Upload with huggingface_hub
Browse files- config.json +239 -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 +652 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,239 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "jonatasgrosman/wav2vec2-large-xlsr-53-italian",
|
| 3 |
+
"activation_dropout": 0.05,
|
| 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": "mean",
|
| 45 |
+
"ctc_zero_infinity": true,
|
| 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": 0.05,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"hidden_act": "gelu",
|
| 56 |
+
"hidden_dropout": 0.05,
|
| 57 |
+
"hidden_size": 1024,
|
| 58 |
+
"id2label": {
|
| 59 |
+
"0": "cooking_query",
|
| 60 |
+
"1": "iot_hue_lightdim",
|
| 61 |
+
"10": "news_query",
|
| 62 |
+
"11": "music_likeness",
|
| 63 |
+
"12": "iot_wemo_off",
|
| 64 |
+
"13": "iot_hue_lighton",
|
| 65 |
+
"14": "iot_wemo_on",
|
| 66 |
+
"15": "iot_coffee",
|
| 67 |
+
"16": "social_post",
|
| 68 |
+
"17": "qa_maths",
|
| 69 |
+
"18": "iot_hue_lightup",
|
| 70 |
+
"19": "music_query",
|
| 71 |
+
"2": "cooking_recipe",
|
| 72 |
+
"20": "transport_taxi",
|
| 73 |
+
"21": "qa_stock",
|
| 74 |
+
"22": "play_radio",
|
| 75 |
+
"23": "play_music",
|
| 76 |
+
"24": "audio_volume_up",
|
| 77 |
+
"25": "calendar_remove",
|
| 78 |
+
"26": "alarm_remove",
|
| 79 |
+
"27": "recommendation_locations",
|
| 80 |
+
"28": "general_joke",
|
| 81 |
+
"29": "email_sendemail",
|
| 82 |
+
"3": "email_addcontact",
|
| 83 |
+
"30": "audio_volume_mute",
|
| 84 |
+
"31": "iot_cleaning",
|
| 85 |
+
"32": "social_query",
|
| 86 |
+
"33": "play_game",
|
| 87 |
+
"34": "recommendation_events",
|
| 88 |
+
"35": "alarm_set",
|
| 89 |
+
"36": "calendar_set",
|
| 90 |
+
"37": "audio_volume_down",
|
| 91 |
+
"38": "lists_createoradd",
|
| 92 |
+
"39": "calendar_query",
|
| 93 |
+
"4": "weather_query",
|
| 94 |
+
"40": "general_greet",
|
| 95 |
+
"41": "qa_factoid",
|
| 96 |
+
"42": "takeaway_order",
|
| 97 |
+
"43": "transport_query",
|
| 98 |
+
"44": "iot_hue_lightchange",
|
| 99 |
+
"45": "play_podcasts",
|
| 100 |
+
"46": "transport_ticket",
|
| 101 |
+
"47": "qa_definition",
|
| 102 |
+
"48": "lists_query",
|
| 103 |
+
"49": "email_query",
|
| 104 |
+
"5": "takeaway_query",
|
| 105 |
+
"50": "datetime_convert",
|
| 106 |
+
"51": "iot_hue_lightoff",
|
| 107 |
+
"52": "alarm_query",
|
| 108 |
+
"53": "audio_volume_other",
|
| 109 |
+
"54": "lists_remove",
|
| 110 |
+
"55": "datetime_query",
|
| 111 |
+
"56": "qa_currency",
|
| 112 |
+
"57": "transport_traffic",
|
| 113 |
+
"58": "music_dislikeness",
|
| 114 |
+
"59": "music_settings",
|
| 115 |
+
"6": "email_querycontact",
|
| 116 |
+
"7": "play_audiobook",
|
| 117 |
+
"8": "recommendation_movies",
|
| 118 |
+
"9": "general_quirky"
|
| 119 |
+
},
|
| 120 |
+
"initializer_range": 0.02,
|
| 121 |
+
"intermediate_size": 4096,
|
| 122 |
+
"label2id": {
|
| 123 |
+
"alarm_query": "52",
|
| 124 |
+
"alarm_remove": "26",
|
| 125 |
+
"alarm_set": "35",
|
| 126 |
+
"audio_volume_down": "37",
|
| 127 |
+
"audio_volume_mute": "30",
|
| 128 |
+
"audio_volume_other": "53",
|
| 129 |
+
"audio_volume_up": "24",
|
| 130 |
+
"calendar_query": "39",
|
| 131 |
+
"calendar_remove": "25",
|
| 132 |
+
"calendar_set": "36",
|
| 133 |
+
"cooking_query": "0",
|
| 134 |
+
"cooking_recipe": "2",
|
| 135 |
+
"datetime_convert": "50",
|
| 136 |
+
"datetime_query": "55",
|
| 137 |
+
"email_addcontact": "3",
|
