Upload folder using huggingface_hub
Browse files- config.json +232 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- trainer_state.json +670 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,232 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-xls-r-300m",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2ForSequenceClassification"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.1,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 256,
|
| 14 |
+
"codevector_dim": 768,
|
| 15 |
+
"contrastive_logits_temperature": 0.1,
|
| 16 |
+
"conv_bias": true,
|
| 17 |
+
"conv_dim": [
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512
|
| 25 |
+
],
|
| 26 |
+
"conv_kernel": [
|
| 27 |
+
10,
|
| 28 |
+
3,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
2,
|
| 33 |
+
2
|
| 34 |
+
],
|
| 35 |
+
"conv_stride": [
|
| 36 |
+
5,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2
|
| 43 |
+
],
|
| 44 |
+
"ctc_loss_reduction": "sum",
|
| 45 |
+
"ctc_zero_infinity": false,
|
| 46 |
+
"diversity_loss_weight": 0.1,
|
| 47 |
+
"do_stable_layer_norm": true,
|
| 48 |
+
"eos_token_id": 2,
|
| 49 |
+
"feat_extract_activation": "gelu",
|
| 50 |
+
"feat_extract_dropout": 0.0,
|
| 51 |
+
"feat_extract_norm": "layer",
|
| 52 |
+
"feat_proj_dropout": 0.1,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"gradient_checkpointing": false,
|
| 56 |
+
"hidden_act": "gelu",
|
| 57 |
+
"hidden_dropout": 0.1,
|
| 58 |
+
"hidden_size": 1024,
|
| 59 |
+
"id2label": {
|
| 60 |
+
"0": "iot_wemo_off",
|
| 61 |
+
"1": "qa_maths",
|
| 62 |
+
"10": "alarm_remove",
|
| 63 |
+
"11": "qa_stock",
|
| 64 |
+
"12": "email_sendemail",
|
| 65 |
+
"13": "music_settings",
|
| 66 |
+
"14": "transport_taxi",
|
| 67 |
+
"15": "iot_hue_lightoff",
|
| 68 |
+
"16": "email_querycontact",
|
| 69 |
+
"17": "audio_volume_other",
|
| 70 |
+
"18": "play_audiobook",
|
| 71 |
+
"19": "calendar_query",
|
| 72 |
+
"2": "recommendation_movies",
|
| 73 |
+
"20": "iot_wemo_on",
|
| 74 |
+
"21": "datetime_convert",
|
| 75 |
+
"22": "music_likeness",
|
| 76 |
+
"23": "cooking_query",
|
| 77 |
+
"24": "play_podcasts",
|
| 78 |
+
"25": "lists_createoradd",
|
| 79 |
+
"26": "lists_query",
|
| 80 |
+
"27": "iot_cleaning",
|
| 81 |
+
"28": "cooking_recipe",
|
| 82 |
+
"29": "audio_volume_mute",
|
| 83 |
+
"3": "audio_volume_up",
|
| 84 |
+
"30": "qa_currency",
|
| 85 |
+
"31": "email_query",
|
| 86 |
+
"32": "qa_definition",
|
| 87 |
+
"33": "iot_hue_lighton",
|
| 88 |
+
"34": "calendar_remove",
|
| 89 |
+
"35": "iot_coffee",
|
| 90 |
+
"36": "calendar_set",
|
| 91 |
+
"37": "general_joke",
|
| 92 |
+
"38": "takeaway_order",
|
| 93 |
+
"39": "iot_hue_lightchange",
|
| 94 |
+
"4": "general_greet",
|
| 95 |
+
"40": "recommendation_events",
|
| 96 |
+
"41": "iot_hue_lightdim",
|
| 97 |
+
"42": "datetime_query",
|
| 98 |
+
"43": "music_dislikeness",
|
| 99 |
+
"44": "transport_traffic",
|
| 100 |
+
"45": "general_quirky",
|
| 101 |
+
"46": "qa_factoid",
|
| 102 |
+
"47": "alarm_query",
|
| 103 |
+
"48": "weather_query",
|
| 104 |
+
"49": "social_query",
|
| 105 |
+
"5": "iot_hue_lightup",
|
| 106 |
+
"50": "transport_query",
|
| 107 |
+
"51": "play_music",
|
| 108 |
+
"52": "news_query",
|
| 109 |
+
"53": "social_post",
|
| 110 |
+
"54": "lists_remove",
|
| 111 |
+
"55": "play_game",
|
| 112 |
+
"56": "alarm_set",
|
| 113 |
+
"57": "transport_ticket",
|
| 114 |
+
"58": "takeaway_query",
|
| 115 |
+
"59": "music_query",
|
| 116 |
+
"6": "email_addcontact",
|
| 117 |
+
"7": "play_radio",
|
| 118 |
+
"8": "audio_volume_down",
