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 +709 -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": "play_audiobook",
|
| 61 |
+
"1": "iot_hue_lightup",
|
| 62 |
+
"10": "music_settings",
|
| 63 |
+
"11": "qa_maths",
|
| 64 |
+
"12": "calendar_remove",
|
| 65 |
+
"13": "music_likeness",
|
| 66 |
+
"14": "general_quirky",
|
| 67 |
+
"15": "news_query",
|
| 68 |
+
"16": "audio_volume_mute",
|
| 69 |
+
"17": "takeaway_order",
|
| 70 |
+
"18": "recommendation_locations",
|
| 71 |
+
"19": "recommendation_movies",
|
| 72 |
+
"2": "iot_coffee",
|
| 73 |
+
"20": "datetime_query",
|
| 74 |
+
"21": "transport_ticket",
|
| 75 |
+
"22": "lists_remove",
|
| 76 |
+
"23": "play_podcasts",
|
| 77 |
+
"24": "audio_volume_other",
|
| 78 |
+
"25": "iot_hue_lighton",
|
| 79 |
+
"26": "alarm_set",
|
| 80 |
+
"27": "general_greet",
|
| 81 |
+
"28": "weather_query",
|
| 82 |
+
"29": "qa_factoid",
|
| 83 |
+
"3": "general_joke",
|
| 84 |
+
"30": "qa_definition",
|
| 85 |
+
"31": "qa_stock",
|
| 86 |
+
"32": "music_dislikeness",
|
| 87 |
+
"33": "music_query",
|
| 88 |
+
"34": "iot_wemo_on",
|
| 89 |
+
"35": "cooking_query",
|
| 90 |
+
"36": "audio_volume_up",
|
| 91 |
+
"37": "transport_query",
|
| 92 |
+
"38": "audio_volume_down",
|
| 93 |
+
"39": "email_querycontact",
|
| 94 |
+
"4": "alarm_query",
|
| 95 |
+
"40": "email_sendemail",
|
| 96 |
+
"41": "iot_wemo_off",
|
| 97 |
+
"42": "takeaway_query",
|
| 98 |
+
"43": "play_music",
|
| 99 |
+
"44": "transport_taxi",
|
| 100 |
+
"45": "calendar_query",
|
| 101 |
+
"46": "cooking_recipe",
|
| 102 |
+
"47": "iot_hue_lightchange",
|
| 103 |
+
"48": "recommendation_events",
|
| 104 |
+
"49": "play_game",
|
| 105 |
+
"5": "social_post",
|
| 106 |
+
"50": "iot_hue_lightdim",
|
| 107 |
+
"51": "qa_currency",
|
| 108 |
+
"52": "email_query",
|
| 109 |
+
"53": "iot_cleaning",
|
| 110 |
+
"54": "transport_traffic",
|
| 111 |
+
"55": "social_query",
|
| 112 |
+
"56": "email_addcontact",
|
| 113 |
+
"57": "calendar_set",
|
| 114 |
+
"58": "lists_createoradd",
|
| 115 |
+
"59": "lists_query",
|
| 116 |
+
"6": "datetime_convert",
|
| 117 |
+
"7": "iot_hue_lightoff",
|
| 118 |
+
"8": "alarm_remove",
|
| 119 |
+
"9": "play_radio"
|
| 120 |
+
},
|
| 121 |
+
"initializer_range": 0.02,
|
| 122 |
+
"intermediate_size": 4096,
|
| 123 |
+
"label2id": {
|
| 124 |
+
"alarm_query": "4",
|
| 125 |
+
"alarm_remove": "8",
|
| 126 |
+
"alarm_set": "26",
|
| 127 |
+
"audio_volume_down": "38",
|
| 128 |
+
"audio_volume_mute": "16",
|
| 129 |
+
"audio_volume_other": "24",
|
| 130 |
+
"audio_volume_up": "36",
|
| 131 |
+
"calendar_query": "45",
|
| 132 |
+
"calendar_remove": "12",
|
| 133 |
+
"calendar_set": "57",
|
| 134 |
+
"cooking_query": "35",
|
| 135 |
+
"cooking_recipe": "46",
|
| 136 |
+
"datetime_convert": "6",
|
| 137 |
+
"datetime_query": "20",
|
| 138 |
+
"email_addcontact": "56",
|
| 139 |
+
"email_query": "52",
|
| 140 |
+
"email_querycontact": "39",
|
| 141 |
+
"email_sendemail": "40",
|
| 142 |
+
"general_greet": "27",
|
| 143 |
+
"general_joke": "3",
|
| 144 |
+
"general_quirky": "14",
|
| 145 |
+
"iot_cleaning": "53",
|
| 146 |
+
"iot_coffee": "2",
|
| 147 |
+
"iot_hue_lightchange": "47",
|
| 148 |
+
"iot_hue_lightdim": "50",
|
| 149 |
+
"iot_hue_lightoff": "7",
|
| 150 |
+
"iot_hue_lighton": "25",
|
| 151 |
+
"iot_hue_lightup": "1",
|
| 152 |
+
"iot_wemo_off": "41",