| 138 |
+
"email_query": "49",
|
| 139 |
+
"email_querycontact": "6",
|
| 140 |
+
"email_sendemail": "29",
|
| 141 |
+
"general_greet": "40",
|
| 142 |
+
"general_joke": "28",
|
| 143 |
+
"general_quirky": "9",
|
| 144 |
+
"iot_cleaning": "31",
|
| 145 |
+
"iot_coffee": "15",
|
| 146 |
+
"iot_hue_lightchange": "44",
|
| 147 |
+
"iot_hue_lightdim": "1",
|
| 148 |
+
"iot_hue_lightoff": "51",
|
| 149 |
+
"iot_hue_lighton": "13",
|
| 150 |
+
"iot_hue_lightup": "18",
|
| 151 |
+
"iot_wemo_off": "12",
|
| 152 |
+
"iot_wemo_on": "14",
|
| 153 |
+
"lists_createoradd": "38",
|
| 154 |
+
"lists_query": "48",
|
| 155 |
+
"lists_remove": "54",
|
| 156 |
+
"music_dislikeness": "58",
|
| 157 |
+
"music_likeness": "11",
|
| 158 |
+
"music_query": "19",
|
| 159 |
+
"music_settings": "59",
|
| 160 |
+
"news_query": "10",
|
| 161 |
+
"play_audiobook": "7",
|
| 162 |
+
"play_game": "33",
|
| 163 |
+
"play_music": "23",
|
| 164 |
+
"play_podcasts": "45",
|
| 165 |
+
"play_radio": "22",
|
| 166 |
+
"qa_currency": "56",
|
| 167 |
+
"qa_definition": "47",
|
| 168 |
+
"qa_factoid": "41",
|
| 169 |
+
"qa_maths": "17",
|
| 170 |
+
"qa_stock": "21",
|
| 171 |
+
"recommendation_events": "34",
|
| 172 |
+
"recommendation_locations": "27",
|
| 173 |
+
"recommendation_movies": "8",
|
| 174 |
+
"social_post": "16",
|
| 175 |
+
"social_query": "32",
|
| 176 |
+
"takeaway_order": "42",
|
| 177 |
+
"takeaway_query": "5",
|
| 178 |
+
"transport_query": "43",
|
| 179 |
+
"transport_taxi": "20",
|
| 180 |
+
"transport_ticket": "46",
|
| 181 |
+
"transport_traffic": "57",
|
| 182 |
+
"weather_query": "4"
|
| 183 |
+
},
|
| 184 |
+
"layer_norm_eps": 1e-05,
|
| 185 |
+
"layerdrop": 0.05,
|
| 186 |
+
"mask_channel_length": 10,
|
| 187 |
+
"mask_channel_min_space": 1,
|
| 188 |
+
"mask_channel_other": 0.0,
|
| 189 |
+
"mask_channel_prob": 0.0,
|
| 190 |
+
"mask_channel_selection": "static",
|
| 191 |
+
"mask_feature_length": 10,
|
| 192 |
+
"mask_feature_min_masks": 0,
|
| 193 |
+
"mask_feature_prob": 0.0,
|
| 194 |
+
"mask_time_length": 10,
|
| 195 |
+
"mask_time_min_masks": 2,
|
| 196 |
+
"mask_time_min_space": 1,
|
| 197 |
+
"mask_time_other": 0.0,
|
| 198 |
+
"mask_time_prob": 0.05,
|
| 199 |
+
"mask_time_selection": "static",
|
| 200 |
+
"model_type": "wav2vec2",
|
| 201 |
+
"num_adapter_layers": 3,
|
| 202 |
+
"num_attention_heads": 16,
|
| 203 |
+
"num_codevector_groups": 2,
|
| 204 |
+
"num_codevectors_per_group": 320,
|
| 205 |
+
"num_conv_pos_embedding_groups": 16,
|
| 206 |
+
"num_conv_pos_embeddings": 128,
|
| 207 |
+
"num_feat_extract_layers": 7,
|
| 208 |
+
"num_hidden_layers": 24,
|
| 209 |
+
"num_negatives": 100,
|
| 210 |
+
"output_hidden_size": 1024,
|
| 211 |
+
"pad_token_id": 0,
|
| 212 |
+
"proj_codevector_dim": 768,
|
| 213 |
+
"tdnn_dilation": [
|
| 214 |
+
1,
|
| 215 |
+
2,
|
| 216 |
+
3,
|
| 217 |
+
1,
|
| 218 |
+
1
|
| 219 |
+
],
|
| 220 |
+
"tdnn_dim": [
|
| 221 |
+
512,
|
| 222 |
+
512,
|
| 223 |
+
512,
|
| 224 |
+
512,
|
| 225 |
+
1500
|
| 226 |
+
],
|
| 227 |
+
"tdnn_kernel": [
|
| 228 |
+
5,
|
| 229 |
+
3,
|
| 230 |
+
3,
|
| 231 |
+
1,
|
| 232 |
+
1
|
| 233 |
+
],
|
| 234 |
+
"torch_dtype": "float32",
|
| 235 |
+
"transformers_version": "4.26.0.dev0",
|
| 236 |
+
"use_weighted_layer_sum": false,
|
| 237 |
+
"vocab_size": 44,
|
| 238 |
+
"xvector_output_dim": 512
|
| 239 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b0308b2b01be06b9079293437bdfd77e671e51d49daeae58acb4c57389bc20f