|
| 119 |
+
"9": "recommendation_locations"
|
| 120 |
+
},
|
| 121 |
+
"initializer_range": 0.02,
|
| 122 |
+
"intermediate_size": 4096,
|
| 123 |
+
"label2id": {
|
| 124 |
+
"alarm_query": "47",
|
| 125 |
+
"alarm_remove": "10",
|
| 126 |
+
"alarm_set": "56",
|
| 127 |
+
"audio_volume_down": "8",
|
| 128 |
+
"audio_volume_mute": "29",
|
| 129 |
+
"audio_volume_other": "17",
|
| 130 |
+
"audio_volume_up": "3",
|
| 131 |
+
"calendar_query": "19",
|
| 132 |
+
"calendar_remove": "34",
|
| 133 |
+
"calendar_set": "36",
|
| 134 |
+
"cooking_query": "23",
|
| 135 |
+
"cooking_recipe": "28",
|
| 136 |
+
"datetime_convert": "21",
|
| 137 |
+
"datetime_query": "42",
|
| 138 |
+
"email_addcontact": "6",
|
| 139 |
+
"email_query": "31",
|
| 140 |
+
"email_querycontact": "16",
|
| 141 |
+
"email_sendemail": "12",
|
| 142 |
+
"general_greet": "4",
|
| 143 |
+
"general_joke": "37",
|
| 144 |
+
"general_quirky": "45",
|
| 145 |
+
"iot_cleaning": "27",
|
| 146 |
+
"iot_coffee": "35",
|
| 147 |
+
"iot_hue_lightchange": "39",
|
| 148 |
+
"iot_hue_lightdim": "41",
|
| 149 |
+
"iot_hue_lightoff": "15",
|
| 150 |
+
"iot_hue_lighton": "33",
|
| 151 |
+
"iot_hue_lightup": "5",
|
| 152 |
+
"iot_wemo_off": "0",
|
| 153 |
+
"iot_wemo_on": "20",
|
| 154 |
+
"lists_createoradd": "25",
|
| 155 |
+
"lists_query": "26",
|
| 156 |
+
"lists_remove": "54",
|
| 157 |
+
"music_dislikeness": "43",
|
| 158 |
+
"music_likeness": "22",
|
| 159 |
+
"music_query": "59",
|
| 160 |
+
"music_settings": "13",
|
| 161 |
+
"news_query": "52",
|
| 162 |
+
"play_audiobook": "18",
|
| 163 |
+
"play_game": "55",
|
| 164 |
+
"play_music": "51",
|
| 165 |
+
"play_podcasts": "24",
|
| 166 |
+
"play_radio": "7",
|
| 167 |
+
"qa_currency": "30",
|
| 168 |
+
"qa_definition": "32",
|
| 169 |
+
"qa_factoid": "46",
|
| 170 |
+
"qa_maths": "1",
|
| 171 |
+
"qa_stock": "11",
|
| 172 |
+
"recommendation_events": "40",
|
| 173 |
+
"recommendation_locations": "9",
|
| 174 |
+
"recommendation_movies": "2",
|
| 175 |
+
"social_post": "53",
|
| 176 |
+
"social_query": "49",
|
| 177 |
+
"takeaway_order": "38",
|
| 178 |
+
"takeaway_query": "58",
|
| 179 |
+
"transport_query": "50",
|
| 180 |
+
"transport_taxi": "14",
|
| 181 |
+
"transport_ticket": "57",
|
| 182 |
+
"transport_traffic": "44",
|
| 183 |
+
"weather_query": "48"
|
| 184 |
+
},
|
| 185 |
+
"layer_norm_eps": 1e-05,
|
| 186 |
+
"layerdrop": 0.1,
|
| 187 |
+
"mask_feature_length": 10,
|
| 188 |
+
"mask_feature_min_masks": 0,
|
| 189 |
+
"mask_feature_prob": 0.0,
|
| 190 |
+
"mask_time_length": 10,
|
| 191 |
+
"mask_time_min_masks": 2,
|
| 192 |
+
"mask_time_prob": 0.075,
|
| 193 |
+
"model_type": "wav2vec2",
|
| 194 |
+
"num_adapter_layers": 3,
|
| 195 |
+
"num_attention_heads": 16,
|
| 196 |
+
"num_codevector_groups": 2,
|
| 197 |
+
"num_codevectors_per_group": 320,
|
| 198 |
+
"num_conv_pos_embedding_groups": 16,
|
| 199 |
+
"num_conv_pos_embeddings": 128,
|
| 200 |
+
"num_feat_extract_layers": 7,
|
| 201 |
+
"num_hidden_layers": 24,
|
| 202 |
+
"num_negatives": 100,
|
| 203 |
+
"output_hidden_size": 1024,
|
| 204 |
+
"pad_token_id": 0,
|
| 205 |
+
"proj_codevector_dim": 768,
|
| 206 |
+
"tdnn_dilation": [
|
| 207 |
+
1,
|
| 208 |
+
2,
|
| 209 |
+
3,
|
| 210 |
+
1,
|
| 211 |
+
1
|
| 212 |
+
],
|
| 213 |
+
"tdnn_dim": [
|
| 214 |
+
512,
|
| 215 |
+
512,
|
| 216 |
+
512,
|
| 217 |
+
512,
|
| 218 |
+
1500
|
| 219 |
+
],
|
| 220 |
+
"tdnn_kernel": [
|
| 221 |
+
5,
|
| 222 |
+
3,
|
| 223 |
+
3,
|
| 224 |
+
1,
|
| 225 |
+
1
|
| 226 |
+
],
|
| 227 |
+
"torch_dtype": "float32",
