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 +1053 -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 sha256:8640b72dec981e1e9d15163a540aacbf60b0430996b248b29177d13b5f5ed6c2
|
| 3 |
+
size 378390784
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3de295f0becb4470dd3deb6c508572cb0800f7fc1a1668cce65ff8dff3adf2e
|
| 3 |
+
size 756909370
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9358e3f0a2f8d4ce5a568834d1a3a228d073c539937ef7f387714f29e59d9ab7
|
| 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 +1,1053 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8324182680309459,
|
| 3 |
+
"best_model_checkpoint": "results_retain/facebook/wav2vec2-base/42/checkpoint-30000",
|
| 4 |
+
"epoch": 69.20415224913495,
|
| 5 |
+
"eval_steps": 500,
|
| 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": 1.1534025374855825,
|
| 13 |
+
"grad_norm": 4.369956970214844,
|
| 14 |
+
"learning_rate": 8.333333333333333e-05,
|
| 15 |
+
"loss": 3.9309,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.1534025374855825,
|
| 20 |
+
"eval_accuracy": 0.15310706264037935,
|
| 21 |
+
"eval_f1_macro": 0.019883818836519816,
|
| 22 |
+
"eval_loss": 3.409226894378662,
|
| 23 |
+
"eval_runtime": 35.4301,
|
| 24 |
+
"eval_samples_per_second": 226.192,
|
| 25 |
+
"eval_steps_per_second": 7.084,
|
| 26 |
+
"step": 500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 2.306805074971165,
|
| 30 |
+
"grad_norm": 14.16882610321045,
|
| 31 |
+
"learning_rate": 0.00016666666666666666,
|
| 32 |
+
"loss": 2.5622,
|
| 33 |
+
"step": 1000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 2.306805074971165,
|
| 37 |
+
"eval_accuracy": 0.541177938607437,
|
| 38 |
+
"eval_f1_macro": 0.26987730386621905,
|
| 39 |
+
"eval_loss": 1.8982651233673096,
|
| 40 |
+
"eval_runtime": 35.0851,
|
| 41 |
+
"eval_samples_per_second": 228.416,
|
| 42 |
+
"eval_steps_per_second": 7.154,
|
| 43 |
+
"step": 1000
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.4602076124567476,
|
| 47 |
+
"grad_norm": 9.389366149902344,
|
| 48 |
+
"learning_rate": 0.00025,
|
| 49 |
+
"loss": 1.6435,
|
| 50 |
+
"step": 1500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.4602076124567476,
|
| 54 |
+
"eval_accuracy": 0.6169203893186923,
|
| 55 |
+
"eval_f1_macro": 0.37606688616805145,
|
| 56 |
+
"eval_loss": 1.6092511415481567,
|
| 57 |
+
"eval_runtime": 35.1816,
|
| 58 |
+
"eval_samples_per_second": 227.79,
|
| 59 |
+
"eval_steps_per_second": 7.134,
|
| 60 |
+
"step": 1500
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 4.61361014994233,
|
| 64 |
+
"grad_norm": 11.559647560119629,
|
| 65 |
+
"learning_rate": 0.0003333333333333333,
|
| 66 |
+
"loss": 1.4186,
|
| 67 |
+
"step": 2000
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 4.61361014994233,
|
| 71 |
+
"eval_accuracy": 0.6561018218118293,
|
| 72 |
+
"eval_f1_macro": 0.4408940890295375,
|
| 73 |
+
"eval_loss": 1.540195345878601,
|
| 74 |
+
"eval_runtime": 35.3339,
|
| 75 |
+
"eval_samples_per_second": 226.808,
|
| 76 |
+
"eval_steps_per_second": 7.104,
|
| 77 |
+
"step": 2000
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 5.767012687427912,
|
| 81 |
+
"grad_norm": 11.722562789916992,
|
| 82 |
+
"learning_rate": 0.0004166666666666667,
|
| 83 |
+
"loss": 1.3754,
|
| 84 |
+
"step": 2500
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 5.767012687427912,
|
| 88 |
+
"eval_accuracy": 0.6512353381582231,
|
| 89 |
+
"eval_f1_macro": 0.4529040361609232,
|
| 90 |
+
"eval_loss": 1.5076924562454224,
|
| 91 |
+
"eval_runtime": 35.3118,
|
| 92 |
+
"eval_samples_per_second": 226.95,
|
| 93 |
+
"eval_steps_per_second": 7.108,
|
| 94 |
+
"step": 2500
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 6.920415224913495,
|
| 98 |
+
"grad_norm": 10.16735553741455,
|
| 99 |
+
"learning_rate": 0.0005,
|
| 100 |
+
"loss": 1.3999,
|
| 101 |
+
"step": 3000
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 6.920415224913495,
|
| 105 |
+
"eval_accuracy": 0.6501123034689293,
|
| 106 |
+
"eval_f1_macro": 0.4559819632486863,
|
| 107 |
+
"eval_loss": 1.5159275531768799,
|
| 108 |
+
"eval_runtime": 35.4302,
|
| 109 |
+
"eval_samples_per_second": 226.191,
|
| 110 |
+
"eval_steps_per_second": 7.084,
|
| 111 |
+
"step": 3000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 8.073817762399077,
|
| 115 |
+
"grad_norm": 8.510236740112305,
|
| 116 |
+
"learning_rate": 0.0004907407407407408,
|
| 117 |
+
"loss": 1.3843,
|
| 118 |
+
"step": 3500
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 8.073817762399077,
|
| 122 |
+
"eval_accuracy": 0.6425006239081608,
|
| 123 |
+
"eval_f1_macro": 0.45380481877858314,
|
| 124 |
+
"eval_loss": 1.531354308128357,