|
| 3 |
+
size 2525980249
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92127f1109fbcbd91a1f28142eeb0ff19a40bb41c016066abd1745f703a0dd84
|
| 3 |
+
size 1263010605
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0989a07dca285a3932be517629db5ee632c087a806a2c8908d8e0a4102715645
|
| 3 |
+
size 14567
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:736ac057cd1baa6f655c41cfb89657fd1618a9438af88afef35199ebb51c3f02
|
| 3 |
+
size 559
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2331be09e7bc668b4ed23bb51a2b651cad51f57e9082333fbf37cc11aefe1553
|
| 3 |
+
size 623
|
trainer_state.json
ADDED
|
@@ -0,0 +1,652 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.824220746039857,
|
| 3 |
+
"best_model_checkpoint": "wav2vec2-large-xlsr-53-italian-ic-finetuning-hard-speaker/checkpoint-2346",
|
| 4 |
+
"epoch": 29.99270072992701,
|
| 5 |
+
"global_step": 3060,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.49,
|
| 12 |
+
"learning_rate": 1.6013071895424836e-05,
|
| 13 |
+
"loss": 4.0662,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.97,
|
| 18 |
+
"learning_rate": 3.169934640522876e-05,
|
| 19 |
+
"loss": 3.9284,
|
| 20 |
+
"step": 100
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.99,
|
| 24 |
+
"eval_accuracy": 0.057230454777721,
|
| 25 |
+
"eval_loss": 3.828322649002075,
|
| 26 |
+
"eval_runtime": 29.2471,
|
| 27 |
+
"eval_samples_per_second": 66.913,
|
| 28 |
+
"eval_steps_per_second": 2.12,
|
| 29 |
+
"step": 102
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.47,
|
| 33 |
+
"learning_rate": 4.803921568627452e-05,
|
| 34 |
+
"loss": 3.8316,
|
| 35 |
+
"step": 150
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 1.95,
|
| 39 |
+
"learning_rate": 6.405228758169934e-05,
|
| 40 |
+
"loss": 3.6033,
|
| 41 |
+
"step": 200
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 1.99,
|
| 45 |
+
"eval_accuracy": 0.18446601941747573,
|
| 46 |
+
"eval_loss": 3.325474262237549,
|
| 47 |
+
"eval_runtime": 29.2457,
|
| 48 |
+
"eval_samples_per_second": 66.916,
|
| 49 |
+
"eval_steps_per_second": 2.12,
|
| 50 |
+
"step": 204
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 2.45,
|
| 54 |
+
"learning_rate": 8.006535947712419e-05,
|
| 55 |
+
"loss": 3.2263,
|
| 56 |
+
"step": 250
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.93,
|
| 60 |
+
"learning_rate": 9.640522875816994e-05,
|
| 61 |
+
"loss": 2.7337,
|
| 62 |
+
"step": 300
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 2.99,
|
| 66 |
+
"eval_accuracy": 0.455799693408278,
|
| 67 |
+
"eval_loss": 2.375624895095825,
|
| 68 |
+
"eval_runtime": 29.1511,
|
| 69 |
+
"eval_samples_per_second": 67.133,
|
| 70 |
+
"eval_steps_per_second": 2.127,
|
| 71 |
+
"step": 306
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 3.43,
|
| 75 |
+
"learning_rate": 9.85838779956427e-05,
|
| 76 |
+
"loss": 2.2867,
|
| 77 |
+
"step": 350
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 3.91,
|
| 81 |
+
"learning_rate": 9.67683369644154e-05,
|
| 82 |
+
"loss": 1.8948,
|
| 83 |
+
"step": 400
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 3.99,
|
| 87 |
+
"eval_accuracy": 0.6075625958099131,
|
| 88 |
+
"eval_loss": 1.6275904178619385,
|
| 89 |
+
"eval_runtime": 29.219,
|
| 90 |
+
"eval_samples_per_second": 66.977,
|
| 91 |
+
"eval_steps_per_second": 2.122,
|
| 92 |
+
"step": 408
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 4.41,