|
| 153 |
+
"iot_wemo_on": "34",
|
| 154 |
+
"lists_createoradd": "58",
|
| 155 |
+
"lists_query": "59",
|
| 156 |
+
"lists_remove": "22",
|
| 157 |
+
"music_dislikeness": "32",
|
| 158 |
+
"music_likeness": "13",
|
| 159 |
+
"music_query": "33",
|
| 160 |
+
"music_settings": "10",
|
| 161 |
+
"news_query": "15",
|
| 162 |
+
"play_audiobook": "0",
|
| 163 |
+
"play_game": "49",
|
| 164 |
+
"play_music": "43",
|
| 165 |
+
"play_podcasts": "23",
|
| 166 |
+
"play_radio": "9",
|
| 167 |
+
"qa_currency": "51",
|
| 168 |
+
"qa_definition": "30",
|
| 169 |
+
"qa_factoid": "29",
|
| 170 |
+
"qa_maths": "11",
|
| 171 |
+
"qa_stock": "31",
|
| 172 |
+
"recommendation_events": "48",
|
| 173 |
+
"recommendation_locations": "18",
|
| 174 |
+
"recommendation_movies": "19",
|
| 175 |
+
"social_post": "5",
|
| 176 |
+
"social_query": "55",
|
| 177 |
+
"takeaway_order": "17",
|
| 178 |
+
"takeaway_query": "42",
|
| 179 |
+
"transport_query": "37",
|
| 180 |
+
"transport_taxi": "44",
|
| 181 |
+
"transport_ticket": "21",
|
| 182 |
+
"transport_traffic": "54",
|
| 183 |
+
"weather_query": "28"
|
| 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 sha256:3a1be98ce10152cb697151747e63cad2caccef76d90bb03ece09ead8286ecb62
|
| 3 |
+
size 2525982533
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b1a710c802aa94006203a54fa98453437cfe5ae62c78d462d9f547e8572549f5
|
| 3 |
+
size 1263013749
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad698817146e43247b37597d8bf5d84322141eca8b79d34115ee4ffab84b6f21
|
| 3 |
+
size 15597
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5d028d09bd0db1f8dd260d72c3333dcaebcb2b8372827556976d9e7e8143ac02
|
| 3 |
+
size 557
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d93c6859f27491212f5601d32196360c1fcad10ff5f0f2a08fe50124d51c7992
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,709 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7855386128873586,
|
| 3 |
+
"best_model_checkpoint": "wav2vec2-xls-r-300m-ic-finetuning/checkpoint-2790",
|
| 4 |
+
"epoch": 35.0,
|
| 5 |
+
"global_step": 3150,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.56,
|
| 12 |
+
"learning_rate": 1.5555555555555555e-05,
|
| 13 |
+
"loss": 4.0884,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.0,
|
| 18 |
+
"eval_accuracy": 0.06443679291687161,
|
| 19 |
+
"eval_loss": 3.94179630279541,
|
| 20 |
+
"eval_runtime": 17.2679,
|
| 21 |
+
"eval_samples_per_second": 117.733,
|
| 22 |
+
"eval_steps_per_second": 0.927,
|
| 23 |
+
"step": 90
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 1.11,
|
| 27 |
+
"learning_rate": 3.142857142857143e-05,
|
| 28 |
+
"loss": 4.0116,
|
| 29 |
+
"step": 100
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.67,
|
| 33 |
+
"learning_rate": 4.73015873015873e-05,
|
| 34 |
+
"loss": 3.8285,
|
| 35 |
+
"step": 150
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 2.0,
|
| 39 |
+
"eval_accuracy": 0.06443679291687161,
|
| 40 |
+
"eval_loss": 3.742227792739868,
|
| 41 |
+
"eval_runtime": 17.4046,
|
| 42 |
+
"eval_samples_per_second": 116.808,
|
| 43 |
+
"eval_steps_per_second": 0.919,
|
| 44 |
+
"step": 180
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.22,
|
| 48 |
+
"learning_rate": 6.317460317460318e-05,
|
| 49 |
+
"loss": 3.7787,
|
| 50 |
+
"step": 200
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 2.78,
|
| 54 |
+