|
| 125 |
+
"eval_runtime": 35.0783,
|
| 126 |
+
"eval_samples_per_second": 228.461,
|
| 127 |
+
"eval_steps_per_second": 7.155,
|
| 128 |
+
"step": 3500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 9.22722029988466,
|
| 132 |
+
"grad_norm": 6.111170768737793,
|
| 133 |
+
"learning_rate": 0.00048148148148148144,
|
| 134 |
+
"loss": 1.2362,
|
| 135 |
+
"step": 4000
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 9.22722029988466,
|
| 139 |
+
"eval_accuracy": 0.6679560768654854,
|
| 140 |
+
"eval_f1_macro": 0.49259444116250145,
|
| 141 |
+
"eval_loss": 1.4102381467819214,
|
| 142 |
+
"eval_runtime": 35.1989,
|
| 143 |
+
"eval_samples_per_second": 227.678,
|
| 144 |
+
"eval_steps_per_second": 7.131,
|
| 145 |
+
"step": 4000
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 10.380622837370241,
|
| 149 |
+
"grad_norm": 7.980649948120117,
|
| 150 |
+
"learning_rate": 0.00047222222222222224,
|
| 151 |
+
"loss": 1.111,
|
| 152 |
+
"step": 4500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 10.380622837370241,
|
| 156 |
+
"eval_accuracy": 0.6789368604941353,
|
| 157 |
+
"eval_f1_macro": 0.49662640502149236,
|
| 158 |
+
"eval_loss": 1.385581612586975,
|
| 159 |
+
"eval_runtime": 35.3766,
|
| 160 |
+
"eval_samples_per_second": 226.534,
|
| 161 |
+
"eval_steps_per_second": 7.095,
|
| 162 |
+
"step": 4500
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 11.534025374855824,
|
| 166 |
+
"grad_norm": 7.416294574737549,
|
| 167 |
+
"learning_rate": 0.000462962962962963,
|
| 168 |
+
"loss": 1.018,
|
| 169 |
+
"step": 5000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 11.534025374855824,
|
| 173 |
+
"eval_accuracy": 0.6690791115547792,
|
| 174 |
+
"eval_f1_macro": 0.4911372457550657,
|
| 175 |
+
"eval_loss": 1.451659917831421,
|
| 176 |
+
"eval_runtime": 35.349,
|
| 177 |
+
"eval_samples_per_second": 226.711,
|
| 178 |
+
"eval_steps_per_second": 7.101,
|
| 179 |
+
"step": 5000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 12.687427912341407,
|
| 183 |
+
"grad_norm": 6.453460216522217,
|
| 184 |
+
"learning_rate": 0.0004537037037037037,
|
| 185 |
+
"loss": 0.9373,
|
| 186 |
+
"step": 5500
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 12.687427912341407,
|
| 190 |
+
"eval_accuracy": 0.697903668579985,
|
| 191 |
+
"eval_f1_macro": 0.5213519190207141,
|
| 192 |
+
"eval_loss": 1.3074634075164795,
|
| 193 |
+
"eval_runtime": 35.3636,
|
| 194 |
+
"eval_samples_per_second": 226.617,
|
| 195 |
+
"eval_steps_per_second": 7.098,
|
| 196 |
+
"step": 5500
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 13.84083044982699,
|
| 200 |
+
"grad_norm": 5.483336925506592,
|
| 201 |
+
"learning_rate": 0.0004444444444444444,
|
| 202 |
+
"loss": 0.8447,
|
| 203 |
+
"step": 6000
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 13.84083044982699,
|
| 207 |
+
"eval_accuracy": 0.7037684052907413,
|
| 208 |
+
"eval_f1_macro": 0.5403759594538143,
|
| 209 |
+
"eval_loss": 1.2658469676971436,
|
| 210 |
+
"eval_runtime": 35.3101,
|
| 211 |
+
"eval_samples_per_second": 226.961,
|
| 212 |
+
"eval_steps_per_second": 7.108,
|
| 213 |
+
"step": 6000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 14.994232987312571,
|
| 217 |
+
"grad_norm": 8.091373443603516,
|
| 218 |
+
"learning_rate": 0.0004351851851851852,
|
| 219 |
+
"loss": 0.7729,
|
| 220 |
+
"step": 6500
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 14.994232987312571,
|
| 224 |
+
"eval_accuracy": 0.7111305215872223,
|
| 225 |
+
"eval_f1_macro": 0.5506675562830741,
|
| 226 |
+
"eval_loss": 1.2573933601379395,
|
| 227 |
+
"eval_runtime": 35.2856,
|
| 228 |
+
"eval_samples_per_second": 227.118,
|
| 229 |
+
"eval_steps_per_second": 7.113,
|
| 230 |
+
"step": 6500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 16.147635524798154,
|
| 234 |
+
"grad_norm": 7.302002906799316,
|
| 235 |
+
"learning_rate": 0.00042592592592592595,
|
| 236 |
+
"loss": 0.6963,
|
| 237 |
+
"step": 7000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 16.147635524798154,
|
| 241 |
+
"eval_accuracy": 0.7262290990766159,
|
| 242 |
+
"eval_f1_macro": 0.5559240421705101,
|
| 243 |
+
"eval_loss": 1.2266888618469238,
|
| 244 |
+
"eval_runtime": 36.5656,
|
| 245 |
+
"eval_samples_per_second": 219.168,
|
| 246 |
+
"eval_steps_per_second": 6.864,
|
| 247 |
+
"step": 7000
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 17.301038062283737,
|
| 251 |
+
"grad_norm": 6.176177501678467,
|
| 252 |
+
"learning_rate": 0.0004166666666666667,
|
| 253 |
+
"loss": 0.6499,
|
| 254 |
+
"step": 7500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 17.301038062283737,
|
| 258 |
+
"eval_accuracy": 0.7266034439730472,
