
+
"eval_samples_per_second": 79.441,
|
| 246 |
+
"eval_steps_per_second": 2.487,
|
| 247 |
+
"step": 5600
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 15.161086544535692,
|
| 251 |
+
"grad_norm": 1.8373284339904785,
|
| 252 |
+
"learning_rate": 0.0004444444444444444,
|
| 253 |
+
"loss": 0.1701,
|
| 254 |
+
"step": 6000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 15.161086544535692,
|
| 258 |
+
"eval_accuracy": 0.7680092059838896,
|
| 259 |
+
"eval_f1_macro": 0.5878361109327175,
|
| 260 |
+
"eval_loss": 1.5120900869369507,
|
| 261 |
+
"eval_runtime": 109.6944,
|
| 262 |
+
"eval_samples_per_second": 79.22,
|
| 263 |
+
"eval_steps_per_second": 2.48,
|
| 264 |
+
"step": 6000
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 16.171825647504736,
|
| 268 |
+
"grad_norm": 2.9617397785186768,
|
| 269 |
+
"learning_rate": 0.00043703703703703705,
|
| 270 |
+
"loss": 0.1504,
|
| 271 |
+
"step": 6400
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 16.171825647504736,
|
| 275 |
+
"eval_accuracy": 0.7609896432681242,
|
| 276 |
+
"eval_f1_macro": 0.6017434401264726,
|
| 277 |
+
"eval_loss": 1.5701994895935059,
|
| 278 |
+
"eval_runtime": 108.7867,
|
| 279 |
+
"eval_samples_per_second": 79.881,
|
| 280 |
+
"eval_steps_per_second": 2.5,
|
| 281 |
+
"step": 6400
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 17.182564750473784,
|
| 285 |
+
"grad_norm": 1.9067094326019287,
|
| 286 |
+
"learning_rate": 0.00042962962962962963,
|
| 287 |
+
"loss": 0.1416,
|
| 288 |
+
"step": 6800
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 17.182564750473784,
|
| 292 |
+
"eval_accuracy": 0.7680092059838896,
|
| 293 |
+
"eval_f1_macro": 0.5846132297229183,
|
| 294 |
+
"eval_loss": 1.6262372732162476,
|
| 295 |
+
"eval_runtime": 109.3355,
|
| 296 |
+
"eval_samples_per_second": 79.48,
|
| 297 |
+
"eval_steps_per_second": 2.488,
|
| 298 |
+
"step": 6800
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 18.19330385344283,
|
| 302 |
+
"grad_norm": 1.788485050201416,
|
| 303 |
+
"learning_rate": 0.0004222222222222222,
|
| 304 |
+
"loss": 0.1345,
|
| 305 |
+
"step": 7200
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 18.19330385344283,
|
| 309 |
+
"eval_accuracy": 0.7582278481012659,
|
| 310 |
+
"eval_f1_macro": 0.606730101292868,
|
| 311 |
+
"eval_loss": 1.6317014694213867,
|
| 312 |
+
"eval_runtime": 109.1193,
|
| 313 |
+
"eval_samples_per_second": 79.638,
|
| 314 |
+
"eval_steps_per_second": 2.493,
|
| 315 |
+
"step": 7200
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 19.204042956411875,
|
| 319 |
+
"grad_norm": 3.0378000736236572,
|
| 320 |
+
"learning_rate": 0.0004148148148148148,
|
| 321 |
+
"loss": 0.1226,
|
| 322 |
+
"step": 7600
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 19.204042956411875,
|
| 326 |
+
"eval_accuracy": 0.7739930955120828,
|
| 327 |
+
"eval_f1_macro": 0.6193094447560485,
|
| 328 |
+
"eval_loss": 1.486433982849121,
|
| 329 |
+
"eval_runtime": 109.0558,
|
| 330 |
+
"eval_samples_per_second": 79.684,
|
| 331 |
+
"eval_steps_per_second": 2.494,
|
| 332 |
+
"step": 7600
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 20.214782059380923,
|
| 336 |
+
"grad_norm": 3.1991524696350098,
|
| 337 |
+
"learning_rate": 0.0004074074074074074,
|
| 338 |
+
"loss": 0.114,
|
| 339 |
+
"step": 8000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 20.214782059380923,
|
| 343 |
+
"eval_accuracy": 0.774108170310702,
|
| 344 |
+
"eval_f1_macro": 0.6157091732739274,
|
| 345 |
+
"eval_loss": 1.5931099653244019,
|
| 346 |
+
"eval_runtime": 109.0943,
|
| 347 |
+
"eval_samples_per_second": 79.656,
|
| 348 |
+
"eval_steps_per_second": 2.493,
|
| 349 |
+
"step": 8000
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 21.225521162349967,
|
| 353 |
+
"grad_norm": 2.1036899089813232,
|
| 354 |
+
"learning_rate": 0.0004,
|
| 355 |
+
"loss": 0.1064,
|
| 356 |
+
"step": 8400
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 21.225521162349967,
|
| 360 |
+
"eval_accuracy": 0.7730724971231301,
|
| 361 |
+
"eval_f1_macro": 0.6020232192562277,
|
| 362 |
+
"eval_loss": 1.7101207971572876,
|
| 363 |
+
"eval_runtime": 108.899,
|
| 364 |
+
"eval_samples_per_second": 79.799,
|
| 365 |
+
"eval_steps_per_second": 2.498,
|
| 366 |
+
"step": 8400
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 22.236260265319014,
|
| 370 |
+
"grad_norm": 2.786360025405884,
|
| 371 |
+
"learning_rate": 0.0003925925925925926,
|
| 372 |
+
"loss": 0.1009,
|
| 373 |
+
"step": 8800
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 22.236260265319014,
|
| 377 |
+
"eval_accuracy": 0.7655926352128883,
|
| 378 |
+
"eval_f1_macro": 0.5794753743607411,
|
| 379 |
+
"eval_loss": 1.6664392948150635,