"learning_rate": 7.904761904761905e-05,
|
| 55 |
+
"loss": 3.7603,
|
| 56 |
+
"step": 250
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 3.0,
|
| 60 |
+
"eval_accuracy": 0.06443679291687161,
|
| 61 |
+
"eval_loss": 3.730922222137451,
|
| 62 |
+
"eval_runtime": 17.2612,
|
| 63 |
+
"eval_samples_per_second": 117.779,
|
| 64 |
+
"eval_steps_per_second": 0.927,
|
| 65 |
+
"step": 270
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 3.33,
|
| 69 |
+
"learning_rate": 9.46031746031746e-05,
|
| 70 |
+
"loss": 3.7493,
|
| 71 |
+
"step": 300
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 3.89,
|
| 75 |
+
"learning_rate": 9.883597883597884e-05,
|
| 76 |
+
"loss": 3.7475,
|
| 77 |
+
"step": 350
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 4.0,
|
| 81 |
+
"eval_accuracy": 0.08460403344810624,
|
| 82 |
+
"eval_loss": 3.719400405883789,
|
| 83 |
+
"eval_runtime": 17.1221,
|
| 84 |
+
"eval_samples_per_second": 118.736,
|
| 85 |
+
"eval_steps_per_second": 0.934,
|
| 86 |
+
"step": 360
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 4.44,
|
| 90 |
+
"learning_rate": 9.707231040564374e-05,
|
| 91 |
+
"loss": 3.7259,
|
| 92 |
+
"step": 400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 5.0,
|
| 96 |
+
"learning_rate": 9.530864197530865e-05,
|
| 97 |
+
"loss": 3.6843,
|
| 98 |
+
"step": 450
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 5.0,
|
| 102 |
+
"eval_accuracy": 0.102803738317757,
|
| 103 |
+
"eval_loss": 3.6270625591278076,
|
| 104 |
+
"eval_runtime": 16.8572,
|
| 105 |
+
"eval_samples_per_second": 120.601,
|
| 106 |
+
"eval_steps_per_second": 0.949,
|
| 107 |
+
"step": 450
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 5.56,
|
| 111 |
+
"learning_rate": 9.354497354497354e-05,
|
| 112 |
+
"loss": 3.6367,
|
| 113 |
+
"step": 500
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 6.0,
|
| 117 |
+
"eval_accuracy": 0.12690605017215936,
|
| 118 |
+
"eval_loss": 3.4777231216430664,
|
| 119 |
+
"eval_runtime": 16.8395,
|
| 120 |
+
"eval_samples_per_second": 120.728,
|
| 121 |
+
"eval_steps_per_second": 0.95,
|
| 122 |
+
"step": 540
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 6.11,
|
| 126 |
+
"learning_rate": 9.178130511463845e-05,
|
| 127 |
+
"loss": 3.5853,
|
| 128 |
+
"step": 550
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 6.67,
|
| 132 |
+
"learning_rate": 9.001763668430336e-05,
|
| 133 |
+
"loss": 3.4016,
|
| 134 |
+
"step": 600
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 7.0,
|
| 138 |
+
"eval_accuracy": 0.21446138711264143,
|
| 139 |
+
"eval_loss": 3.1104345321655273,
|
| 140 |
+
"eval_runtime": 16.8901,
|
| 141 |
+
"eval_samples_per_second": 120.366,
|
| 142 |
+
"eval_steps_per_second": 0.947,
|
| 143 |
+
"step": 630
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 7.22,
|
| 147 |
+
"learning_rate": 8.825396825396825e-05,
|
| 148 |
+
"loss": 3.213,
|
| 149 |
+
"step": 650
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 7.78,
|
| 153 |
+
"learning_rate": 8.649029982363316e-05,
|
| 154 |
+
"loss": 2.946,
|
| 155 |
+
"step": 700
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 8.0,
|
| 159 |
+
"eval_accuracy": 0.31234628627643873,
|
| 160 |
+
"eval_loss": 2.6535961627960205,
|
| 161 |
+
"eval_runtime": 16.8793,
|
| 162 |
+
"eval_samples_per_second": 120.444,
|
| 163 |
+
"eval_steps_per_second": 0.948,
|
| 164 |
+