|
| 259 |
+
"eval_f1_macro": 0.549456325002494,
|
| 260 |
+
"eval_loss": 1.245578646659851,
|
| 261 |
+
"eval_runtime": 35.3365,
|
| 262 |
+
"eval_samples_per_second": 226.791,
|
| 263 |
+
"eval_steps_per_second": 7.103,
|
| 264 |
+
"step": 7500
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 18.45444059976932,
|
| 268 |
+
"grad_norm": 6.318857192993164,
|
| 269 |
+
"learning_rate": 0.0004074074074074074,
|
| 270 |
+
"loss": 0.5983,
|
| 271 |
+
"step": 8000
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 18.45444059976932,
|
| 275 |
+
"eval_accuracy": 0.7292238582480659,
|
| 276 |
+
"eval_f1_macro": 0.575232140696358,
|
| 277 |
+
"eval_loss": 1.2900265455245972,
|
| 278 |
+
"eval_runtime": 35.5024,
|
| 279 |
+
"eval_samples_per_second": 225.731,
|
| 280 |
+
"eval_steps_per_second": 7.07,
|
| 281 |
+
"step": 8000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 19.607843137254903,
|
| 285 |
+
"grad_norm": 10.460654258728027,
|
| 286 |
+
"learning_rate": 0.0003981481481481481,
|
| 287 |
+
"loss": 0.5503,
|
| 288 |
+
"step": 8500
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 19.607843137254903,
|
| 292 |
+
"eval_accuracy": 0.7196156725729973,
|
| 293 |
+
"eval_f1_macro": 0.5556217816058379,
|
| 294 |
+
"eval_loss": 1.3130799531936646,
|
| 295 |
+
"eval_runtime": 35.1865,
|
| 296 |
+
"eval_samples_per_second": 227.758,
|
| 297 |
+
"eval_steps_per_second": 7.133,
|
| 298 |
+
"step": 8500
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 20.761245674740483,
|
| 302 |
+
"grad_norm": 8.248889923095703,
|
| 303 |
+
"learning_rate": 0.0003888888888888889,
|
| 304 |
+
"loss": 0.511,
|
| 305 |
+
"step": 9000
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 20.761245674740483,
|
| 309 |
+
"eval_accuracy": 0.7455702520588969,
|
| 310 |
+
"eval_f1_macro": 0.5823341220355842,
|
| 311 |
+
"eval_loss": 1.2670824527740479,
|
| 312 |
+
"eval_runtime": 35.4588,
|
| 313 |
+
"eval_samples_per_second": 226.009,
|
| 314 |
+
"eval_steps_per_second": 7.079,
|
| 315 |
+
"step": 9000
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 21.914648212226066,
|
| 319 |
+
"grad_norm": 5.647006988525391,
|
| 320 |
+
"learning_rate": 0.00037962962962962966,
|
| 321 |
+
"loss": 0.4782,
|
| 322 |
+
"step": 9500
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 21.914648212226066,
|
| 326 |
+
"eval_accuracy": 0.7422011479910158,
|
| 327 |
+
"eval_f1_macro": 0.5962421465197172,
|
| 328 |
+
"eval_loss": 1.3255208730697632,
|
| 329 |
+
"eval_runtime": 35.2623,
|
| 330 |
+
"eval_samples_per_second": 227.268,
|
| 331 |
+
"eval_steps_per_second": 7.118,
|
| 332 |
+
"step": 9500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 23.06805074971165,
|
| 336 |
+
"grad_norm": 4.740267753601074,
|
| 337 |
+
"learning_rate": 0.00037037037037037035,
|
| 338 |
+
"loss": 0.4381,
|
| 339 |
+
"step": 10000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 23.06805074971165,
|
| 343 |
+
"eval_accuracy": 0.7499376091839282,
|
| 344 |
+
"eval_f1_macro": 0.5833773217449505,
|
| 345 |
+
"eval_loss": 1.1890977621078491,
|
| 346 |
+
"eval_runtime": 35.3036,
|
| 347 |
+
"eval_samples_per_second": 227.002,
|
| 348 |
+
"eval_steps_per_second": 7.11,
|
| 349 |
+
"step": 10000
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 24.22145328719723,
|
| 353 |
+
"grad_norm": 6.188416481018066,
|
| 354 |
+
"learning_rate": 0.0003611111111111111,
|
| 355 |
+
"loss": 0.3939,
|
| 356 |
+
"step": 10500
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 24.22145328719723,
|
| 360 |
+
"eval_accuracy": 0.7560519091589718,
|
| 361 |
+
"eval_f1_macro": 0.5989834099897938,
|
| 362 |
+
"eval_loss": 1.2394686937332153,
|
| 363 |
+
"eval_runtime": 35.2099,
|
| 364 |
+
"eval_samples_per_second": 227.607,
|
| 365 |
+
"eval_steps_per_second": 7.129,
|
| 366 |
+
"step": 10500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 25.374855824682815,
|
| 370 |
+
"grad_norm": 6.43159294128418,
|
| 371 |
+
"learning_rate": 0.0003518518518518519,
|
| 372 |
+
"loss": 0.3749,
|
| 373 |
+
"step": 11000
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 25.374855824682815,
|
| 377 |
+
"eval_accuracy": 0.7571749438482656,
|
| 378 |
+
"eval_f1_macro": 0.5924283240267194,
|
| 379 |
+
"eval_loss": 1.2047547101974487,
|
| 380 |
+
"eval_runtime": 35.0834,
|
| 381 |
+
"eval_samples_per_second": 228.427,
|
| 382 |
+
"eval_steps_per_second": 7.154,
|
| 383 |
+
"step": 11000
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 26.528258362168398,
|
| 387 |
+
"grad_norm": 5.618147850036621,
|
| 388 |
+
"learning_rate": 0.00034259259259259263,
|
| 389 |
+
"loss": 0.3433,
|
| 390 |
+
"step": 11500