|
| 380 |
+
"eval_runtime": 109.2502,
|
| 381 |
+
"eval_samples_per_second": 79.542,
|
| 382 |
+
"eval_steps_per_second": 2.49,
|
| 383 |
+
"step": 8800
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 23.246999368288062,
|
| 387 |
+
"grad_norm": 1.0751720666885376,
|
| 388 |
+
"learning_rate": 0.0003851851851851852,
|
| 389 |
+
"loss": 0.0941,
|
| 390 |
+
"step": 9200
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 23.246999368288062,
|
| 394 |
+
"eval_accuracy": 0.7772151898734178,
|
| 395 |
+
"eval_f1_macro": 0.5717636011134882,
|
| 396 |
+
"eval_loss": 1.5253993272781372,
|
| 397 |
+
"eval_runtime": 109.0143,
|
| 398 |
+
"eval_samples_per_second": 79.714,
|
| 399 |
+
"eval_steps_per_second": 2.495,
|
| 400 |
+
"step": 9200
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 24.257738471257106,
|
| 404 |
+
"grad_norm": 1.744019865989685,
|
| 405 |
+
"learning_rate": 0.00037777777777777777,
|
| 406 |
+
"loss": 0.0861,
|
| 407 |
+
"step": 9600
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 24.257738471257106,
|
| 411 |
+
"eval_accuracy": 0.777445339470656,
|
| 412 |
+
"eval_f1_macro": 0.625140306336925,
|
| 413 |
+
"eval_loss": 1.6324084997177124,
|
| 414 |
+
"eval_runtime": 108.6336,
|
| 415 |
+
"eval_samples_per_second": 79.994,
|
| 416 |
+
"eval_steps_per_second": 2.504,
|
| 417 |
+
"step": 9600
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 25.268477574226154,
|
| 421 |
+
"grad_norm": 1.838752269744873,
|
| 422 |
+
"learning_rate": 0.00037037037037037035,
|
| 423 |
+
"loss": 0.0807,
|
| 424 |
+
"step": 10000
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 25.268477574226154,
|
| 428 |
+
"eval_accuracy": 0.7728423475258919,
|
| 429 |
+
"eval_f1_macro": 0.5870939911644882,
|
| 430 |
+
"eval_loss": 1.7057673931121826,
|
| 431 |
+
"eval_runtime": 108.6842,
|
| 432 |
+
"eval_samples_per_second": 79.956,
|
| 433 |
+
"eval_steps_per_second": 2.503,
|
| 434 |
+
"step": 10000
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 26.279216677195198,
|
| 438 |
+
"grad_norm": 2.3391871452331543,
|
| 439 |
+
"learning_rate": 0.000362962962962963,
|
| 440 |
+
"loss": 0.0739,
|
| 441 |
+
"step": 10400
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 26.279216677195198,
|
| 445 |
+
"eval_accuracy": 0.774108170310702,
|
| 446 |
+
"eval_f1_macro": 0.6190123341706849,
|
| 447 |
+
"eval_loss": 1.6950148344039917,
|
| 448 |
+
"eval_runtime": 108.9167,
|
| 449 |
+
"eval_samples_per_second": 79.786,
|
| 450 |
+
"eval_steps_per_second": 2.497,
|
| 451 |
+
"step": 10400
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 27.289955780164245,
|
| 455 |
+
"grad_norm": 1.3197505474090576,
|
| 456 |
+
"learning_rate": 0.00035555555555555557,
|
| 457 |
+
"loss": 0.0685,
|
| 458 |
+
"step": 10800
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 27.289955780164245,
|
| 462 |
+
"eval_accuracy": 0.7652474108170311,
|
| 463 |
+
"eval_f1_macro": 0.5984200620053731,
|
| 464 |
+
"eval_loss": 1.8148038387298584,
|
| 465 |
+
"eval_runtime": 108.998,
|
| 466 |
+
"eval_samples_per_second": 79.726,
|
| 467 |
+
"eval_steps_per_second": 2.495,
|
| 468 |
+
"step": 10800
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 28.300694883133293,
|
| 472 |
+
"grad_norm": 0.8027063608169556,
|
| 473 |
+
"learning_rate": 0.00034814814814814816,
|
| 474 |
+
"loss": 0.0692,
|
| 475 |
+
"step": 11200
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 28.300694883133293,
|
| 479 |
+
"eval_accuracy": 0.776409666283084,
|
| 480 |
+
"eval_f1_macro": 0.6002766778970904,
|
| 481 |
+
"eval_loss": 1.6219606399536133,
|
| 482 |
+
"eval_runtime": 108.9613,
|
| 483 |
+
"eval_samples_per_second": 79.753,
|
| 484 |
+
"eval_steps_per_second": 2.496,
|
| 485 |
+
"step": 11200
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 29.311433986102337,
|
| 489 |
+
"grad_norm": 0.8713662028312683,
|
| 490 |
+
"learning_rate": 0.00034074074074074074,
|
| 491 |
+
"loss": 0.0662,
|
| 492 |
+
"step": 11600
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 29.311433986102337,
|
| 496 |
+
"eval_accuracy": 0.7794016110471806,
|
| 497 |
+
"eval_f1_macro": 0.6123819840203646,
|
| 498 |
+
"eval_loss": 1.6953762769699097,
|
| 499 |
+
"eval_runtime": 109.1585,
|
| 500 |
+
"eval_samples_per_second": 79.609,
|
| 501 |
+
"eval_steps_per_second": 2.492,
|
| 502 |
+
"step": 11600
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 30.322173089071384,
|
| 506 |
+
"grad_norm": 0.9094525575637817,
|
| 507 |
+
"learning_rate": 0.0003333333333333333,
|
| 508 |
+
"loss": 0.0639,
|
| 509 |
+
"step": 12000
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 30.322173089071384,
|
| 513 |
+
"eval_accuracy": 0.7785960874568469,