"step": 720
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 8.33,
|
| 168 |
+
"learning_rate": 8.472663139329807e-05,
|
| 169 |
+
"loss": 2.7358,
|
| 170 |
+
"step": 750
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 8.89,
|
| 174 |
+
"learning_rate": 8.296296296296296e-05,
|
| 175 |
+
"loss": 2.514,
|
| 176 |
+
"step": 800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 9.0,
|
| 180 |
+
"eval_accuracy": 0.4176094441711756,
|
| 181 |
+
"eval_loss": 2.3149795532226562,
|
| 182 |
+
"eval_runtime": 16.8771,
|
| 183 |
+
"eval_samples_per_second": 120.459,
|
| 184 |
+
"eval_steps_per_second": 0.948,
|
| 185 |
+
"step": 810
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 9.44,
|
| 189 |
+
"learning_rate": 8.119929453262787e-05,
|
| 190 |
+
"loss": 2.3267,
|
| 191 |
+
"step": 850
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 10.0,
|
| 195 |
+
"learning_rate": 7.947089947089948e-05,
|
| 196 |
+
"loss": 2.163,
|
| 197 |
+
"step": 900
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 10.0,
|
| 201 |
+
"eval_accuracy": 0.48106246925725527,
|
| 202 |
+
"eval_loss": 2.0448246002197266,
|
| 203 |
+
"eval_runtime": 16.8844,
|
| 204 |
+
"eval_samples_per_second": 120.407,
|
| 205 |
+
"eval_steps_per_second": 0.948,
|
| 206 |
+
"step": 900
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 10.56,
|
| 210 |
+
"learning_rate": 7.770723104056437e-05,
|
| 211 |
+
"loss": 1.947,
|
| 212 |
+
"step": 950
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 11.0,
|
| 216 |
+
"eval_accuracy": 0.5700934579439252,
|
| 217 |
+
"eval_loss": 1.7913175821304321,
|
| 218 |
+
"eval_runtime": 16.8279,
|
| 219 |
+
"eval_samples_per_second": 120.811,
|
| 220 |
+
"eval_steps_per_second": 0.951,
|
| 221 |
+
"step": 990
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 11.11,
|
| 225 |
+
"learning_rate": 7.594356261022928e-05,
|
| 226 |
+
"loss": 1.8568,
|
| 227 |
+
"step": 1000
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 11.67,
|
| 231 |
+
"learning_rate": 7.417989417989419e-05,
|
| 232 |
+
"loss": 1.7119,
|
| 233 |
+
"step": 1050
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 12.0,
|
| 237 |
+
"eval_accuracy": 0.6242006886374816,
|
| 238 |
+
"eval_loss": 1.616714596748352,
|
| 239 |
+
"eval_runtime": 16.8545,
|
| 240 |
+
"eval_samples_per_second": 120.62,
|
| 241 |
+
"eval_steps_per_second": 0.949,
|
| 242 |
+
"step": 1080
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 12.22,
|
| 246 |
+
"learning_rate": 7.241622574955908e-05,
|
| 247 |
+
"loss": 1.5846,
|
| 248 |
+
"step": 1100
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 12.78,
|
| 252 |
+
"learning_rate": 7.065255731922399e-05,
|
| 253 |
+
"loss": 1.4849,
|
| 254 |
+
"step": 1150
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 13.0,
|
| 258 |
+
"eval_accuracy": 0.6620757501229709,
|
| 259 |
+
"eval_loss": 1.5006158351898193,
|
| 260 |
+
"eval_runtime": 16.8739,
|
| 261 |
+
"eval_samples_per_second": 120.482,
|
| 262 |
+
"eval_steps_per_second": 0.948,
|
| 263 |
+
"step": 1170
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 13.33,
|
| 267 |
+
"learning_rate": 6.88888888888889e-05,
|
| 268 |
+
"loss": 1.3323,
|
| 269 |
+
"step": 1200
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 13.89,
|
| 273 |
+
"learning_rate": 6.712522045855379e-05,
|
| 274 |
+
"loss": 1.3281,
|
| 275 |
+
"step": 1250