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 26.528258362168398,
|
| 394 |
+
"eval_accuracy": 0.7448215622660345,
|
| 395 |
+
"eval_f1_macro": 0.5914660771680106,
|
| 396 |
+
"eval_loss": 1.2851983308792114,
|
| 397 |
+
"eval_runtime": 34.8189,
|
| 398 |
+
"eval_samples_per_second": 230.162,
|
| 399 |
+
"eval_steps_per_second": 7.209,
|
| 400 |
+
"step": 11500
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 27.68166089965398,
|
| 404 |
+
"grad_norm": 5.279151439666748,
|
| 405 |
+
"learning_rate": 0.0003333333333333333,
|
| 406 |
+
"loss": 0.3211,
|
| 407 |
+
"step": 12000
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 27.68166089965398,
|
| 411 |
+
"eval_accuracy": 0.7535562765160968,
|
| 412 |
+
"eval_f1_macro": 0.6027492509194426,
|
| 413 |
+
"eval_loss": 1.268044114112854,
|
| 414 |
+
"eval_runtime": 34.8176,
|
| 415 |
+
"eval_samples_per_second": 230.171,
|
| 416 |
+
"eval_steps_per_second": 7.209,
|
| 417 |
+
"step": 12000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 28.83506343713956,
|
| 421 |
+
"grad_norm": 4.73464298248291,
|
| 422 |
+
"learning_rate": 0.00032407407407407406,
|
| 423 |
+
"loss": 0.3006,
|
| 424 |
+
"step": 12500
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 28.83506343713956,
|
| 428 |
+
"eval_accuracy": 0.761417519341153,
|
| 429 |
+
"eval_f1_macro": 0.6132345857728708,
|
| 430 |
+
"eval_loss": 1.2127602100372314,
|
| 431 |
+
"eval_runtime": 34.7848,
|
| 432 |
+
"eval_samples_per_second": 230.388,
|
| 433 |
+
"eval_steps_per_second": 7.216,
|
| 434 |
+
"step": 12500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 29.988465974625143,
|
| 438 |
+
"grad_norm": 5.955222129821777,
|
| 439 |
+
"learning_rate": 0.0003148148148148148,
|
| 440 |
+
"loss": 0.2799,
|
| 441 |
+
"step": 13000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 29.988465974625143,
|
| 445 |
+
"eval_accuracy": 0.7632892438233092,
|
| 446 |
+
"eval_f1_macro": 0.6125965461682634,
|
| 447 |
+
"eval_loss": 1.2151060104370117,
|
| 448 |
+
"eval_runtime": 34.8242,
|
| 449 |
+
"eval_samples_per_second": 230.128,
|
| 450 |
+
"eval_steps_per_second": 7.208,
|
| 451 |
+
"step": 13000
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 31.141868512110726,
|
| 455 |
+
"grad_norm": 8.98561954498291,
|
| 456 |
+
"learning_rate": 0.0003055555555555556,
|
| 457 |
+
"loss": 0.2657,
|
| 458 |
+
"step": 13500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 31.141868512110726,
|
| 462 |
+
"eval_accuracy": 0.7513102071375094,
|
| 463 |
+
"eval_f1_macro": 0.6058015188813233,
|
| 464 |
+
"eval_loss": 1.3033212423324585,
|
| 465 |
+
"eval_runtime": 34.82,
|
| 466 |
+
"eval_samples_per_second": 230.155,
|
| 467 |
+
"eval_steps_per_second": 7.208,
|
| 468 |
+
"step": 13500
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 32.29527104959631,
|
| 472 |
+
"grad_norm": 5.852510929107666,
|
| 473 |
+
"learning_rate": 0.0002962962962962963,
|
| 474 |
+
"loss": 0.2341,
|
| 475 |
+
"step": 14000
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 32.29527104959631,
|
| 479 |
+
"eval_accuracy": 0.7716496131769404,
|
| 480 |
+
"eval_f1_macro": 0.6079228678531986,
|
| 481 |
+
"eval_loss": 1.2467527389526367,
|
| 482 |
+
"eval_runtime": 34.7926,
|
| 483 |
+
"eval_samples_per_second": 230.336,
|
| 484 |
+
"eval_steps_per_second": 7.214,
|
| 485 |
+
"step": 14000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 33.44867358708189,
|
| 489 |
+
"grad_norm": 3.7392752170562744,
|
| 490 |
+
"learning_rate": 0.00028703703703703703,
|
| 491 |
+
"loss": 0.2268,
|
| 492 |
+
"step": 14500
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 33.44867358708189,
|
| 496 |
+
"eval_accuracy": 0.7672822560519091,
|
| 497 |
+
"eval_f1_macro": 0.6200186960995067,
|
| 498 |
+
"eval_loss": 1.173102855682373,
|
| 499 |
+
"eval_runtime": 34.7778,
|
| 500 |
+
"eval_samples_per_second": 230.434,
|
| 501 |
+
"eval_steps_per_second": 7.217,
|
| 502 |
+
"step": 14500
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 34.602076124567475,
|
| 506 |
+
"grad_norm": 3.8082404136657715,
|
| 507 |
+
"learning_rate": 0.0002777777777777778,
|
| 508 |
+
"loss": 0.2138,
|
| 509 |
+
"step": 15000
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 34.602076124567475,
|
| 513 |
+
"eval_accuracy": 0.7682805091090591,
|
| 514 |
+
"eval_f1_macro": 0.6089872489834017,
|
| 515 |
+
"eval_loss": 1.2769969701766968,
|
| 516 |
+
"eval_runtime": 34.8228,
|
| 517 |
+
"eval_samples_per_second": 230.137,
|
| 518 |
+
"eval_steps_per_second": 7.208,
|
| 519 |
+
"step": 15000
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 35.75547866205306,
|
| 523 |
+
"grad_norm": 4.899130821228027,
|
| 524 |
+