|
| 514 |
+
"eval_f1_macro": 0.5900178041075752,
|
| 515 |
+
"eval_loss": 1.7562154531478882,
|
| 516 |
+
"eval_runtime": 108.917,
|
| 517 |
+
"eval_samples_per_second": 79.786,
|
| 518 |
+
"eval_steps_per_second": 2.497,
|
| 519 |
+
"step": 12000
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 31.33291219204043,
|
| 523 |
+
"grad_norm": 2.3824515342712402,
|
| 524 |
+
"learning_rate": 0.00032592592592592596,
|
| 525 |
+
"loss": 0.0613,
|
| 526 |
+
"step": 12400
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 31.33291219204043,
|
| 530 |
+
"eval_accuracy": 0.7708860759493671,
|
| 531 |
+
"eval_f1_macro": 0.5886611331241638,
|
| 532 |
+
"eval_loss": 1.7263332605361938,
|
| 533 |
+
"eval_runtime": 109.2037,
|
| 534 |
+
"eval_samples_per_second": 79.576,
|
| 535 |
+
"eval_steps_per_second": 2.491,
|
| 536 |
+
"step": 12400
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 32.34365129500947,
|
| 540 |
+
"grad_norm": 1.1265066862106323,
|
| 541 |
+
"learning_rate": 0.00031851851851851854,
|
| 542 |
+
"loss": 0.0562,
|
| 543 |
+
"step": 12800
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 32.34365129500947,
|
| 547 |
+
"eval_accuracy": 0.777445339470656,
|
| 548 |
+
"eval_f1_macro": 0.6069323146272442,
|
| 549 |
+
"eval_loss": 1.595489263534546,
|
| 550 |
+
"eval_runtime": 110.1086,
|
| 551 |
+
"eval_samples_per_second": 78.922,
|
| 552 |
+
"eval_steps_per_second": 2.47,
|
| 553 |
+
"step": 12800
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 33.35439039797852,
|
| 557 |
+
"grad_norm": 0.765870988368988,
|
| 558 |
+
"learning_rate": 0.0003111111111111111,
|
| 559 |
+
"loss": 0.0482,
|
| 560 |
+
"step": 13200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 33.35439039797852,
|
| 564 |
+
"eval_accuracy": 0.7858457997698504,
|
| 565 |
+
"eval_f1_macro": 0.6152260699722518,
|
| 566 |
+
"eval_loss": 1.6528053283691406,
|
| 567 |
+
"eval_runtime": 109.0363,
|
| 568 |
+
"eval_samples_per_second": 79.698,
|
| 569 |
+
"eval_steps_per_second": 2.495,
|
| 570 |
+
"step": 13200
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 34.36512950094757,
|
| 574 |
+
"grad_norm": 2.386359930038452,
|
| 575 |
+
"learning_rate": 0.0003037037037037037,
|
| 576 |
+
"loss": 0.0516,
|
| 577 |
+
"step": 13600
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 34.36512950094757,
|
| 581 |
+
"eval_accuracy": 0.7713463751438435,
|
| 582 |
+
"eval_f1_macro": 0.5894778786253475,
|
| 583 |
+
"eval_loss": 1.65277099609375,
|
| 584 |
+
"eval_runtime": 109.1673,
|
| 585 |
+
"eval_samples_per_second": 79.603,
|
| 586 |
+
"eval_steps_per_second": 2.492,
|
| 587 |
+
"step": 13600
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 35.375868603916615,
|
| 591 |
+
"grad_norm": 1.8987774848937988,
|
| 592 |
+
"learning_rate": 0.0002962962962962963,
|
| 593 |
+
"loss": 0.0447,
|
| 594 |
+
"step": 14000
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 35.375868603916615,
|
| 598 |
+
"eval_accuracy": 0.7799769850402761,
|
| 599 |
+
"eval_f1_macro": 0.6297477374058172,
|
| 600 |
+
"eval_loss": 1.813390851020813,
|
| 601 |
+
"eval_runtime": 109.6977,
|
| 602 |
+
"eval_samples_per_second": 79.218,
|
| 603 |
+
"eval_steps_per_second": 2.48,
|
| 604 |
+
"step": 14000
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 36.38660770688566,
|
| 608 |
+
"grad_norm": 1.353411078453064,
|
| 609 |
+
"learning_rate": 0.0002888888888888889,
|
| 610 |
+
"loss": 0.047,
|
| 611 |
+
"step": 14400
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 36.38660770688566,
|
| 615 |
+
"eval_accuracy": 0.7795166858457998,
|
| 616 |
+
"eval_f1_macro": 0.5795862617467612,
|
| 617 |
+
"eval_loss": 1.663203477859497,
|
| 618 |
+
"eval_runtime": 109.0323,
|
| 619 |
+
"eval_samples_per_second": 79.701,
|
| 620 |
+
"eval_steps_per_second": 2.495,
|
| 621 |
+
"step": 14400
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 37.3973468098547,
|
| 625 |
+
"grad_norm": 1.1114296913146973,
|
| 626 |
+
"learning_rate": 0.0002814814814814815,
|
| 627 |
+
"loss": 0.0436,
|
| 628 |
+
"step": 14800
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 37.3973468098547,
|
| 632 |
+
"eval_accuracy": 0.784234752589183,
|
| 633 |
+
"eval_f1_macro": 0.5995152264247978,
|
| 634 |
+
"eval_loss": 1.783818006515503,
|
| 635 |
+
"eval_runtime": 109.4106,
|
| 636 |
+
"eval_samples_per_second": 79.426,
|
| 637 |
+
"eval_steps_per_second": 2.486,
|
| 638 |
+
"step": 14800
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 38.40808591282375,
|
| 642 |
+
"grad_norm": 1.3422303199768066,
|
| 643 |
+
"learning_rate": 0.0002740740740740741,
|
| 644 |
+
"loss": 0.0422,
|
| 645 |
+
"step": 15200
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 38.40808591282375,