|
| 228 |
+
"transformers_version": "4.26.0.dev0",
|
| 229 |
+
"use_weighted_layer_sum": false,
|
| 230 |
+
"vocab_size": 32,
|
| 231 |
+
"xvector_output_dim": 512
|
| 232 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5b0528a58a33b3e7f9f625a688f3120b1b219ee9cfbdf5ef998a6ee0069a93d3
|
| 3 |
+
size 2525980249
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2001393616260a2451f54ce8eed437122326c6efcd9eccabb0a896fface9316f
|
| 3 |
+
size 1263010605
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1cb8ac73d9152bed148974deef5c1ab92db946fbb30d44e16d7dc628aea3849d
|
| 3 |
+
size 14567
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:093366c8f46dc68efc47501492eaa2b2aac2f9563b6eb039428d08ced56a2d70
|
| 3 |
+
size 559
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:82e44ac207bea40c1c5d082a1a24bbc7b8e0a079cb4ee6a13288dd13795b56b0
|
| 3 |
+
size 623
|
trainer_state.json
ADDED
|
@@ -0,0 +1,670 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8040629095674967,
|
| 3 |
+
"best_model_checkpoint": "wav2vec2-xls-r-300m-ic-finetuning/checkpoint-2675",
|
| 4 |
+
"epoch": 29.995348837209303,
|
| 5 |
+
"global_step": 3210,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.47,
|
| 12 |
+
"learning_rate": 1.5264797507788163e-05,
|
| 13 |
+
"loss": 4.0833,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.93,
|
| 18 |
+
"learning_rate": 3.08411214953271e-05,
|
| 19 |
+
"loss": 4.013,
|
| 20 |
+
"step": 100
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 1.0,
|
| 24 |
+
"eval_accuracy": 0.06946264744429882,
|
| 25 |
+
"eval_loss": 3.8936545848846436,
|
| 26 |
+
"eval_runtime": 23.3839,
|
| 27 |
+
"eval_samples_per_second": 65.259,
|
| 28 |
+
"eval_steps_per_second": 2.053,
|
| 29 |
+
"step": 107
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.4,
|
| 33 |
+
"learning_rate": 4.641744548286604e-05,
|
| 34 |
+
"loss": 3.8855,
|
| 35 |
+
"step": 150
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 1.87,
|
| 39 |
+
"learning_rate": 6.199376947040498e-05,
|
| 40 |
+
"loss": 3.7509,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 2.0,
|
| 45 |
+
"eval_accuracy": 0.06946264744429882,
|
| 46 |
+
"eval_loss": 3.778992176055908,
|
| 47 |
+
"eval_runtime": 23.2928,
|
| 48 |
+
"eval_samples_per_second": 65.514,
|
| 49 |
+
"eval_steps_per_second": 2.061,
|
| 50 |
+
"step": 214
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 2.33,
|
| 54 |
+
"learning_rate": 7.757009345794393e-05,
|
| 55 |
+
"loss": 3.7888,
|
| 56 |
+
"step": 250
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.8,
|
| 60 |
+
"learning_rate": 9.314641744548286e-05,
|
| 61 |
+
"loss": 3.7477,
|
| 62 |
+
"step": 300
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 3.0,
|
| 66 |
+
"eval_accuracy": 0.06422018348623854,
|
| 67 |
+
"eval_loss": 3.760488748550415,
|
| 68 |
+
"eval_runtime": 23.2905,
|
| 69 |
+
"eval_samples_per_second": 65.52,
|
| 70 |
+
"eval_steps_per_second": 2.061,
|
| 71 |
+
"step": 321
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 3.27,
|
| 75 |
+
"learning_rate": 9.903080650744203e-05,
|
| 76 |
+
"loss": 3.7551,
|
| 77 |
+
"step": 350
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 3.73,
|
| 81 |
+
"learning_rate": 9.730010384215992e-05,
|
| 82 |
+
"loss": 3.7013,
|
| 83 |
+
"step": 400
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 4.0,
|
| 87 |
+
"eval_accuracy": 0.10288335517693316,
|
| 88 |
+
"eval_loss": 3.6565122604370117,
|
| 89 |
+
"eval_runtime": 23.3565,
|
| 90 |
+
"eval_samples_per_second": 65.335,
|
| 91 |
+
