Training in progress, step 11000, checkpoint
Browse files- checkpoint-11000/config.json +65 -0
- checkpoint-11000/generation_config.json +265 -0
- checkpoint-11000/model.safetensors +3 -0
- checkpoint-11000/optimizer.pt +3 -0
- checkpoint-11000/preprocessor_config.json +14 -0
- checkpoint-11000/rng_state.pth +3 -0
- checkpoint-11000/scheduler.pt +3 -0
- checkpoint-11000/trainer_state.json +3299 -0
- checkpoint-11000/training_args.bin +3 -0
checkpoint-11000/config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/kaggle/input/whisper-small-final/whisper-small-final-v2",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"WhisperForConditionalGeneration"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"begin_suppress_tokens": [
|
| 11 |
+
220,
|
| 12 |
+
50257
|
| 13 |
+
],
|
| 14 |
+
"bos_token_id": 50257,
|
| 15 |
+
"classifier_proj_size": 256,
|
| 16 |
+
"d_model": 768,
|
| 17 |
+
"decoder_attention_heads": 12,
|
| 18 |
+
"decoder_ffn_dim": 3072,
|
| 19 |
+
"decoder_layerdrop": 0.0,
|
| 20 |
+
"decoder_layers": 12,
|
| 21 |
+
"decoder_start_token_id": 50258,
|
| 22 |
+
"dropout": 0.0,
|
| 23 |
+
"encoder_attention_heads": 12,
|
| 24 |
+
"encoder_ffn_dim": 3072,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 12,
|
| 27 |
+
"eos_token_id": 50257,
|
| 28 |
+
"forced_decoder_ids": [
|
| 29 |
+
[
|
| 30 |
+
1,
|
| 31 |
+
50272
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
2,
|
| 35 |
+
50359
|
| 36 |
+
],
|
| 37 |
+
[
|
| 38 |
+
3,
|
| 39 |
+
50363
|
| 40 |
+
]
|
| 41 |
+
],
|
| 42 |
+
"init_std": 0.02,
|
| 43 |
+
"is_encoder_decoder": true,
|
| 44 |
+
"mask_feature_length": 10,
|
| 45 |
+
"mask_feature_min_masks": 0,
|
| 46 |
+
"mask_feature_prob": 0.0,
|
| 47 |
+
"mask_time_length": 10,
|
| 48 |
+
"mask_time_min_masks": 2,
|
| 49 |
+
"mask_time_prob": 0.05,
|
| 50 |
+
"max_length": 448,
|
| 51 |
+
"max_source_positions": 1500,
|
| 52 |
+
"max_target_positions": 448,
|
| 53 |
+
"median_filter_width": 7,
|
| 54 |
+
"model_type": "whisper",
|
| 55 |
+
"num_hidden_layers": 12,
|
| 56 |
+
"num_mel_bins": 80,
|
| 57 |
+
"pad_token_id": 50257,
|
| 58 |
+
"scale_embedding": false,
|
| 59 |
+
"suppress_tokens": [],
|
| 60 |
+
"torch_dtype": "float32",
|
| 61 |
+
"transformers_version": "4.40.0.dev0",
|
| 62 |
+
"use_cache": true,
|
| 63 |
+
"use_weighted_layer_sum": false,
|
| 64 |
+
"vocab_size": 51865
|
| 65 |
+
}
|
checkpoint-11000/generation_config.json
ADDED
|
@@ -0,0 +1,265 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alignment_heads": [
|
| 3 |
+
[
|
| 4 |
+
5,
|
| 5 |
+
3
|
| 6 |
+
],
|
| 7 |
+
[
|
| 8 |
+
5,
|
| 9 |
+
9
|
| 10 |
+
],
|
| 11 |
+
[
|
| 12 |
+
8,
|
| 13 |
+
0
|
| 14 |
+
],
|
| 15 |
+
[
|
| 16 |
+
8,
|
| 17 |
+
4
|
| 18 |
+
],
|
| 19 |
+
[
|
| 20 |
+
8,
|
| 21 |
+
7
|
| 22 |
+
],
|
| 23 |
+
[
|
| 24 |
+
8,
|
| 25 |
+
8
|