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 14.0,
|
| 279 |
+
"eval_accuracy": 0.6699458927693065,
|
| 280 |
+
"eval_loss": 1.4369958639144897,
|
| 281 |
+
"eval_runtime": 16.7964,
|
| 282 |
+
"eval_samples_per_second": 121.038,
|
| 283 |
+
"eval_steps_per_second": 0.953,
|
| 284 |
+
"step": 1260
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 14.44,
|
| 288 |
+
"learning_rate": 6.53615520282187e-05,
|
| 289 |
+
"loss": 1.1692,
|
| 290 |
+
"step": 1300
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 15.0,
|
| 294 |
+
"learning_rate": 6.35978835978836e-05,
|
| 295 |
+
"loss": 1.1518,
|
| 296 |
+
"step": 1350
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 15.0,
|
| 300 |
+
"eval_accuracy": 0.6984751598622725,
|
| 301 |
+
"eval_loss": 1.3894743919372559,
|
| 302 |
+
"eval_runtime": 16.9538,
|
| 303 |
+
"eval_samples_per_second": 119.914,
|
| 304 |
+
"eval_steps_per_second": 0.944,
|
| 305 |
+
"step": 1350
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 15.56,
|
| 309 |
+
"learning_rate": 6.183421516754851e-05,
|
| 310 |
+
"loss": 0.9798,
|
| 311 |
+
"step": 1400
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 16.0,
|
| 315 |
+
"eval_accuracy": 0.7329070339399901,
|
| 316 |
+
"eval_loss": 1.3057465553283691,
|
| 317 |
+
"eval_runtime": 16.8273,
|
| 318 |
+
"eval_samples_per_second": 120.816,
|
| 319 |
+
"eval_steps_per_second": 0.951,
|
| 320 |
+
"step": 1440
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 16.11,
|
| 324 |
+
"learning_rate": 6.0070546737213404e-05,
|
| 325 |
+
"loss": 1.0168,
|
| 326 |
+
"step": 1450
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 16.67,
|
| 330 |
+
"learning_rate": 5.830687830687831e-05,
|
| 331 |
+
"loss": 0.9296,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 17.0,
|
| 336 |
+
"eval_accuracy": 0.7353664535169699,
|
| 337 |
+
"eval_loss": 1.2697882652282715,
|
| 338 |
+
"eval_runtime": 16.8455,
|
| 339 |
+
"eval_samples_per_second": 120.685,
|
| 340 |
+
"eval_steps_per_second": 0.95,
|
| 341 |
+
"step": 1530
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 17.22,
|
| 345 |
+
"learning_rate": 5.654320987654321e-05,
|
| 346 |
+
"loss": 0.8497,
|
| 347 |
+
"step": 1550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 17.78,
|
| 351 |
+
"learning_rate": 5.477954144620812e-05,
|
| 352 |
+
"loss": 0.7886,
|
| 353 |
+
"step": 1600
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 18.0,
|
| 357 |
+
"eval_accuracy": 0.750122970978849,
|
| 358 |
+
"eval_loss": 1.2826693058013916,
|
| 359 |
+
"eval_runtime": 16.966,
|
| 360 |
+
"eval_samples_per_second": 119.828,
|
| 361 |
+
"eval_steps_per_second": 0.943,
|
| 362 |
+
"step": 1620
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 18.33,
|
| 366 |
+
"learning_rate": 5.301587301587302e-05,
|
| 367 |
+
"loss": 0.7637,
|
| 368 |
+
"step": 1650
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 18.89,
|
| 372 |
+
"learning_rate": 5.125220458553792e-05,
|
| 373 |
+
"loss": 0.7191,
|
| 374 |
+
"step": 1700
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 19.0,
|
| 378 |
+
"eval_accuracy": 0.749139203148057,
|
| 379 |
+
"eval_loss": 1.3038619756698608,
|
| 380 |
+
"eval_runtime": 16.7983,
|
| 381 |
+
"eval_samples_per_second": 121.024,
|
| 382 |
+
"eval_steps_per_second": 0.952,
|
| 383 |
+
"step": 1710
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 19.44,