|
| 649 |
+
"eval_accuracy": 0.7838895281933257,
|
| 650 |
+
"eval_f1_macro": 0.6189287691248615,
|
| 651 |
+
"eval_loss": 1.7172709703445435,
|
| 652 |
+
"eval_runtime": 108.6629,
|
| 653 |
+
"eval_samples_per_second": 79.972,
|
| 654 |
+
"eval_steps_per_second": 2.503,
|
| 655 |
+
"step": 15200
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 39.4188250157928,
|
| 659 |
+
"grad_norm": 1.8279023170471191,
|
| 660 |
+
"learning_rate": 0.0002666666666666667,
|
| 661 |
+
"loss": 0.0377,
|
| 662 |
+
"step": 15600
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 39.4188250157928,
|
| 666 |
+
"eval_accuracy": 0.7834292289988493,
|
| 667 |
+
"eval_f1_macro": 0.5814739153081228,
|
| 668 |
+
"eval_loss": 1.7523770332336426,
|
| 669 |
+
"eval_runtime": 108.9839,
|
| 670 |
+
"eval_samples_per_second": 79.737,
|
| 671 |
+
"eval_steps_per_second": 2.496,
|
| 672 |
+
"step": 15600
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 40.429564118761846,
|
| 676 |
+
"grad_norm": 2.154459238052368,
|
| 677 |
+
"learning_rate": 0.00025925925925925926,
|
| 678 |
+
"loss": 0.0359,
|
| 679 |
+
"step": 16000
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 40.429564118761846,
|
| 683 |
+
"eval_accuracy": 0.7886075949367088,
|
| 684 |
+
"eval_f1_macro": 0.6293741181702724,
|
| 685 |
+
"eval_loss": 1.623598337173462,
|
| 686 |
+
"eval_runtime": 108.8195,
|
| 687 |
+
"eval_samples_per_second": 79.857,
|
| 688 |
+
"eval_steps_per_second": 2.5,
|
| 689 |
+
"step": 16000
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 41.44030322173089,
|
| 693 |
+
"grad_norm": 0.8551483154296875,
|
| 694 |
+
"learning_rate": 0.00025185185185185185,
|
| 695 |
+
"loss": 0.0344,
|
| 696 |
+
"step": 16400
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 41.44030322173089,
|
| 700 |
+
"eval_accuracy": 0.7815880322209436,
|
| 701 |
+
"eval_f1_macro": 0.6087804648227756,
|
| 702 |
+
"eval_loss": 1.7353272438049316,
|
| 703 |
+
"eval_runtime": 109.2273,
|
| 704 |
+
"eval_samples_per_second": 79.559,
|
| 705 |
+
"eval_steps_per_second": 2.49,
|
| 706 |
+
"step": 16400
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 42.451042324699934,
|
| 710 |
+
"grad_norm": 0.5178919434547424,
|
| 711 |
+
"learning_rate": 0.00024444444444444443,
|
| 712 |
+
"loss": 0.033,
|
| 713 |
+
"step": 16800
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 42.451042324699934,
|
| 717 |
+
"eval_accuracy": 0.7820483314154201,
|
| 718 |
+
"eval_f1_macro": 0.6001569016578011,
|
| 719 |
+
"eval_loss": 1.727620244026184,
|
| 720 |
+
"eval_runtime": 109.4385,
|
| 721 |
+
"eval_samples_per_second": 79.405,
|
| 722 |
+
"eval_steps_per_second": 2.485,
|
| 723 |
+
"step": 16800
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 43.46178142766898,
|
| 727 |
+
"grad_norm": 0.4940205514431,
|
| 728 |
+
"learning_rate": 0.00023703703703703704,
|
| 729 |
+
"loss": 0.0325,
|
| 730 |
+
"step": 17200
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 43.46178142766898,
|
| 734 |
+
"eval_accuracy": 0.7783659378596087,
|
| 735 |
+
"eval_f1_macro": 0.6283289368126677,
|
| 736 |
+
"eval_loss": 1.7798371315002441,
|
| 737 |
+
"eval_runtime": 109.2576,
|
| 738 |
+
"eval_samples_per_second": 79.537,
|
| 739 |
+
"eval_steps_per_second": 2.49,
|
| 740 |
+
"step": 17200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 44.47252053063803,
|
| 744 |
+
"grad_norm": 0.8661497235298157,
|
| 745 |
+
"learning_rate": 0.00022962962962962962,
|
| 746 |
+
"loss": 0.0302,
|
| 747 |
+
"step": 17600
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 44.47252053063803,
|
| 751 |
+
"eval_accuracy": 0.7828538550057538,
|
| 752 |
+
"eval_f1_macro": 0.6164776778280789,
|
| 753 |
+
"eval_loss": 1.7507109642028809,
|
| 754 |
+
"eval_runtime": 109.1869,
|
| 755 |
+
"eval_samples_per_second": 79.588,
|
| 756 |
+
"eval_steps_per_second": 2.491,
|
| 757 |
+
"step": 17600
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 45.48325963360708,
|
| 761 |
+
"grad_norm": 0.015332411043345928,
|
| 762 |
+
"learning_rate": 0.0002222222222222222,
|
| 763 |
+
"loss": 0.0268,
|
| 764 |
+
"step": 18000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 45.48325963360708,
|
| 768 |
+
"eval_accuracy": 0.7826237054085156,
|
| 769 |
+
"eval_f1_macro": 0.6031617249417177,
|
| 770 |
+
"eval_loss": 1.7825220823287964,
|
| 771 |
+
"eval_runtime": 109.3518,
|
| 772 |
+
"eval_samples_per_second": 79.468,
|
| 773 |
+
"eval_steps_per_second": 2.487,
|
| 774 |
+
"step": 18000
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 46.493998736576124,
|
| 778 |
+
"grad_norm": 0.5325392484664917,
|
| 779 |
+
"learning_rate": 0.00021481481481481482,
|
| 780 |
+
"loss": 0.0287,