"eval_steps_per_second": 2.055,
|
| 92 |
+
"step": 428
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 4.2,
|
| 96 |
+
"learning_rate": 9.556940117687782e-05,
|
| 97 |
+
"loss": 3.7122,
|
| 98 |
+
"step": 450
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 4.67,
|
| 102 |
+
"learning_rate": 9.383869851159571e-05,
|
| 103 |
+
"loss": 3.5953,
|
| 104 |
+
"step": 500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 5.0,
|
| 108 |
+
"eval_accuracy": 0.12647444298820446,
|
| 109 |
+
"eval_loss": 3.4294865131378174,
|
| 110 |
+
"eval_runtime": 23.3639,
|
| 111 |
+
"eval_samples_per_second": 65.315,
|
| 112 |
+
"eval_steps_per_second": 2.054,
|
| 113 |
+
"step": 535
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 5.14,
|
| 117 |
+
"learning_rate": 9.21079958463136e-05,
|
| 118 |
+
"loss": 3.5112,
|
| 119 |
+
"step": 550
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 5.6,
|
| 123 |
+
"learning_rate": 9.041190723433715e-05,
|
| 124 |
+
"loss": 3.35,
|
| 125 |
+
"step": 600
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 6.0,
|
| 129 |
+
"eval_accuracy": 0.19855832241153343,
|
| 130 |
+
"eval_loss": 3.083688974380493,
|
| 131 |
+
"eval_runtime": 23.4304,
|
| 132 |
+
"eval_samples_per_second": 65.129,
|
| 133 |
+
"eval_steps_per_second": 2.049,
|
| 134 |
+
"step": 642
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 6.07,
|
| 138 |
+
"learning_rate": 8.868120456905504e-05,
|
| 139 |
+
"loss": 3.2423,
|
| 140 |
+
"step": 650
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 6.54,
|
| 144 |
+
"learning_rate": 8.695050190377294e-05,
|
| 145 |
+
"loss": 2.9948,
|
| 146 |
+
"step": 700
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 7.0,
|
| 150 |
+
"eval_accuracy": 0.32044560943643513,
|
| 151 |
+
"eval_loss": 2.5826685428619385,
|
| 152 |
+
"eval_runtime": 23.3418,
|
| 153 |
+
"eval_samples_per_second": 65.376,
|
| 154 |
+
"eval_steps_per_second": 2.056,
|
| 155 |
+
"step": 749
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 7.01,
|
| 159 |
+
"learning_rate": 8.521979923849083e-05,
|
| 160 |
+
"loss": 2.8101,
|
| 161 |
+
"step": 750
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 7.47,
|
| 165 |
+
"learning_rate": 8.352371062651437e-05,
|
| 166 |
+
"loss": 2.5137,
|
| 167 |
+
"step": 800
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 7.94,
|
| 171 |
+
"learning_rate": 8.179300796123225e-05,
|
| 172 |
+
"loss": 2.3776,
|
| 173 |
+
"step": 850
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 8.0,
|
| 177 |
+
"eval_accuracy": 0.4619921363040629,
|
| 178 |
+
"eval_loss": 2.1507654190063477,
|
| 179 |
+
"eval_runtime": 23.378,
|
| 180 |
+
"eval_samples_per_second": 65.275,
|
| 181 |
+
"eval_steps_per_second": 2.053,
|
| 182 |
+
"step": 856
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 8.41,
|
| 186 |
+
"learning_rate": 8.006230529595016e-05,
|
| 187 |
+
"loss": 2.1621,
|
| 188 |
+
"step": 900
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 8.87,
|
| 192 |
+
"learning_rate": 7.833160263066806e-05,
|
| 193 |
+
"loss": 2.0058,
|
| 194 |
+
"step": 950
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 9.0,
|
| 198 |
+
"eval_accuracy": 0.528178243774574,
|
| 199 |
+
"eval_loss": 1.857922911643982,
|
| 200 |
+
"eval_runtime": 23.3965,
|
| 201 |
+
"eval_samples_per_second": 65.223,
|
| 202 |
+
"eval_steps_per_second": 2.052,
|
| 203 |
+
"step": 963
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 9.34,
|
| 207 |
+
"learning_rate": 7.660089996538595e-05,
|
| 208 |
+
"loss": 1.8365,
|
| 209 |
+