"learning_rate": 0.0002685185185185186,
|
| 525 |
+
"loss": 0.2019,
|
| 526 |
+
"step": 15500
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 35.75547866205306,
|
| 530 |
+
"eval_accuracy": 0.7803843274270027,
|
| 531 |
+
"eval_f1_macro": 0.6410356899857027,
|
| 532 |
+
"eval_loss": 1.182688593864441,
|
| 533 |
+
"eval_runtime": 34.7724,
|
| 534 |
+
"eval_samples_per_second": 230.47,
|
| 535 |
+
"eval_steps_per_second": 7.218,
|
| 536 |
+
"step": 15500
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 36.90888119953864,
|
| 540 |
+
"grad_norm": 4.297879219055176,
|
| 541 |
+
"learning_rate": 0.00025925925925925926,
|
| 542 |
+
"loss": 0.1832,
|
| 543 |
+
"step": 16000
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 36.90888119953864,
|
| 547 |
+
"eval_accuracy": 0.7696531070626403,
|
| 548 |
+
"eval_f1_macro": 0.628787412236927,
|
| 549 |
+
"eval_loss": 1.2591631412506104,
|
| 550 |
+
"eval_runtime": 34.7726,
|
| 551 |
+
"eval_samples_per_second": 230.469,
|
| 552 |
+
"eval_steps_per_second": 7.218,
|
| 553 |
+
"step": 16000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 38.062283737024224,
|
| 557 |
+
"grad_norm": 4.6692214012146,
|
| 558 |
+
"learning_rate": 0.00025,
|
| 559 |
+
"loss": 0.1737,
|
| 560 |
+
"step": 16500
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 38.062283737024224,
|
| 564 |
+
"eval_accuracy": 0.7709009233840779,
|
| 565 |
+
"eval_f1_macro": 0.6107952760204466,
|
| 566 |
+
"eval_loss": 1.2620958089828491,
|
| 567 |
+
"eval_runtime": 36.0409,
|
| 568 |
+
"eval_samples_per_second": 222.358,
|
| 569 |
+
"eval_steps_per_second": 6.964,
|
| 570 |
+
"step": 16500
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 39.21568627450981,
|
| 574 |
+
"grad_norm": 3.8761215209960938,
|
| 575 |
+
"learning_rate": 0.00024074074074074072,
|
| 576 |
+
"loss": 0.1574,
|
| 577 |
+
"step": 17000
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 39.21568627450981,
|
| 581 |
+
"eval_accuracy": 0.7748939356126778,
|
| 582 |
+
"eval_f1_macro": 0.6391500982858909,
|
| 583 |
+
"eval_loss": 1.3058395385742188,
|
| 584 |
+
"eval_runtime": 36.1321,
|
| 585 |
+
"eval_samples_per_second": 221.797,
|
| 586 |
+
"eval_steps_per_second": 6.947,
|
| 587 |
+
"step": 17000
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 40.36908881199539,
|
| 591 |
+
"grad_norm": 3.7799384593963623,
|
| 592 |
+
"learning_rate": 0.0002314814814814815,
|
| 593 |
+
"loss": 0.1494,
|
| 594 |
+
"step": 17500
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 40.36908881199539,
|
| 598 |
+
"eval_accuracy": 0.7775143498876965,
|
| 599 |
+
"eval_f1_macro": 0.634463633421762,
|
| 600 |
+
"eval_loss": 1.2541757822036743,
|
| 601 |
+
"eval_runtime": 36.0992,
|
| 602 |
+
"eval_samples_per_second": 222.0,
|
| 603 |
+
"eval_steps_per_second": 6.953,
|
| 604 |
+
"step": 17500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 41.522491349480966,
|
| 608 |
+
"grad_norm": 3.6442158222198486,
|
| 609 |
+
"learning_rate": 0.0002222222222222222,
|
| 610 |
+
"loss": 0.1428,
|
| 611 |
+
"step": 18000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 41.522491349480966,
|
| 615 |
+
"eval_accuracy": 0.7815073621162965,
|
| 616 |
+
"eval_f1_macro": 0.6354469092694762,
|
| 617 |
+
"eval_loss": 1.2337605953216553,
|
| 618 |
+
"eval_runtime": 36.0255,
|
| 619 |
+
"eval_samples_per_second": 222.453,
|
| 620 |
+
"eval_steps_per_second": 6.967,
|
| 621 |
+
"step": 18000
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 42.67589388696655,
|
| 625 |
+
"grad_norm": 3.6234309673309326,
|
| 626 |
+
"learning_rate": 0.00021296296296296298,
|
| 627 |
+
"loss": 0.1273,
|
| 628 |
+
"step": 18500
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 42.67589388696655,
|
| 632 |
+
"eval_accuracy": 0.7836286498627402,
|
| 633 |
+
"eval_f1_macro": 0.6407864496886617,
|
| 634 |
+
"eval_loss": 1.2769566774368286,
|
| 635 |
+
"eval_runtime": 35.6646,
|
| 636 |
+
"eval_samples_per_second": 224.705,
|
| 637 |
+
"eval_steps_per_second": 7.038,
|
| 638 |
+
"step": 18500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 43.82929642445213,
|
| 642 |
+
"grad_norm": 2.8652567863464355,
|
| 643 |
+
"learning_rate": 0.0002037037037037037,
|
| 644 |
+
"loss": 0.1181,
|
| 645 |
+
"step": 19000
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 43.82929642445213,
|
| 649 |
+
"eval_accuracy": 0.7831295233341652,
|
| 650 |
+
"eval_f1_macro": 0.6496093521953449,
|
| 651 |
+
"eval_loss": 1.3619425296783447,
|
| 652 |
+
"eval_runtime": 36.0707,
|
| 653 |
+
"eval_samples_per_second": 222.175,
|
| 654 |
+
"eval_steps_per_second": 6.959,
|
| 655 |
+
"step": 19000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 44.982698961937714,