|
| 387 |
+
"learning_rate": 4.948853615520283e-05,
|
| 388 |
+
"loss": 0.6952,
|
| 389 |
+
"step": 1750
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 20.0,
|
| 393 |
+
"learning_rate": 4.772486772486773e-05,
|
| 394 |
+
"loss": 0.609,
|
| 395 |
+
"step": 1800
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 20.0,
|
| 399 |
+
"eval_accuracy": 0.7594687653713723,
|
| 400 |
+
"eval_loss": 1.2997246980667114,
|
| 401 |
+
"eval_runtime": 16.7476,
|
| 402 |
+
"eval_samples_per_second": 121.39,
|
| 403 |
+
"eval_steps_per_second": 0.955,
|
| 404 |
+
"step": 1800
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 20.56,
|
| 408 |
+
"learning_rate": 4.596119929453263e-05,
|
| 409 |
+
"loss": 0.5703,
|
| 410 |
+
"step": 1850
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 21.0,
|
| 414 |
+
"eval_accuracy": 0.76586325627152,
|
| 415 |
+
"eval_loss": 1.2955206632614136,
|
| 416 |
+
"eval_runtime": 16.7956,
|
| 417 |
+
"eval_samples_per_second": 121.044,
|
| 418 |
+
"eval_steps_per_second": 0.953,
|
| 419 |
+
"step": 1890
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 21.11,
|
| 423 |
+
"learning_rate": 4.4197530864197535e-05,
|
| 424 |
+
"loss": 0.5608,
|
| 425 |
+
"step": 1900
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 21.67,
|
| 429 |
+
"learning_rate": 4.2433862433862435e-05,
|
| 430 |
+
"loss": 0.5251,
|
| 431 |
+
"step": 1950
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 22.0,
|
| 435 |
+
"eval_accuracy": 0.7614363010329562,
|
| 436 |
+
"eval_loss": 1.3140268325805664,
|
| 437 |
+
"eval_runtime": 16.7643,
|
| 438 |
+
"eval_samples_per_second": 121.27,
|
| 439 |
+
"eval_steps_per_second": 0.954,
|
| 440 |
+
"step": 1980
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 22.22,
|
| 444 |
+
"learning_rate": 4.067019400352734e-05,
|
| 445 |
+
"loss": 0.5351,
|
| 446 |
+
"step": 2000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 22.78,
|
| 450 |
+
"learning_rate": 3.890652557319224e-05,
|
| 451 |
+
"loss": 0.5041,
|
| 452 |
+
"step": 2050
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 23.0,
|
| 456 |
+
"eval_accuracy": 0.7668470241023119,
|
| 457 |
+
"eval_loss": 1.319765567779541,
|
| 458 |
+
"eval_runtime": 16.8549,
|
| 459 |
+
"eval_samples_per_second": 120.618,
|
| 460 |
+
"eval_steps_per_second": 0.949,
|
| 461 |
+
"step": 2070
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 23.33,
|
| 465 |
+
"learning_rate": 3.717813051146385e-05,
|
| 466 |
+
"loss": 0.435,
|
| 467 |
+
"step": 2100
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 23.89,
|
| 471 |
+
"learning_rate": 3.5414462081128755e-05,
|
| 472 |
+
"loss": 0.4455,
|
| 473 |
+
"step": 2150
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 24.0,
|
| 477 |
+
"eval_accuracy": 0.7816035415641909,
|
| 478 |
+
"eval_loss": 1.3251405954360962,
|
| 479 |
+
"eval_runtime": 16.7377,
|
| 480 |
+
"eval_samples_per_second": 121.462,
|
| 481 |
+
"eval_steps_per_second": 0.956,
|
| 482 |
+
"step": 2160
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 24.44,
|
| 486 |
+
"learning_rate": 3.3650793650793656e-05,
|
| 487 |
+
"loss": 0.4152,
|
| 488 |
+
"step": 2200
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 25.0,
|
| 492 |
+
"learning_rate": 3.1887125220458556e-05,
|
| 493 |
+
"loss": 0.4046,
|
| 494 |
+
"step": 2250
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 25.0,