|
| 96 |
+
"learning_rate": 9.49527959331881e-05,
|
| 97 |
+
"loss": 1.5128,
|
| 98 |
+
"step": 450
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 4.9,
|
| 102 |
+
"learning_rate": 9.313725490196079e-05,
|
| 103 |
+
"loss": 1.3163,
|
| 104 |
+
"step": 500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 4.99,
|
| 108 |
+
"eval_accuracy": 0.7189575881451201,
|
| 109 |
+
"eval_loss": 1.2288696765899658,
|
| 110 |
+
"eval_runtime": 29.2646,
|
| 111 |
+
"eval_samples_per_second": 66.872,
|
| 112 |
+
"eval_steps_per_second": 2.119,
|
| 113 |
+
"step": 510
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 5.39,
|
| 117 |
+
"learning_rate": 9.132171387073348e-05,
|
| 118 |
+
"loss": 1.1049,
|
| 119 |
+
"step": 550
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 5.88,
|
| 123 |
+
"learning_rate": 8.950617283950617e-05,
|
| 124 |
+
"loss": 0.9525,
|
| 125 |
+
"step": 600
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 5.99,
|
| 129 |
+
"eval_accuracy": 0.7567705671946857,
|
| 130 |
+
"eval_loss": 1.070212960243225,
|
| 131 |
+
"eval_runtime": 29.2153,
|
| 132 |
+
"eval_samples_per_second": 66.985,
|
| 133 |
+
"eval_steps_per_second": 2.122,
|
| 134 |
+
"step": 612
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 6.37,
|
| 138 |
+
"learning_rate": 8.769063180827887e-05,
|
| 139 |
+
"loss": 0.7733,
|
| 140 |
+
"step": 650
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 6.86,
|
| 144 |
+
"learning_rate": 8.587509077705157e-05,
|
| 145 |
+
"loss": 0.7063,
|
| 146 |
+
"step": 700
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 6.99,
|
| 150 |
+
"eval_accuracy": 0.7797649463464487,
|
| 151 |
+
"eval_loss": 0.9915740489959717,
|
| 152 |
+
"eval_runtime": 29.2648,
|
| 153 |
+
"eval_samples_per_second": 66.872,
|
| 154 |
+
"eval_steps_per_second": 2.119,
|
| 155 |
+
"step": 714
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 7.35,
|
| 159 |
+
"learning_rate": 8.405954974582426e-05,
|
| 160 |
+
"loss": 0.6261,
|
| 161 |
+
"step": 750
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 7.84,
|
| 165 |
+
"learning_rate": 8.224400871459695e-05,
|
| 166 |
+
"loss": 0.5398,
|
| 167 |
+
"step": 800
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 7.99,
|
| 171 |
+
"eval_accuracy": 0.7797649463464487,
|
| 172 |
+
"eval_loss": 0.9585044980049133,
|
| 173 |
+
"eval_runtime": 29.2234,
|
| 174 |
+
"eval_samples_per_second": 66.967,
|
| 175 |
+
"eval_steps_per_second": 2.122,
|
| 176 |
+
"step": 816
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 8.33,
|
| 180 |
+
"learning_rate": 8.042846768336965e-05,
|
| 181 |
+
"loss": 0.4993,
|
| 182 |
+
"step": 850
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 8.82,
|
| 186 |
+
"learning_rate": 7.861292665214233e-05,
|
| 187 |
+
"loss": 0.4401,
|
| 188 |
+
"step": 900
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 8.99,
|
| 192 |
+
"eval_accuracy": 0.7894736842105263,
|
| 193 |
+
"eval_loss": 0.9442699551582336,
|
| 194 |
+
"eval_runtime": 29.2708,
|
| 195 |
+
"eval_samples_per_second": 66.858,
|
| 196 |
+
"eval_steps_per_second": 2.118,
|
| 197 |
+
"step": 918
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 9.31,
|
| 201 |
+
"learning_rate": 7.679738562091505e-05,
|
| 202 |
+
"loss": 0.3974,
|
| 203 |
+
"step": 950
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 9.8,
|
| 207 |
+
"learning_rate": 7.498184458968773e-05,
|
| 208 |
+
"loss": 0.3377,