"step": 1000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 9.81,
|
| 213 |
+
"learning_rate": 7.487019730010385e-05,
|
| 214 |
+
"loss": 1.683,
|
| 215 |
+
"step": 1050
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 10.0,
|
| 219 |
+
"eval_accuracy": 0.6133682830930537,
|
| 220 |
+
"eval_loss": 1.5978834629058838,
|
| 221 |
+
"eval_runtime": 23.2913,
|
| 222 |
+
"eval_samples_per_second": 65.518,
|
| 223 |
+
"eval_steps_per_second": 2.061,
|
| 224 |
+
"step": 1070
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 10.28,
|
| 228 |
+
"learning_rate": 7.313949463482174e-05,
|
| 229 |
+
"loss": 1.5977,
|
| 230 |
+
"step": 1100
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 10.74,
|
| 234 |
+
"learning_rate": 7.140879196953964e-05,
|
| 235 |
+
"loss": 1.4444,
|
| 236 |
+
"step": 1150
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 11.0,
|
| 240 |
+
"eval_accuracy": 0.663826998689384,
|
| 241 |
+
"eval_loss": 1.4753053188323975,
|
| 242 |
+
"eval_runtime": 23.3735,
|
| 243 |
+
"eval_samples_per_second": 65.288,
|
| 244 |
+
"eval_steps_per_second": 2.054,
|
| 245 |
+
"step": 1177
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 11.21,
|
| 249 |
+
"learning_rate": 6.967808930425753e-05,
|
| 250 |
+
"loss": 1.3652,
|
| 251 |
+
"step": 1200
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 11.68,
|
| 255 |
+
"learning_rate": 6.794738663897543e-05,
|
| 256 |
+
"loss": 1.232,
|
| 257 |
+
"step": 1250
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 12.0,
|
| 261 |
+
"eval_accuracy": 0.6933158584534731,
|
| 262 |
+
"eval_loss": 1.316873550415039,
|
| 263 |
+
"eval_runtime": 23.4493,
|
| 264 |
+
"eval_samples_per_second": 65.077,
|
| 265 |
+
"eval_steps_per_second": 2.047,
|
| 266 |
+
"step": 1284
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 12.15,
|
| 270 |
+
"learning_rate": 6.621668397369333e-05,
|
| 271 |
+
"loss": 1.2344,
|
| 272 |
+
"step": 1300
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 12.61,
|
| 276 |
+
"learning_rate": 6.448598130841122e-05,
|
| 277 |
+
"loss": 1.0976,
|
| 278 |
+
"step": 1350
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 13.0,
|
| 282 |
+
"eval_accuracy": 0.7319790301441678,
|
| 283 |
+
"eval_loss": 1.229066252708435,
|
| 284 |
+
"eval_runtime": 23.3919,
|
| 285 |
+
"eval_samples_per_second": 65.236,
|
| 286 |
+
"eval_steps_per_second": 2.052,
|
| 287 |
+
"step": 1391
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 13.08,
|
| 291 |
+
"learning_rate": 6.275527864312911e-05,
|
| 292 |
+
"loss": 1.0488,
|
| 293 |
+
"step": 1400
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 13.55,
|
| 297 |
+
"learning_rate": 6.102457597784701e-05,
|
| 298 |
+
"loss": 0.9466,
|
| 299 |
+
"step": 1450
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 14.0,
|
| 303 |
+
"eval_accuracy": 0.744429882044561,
|
| 304 |
+
"eval_loss": 1.184418797492981,
|
| 305 |
+
"eval_runtime": 23.4293,
|
| 306 |
+
"eval_samples_per_second": 65.132,
|
| 307 |
+
"eval_steps_per_second": 2.049,
|
| 308 |
+
"step": 1498
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 14.02,
|
| 312 |
+
"learning_rate": 5.9293873312564906e-05,
|
| 313 |
+
"loss": 0.9305,
|
| 314 |
+
"step": 1500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 14.48,
|
| 318 |
+
"learning_rate": 5.756317064728281e-05,
|
| 319 |
+
"loss": 0.833,
|
| 320 |
+
"step": 1550
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 14.95,
|
| 324 |
+
"learning_rate": 5.583246798200069e-05,
|
| 325 |
+
"loss": 0.839,
|
| 326 |
+
"step": 1600