|
| 498 |
+
"eval_accuracy": 0.7825873093949828,
|
| 499 |
+
"eval_loss": 1.3233106136322021,
|
| 500 |
+
"eval_runtime": 16.7729,
|
| 501 |
+
"eval_samples_per_second": 121.207,
|
| 502 |
+
"eval_steps_per_second": 0.954,
|
| 503 |
+
"step": 2250
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 25.56,
|
| 507 |
+
"learning_rate": 3.012345679012346e-05,
|
| 508 |
+
"loss": 0.3861,
|
| 509 |
+
"step": 2300
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 26.0,
|
| 513 |
+
"eval_accuracy": 0.7737333989178554,
|
| 514 |
+
"eval_loss": 1.3713818788528442,
|
| 515 |
+
"eval_runtime": 16.8666,
|
| 516 |
+
"eval_samples_per_second": 120.534,
|
| 517 |
+
"eval_steps_per_second": 0.949,
|
| 518 |
+
"step": 2340
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 26.11,
|
| 522 |
+
"learning_rate": 2.8359788359788357e-05,
|
| 523 |
+
"loss": 0.3571,
|
| 524 |
+
"step": 2350
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 26.67,
|
| 528 |
+
"learning_rate": 2.659611992945326e-05,
|
| 529 |
+
"loss": 0.3798,
|
| 530 |
+
"step": 2400
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 27.0,
|
| 534 |
+
"eval_accuracy": 0.7747171667486473,
|
| 535 |
+
"eval_loss": 1.3577079772949219,
|
| 536 |
+
"eval_runtime": 16.7848,
|
| 537 |
+
"eval_samples_per_second": 121.122,
|
| 538 |
+
"eval_steps_per_second": 0.953,
|
| 539 |
+
"step": 2430
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 27.22,
|
| 543 |
+
"learning_rate": 2.4832451499118168e-05,
|
| 544 |
+
"loss": 0.3686,
|
| 545 |
+
"step": 2450
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 27.78,
|
| 549 |
+
"learning_rate": 2.3068783068783072e-05,
|
| 550 |
+
"loss": 0.3402,
|
| 551 |
+
"step": 2500
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 28.0,
|
| 555 |
+
"eval_accuracy": 0.7757009345794392,
|
| 556 |
+
"eval_loss": 1.3933676481246948,
|
| 557 |
+
"eval_runtime": 16.7209,
|
| 558 |
+
"eval_samples_per_second": 121.585,
|
| 559 |
+
"eval_steps_per_second": 0.957,
|
| 560 |
+
"step": 2520
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 28.33,
|
| 564 |
+
"learning_rate": 2.1305114638447972e-05,
|
| 565 |
+
"loss": 0.3556,
|
| 566 |
+
"step": 2550
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 28.89,
|
| 570 |
+
"learning_rate": 1.9541446208112876e-05,
|
| 571 |
+
"loss": 0.2951,
|
| 572 |
+
"step": 2600
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 29.0,
|
| 576 |
+
"eval_accuracy": 0.779636005902607,
|
| 577 |
+
"eval_loss": 1.3655041456222534,
|
| 578 |
+
"eval_runtime": 16.7839,
|
| 579 |
+
"eval_samples_per_second": 121.128,
|
| 580 |
+
"eval_steps_per_second": 0.953,
|
| 581 |
+
"step": 2610
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 29.44,
|
| 585 |
+
"learning_rate": 1.777777777777778e-05,
|
| 586 |
+
"loss": 0.3126,
|
| 587 |
+
"step": 2650
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 30.0,
|
| 591 |
+
"learning_rate": 1.6014109347442684e-05,
|
| 592 |
+
"loss": 0.2953,
|
| 593 |
+
"step": 2700
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 30.0,
|
| 597 |
+
"eval_accuracy": 0.7776684702410231,
|
| 598 |
+
"eval_loss": 1.3812841176986694,
|
| 599 |
+
"eval_runtime": 16.7955,
|
| 600 |
+
"eval_samples_per_second": 121.044,
|
| 601 |
+
"eval_steps_per_second": 0.953,
|
| 602 |
+
"step": 2700
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 30.56,