|
| 209 |
+
"step": 1000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 9.99,
|
| 213 |
+
"eval_accuracy": 0.7966274910577414,
|
| 214 |
+
"eval_loss": 0.9571655988693237,
|
| 215 |
+
"eval_runtime": 29.1828,
|
| 216 |
+
"eval_samples_per_second": 67.06,
|
| 217 |
+
"eval_steps_per_second": 2.125,
|
| 218 |
+
"step": 1020
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 10.29,
|
| 222 |
+
"learning_rate": 7.316630355846043e-05,
|
| 223 |
+
"loss": 0.2958,
|
| 224 |
+
"step": 1050
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 10.78,
|
| 228 |
+
"learning_rate": 7.135076252723311e-05,
|
| 229 |
+
"loss": 0.2903,
|
| 230 |
+
"step": 1100
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 10.99,
|
| 234 |
+
"eval_accuracy": 0.8114460909555442,
|
| 235 |
+
"eval_loss": 0.9565660357475281,
|
| 236 |
+
"eval_runtime": 29.2796,
|
| 237 |
+
"eval_samples_per_second": 66.838,
|
| 238 |
+
"eval_steps_per_second": 2.118,
|
| 239 |
+
"step": 1122
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 11.27,
|
| 243 |
+
"learning_rate": 6.953522149600581e-05,
|
| 244 |
+
"loss": 0.27,
|
| 245 |
+
"step": 1150
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 11.76,
|
| 249 |
+
"learning_rate": 6.771968046477851e-05,
|
| 250 |
+
"loss": 0.2313,
|
| 251 |
+
"step": 1200
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 11.99,
|
| 255 |
+
"eval_accuracy": 0.8160449667858968,
|
| 256 |
+
"eval_loss": 0.9552507996559143,
|
| 257 |
+
"eval_runtime": 29.3904,
|
| 258 |
+
"eval_samples_per_second": 66.586,
|
| 259 |
+
"eval_steps_per_second": 2.11,
|
| 260 |
+
"step": 1224
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 12.25,
|
| 264 |
+
"learning_rate": 6.590413943355121e-05,
|
| 265 |
+
"loss": 0.2222,
|
| 266 |
+
"step": 1250
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 12.74,
|
| 270 |
+
"learning_rate": 6.40885984023239e-05,
|
| 271 |
+
"loss": 0.196,
|
| 272 |
+
"step": 1300
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 12.99,
|
| 276 |
+
"eval_accuracy": 0.7976494634644865,
|
| 277 |
+
"eval_loss": 1.0041111707687378,
|
| 278 |
+
"eval_runtime": 29.2891,
|
| 279 |
+
"eval_samples_per_second": 66.817,
|
| 280 |
+
"eval_steps_per_second": 2.117,
|
| 281 |
+
"step": 1326
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 13.23,
|
| 285 |
+
"learning_rate": 6.22730573710966e-05,
|
| 286 |
+
"loss": 0.186,
|
| 287 |
+
"step": 1350
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 13.72,
|
| 291 |
+
"learning_rate": 6.045751633986928e-05,
|
| 292 |
+
"loss": 0.189,
|
| 293 |
+
"step": 1400
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 13.99,
|
| 297 |
+
"eval_accuracy": 0.8109351047521717,
|
| 298 |
+
"eval_loss": 1.0098705291748047,
|
| 299 |
+
"eval_runtime": 29.5324,
|
| 300 |
+
"eval_samples_per_second": 66.266,
|
| 301 |
+
"eval_steps_per_second": 2.099,
|
| 302 |
+
"step": 1428
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 14.21,
|
| 306 |
+
"learning_rate": 5.8641975308641984e-05,
|
| 307 |
+
"loss": 0.1511,
|
| 308 |
+
"step": 1450
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 14.7,
|
| 312 |
+
"learning_rate": 5.6826434277414676e-05,
|
| 313 |
+
"loss": 0.1253,
|
| 314 |
+
"step": 1500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 14.99,
|
| 318 |
+
"eval_accuracy": 0.8119570771589167,
|
| 319 |
+
"eval_loss": 1.030671238899231,
|
| 320 |
+
"eval_runtime": 30.8165,
|