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 15.0,
|
| 330 |
+
"eval_accuracy": 0.755570117955439,
|
| 331 |
+
"eval_loss": 1.1289950609207153,
|
| 332 |
+
"eval_runtime": 23.407,
|
| 333 |
+
"eval_samples_per_second": 65.194,
|
| 334 |
+
"eval_steps_per_second": 2.051,
|
| 335 |
+
"step": 1605
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 15.42,
|
| 339 |
+
"learning_rate": 5.410176531671859e-05,
|
| 340 |
+
"loss": 0.7683,
|
| 341 |
+
"step": 1650
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 15.88,
|
| 345 |
+
"learning_rate": 5.2371062651436485e-05,
|
| 346 |
+
"loss": 0.6801,
|
| 347 |
+
"step": 1700
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 16.0,
|
| 351 |
+
"eval_accuracy": 0.7693315858453473,
|
| 352 |
+
"eval_loss": 1.125131368637085,
|
| 353 |
+
"eval_runtime": 23.6242,
|
| 354 |
+
"eval_samples_per_second": 64.595,
|
| 355 |
+
"eval_steps_per_second": 2.032,
|
| 356 |
+
"step": 1712
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 16.35,
|
| 360 |
+
"learning_rate": 5.064035998615439e-05,
|
| 361 |
+
"loss": 0.6717,
|
| 362 |
+
"step": 1750
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 16.82,
|
| 366 |
+
"learning_rate": 4.8909657320872275e-05,
|
| 367 |
+
"loss": 0.6781,
|
| 368 |
+
"step": 1800
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 17.0,
|
| 372 |
+
"eval_accuracy": 0.7745740498034076,
|
| 373 |
+
"eval_loss": 1.1027562618255615,
|
| 374 |
+
"eval_runtime": 23.455,
|
| 375 |
+
"eval_samples_per_second": 65.061,
|
| 376 |
+
"eval_steps_per_second": 2.046,
|
| 377 |
+
"step": 1819
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 17.29,
|
| 381 |
+
"learning_rate": 4.7178954655590177e-05,
|
| 382 |
+
"loss": 0.6023,
|
| 383 |
+
"step": 1850
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 17.75,
|
| 387 |
+
"learning_rate": 4.5448251990308065e-05,
|
| 388 |
+
"loss": 0.5996,
|
| 389 |
+
"step": 1900
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 18.0,
|
| 393 |
+
"eval_accuracy": 0.7804718217562254,
|
| 394 |
+
"eval_loss": 1.0976942777633667,
|
| 395 |
+
"eval_runtime": 23.4288,
|
| 396 |
+
"eval_samples_per_second": 65.134,
|
| 397 |
+
"eval_steps_per_second": 2.049,
|
| 398 |
+
"step": 1926
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 18.22,
|
| 402 |
+
"learning_rate": 4.3717549325025966e-05,
|
| 403 |
+
"loss": 0.5499,
|
| 404 |
+
"step": 1950
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 18.69,
|
| 408 |
+
"learning_rate": 4.1986846659743854e-05,
|
| 409 |
+
"loss": 0.5165,
|
| 410 |
+
"step": 2000
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 19.0,
|
| 414 |
+
"eval_accuracy": 0.783748361730013,
|
| 415 |
+
"eval_loss": 1.117361068725586,
|
| 416 |
+
"eval_runtime": 23.3249,
|
| 417 |
+
"eval_samples_per_second": 65.424,
|
| 418 |
+
"eval_steps_per_second": 2.058,
|
| 419 |
+
"step": 2033
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 19.16,
|
| 423 |
+
"learning_rate": 4.0256143994461756e-05,
|
| 424 |
+
"loss": 0.5191,
|
| 425 |
+
"step": 2050
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 19.62,
|
| 429 |
+
"learning_rate": 3.852544132917965e-05,
|
| 430 |
+
"loss": 0.4555,
|
| 431 |
+
"step": 2100
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 20.0,
|
| 435 |
+
"eval_accuracy": 0.7883355176933159,
|
| 436 |
+
"eval_loss": 1.0993103981018066,
|
| 437 |
+
"eval_runtime": 23.3454,
|
| 438 |
+
"eval_samples_per_second": 65.366,
|
| 439 |
+
"eval_steps_per_second": 2.056,
|
| 440 |
+
"step": 2140