|
| 606 |
+
"learning_rate": 1.4250440917107586e-05,
|
| 607 |
+
"loss": 0.3125,
|
| 608 |
+
"step": 2750
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 31.0,
|
| 612 |
+
"eval_accuracy": 0.7855386128873586,
|
| 613 |
+
"eval_loss": 1.3787145614624023,
|
| 614 |
+
"eval_runtime": 16.8143,
|
| 615 |
+
"eval_samples_per_second": 120.909,
|
| 616 |
+
"eval_steps_per_second": 0.952,
|
| 617 |
+
"step": 2790
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 31.11,
|
| 621 |
+
"learning_rate": 1.2486772486772486e-05,
|
| 622 |
+
"loss": 0.2608,
|
| 623 |
+
"step": 2800
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 31.67,
|
| 627 |
+
"learning_rate": 1.072310405643739e-05,
|
| 628 |
+
"loss": 0.2564,
|
| 629 |
+
"step": 2850
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 32.0,
|
| 633 |
+
"eval_accuracy": 0.778652238071815,
|
| 634 |
+
"eval_loss": 1.3935987949371338,
|
| 635 |
+
"eval_runtime": 16.7896,
|
| 636 |
+
"eval_samples_per_second": 121.087,
|
| 637 |
+
"eval_steps_per_second": 0.953,
|
| 638 |
+
"step": 2880
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 32.22,
|
| 642 |
+
"learning_rate": 8.959435626102292e-06,
|
| 643 |
+
"loss": 0.2534,
|
| 644 |
+
"step": 2900
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 32.78,
|
| 648 |
+
"learning_rate": 7.195767195767196e-06,
|
| 649 |
+
"loss": 0.2427,
|
| 650 |
+
"step": 2950
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 33.0,
|
| 654 |
+
"eval_accuracy": 0.7855386128873586,
|
| 655 |
+
"eval_loss": 1.388238787651062,
|
| 656 |
+
"eval_runtime": 16.8064,
|
| 657 |
+
"eval_samples_per_second": 120.966,
|
| 658 |
+
"eval_steps_per_second": 0.952,
|
| 659 |
+
"step": 2970
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 33.33,
|
| 663 |
+
"learning_rate": 5.432098765432099e-06,
|
| 664 |
+
"loss": 0.2691,
|
| 665 |
+
"step": 3000
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 33.89,
|
| 669 |
+
"learning_rate": 3.668430335097002e-06,
|
| 670 |
+
"loss": 0.279,
|
| 671 |
+
"step": 3050
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 34.0,
|
| 675 |
+
"eval_accuracy": 0.7855386128873586,
|
| 676 |
+
"eval_loss": 1.4113564491271973,
|
| 677 |
+
"eval_runtime": 16.7799,
|
| 678 |
+
"eval_samples_per_second": 121.157,
|
| 679 |
+
"eval_steps_per_second": 0.954,
|
| 680 |
+
"step": 3060
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 34.44,
|
| 684 |
+
"learning_rate": 1.9047619047619051e-06,
|
| 685 |
+
"loss": 0.2582,
|
| 686 |
+
"step": 3100
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 35.0,
|
| 690 |
+
"learning_rate": 1.4109347442680778e-07,
|
| 691 |
+
"loss": 0.2308,
|
| 692 |
+
"step": 3150
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 35.0,
|
| 696 |
+
"eval_accuracy": 0.7855386128873586,
|
| 697 |
+
"eval_loss": 1.4130984544754028,
|
| 698 |
+
"eval_runtime": 16.8643,
|
| 699 |
+
"eval_samples_per_second": 120.551,
|
| 700 |
+
"eval_steps_per_second": 0.949,
|
| 701 |
+
"step": 3150
|
| 702 |
+
}
|
| 703 |
+
],
|
| 704 |
+
"max_steps": 3150,
|
| 705 |
+
"num_train_epochs": 35,
|
| 706 |
+
"total_flos": 1.221413841164159e+20,
|
| 707 |
+
"trial_name": null,
|
| 708 |
+
"trial_params": null
|
| 709 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cea5e706194b99364d662769e11fa6178dfd76cfa9cf388f0bd4af214cf6ce92
|
| 3 |
+
size 3515
|