|
| 659 |
+
"grad_norm": 5.1563191413879395,
|
| 660 |
+
"learning_rate": 0.00019444444444444446,
|
| 661 |
+
"loss": 0.1139,
|
| 662 |
+
"step": 19500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 44.982698961937714,
|
| 666 |
+
"eval_accuracy": 0.7891190416770651,
|
| 667 |
+
"eval_f1_macro": 0.639353026374818,
|
| 668 |
+
"eval_loss": 1.231618881225586,
|
| 669 |
+
"eval_runtime": 35.9657,
|
| 670 |
+
"eval_samples_per_second": 222.823,
|
| 671 |
+
"eval_steps_per_second": 6.979,
|
| 672 |
+
"step": 19500
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 46.1361014994233,
|
| 676 |
+
"grad_norm": 1.5734127759933472,
|
| 677 |
+
"learning_rate": 0.00018518518518518518,
|
| 678 |
+
"loss": 0.1036,
|
| 679 |
+
"step": 20000
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 46.1361014994233,
|
| 683 |
+
"eval_accuracy": 0.7962315947092587,
|
| 684 |
+
"eval_f1_macro": 0.6474614824910541,
|
| 685 |
+
"eval_loss": 1.2083463668823242,
|
| 686 |
+
"eval_runtime": 35.5967,
|
| 687 |
+
"eval_samples_per_second": 225.133,
|
| 688 |
+
"eval_steps_per_second": 7.051,
|
| 689 |
+
"step": 20000
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 47.28950403690888,
|
| 693 |
+
"grad_norm": 0.9659342765808105,
|
| 694 |
+
"learning_rate": 0.00017592592592592595,
|
| 695 |
+
"loss": 0.0984,
|
| 696 |
+
"step": 20500
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 47.28950403690888,
|
| 700 |
+
"eval_accuracy": 0.7961068130771151,
|
| 701 |
+
"eval_f1_macro": 0.6731347859037599,
|
| 702 |
+
"eval_loss": 1.2197343111038208,
|
| 703 |
+
"eval_runtime": 40.3877,
|
| 704 |
+
"eval_samples_per_second": 198.427,
|
| 705 |
+
"eval_steps_per_second": 6.215,
|
| 706 |
+
"step": 20500
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 48.44290657439446,
|
| 710 |
+
"grad_norm": 3.5627315044403076,
|
| 711 |
+
"learning_rate": 0.00016666666666666666,
|
| 712 |
+
"loss": 0.0949,
|
| 713 |
+
"step": 21000
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 48.44290657439446,
|
| 717 |
+
"eval_accuracy": 0.7923633641128026,
|
| 718 |
+
"eval_f1_macro": 0.6493884652975611,
|
| 719 |
+
"eval_loss": 1.255406379699707,
|
| 720 |
+
"eval_runtime": 37.0454,
|
| 721 |
+
"eval_samples_per_second": 216.329,
|
| 722 |
+
"eval_steps_per_second": 6.775,
|
| 723 |
+
"step": 21000
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 49.596309111880046,
|
| 727 |
+
"grad_norm": 2.2700695991516113,
|
| 728 |
+
"learning_rate": 0.0001574074074074074,
|
| 729 |
+
"loss": 0.0822,
|
| 730 |
+
"step": 21500
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 49.596309111880046,
|
| 734 |
+
"eval_accuracy": 0.7922385824806588,
|
| 735 |
+
"eval_f1_macro": 0.6543829429647029,
|
| 736 |
+
"eval_loss": 1.2859455347061157,
|
| 737 |
+
"eval_runtime": 36.8794,
|
| 738 |
+
"eval_samples_per_second": 217.303,
|
| 739 |
+
"eval_steps_per_second": 6.806,
|
| 740 |
+
"step": 21500
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 50.74971164936563,
|
| 744 |
+
"grad_norm": 2.4785401821136475,
|
| 745 |
+
"learning_rate": 0.00014814814814814815,
|
| 746 |
+
"loss": 0.0794,
|
| 747 |
+
"step": 22000
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 50.74971164936563,
|
| 751 |
+
"eval_accuracy": 0.7938607436985275,
|
| 752 |
+
"eval_f1_macro": 0.6632807007483084,
|
| 753 |
+
"eval_loss": 1.3257797956466675,
|
| 754 |
+
"eval_runtime": 65.9338,
|
| 755 |
+
"eval_samples_per_second": 121.546,
|
| 756 |
+
"eval_steps_per_second": 3.807,
|
| 757 |
+
"step": 22000
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 51.90311418685121,
|
| 761 |
+
"grad_norm": 5.8285675048828125,
|
| 762 |
+
"learning_rate": 0.0001388888888888889,
|
| 763 |
+
"loss": 0.071,
|
| 764 |
+
"step": 22500
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 51.90311418685121,
|
| 768 |
+
"eval_accuracy": 0.7877464437234839,
|
| 769 |
+
"eval_f1_macro": 0.6514161143660198,
|
| 770 |
+
"eval_loss": 1.3142781257629395,
|
| 771 |
+
"eval_runtime": 63.1498,
|
| 772 |
+
"eval_samples_per_second": 126.905,
|
| 773 |
+
"eval_steps_per_second": 3.975,
|
| 774 |
+
"step": 22500
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 53.056516724336795,
|
| 778 |
+
"grad_norm": 2.3869199752807617,
|
| 779 |
+
"learning_rate": 0.00012962962962962963,
|
| 780 |
+
"loss": 0.0632,
|
| 781 |
+
"step": 23000
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 53.056516724336795,
|
| 785 |
+
"eval_accuracy": 0.7967307212378338,
|
| 786 |
+
"eval_f1_macro": 0.6650145750564478,
|
| 787 |
+
"eval_loss": 1.2448372840881348,
|
| 788 |
+
"eval_runtime": 34.9343,
|
| 789 |
+
"eval_samples_per_second": 229.402,