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 20.09,
|
| 444 |
+
"learning_rate": 3.6794738663897545e-05,
|
| 445 |
+
"loss": 0.4704,
|
| 446 |
+
"step": 2150
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 20.56,
|
| 450 |
+
"learning_rate": 3.506403599861544e-05,
|
| 451 |
+
"loss": 0.4406,
|
| 452 |
+
"step": 2200
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 21.0,
|
| 456 |
+
"eval_accuracy": 0.7942332896461337,
|
| 457 |
+
"eval_loss": 1.0940536260604858,
|
| 458 |
+
"eval_runtime": 23.3681,
|
| 459 |
+
"eval_samples_per_second": 65.303,
|
| 460 |
+
"eval_steps_per_second": 2.054,
|
| 461 |
+
"step": 2247
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 21.03,
|
| 465 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 466 |
+
"loss": 0.4113,
|
| 467 |
+
"step": 2250
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 21.49,
|
| 471 |
+
"learning_rate": 3.160263066805123e-05,
|
| 472 |
+
"loss": 0.3725,
|
| 473 |
+
"step": 2300
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 21.96,
|
| 477 |
+
"learning_rate": 2.9871928002769128e-05,
|
| 478 |
+
"loss": 0.3822,
|
| 479 |
+
"step": 2350
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 22.0,
|
| 483 |
+
"eval_accuracy": 0.7889908256880734,
|
| 484 |
+
"eval_loss": 1.0881223678588867,
|
| 485 |
+
"eval_runtime": 23.2822,
|
| 486 |
+
"eval_samples_per_second": 65.544,
|
| 487 |
+
"eval_steps_per_second": 2.062,
|
| 488 |
+
"step": 2354
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 22.43,
|
| 492 |
+
"learning_rate": 2.814122533748702e-05,
|
| 493 |
+
"loss": 0.3608,
|
| 494 |
+
"step": 2400
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 22.89,
|
| 498 |
+
"learning_rate": 2.6410522672204917e-05,
|
| 499 |
+
"loss": 0.3348,
|
| 500 |
+
"step": 2450
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 23.0,
|
| 504 |
+
"eval_accuracy": 0.8027522935779816,
|
| 505 |
+
"eval_loss": 1.1017186641693115,
|
| 506 |
+
"eval_runtime": 23.2012,
|
| 507 |
+
"eval_samples_per_second": 65.773,
|
| 508 |
+
"eval_steps_per_second": 2.069,
|
| 509 |
+
"step": 2461
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 23.36,
|
| 513 |
+
"learning_rate": 2.4679820006922812e-05,
|
| 514 |
+
"loss": 0.3675,
|
| 515 |
+
"step": 2500
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 23.83,
|
| 519 |
+
"learning_rate": 2.2949117341640707e-05,
|
| 520 |
+
"loss": 0.297,
|
| 521 |
+
"step": 2550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 24.0,
|
| 525 |
+
"eval_accuracy": 0.7981651376146789,
|
| 526 |
+
"eval_loss": 1.0910543203353882,
|
| 527 |
+
"eval_runtime": 23.2808,
|
| 528 |
+
"eval_samples_per_second": 65.547,
|
| 529 |
+
"eval_steps_per_second": 2.062,
|
| 530 |
+
"step": 2568
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 24.3,
|
| 534 |
+
"learning_rate": 2.1218414676358602e-05,
|
| 535 |
+
"loss": 0.2985,
|
| 536 |
+
"step": 2600
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 24.76,
|
| 540 |
+
"learning_rate": 1.9487712011076497e-05,
|
| 541 |
+
"loss": 0.2707,
|
| 542 |
+
"step": 2650
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 25.0,
|
| 546 |
+
"eval_accuracy": 0.8040629095674967,
|
| 547 |
+
"eval_loss": 1.1111068725585938,
|
| 548 |
+
"eval_runtime": 23.5628,
|
| 549 |
+
"eval_samples_per_second": 64.763,
|
| 550 |
+
"eval_steps_per_second": 2.037,
|
| 551 |
+
"step": 2675
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 25.23,
|
| 555 |
+
"learning_rate": 1.775700934579439e-05,
|
| 556 |
+
"loss": 0.2748,
|
| 557 |
+
"step": 2700