|
| 781 |
+
"step": 18400
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 46.493998736576124,
|
| 785 |
+
"eval_accuracy": 0.7882623705408516,
|
| 786 |
+
"eval_f1_macro": 0.6256320010133759,
|
| 787 |
+
"eval_loss": 1.6932624578475952,
|
| 788 |
+
"eval_runtime": 108.513,
|
| 789 |
+
"eval_samples_per_second": 80.083,
|
| 790 |
+
"eval_steps_per_second": 2.507,
|
| 791 |
+
"step": 18400
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 47.504737839545164,
|
| 795 |
+
"grad_norm": 0.5086055994033813,
|
| 796 |
+
"learning_rate": 0.0002074074074074074,
|
| 797 |
+
"loss": 0.0252,
|
| 798 |
+
"step": 18800
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 47.504737839545164,
|
| 802 |
+
"eval_accuracy": 0.7856156501726121,
|
| 803 |
+
"eval_f1_macro": 0.6143416230351354,
|
| 804 |
+
"eval_loss": 1.7501070499420166,
|
| 805 |
+
"eval_runtime": 109.2365,
|
| 806 |
+
"eval_samples_per_second": 79.552,
|
| 807 |
+
"eval_steps_per_second": 2.49,
|
| 808 |
+
"step": 18800
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 48.51547694251421,
|
| 812 |
+
"grad_norm": 1.229317545890808,
|
| 813 |
+
"learning_rate": 0.0002,
|
| 814 |
+
"loss": 0.0283,
|
| 815 |
+
"step": 19200
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 48.51547694251421,
|
| 819 |
+
"eval_accuracy": 0.7843498273878021,
|
| 820 |
+
"eval_f1_macro": 0.6189575264715401,
|
| 821 |
+
"eval_loss": 1.9032423496246338,
|
| 822 |
+
"eval_runtime": 108.2906,
|
| 823 |
+
"eval_samples_per_second": 80.247,
|
| 824 |
+
"eval_steps_per_second": 2.512,
|
| 825 |
+
"step": 19200
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 49.52621604548326,
|
| 829 |
+
"grad_norm": 0.05275914818048477,
|
| 830 |
+
"learning_rate": 0.0001925925925925926,
|
| 831 |
+
"loss": 0.024,
|
| 832 |
+
"step": 19600
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 49.52621604548326,
|
| 836 |
+
"eval_accuracy": 0.7874568469505179,
|
| 837 |
+
"eval_f1_macro": 0.6393370936978522,
|
| 838 |
+
"eval_loss": 1.8691409826278687,
|
| 839 |
+
"eval_runtime": 108.1545,
|
| 840 |
+
"eval_samples_per_second": 80.348,
|
| 841 |
+
"eval_steps_per_second": 2.515,
|
| 842 |
+
"step": 19600
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 50.53695514845231,
|
| 846 |
+
"grad_norm": 0.9653208255767822,
|
| 847 |
+
"learning_rate": 0.00018518518518518518,
|
| 848 |
+
"loss": 0.0229,
|
| 849 |
+
"step": 20000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 50.53695514845231,
|
| 853 |
+
"eval_accuracy": 0.786536248561565,
|
| 854 |
+
"eval_f1_macro": 0.6026385719720891,
|
| 855 |
+
"eval_loss": 1.7541390657424927,
|
| 856 |
+
"eval_runtime": 107.9085,
|
| 857 |
+
"eval_samples_per_second": 80.531,
|
| 858 |
+
"eval_steps_per_second": 2.521,
|
| 859 |
+
"step": 20000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 51.547694251421355,
|
| 863 |
+
"grad_norm": 0.4658529758453369,
|
| 864 |
+
"learning_rate": 0.00017777777777777779,
|
| 865 |
+
"loss": 0.0219,
|
| 866 |
+
"step": 20400
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 51.547694251421355,
|
| 870 |
+
"eval_accuracy": 0.7872266973532797,
|
| 871 |
+
"eval_f1_macro": 0.6309747652348119,
|
| 872 |
+
"eval_loss": 1.7537351846694946,
|
| 873 |
+
"eval_runtime": 107.7743,
|
| 874 |
+
"eval_samples_per_second": 80.632,
|
| 875 |
+
"eval_steps_per_second": 2.524,
|
| 876 |
+
"step": 20400
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 52.558433354390395,
|
| 880 |
+
"grad_norm": 0.32756420969963074,
|
| 881 |
+
"learning_rate": 0.00017037037037037037,
|
| 882 |
+
"loss": 0.0211,
|
| 883 |
+
"step": 20800
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 52.558433354390395,
|
| 887 |
+
"eval_accuracy": 0.7934407364787112,
|
| 888 |
+
"eval_f1_macro": 0.6206166338546538,
|
| 889 |
+
"eval_loss": 1.6842619180679321,
|
| 890 |
+
"eval_runtime": 107.7209,
|
| 891 |
+
"eval_samples_per_second": 80.671,
|
| 892 |
+
"eval_steps_per_second": 2.525,
|
| 893 |
+
"step": 20800
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 53.56917245735944,
|
| 897 |
+
"grad_norm": 0.584701418876648,
|
| 898 |
+
"learning_rate": 0.00016296296296296298,
|
| 899 |
+
"loss": 0.0203,
|
| 900 |
+
"step": 21200
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 53.56917245735944,
|
| 904 |
+
"eval_accuracy": 0.7950517836593786,
|
| 905 |
+
"eval_f1_macro": 0.6206542591204762,
|
| 906 |
+
"eval_loss": 1.699610710144043,
|
| 907 |
+
"eval_runtime": 107.6954,
|
| 908 |
+
"eval_samples_per_second": 80.691,
|
| 909 |
+
"eval_steps_per_second": 2.526,
|
| 910 |
+
"step": 21200
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 54.57991156032849,
|
| 914 |
+
"grad_norm": 0.0553191676735878,
|
| 915 |