|
| 790 |
+
"eval_steps_per_second": 7.185,
|
| 791 |
+
"step": 23000
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 54.20991926182238,
|
| 795 |
+
"grad_norm": 1.7126274108886719,
|
| 796 |
+
"learning_rate": 0.00012037037037037036,
|
| 797 |
+
"loss": 0.0592,
|
| 798 |
+
"step": 23500
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 54.20991926182238,
|
| 802 |
+
"eval_accuracy": 0.8083354130272024,
|
| 803 |
+
"eval_f1_macro": 0.653147976938688,
|
| 804 |
+
"eval_loss": 1.273723840713501,
|
| 805 |
+
"eval_runtime": 35.1779,
|
| 806 |
+
"eval_samples_per_second": 227.813,
|
| 807 |
+
"eval_steps_per_second": 7.135,
|
| 808 |
+
"step": 23500
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 55.36332179930796,
|
| 812 |
+
"grad_norm": 1.4396089315414429,
|
| 813 |
+
"learning_rate": 0.0001111111111111111,
|
| 814 |
+
"loss": 0.0551,
|
| 815 |
+
"step": 24000
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 55.36332179930796,
|
| 819 |
+
"eval_accuracy": 0.8022211130521587,
|
| 820 |
+
"eval_f1_macro": 0.6674628853192668,
|
| 821 |
+
"eval_loss": 1.289405107498169,
|
| 822 |
+
"eval_runtime": 35.37,
|
| 823 |
+
"eval_samples_per_second": 226.576,
|
| 824 |
+
"eval_steps_per_second": 7.096,
|
| 825 |
+
"step": 24000
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 56.516724336793544,
|
| 829 |
+
"grad_norm": 2.4105160236358643,
|
| 830 |
+
"learning_rate": 0.00010185185185185185,
|
| 831 |
+
"loss": 0.0495,
|
| 832 |
+
"step": 24500
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 56.516724336793544,
|
| 836 |
+
"eval_accuracy": 0.803094584477165,
|
| 837 |
+
"eval_f1_macro": 0.6638655285331525,
|
| 838 |
+
"eval_loss": 1.241363525390625,
|
| 839 |
+
"eval_runtime": 35.1653,
|
| 840 |
+
"eval_samples_per_second": 227.895,
|
| 841 |
+
"eval_steps_per_second": 7.138,
|
| 842 |
+
"step": 24500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 57.67012687427912,
|
| 846 |
+
"grad_norm": 3.4061028957366943,
|
| 847 |
+
"learning_rate": 9.259259259259259e-05,
|
| 848 |
+
"loss": 0.0449,
|
| 849 |
+
"step": 25000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 57.67012687427912,
|
| 853 |
+
"eval_accuracy": 0.8077115048664837,
|
| 854 |
+
"eval_f1_macro": 0.6679687949981502,
|
| 855 |
+
"eval_loss": 1.2535569667816162,
|
| 856 |
+
"eval_runtime": 35.1863,
|
| 857 |
+
"eval_samples_per_second": 227.759,
|
| 858 |
+
"eval_steps_per_second": 7.133,
|
| 859 |
+
"step": 25000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 58.8235294117647,
|
| 863 |
+
"grad_norm": 4.6967363357543945,
|
| 864 |
+
"learning_rate": 8.333333333333333e-05,
|
| 865 |
+
"loss": 0.0412,
|
| 866 |
+
"step": 25500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 58.8235294117647,
|
| 870 |
+
"eval_accuracy": 0.8135762415772398,
|
| 871 |
+
"eval_f1_macro": 0.6707472142083073,
|
| 872 |
+
"eval_loss": 1.2510894536972046,
|
| 873 |
+
"eval_runtime": 35.1709,
|
| 874 |
+
"eval_samples_per_second": 227.859,
|
| 875 |
+
"eval_steps_per_second": 7.137,
|
| 876 |
+
"step": 25500
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 59.976931949250286,
|
| 880 |
+
"grad_norm": 1.2903478145599365,
|
| 881 |
+
"learning_rate": 7.407407407407407e-05,
|
| 882 |
+
"loss": 0.0343,
|
| 883 |
+
"step": 26000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 59.976931949250286,
|
| 887 |
+
"eval_accuracy": 0.8100823558772149,
|
| 888 |
+
"eval_f1_macro": 0.6867023078289406,
|
| 889 |
+
"eval_loss": 1.326628565788269,
|
| 890 |
+
"eval_runtime": 35.2142,
|
| 891 |
+
"eval_samples_per_second": 227.579,
|
| 892 |
+
"eval_steps_per_second": 7.128,
|
| 893 |
+
"step": 26000
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 61.13033448673587,
|
| 897 |
+
"grad_norm": 4.929038047790527,
|
| 898 |
+
"learning_rate": 6.481481481481482e-05,
|
| 899 |
+
"loss": 0.0325,
|
| 900 |
+
"step": 26500
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 61.13033448673587,
|
| 904 |
+
"eval_accuracy": 0.8146992762665336,
|
| 905 |
+
"eval_f1_macro": 0.6813981931163788,
|
| 906 |
+
"eval_loss": 1.2899118661880493,
|
| 907 |
+
"eval_runtime": 35.3037,
|
| 908 |
+
"eval_samples_per_second": 227.002,
|
| 909 |
+
"eval_steps_per_second": 7.11,
|
| 910 |
+
"step": 26500
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 62.28373702422145,
|
| 914 |
+
"grad_norm": 0.04143017157912254,
|
| 915 |
+
"learning_rate": 5.555555555555555e-05,
|
| 916 |
+
"loss": 0.0278,
|
| 917 |
+
"step": 27000
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 62.28373702422145,
|
| 921 |
+
"eval_accuracy": 0.818941851759421,
|
| 922 |
+
"eval_f1_macro": 0.6660686006005503,
|
| 923 |
+