| 321 |
+
"eval_samples_per_second": 63.505,
|
| 322 |
+
"eval_steps_per_second": 2.012,
|
| 323 |
+
"step": 1530
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 15.19,
|
| 327 |
+
"learning_rate": 5.501089324618737e-05,
|
| 328 |
+
"loss": 0.1539,
|
| 329 |
+
"step": 1550
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 15.68,
|
| 333 |
+
"learning_rate": 5.319535221496006e-05,
|
| 334 |
+
"loss": 0.1114,
|
| 335 |
+
"step": 1600
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 15.99,
|
| 339 |
+
"eval_accuracy": 0.8048032703117016,
|
| 340 |
+
"eval_loss": 1.1111302375793457,
|
| 341 |
+
"eval_runtime": 30.7697,
|
| 342 |
+
"eval_samples_per_second": 63.601,
|
| 343 |
+
"eval_steps_per_second": 2.015,
|
| 344 |
+
"step": 1632
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 16.18,
|
| 348 |
+
"learning_rate": 5.137981118373275e-05,
|
| 349 |
+
"loss": 0.1331,
|
| 350 |
+
"step": 1650
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 16.66,
|
| 354 |
+
"learning_rate": 4.956427015250545e-05,
|
| 355 |
+
"loss": 0.1241,
|
| 356 |
+
"step": 1700
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 16.99,
|
| 360 |
+
"eval_accuracy": 0.8109351047521717,
|
| 361 |
+
"eval_loss": 1.0672301054000854,
|
| 362 |
+
"eval_runtime": 31.1401,
|
| 363 |
+
"eval_samples_per_second": 62.845,
|
| 364 |
+
"eval_steps_per_second": 1.991,
|
| 365 |
+
"step": 1734
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 17.16,
|
| 369 |
+
"learning_rate": 4.774872912127814e-05,
|
| 370 |
+
"loss": 0.0923,
|
| 371 |
+
"step": 1750
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 17.64,
|
| 375 |
+
"learning_rate": 4.593318809005084e-05,
|
| 376 |
+
"loss": 0.1128,
|
| 377 |
+
"step": 1800
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 17.99,
|
| 381 |
+
"eval_accuracy": 0.8119570771589167,
|
| 382 |
+
"eval_loss": 1.092710018157959,
|
| 383 |
+
"eval_runtime": 31.1388,
|
| 384 |
+
"eval_samples_per_second": 62.848,
|
| 385 |
+
"eval_steps_per_second": 1.991,
|
| 386 |
+
"step": 1836
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 18.14,
|
| 390 |
+
"learning_rate": 4.411764705882353e-05,
|
| 391 |
+
"loss": 0.0897,
|
| 392 |
+
"step": 1850
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 18.62,
|
| 396 |
+
"learning_rate": 4.230210602759623e-05,
|
| 397 |
+
"loss": 0.0932,
|
| 398 |
+
"step": 1900
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 18.99,
|
| 402 |
+
"eval_accuracy": 0.8185998978027593,
|
| 403 |
+
"eval_loss": 1.0788525342941284,
|
| 404 |
+
"eval_runtime": 31.2301,
|
| 405 |
+
"eval_samples_per_second": 62.664,
|
| 406 |
+
"eval_steps_per_second": 1.985,
|
| 407 |
+
"step": 1938
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 19.12,
|
| 411 |
+
"learning_rate": 4.048656499636892e-05,
|
| 412 |
+
"loss": 0.0849,
|
| 413 |
+
"step": 1950
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 19.6,
|
| 417 |
+
"learning_rate": 3.8671023965141615e-05,
|
| 418 |
+
"loss": 0.0889,
|
| 419 |
+
"step": 2000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 19.99,
|
| 423 |
+
"eval_accuracy": 0.8104241185487991,
|
| 424 |
+
"eval_loss": 1.153747320175171,
|
| 425 |
+
"eval_runtime": 31.9386,
|
| 426 |
+
"eval_samples_per_second": 61.274,
|
| 427 |
+
"eval_steps_per_second": 1.941,
|
| 428 |
+
"step": 2040
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 20.1,
|
| 432 |
+
"learning_rate": 3.6855482933914306e-05,