+
"learning_rate": 0.00015555555555555556,
|
| 916 |
+
"loss": 0.0174,
|
| 917 |
+
"step": 21600
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 54.57991156032849,
|
| 921 |
+
"eval_accuracy": 0.7894131185270425,
|
| 922 |
+
"eval_f1_macro": 0.6214961351780512,
|
| 923 |
+
"eval_loss": 1.8445045948028564,
|
| 924 |
+
"eval_runtime": 107.7853,
|
| 925 |
+
"eval_samples_per_second": 80.623,
|
| 926 |
+
"eval_steps_per_second": 2.524,
|
| 927 |
+
"step": 21600
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 55.59065066329754,
|
| 931 |
+
"grad_norm": 0.4328874945640564,
|
| 932 |
+
"learning_rate": 0.00014814814814814815,
|
| 933 |
+
"loss": 0.0197,
|
| 934 |
+
"step": 22000
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 55.59065066329754,
|
| 938 |
+
"eval_accuracy": 0.792059838895282,
|
| 939 |
+
"eval_f1_macro": 0.6308138834712996,
|
| 940 |
+
"eval_loss": 1.8310879468917847,
|
| 941 |
+
"eval_runtime": 107.7421,
|
| 942 |
+
"eval_samples_per_second": 80.656,
|
| 943 |
+
"eval_steps_per_second": 2.525,
|
| 944 |
+
"step": 22000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 56.601389766266585,
|
| 948 |
+
"grad_norm": 0.02704198658466339,
|
| 949 |
+
"learning_rate": 0.00014074074074074076,
|
| 950 |
+
"loss": 0.0169,
|
| 951 |
+
"step": 22400
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 56.601389766266585,
|
| 955 |
+
"eval_accuracy": 0.7879171461449942,
|
| 956 |
+
"eval_f1_macro": 0.5896127682611725,
|
| 957 |
+
"eval_loss": 1.8162003755569458,
|
| 958 |
+
"eval_runtime": 107.8141,
|
| 959 |
+
"eval_samples_per_second": 80.602,
|
| 960 |
+
"eval_steps_per_second": 2.523,
|
| 961 |
+
"step": 22400
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 57.612128869235626,
|
| 965 |
+
"grad_norm": 0.2748865485191345,
|
| 966 |
+
"learning_rate": 0.00013333333333333334,
|
| 967 |
+
"loss": 0.0121,
|
| 968 |
+
"step": 22800
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 57.612128869235626,
|
| 972 |
+
"eval_accuracy": 0.7852704257767549,
|
| 973 |
+
"eval_f1_macro": 0.5951106108532582,
|
| 974 |
+
"eval_loss": 1.924727201461792,
|
| 975 |
+
"eval_runtime": 107.712,
|
| 976 |
+
"eval_samples_per_second": 80.678,
|
| 977 |
+
"eval_steps_per_second": 2.525,
|
| 978 |
+
"step": 22800
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 58.62286797220467,
|
| 982 |
+
"grad_norm": 0.0328911654651165,
|
| 983 |
+
"learning_rate": 0.00012592592592592592,
|
| 984 |
+
"loss": 0.0152,
|
| 985 |
+
"step": 23200
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 58.62286797220467,
|
| 989 |
+
"eval_accuracy": 0.7881472957422324,
|
| 990 |
+
"eval_f1_macro": 0.6063430405057288,
|
| 991 |
+
"eval_loss": 1.8502182960510254,
|
| 992 |
+
"eval_runtime": 107.788,
|
| 993 |
+
"eval_samples_per_second": 80.621,
|
| 994 |
+
"eval_steps_per_second": 2.523,
|
| 995 |
+
"step": 23200
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 59.63360707517372,
|
| 999 |
+
"grad_norm": 0.00955616869032383,
|
| 1000 |
+
"learning_rate": 0.00011851851851851852,
|
| 1001 |
+
"loss": 0.0142,
|
| 1002 |
+
"step": 23600
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 59.63360707517372,
|
| 1006 |
+
"eval_accuracy": 0.789873417721519,
|
| 1007 |
+
"eval_f1_macro": 0.617993825444742,
|
| 1008 |
+
"eval_loss": 1.7803289890289307,
|
| 1009 |
+
"eval_runtime": 107.8043,
|
| 1010 |
+
"eval_samples_per_second": 80.609,
|
| 1011 |
+
"eval_steps_per_second": 2.523,
|
| 1012 |
+
"step": 23600
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 60.64434617814277,
|
| 1016 |
+
"grad_norm": 0.06125176325440407,
|
| 1017 |
+
"learning_rate": 0.0001111111111111111,
|
| 1018 |
+
"loss": 0.0105,
|
| 1019 |
+
"step": 24000
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 60.64434617814277,
|
| 1023 |
+
"eval_accuracy": 0.7861910241657077,
|
| 1024 |
+
"eval_f1_macro": 0.6254018987758924,
|
| 1025 |
+
"eval_loss": 1.916595458984375,
|
| 1026 |
+
"eval_runtime": 107.7673,
|
| 1027 |
+
"eval_samples_per_second": 80.637,
|
| 1028 |
+
"eval_steps_per_second": 2.524,
|
| 1029 |
+
"step": 24000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 61.655085281111816,
|
| 1033 |
+
"grad_norm": 0.10605888813734055,
|
| 1034 |
+
"learning_rate": 0.0001037037037037037,
|
| 1035 |
+
"loss": 0.0116,
|
| 1036 |
+
"step": 24400
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 61.655085281111816,
|
| 1040 |
+
"eval_accuracy": 0.7858457997698504,
|
| 1041 |
+
"eval_f1_macro": 0.5961002471321352,
|
| 1042 |
+
"eval_loss": 1.9204109907150269,
|
| 1043 |
+
"eval_runtime": 107.7648,
|
| 1044 |
+
"eval_samples_per_second": 80.639,
|
| 1045 |
+
"eval_steps_per_second": 2.524,
|
| 1046 |
+
"step": 24400