"eval_loss": 1.2699165344238281,
|
| 924 |
+
"eval_runtime": 35.3107,
|
| 925 |
+
"eval_samples_per_second": 226.957,
|
| 926 |
+
"eval_steps_per_second": 7.108,
|
| 927 |
+
"step": 27000
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 63.437139561707035,
|
| 931 |
+
"grad_norm": 1.8004887104034424,
|
| 932 |
+
"learning_rate": 4.6296296296296294e-05,
|
| 933 |
+
"loss": 0.0265,
|
| 934 |
+
"step": 27500
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 63.437139561707035,
|
| 938 |
+
"eval_accuracy": 0.8170701272772648,
|
| 939 |
+
"eval_f1_macro": 0.6702275226911916,
|
| 940 |
+
"eval_loss": 1.2782015800476074,
|
| 941 |
+
"eval_runtime": 35.0186,
|
| 942 |
+
"eval_samples_per_second": 228.85,
|
| 943 |
+
"eval_steps_per_second": 7.168,
|
| 944 |
+
"step": 27500
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 64.59054209919262,
|
| 948 |
+
"grad_norm": 0.07229655981063843,
|
| 949 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 950 |
+
"loss": 0.0224,
|
| 951 |
+
"step": 28000
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 64.59054209919262,
|
| 955 |
+
"eval_accuracy": 0.8225605190915897,
|
| 956 |
+
"eval_f1_macro": 0.6808030174100609,
|
| 957 |
+
"eval_loss": 1.2255030870437622,
|
| 958 |
+
"eval_runtime": 34.8632,
|
| 959 |
+
"eval_samples_per_second": 229.87,
|
| 960 |
+
"eval_steps_per_second": 7.2,
|
| 961 |
+
"step": 28000
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 65.7439446366782,
|
| 965 |
+
"grad_norm": 1.0921183824539185,
|
| 966 |
+
"learning_rate": 2.7777777777777776e-05,
|
| 967 |
+
"loss": 0.0182,
|
| 968 |
+
"step": 28500
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 65.7439446366782,
|
| 972 |
+
"eval_accuracy": 0.8276765660094834,
|
| 973 |
+
"eval_f1_macro": 0.6838139982176796,
|
| 974 |
+
"eval_loss": 1.2417439222335815,
|
| 975 |
+
"eval_runtime": 34.79,
|
| 976 |
+
"eval_samples_per_second": 230.353,
|
| 977 |
+
"eval_steps_per_second": 7.215,
|
| 978 |
+
"step": 28500
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 66.89734717416378,
|
| 982 |
+
"grad_norm": 0.2541353404521942,
|
| 983 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 984 |
+
"loss": 0.0189,
|
| 985 |
+
"step": 29000
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 66.89734717416378,
|
| 989 |
+
"eval_accuracy": 0.8255552782630396,
|
| 990 |
+
"eval_f1_macro": 0.6846466970174251,
|
| 991 |
+
"eval_loss": 1.1945445537567139,
|
| 992 |
+
"eval_runtime": 35.2088,
|
| 993 |
+
"eval_samples_per_second": 227.613,
|
| 994 |
+
"eval_steps_per_second": 7.129,
|
| 995 |
+
"step": 29000
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 68.05074971164936,
|
| 999 |
+
"grad_norm": 0.49906083941459656,
|
| 1000 |
+
"learning_rate": 9.259259259259259e-06,
|
| 1001 |
+
"loss": 0.0173,
|
| 1002 |
+
"step": 29500
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 68.05074971164936,
|
| 1006 |
+
"eval_accuracy": 0.8311704517095083,
|
| 1007 |
+
"eval_f1_macro": 0.6982812455994792,
|
| 1008 |
+
"eval_loss": 1.1982783079147339,
|
| 1009 |
+
"eval_runtime": 35.2862,
|
| 1010 |
+
"eval_samples_per_second": 227.114,
|
| 1011 |
+
"eval_steps_per_second": 7.113,
|
| 1012 |
+
"step": 29500
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 69.20415224913495,
|
| 1016 |
+
"grad_norm": 0.035026829689741135,
|
| 1017 |
+
"learning_rate": 0.0,
|
| 1018 |
+
"loss": 0.0139,
|
| 1019 |
+
"step": 30000
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 69.20415224913495,
|
| 1023 |
+
"eval_accuracy": 0.8324182680309459,
|
| 1024 |
+
"eval_f1_macro": 0.7059020357212862,
|
| 1025 |
+
"eval_loss": 1.1972322463989258,
|
| 1026 |
+
"eval_runtime": 35.1374,
|
| 1027 |
+
"eval_samples_per_second": 228.076,
|
| 1028 |
+
"eval_steps_per_second": 7.143,
|
| 1029 |
+
"step": 30000
|
| 1030 |
+
}
|
| 1031 |
+
],
|
| 1032 |
+
"logging_steps": 500,
|
| 1033 |
+
"max_steps": 30000,
|
| 1034 |
+
"num_input_tokens_seen": 0,
|
| 1035 |
+
"num_train_epochs": 70,
|
| 1036 |
+
"save_steps": 1000,
|
| 1037 |
+
"stateful_callbacks": {
|
| 1038 |
+
"TrainerControl": {
|
| 1039 |
+
"args": {
|
| 1040 |
+
"should_epoch_stop": false,
|
| 1041 |
+
"should_evaluate": false,
|
| 1042 |
+
"should_log": false,
|
| 1043 |
+
"should_save": true,
|
| 1044 |
+
"should_training_stop": true
|
| 1045 |
+
},
|
| 1046 |
+
"attributes": {}
|
| 1047 |
+
}
|
| 1048 |
+
},
|
| 1049 |
+
"total_flos": 1.7427529644770302e+20,
|
| 1050 |
+
"train_batch_size": 32,
|
| 1051 |
+
"trial_name": null,
|
| 1052 |
+
"trial_params": null
|
| 1053 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b5da9191fb9de5530dc9456b1064bef7072460bea45a07406ee1fc5467299c2
|
| 3 |
+
size 5240
|