|
| 433 |
+
"loss": 0.0732,
|
| 434 |
+
"step": 2050
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 20.58,
|
| 438 |
+
"learning_rate": 3.5039941902687e-05,
|
| 439 |
+
"loss": 0.0649,
|
| 440 |
+
"step": 2100
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 20.99,
|
| 444 |
+
"eval_accuracy": 0.8088911599386817,
|
| 445 |
+
"eval_loss": 1.1140419244766235,
|
| 446 |
+
"eval_runtime": 32.7576,
|
| 447 |
+
"eval_samples_per_second": 59.742,
|
| 448 |
+
"eval_steps_per_second": 1.893,
|
| 449 |
+
"step": 2142
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 21.08,
|
| 453 |
+
"learning_rate": 3.32244008714597e-05,
|
| 454 |
+
"loss": 0.061,
|
| 455 |
+
"step": 2150
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 21.56,
|
| 459 |
+
"learning_rate": 3.140885984023239e-05,
|
| 460 |
+
"loss": 0.0645,
|
| 461 |
+
"step": 2200
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 21.99,
|
| 465 |
+
"eval_accuracy": 0.8150229943791517,
|
| 466 |
+
"eval_loss": 1.133326530456543,
|
| 467 |
+
"eval_runtime": 31.5293,
|
| 468 |
+
"eval_samples_per_second": 62.069,
|
| 469 |
+
"eval_steps_per_second": 1.966,
|
| 470 |
+
"step": 2244
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 22.06,
|
| 474 |
+
"learning_rate": 2.9593318809005084e-05,
|
| 475 |
+
"loss": 0.0619,
|
| 476 |
+
"step": 2250
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 22.55,
|
| 480 |
+
"learning_rate": 2.777777777777778e-05,
|
| 481 |
+
"loss": 0.0637,
|
| 482 |
+
"step": 2300
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 22.99,
|
| 486 |
+
"eval_accuracy": 0.824220746039857,
|
| 487 |
+
"eval_loss": 1.1310803890228271,
|
| 488 |
+
"eval_runtime": 29.6669,
|
| 489 |
+
"eval_samples_per_second": 65.966,
|
| 490 |
+
"eval_steps_per_second": 2.09,
|
| 491 |
+
"step": 2346
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 23.04,
|
| 495 |
+
"learning_rate": 2.596223674655047e-05,
|
| 496 |
+
"loss": 0.0453,
|
| 497 |
+
"step": 2350
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 23.53,
|
| 501 |
+
"learning_rate": 2.4183006535947712e-05,
|
| 502 |
+
"loss": 0.0767,
|
| 503 |
+
"step": 2400
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 23.99,
|
| 507 |
+
"eval_accuracy": 0.8165559529892693,
|
| 508 |
+
"eval_loss": 1.1296130418777466,
|
| 509 |
+
"eval_runtime": 29.5997,
|
| 510 |
+
"eval_samples_per_second": 66.115,
|
| 511 |
+
"eval_steps_per_second": 2.095,
|
| 512 |
+
"step": 2448
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 24.02,
|
| 516 |
+
"learning_rate": 2.2367465504720407e-05,
|
| 517 |
+
"loss": 0.0446,
|
| 518 |
+
"step": 2450
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 24.51,
|
| 522 |
+
"learning_rate": 2.0551924473493103e-05,
|
| 523 |
+
"loss": 0.0571,
|
| 524 |
+
"step": 2500
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 24.99,
|
| 528 |
+
"learning_rate": 1.877269426289034e-05,
|
| 529 |
+
"loss": 0.0457,
|
| 530 |
+
"step": 2550
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 24.99,
|
| 534 |
+
"eval_accuracy": 0.8083801737353091,
|
| 535 |
+
"eval_loss": 1.168392539024353,
|
| 536 |
+
"eval_runtime": 29.5486,
|
| 537 |
+
"eval_samples_per_second": 66.23,
|
| 538 |
+
"eval_steps_per_second": 2.098,
|
| 539 |
+
"step": 2550
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 25.49,
|
| 543 |
+
"learning_rate": 1.6957153231663036e-05,
|
| 544 |
+
"loss": 0.0424,
|