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 25.7,
|
| 561 |
+
"learning_rate": 1.6026306680512286e-05,
|
| 562 |
+
"loss": 0.2834,
|
| 563 |
+
"step": 2750
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 26.0,
|
| 567 |
+
"eval_accuracy": 0.7975098296199213,
|
| 568 |
+
"eval_loss": 1.1192971467971802,
|
| 569 |
+
"eval_runtime": 23.4535,
|
| 570 |
+
"eval_samples_per_second": 65.065,
|
| 571 |
+
"eval_steps_per_second": 2.047,
|
| 572 |
+
"step": 2782
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 26.17,
|
| 576 |
+
"learning_rate": 1.4295604015230184e-05,
|
| 577 |
+
"loss": 0.2589,
|
| 578 |
+
"step": 2800
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 26.63,
|
| 582 |
+
"learning_rate": 1.256490134994808e-05,
|
| 583 |
+
"loss": 0.2662,
|
| 584 |
+
"step": 2850
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 27.0,
|
| 588 |
+
"eval_accuracy": 0.8034076015727392,
|
| 589 |
+
"eval_loss": 1.1238927841186523,
|
| 590 |
+
"eval_runtime": 23.4418,
|
| 591 |
+
"eval_samples_per_second": 65.097,
|
| 592 |
+
"eval_steps_per_second": 2.048,
|
| 593 |
+
"step": 2889
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 27.1,
|
| 597 |
+
"learning_rate": 1.0834198684665976e-05,
|
| 598 |
+
"loss": 0.2647,
|
| 599 |
+
"step": 2900
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 27.57,
|
| 603 |
+
"learning_rate": 9.10349601938387e-06,
|
| 604 |
+
"loss": 0.2495,
|
| 605 |
+
"step": 2950
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 28.0,
|
| 609 |
+
"eval_accuracy": 0.7961992136304062,
|
| 610 |
+
"eval_loss": 1.1401643753051758,
|
| 611 |
+
"eval_runtime": 23.4674,
|
| 612 |
+
"eval_samples_per_second": 65.026,
|
| 613 |
+
"eval_steps_per_second": 2.045,
|
| 614 |
+
"step": 2996
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 28.04,
|
| 618 |
+
"learning_rate": 7.372793354101766e-06,
|
| 619 |
+
"loss": 0.2434,
|
| 620 |
+
"step": 3000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 28.5,
|
| 624 |
+
"learning_rate": 5.642090688819661e-06,
|
| 625 |
+
"loss": 0.2345,
|
| 626 |
+
"step": 3050
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 28.97,
|
| 630 |
+
"learning_rate": 3.9113880235375566e-06,
|
| 631 |
+
"loss": 0.2467,
|
| 632 |
+
"step": 3100
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 29.0,
|
| 636 |
+
"eval_accuracy": 0.7981651376146789,
|
| 637 |
+
"eval_loss": 1.1323330402374268,
|
| 638 |
+
"eval_runtime": 23.4948,
|
| 639 |
+
"eval_samples_per_second": 64.951,
|
| 640 |
+
"eval_steps_per_second": 2.043,
|
| 641 |
+
"step": 3103
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 29.44,
|
| 645 |
+
"learning_rate": 2.1806853582554518e-06,
|
| 646 |
+
"loss": 0.2289,
|
| 647 |
+
"step": 3150
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 29.9,
|
| 651 |
+
"learning_rate": 4.845967462789893e-07,
|
| 652 |
+
"loss": 0.2438,
|
| 653 |
+
"step": 3200
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 30.0,
|
| 657 |
+
"eval_accuracy": 0.8001310615989515,
|
| 658 |
+
"eval_loss": 1.138134479522705,
|
| 659 |
+
"eval_runtime": 23.5462,
|
| 660 |
+
"eval_samples_per_second": 64.809,
|
| 661 |
+
"eval_steps_per_second": 2.039,
|
| 662 |
+
"step": 3210
|
| 663 |
+
}
|
| 664 |
+
],
|
| 665 |
+
"max_steps": 3210,
|
| 666 |
+
"num_train_epochs": 30,
|
| 667 |
+
"total_flos": 1.2493706670677769e+20,
|
| 668 |
+
"trial_name": null,
|
| 669 |
+
"trial_params": null
|
| 670 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b6b5f205b08f6f779bd7a788849ccfab4ae877cb2e30618172a56f0c71b8be1
|
| 3 |
+
size 3503
|