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 62.66582438408086,
|
| 1050 |
+
"grad_norm": 0.044181693345308304,
|
| 1051 |
+
"learning_rate": 9.62962962962963e-05,
|
| 1052 |
+
"loss": 0.0112,
|
| 1053 |
+
"step": 24800
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 62.66582438408086,
|
| 1057 |
+
"eval_accuracy": 0.7878020713463751,
|
| 1058 |
+
"eval_f1_macro": 0.6235710102313945,
|
| 1059 |
+
"eval_loss": 1.9822152853012085,
|
| 1060 |
+
"eval_runtime": 107.735,
|
| 1061 |
+
"eval_samples_per_second": 80.661,
|
| 1062 |
+
"eval_steps_per_second": 2.525,
|
| 1063 |
+
"step": 24800
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 63.676563487049904,
|
| 1067 |
+
"grad_norm": 0.023459970951080322,
|
| 1068 |
+
"learning_rate": 8.888888888888889e-05,
|
| 1069 |
+
"loss": 0.0102,
|
| 1070 |
+
"step": 25200
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 63.676563487049904,
|
| 1074 |
+
"eval_accuracy": 0.7840046029919447,
|
| 1075 |
+
"eval_f1_macro": 0.6155669395709024,
|
| 1076 |
+
"eval_loss": 1.9653674364089966,
|
| 1077 |
+
"eval_runtime": 107.7821,
|
| 1078 |
+
"eval_samples_per_second": 80.626,
|
| 1079 |
+
"eval_steps_per_second": 2.524,
|
| 1080 |
+
"step": 25200
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 64.68730259001894,
|
| 1084 |
+
"grad_norm": 1.9076263904571533,
|
| 1085 |
+
"learning_rate": 8.148148148148149e-05,
|
| 1086 |
+
"loss": 0.01,
|
| 1087 |
+
"step": 25600
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 64.68730259001894,
|
| 1091 |
+
"eval_accuracy": 0.7880322209436134,
|
| 1092 |
+
"eval_f1_macro": 0.6226637633596005,
|
| 1093 |
+
"eval_loss": 1.938231348991394,
|
| 1094 |
+
"eval_runtime": 107.7205,
|
| 1095 |
+
"eval_samples_per_second": 80.672,
|
| 1096 |
+
"eval_steps_per_second": 2.525,
|
| 1097 |
+
"step": 25600
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 65.698041692988,
|
| 1101 |
+
"grad_norm": 0.4948989748954773,
|
| 1102 |
+
"learning_rate": 7.407407407407407e-05,
|
| 1103 |
+
"loss": 0.0101,
|
| 1104 |
+
"step": 26000
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 65.698041692988,
|
| 1108 |
+
"eval_accuracy": 0.7960874568469505,
|
| 1109 |
+
"eval_f1_macro": 0.6277935659004009,
|
| 1110 |
+
"eval_loss": 1.8299671411514282,
|
| 1111 |
+
"eval_runtime": 107.7348,
|
| 1112 |
+
"eval_samples_per_second": 80.661,
|
| 1113 |
+
"eval_steps_per_second": 2.525,
|
| 1114 |
+
"step": 26000
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 66.70878079595704,
|
| 1118 |
+
"grad_norm": 0.00608784519135952,
|
| 1119 |
+
"learning_rate": 6.666666666666667e-05,
|
| 1120 |
+
"loss": 0.0086,
|
| 1121 |
+
"step": 26400
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 66.70878079595704,
|
| 1125 |
+
"eval_accuracy": 0.7968929804372842,
|
| 1126 |
+
"eval_f1_macro": 0.6234372893298947,
|
| 1127 |
+
"eval_loss": 1.9254202842712402,
|
| 1128 |
+
"eval_runtime": 108.035,
|
| 1129 |
+
"eval_samples_per_second": 80.437,
|
| 1130 |
+
"eval_steps_per_second": 2.518,
|
| 1131 |
+
"step": 26400
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 67.7195198989261,
|
| 1135 |
+
"grad_norm": 0.08328448981046677,
|
| 1136 |
+
"learning_rate": 5.925925925925926e-05,
|
| 1137 |
+
"loss": 0.0073,
|
| 1138 |
+
"step": 26800
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 67.7195198989261,
|
| 1142 |
+
"eval_accuracy": 0.7915995397008055,
|
| 1143 |
+
"eval_f1_macro": 0.6320923241131308,
|
| 1144 |
+
"eval_loss": 1.8887046575546265,
|
| 1145 |
+
"eval_runtime": 107.8399,
|
| 1146 |
+
"eval_samples_per_second": 80.582,
|
| 1147 |
+
"eval_steps_per_second": 2.522,
|
| 1148 |
+
"step": 26800
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 68.73025900189513,
|
| 1152 |
+
"grad_norm": 0.02061997540295124,
|
| 1153 |
+
"learning_rate": 5.185185185185185e-05,
|
| 1154 |
+
"loss": 0.0069,
|
| 1155 |
+
"step": 27200
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 68.73025900189513,
|
| 1159 |
+
"eval_accuracy": 0.794361334867664,
|
| 1160 |
+
"eval_f1_macro": 0.636665979654867,
|
| 1161 |
+
"eval_loss": 1.9074466228485107,
|
| 1162 |
+
"eval_runtime": 107.6829,
|
| 1163 |
+
"eval_samples_per_second": 80.7,
|
| 1164 |
+
"eval_steps_per_second": 2.526,
|
| 1165 |
+
"step": 27200
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 69.74099810486418,
|
| 1169 |
+
"grad_norm": 0.012987918220460415,
|
| 1170 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 1171 |
+
"loss": 0.0059,
|
| 1172 |
+
"step": 27600
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 69.74099810486418,
|
| 1176 |
+
"eval_accuracy": 0.792059838895282,
|
| 1177 |
+
"eval_f1_macro": 0.6315720450251525,
|
| 1178 |
+
"eval_loss": 1.9398057460784912,