| 545 |
+
"step": 2600
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 25.97,
|
| 549 |
+
"learning_rate": 1.5141612200435731e-05,
|
| 550 |
+
"loss": 0.0405,
|
| 551 |
+
"step": 2650
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 25.99,
|
| 555 |
+
"eval_accuracy": 0.8160449667858968,
|
| 556 |
+
"eval_loss": 1.1332188844680786,
|
| 557 |
+
"eval_runtime": 29.6263,
|
| 558 |
+
"eval_samples_per_second": 66.056,
|
| 559 |
+
"eval_steps_per_second": 2.093,
|
| 560 |
+
"step": 2652
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 26.47,
|
| 564 |
+
"learning_rate": 1.3326071169208426e-05,
|
| 565 |
+
"loss": 0.0454,
|
| 566 |
+
"step": 2700
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 26.95,
|
| 570 |
+
"learning_rate": 1.151053013798112e-05,
|
| 571 |
+
"loss": 0.0362,
|
| 572 |
+
"step": 2750
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 26.99,
|
| 576 |
+
"eval_accuracy": 0.8201328564128768,
|
| 577 |
+
"eval_loss": 1.131385087966919,
|
| 578 |
+
"eval_runtime": 29.8487,
|
| 579 |
+
"eval_samples_per_second": 65.564,
|
| 580 |
+
"eval_steps_per_second": 2.077,
|
| 581 |
+
"step": 2754
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 27.45,
|
| 585 |
+
"learning_rate": 9.694989106753813e-06,
|
| 586 |
+
"loss": 0.0337,
|
| 587 |
+
"step": 2800
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 27.93,
|
| 591 |
+
"learning_rate": 7.879448075526508e-06,
|
| 592 |
+
"loss": 0.0391,
|
| 593 |
+
"step": 2850
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 27.99,
|
| 597 |
+
"eval_accuracy": 0.8211548288196219,
|
| 598 |
+
"eval_loss": 1.1564393043518066,
|
| 599 |
+
"eval_runtime": 29.8613,
|
| 600 |
+
"eval_samples_per_second": 65.536,
|
| 601 |
+
"eval_steps_per_second": 2.076,
|
| 602 |
+
"step": 2856
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 28.43,
|
| 606 |
+
"learning_rate": 6.063907044299202e-06,
|
| 607 |
+
"loss": 0.0435,
|
| 608 |
+
"step": 2900
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 28.91,
|
| 612 |
+
"learning_rate": 4.2483660130718954e-06,
|
| 613 |
+
"loss": 0.0452,
|
| 614 |
+
"step": 2950
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 28.99,
|
| 618 |
+
"eval_accuracy": 0.8221768012263669,
|
| 619 |
+
"eval_loss": 1.1484907865524292,
|
| 620 |
+
"eval_runtime": 29.975,
|
| 621 |
+
"eval_samples_per_second": 65.288,
|
| 622 |
+
"eval_steps_per_second": 2.068,
|
| 623 |
+
"step": 2958
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 29.41,
|
| 627 |
+
"learning_rate": 2.43282498184459e-06,
|
| 628 |
+
"loss": 0.0484,
|
| 629 |
+
"step": 3000
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 29.9,
|
| 633 |
+
"learning_rate": 6.17283950617284e-07,
|
| 634 |
+
"loss": 0.0343,
|
| 635 |
+
"step": 3050
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 29.99,
|
| 639 |
+
"eval_accuracy": 0.8201328564128768,
|
| 640 |
+
"eval_loss": 1.1464686393737793,
|
| 641 |
+
"eval_runtime": 29.8646,
|
| 642 |
+
"eval_samples_per_second": 65.529,
|
| 643 |
+
"eval_steps_per_second": 2.076,
|
| 644 |
+
"step": 3060
|
| 645 |
+
}
|
| 646 |
+
],
|
| 647 |
+
"max_steps": 3060,
|
| 648 |
+
"num_train_epochs": 30,
|
| 649 |
+
"total_flos": 1.1930235995944326e+20,
|
| 650 |
+
"trial_name": null,
|
| 651 |
+
"trial_params": null
|
| 652 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40d88c79085c5cfec03e1ac580425ac76d3fe307b9edff2b4cd3cb2e6336d157
|
| 3 |
+
size 3567
|