|
| 1179 |
+
"eval_runtime": 107.8991,
|
| 1180 |
+
"eval_samples_per_second": 80.538,
|
| 1181 |
+
"eval_steps_per_second": 2.521,
|
| 1182 |
+
"step": 27600
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 70.75173720783323,
|
| 1186 |
+
"grad_norm": 0.005101632326841354,
|
| 1187 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 1188 |
+
"loss": 0.0066,
|
| 1189 |
+
"step": 28000
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 70.75173720783323,
|
| 1193 |
+
"eval_accuracy": 0.794361334867664,
|
| 1194 |
+
"eval_f1_macro": 0.6349818220797456,
|
| 1195 |
+
"eval_loss": 1.8699119091033936,
|
| 1196 |
+
"eval_runtime": 109.2809,
|
| 1197 |
+
"eval_samples_per_second": 79.52,
|
| 1198 |
+
"eval_steps_per_second": 2.489,
|
| 1199 |
+
"step": 28000
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 71.76247631080227,
|
| 1203 |
+
"grad_norm": 0.6047748923301697,
|
| 1204 |
+
"learning_rate": 2.962962962962963e-05,
|
| 1205 |
+
"loss": 0.0062,
|
| 1206 |
+
"step": 28400
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 71.76247631080227,
|
| 1210 |
+
"eval_accuracy": 0.7951668584579977,
|
| 1211 |
+
"eval_f1_macro": 0.6343250573277666,
|
| 1212 |
+
"eval_loss": 1.8893409967422485,
|
| 1213 |
+
"eval_runtime": 109.2978,
|
| 1214 |
+
"eval_samples_per_second": 79.508,
|
| 1215 |
+
"eval_steps_per_second": 2.489,
|
| 1216 |
+
"step": 28400
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 72.77321541377133,
|
| 1220 |
+
"grad_norm": 0.012553258799016476,
|
| 1221 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 1222 |
+
"loss": 0.0058,
|
| 1223 |
+
"step": 28800
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 72.77321541377133,
|
| 1227 |
+
"eval_accuracy": 0.7982738780207135,
|
| 1228 |
+
"eval_f1_macro": 0.6409643965446785,
|
| 1229 |
+
"eval_loss": 1.883091926574707,
|
| 1230 |
+
"eval_runtime": 109.2468,
|
| 1231 |
+
"eval_samples_per_second": 79.545,
|
| 1232 |
+
"eval_steps_per_second": 2.49,
|
| 1233 |
+
"step": 28800
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 73.78395451674037,
|
| 1237 |
+
"grad_norm": 0.0007793375989422202,
|
| 1238 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 1239 |
+
"loss": 0.0056,
|
| 1240 |
+
"step": 29200
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 73.78395451674037,
|
| 1244 |
+
"eval_accuracy": 0.7958573072497123,
|
| 1245 |
+
"eval_f1_macro": 0.6356613761441215,
|
| 1246 |
+
"eval_loss": 1.8901586532592773,
|
| 1247 |
+
"eval_runtime": 108.6154,
|
| 1248 |
+
"eval_samples_per_second": 80.007,
|
| 1249 |
+
"eval_steps_per_second": 2.504,
|
| 1250 |
+
"step": 29200
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 74.7946936197094,
|
| 1254 |
+
"grad_norm": 0.14352725446224213,
|
| 1255 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 1256 |
+
"loss": 0.0053,
|
| 1257 |
+
"step": 29600
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 74.7946936197094,
|
| 1261 |
+
"eval_accuracy": 0.7991944764096662,
|
| 1262 |
+
"eval_f1_macro": 0.643747242061282,
|
| 1263 |
+
"eval_loss": 1.888542890548706,
|
| 1264 |
+
"eval_runtime": 108.5316,
|
| 1265 |
+
"eval_samples_per_second": 80.069,
|
| 1266 |
+
"eval_steps_per_second": 2.506,
|
| 1267 |
+
"step": 29600
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 75.80543272267846,
|
| 1271 |
+
"grad_norm": 0.9781034588813782,
|
| 1272 |
+
"learning_rate": 0.0,
|
| 1273 |
+
"loss": 0.0046,
|
| 1274 |
+
"step": 30000
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 75.80543272267846,
|
| 1278 |
+
"eval_accuracy": 0.8005753739930955,
|
| 1279 |
+
"eval_f1_macro": 0.6435443913467072,
|
| 1280 |
+
"eval_loss": 1.888439655303955,
|
| 1281 |
+
"eval_runtime": 108.5256,
|
| 1282 |
+
"eval_samples_per_second": 80.073,
|
| 1283 |
+
"eval_steps_per_second": 2.506,
|
| 1284 |
+
"step": 30000
|
| 1285 |
+
}
|
| 1286 |
+
],
|
| 1287 |
+
"logging_steps": 400,
|
| 1288 |
+
"max_steps": 30000,
|
| 1289 |
+
"num_input_tokens_seen": 0,
|
| 1290 |
+
"num_train_epochs": 76,
|
| 1291 |
+
"save_steps": 1200,
|
| 1292 |
+
"stateful_callbacks": {
|
| 1293 |
+
"TrainerControl": {
|
| 1294 |
+
"args": {
|
| 1295 |
+
"should_epoch_stop": false,
|
| 1296 |
+
"should_evaluate": false,
|
| 1297 |
+
"should_log": false,
|
| 1298 |
+
"should_save": true,
|
| 1299 |
+
"should_training_stop": true
|
| 1300 |
+
},
|
| 1301 |
+
"attributes": {}
|
| 1302 |
+
}
|
| 1303 |
+
},
|
| 1304 |
+
"total_flos": 5.8164789316384843e+20,
|
| 1305 |
+
"train_batch_size": 32,
|
| 1306 |
+
"trial_name": null,
|
| 1307 |
+
"trial_params": null
|
| 1308 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93195b6ce37c327e37144d872765300fb413254de9caea3613dca78eed8e3139
|
| 3 |
+
size 5304
|