| 26 |
+
],
|
| 27 |
+
[
|
| 28 |
+
9,
|
| 29 |
+
0
|
| 30 |
+
],
|
| 31 |
+
[
|
| 32 |
+
9,
|
| 33 |
+
7
|
| 34 |
+
],
|
| 35 |
+
[
|
| 36 |
+
9,
|
| 37 |
+
9
|
| 38 |
+
],
|
| 39 |
+
[
|
| 40 |
+
10,
|
| 41 |
+
5
|
| 42 |
+
]
|
| 43 |
+
],
|
| 44 |
+
"begin_suppress_tokens": [
|
| 45 |
+
220,
|
| 46 |
+
50257
|
| 47 |
+
],
|
| 48 |
+
"bos_token_id": 50257,
|
| 49 |
+
"decoder_start_token_id": 50258,
|
| 50 |
+
"eos_token_id": 50257,
|
| 51 |
+
"forced_decoder_ids": [
|
| 52 |
+
[
|
| 53 |
+
1,
|
| 54 |
+
null
|
| 55 |
+
],
|
| 56 |
+
[
|
| 57 |
+
2,
|
| 58 |
+
50359
|
| 59 |
+
]
|
| 60 |
+
],
|
| 61 |
+
"is_multilingual": true,
|
| 62 |
+
"lang_to_id": {
|
| 63 |
+
"<|af|>": 50327,
|
| 64 |
+
"<|am|>": 50334,
|
| 65 |
+
"<|ar|>": 50272,
|
| 66 |
+
"<|as|>": 50350,
|
| 67 |
+
"<|az|>": 50304,
|
| 68 |
+
"<|ba|>": 50355,
|
| 69 |
+
"<|be|>": 50330,
|
| 70 |
+
"<|bg|>": 50292,
|
| 71 |
+
"<|bn|>": 50302,
|
| 72 |
+
"<|bo|>": 50347,
|
| 73 |
+
"<|br|>": 50309,
|
| 74 |
+
"<|bs|>": 50315,
|
| 75 |
+
"<|ca|>": 50270,
|
| 76 |
+
"<|cs|>": 50283,
|
| 77 |
+
"<|cy|>": 50297,
|
| 78 |
+
"<|da|>": 50285,
|
| 79 |
+
"<|de|>": 50261,
|
| 80 |
+
"<|el|>": 50281,
|
| 81 |
+
"<|en|>": 50259,
|
| 82 |
+
"<|es|>": 50262,
|
| 83 |
+
"<|et|>": 50307,
|
| 84 |
+
"<|eu|>": 50310,
|
| 85 |
+
"<|fa|>": 50300,
|
| 86 |
+
"<|fi|>": 50277,
|
| 87 |
+
"<|fo|>": 50338,
|
| 88 |
+
"<|fr|>": 50265,
|
| 89 |
+
"<|gl|>": 50319,
|
| 90 |
+
"<|gu|>": 50333,
|
| 91 |
+
"<|haw|>": 50352,
|
| 92 |
+
"<|ha|>": 50354,
|
| 93 |
+
"<|he|>": 50279,
|
| 94 |
+
"<|hi|>": 50276,
|
| 95 |
+
"<|hr|>": 50291,
|
| 96 |
+
"<|ht|>": 50339,
|
| 97 |
+
"<|hu|>": 50286,
|
| 98 |
+
"<|hy|>": 50312,
|
| 99 |
+
"<|id|>": 50275,
|
| 100 |
+
"<|is|>": 50311,
|
| 101 |
+
"<|it|>": 50274,
|
| 102 |
+
"<|ja|>": 50266,
|
| 103 |
+
"<|jw|>": 50356,
|
| 104 |
+
"<|ka|>": 50329,
|
| 105 |
+
"<|kk|>": 50316,
|
| 106 |
+
"<|km|>": 50323,
|
| 107 |
+
"<|kn|>": 50306,
|
| 108 |
+
"<|ko|>": 50264,
|
| 109 |
+
"<|la|>": 50294,
|
| 110 |
+
"<|lb|>": 50345,
|
| 111 |
+
"<|ln|>": 50353,
|
| 112 |
+
"<|lo|>": 50336,
|
| 113 |
+
"<|lt|>": 50293,
|
| 114 |
+
"<|lv|>": 50301,
|
| 115 |
+
"<|mg|>": 50349,
|
| 116 |
+
"<|mi|>": 50295,
|
| 117 |
+
"<|mk|>": 50308,
|
| 118 |
+
"<|ml|>": 50296,
|
| 119 |
+
"<|mn|>": 50314,
|
| 120 |
+
"<|mr|>": 50320,
|
| 121 |
+
"<|ms|>": 50282,
|
| 122 |
+
"<|mt|>": 50343,
|
| 123 |
+
"<|my|>": 50346,
|
| 124 |
+
"<|ne|>": 50313,
|
| 125 |
+
"<|nl|>": 50271,
|
| 126 |
+
"<|nn|>": 50342,
|
| 127 |
+
"<|no|>": 50288,
|
| 128 |
+
"<|oc|>": 50328,
|
| 129 |
+
"<|pa|>": 50321,
|
| 130 |
+
"<|pl|>": 50269,
|
| 131 |
+
"<|ps|>": 50340,
|
| 132 |
+
"<|pt|>": 50267,
|
| 133 |
+
"<|ro|>": 50284,
|
| 134 |
+
"<|ru|>": 50263,
|
| 135 |
+
"<|sa|>": 50344,
|
| 136 |
+
"<|sd|>": 50332,
|
| 137 |
+
"<|si|>": 50322,
|
| 138 |
+
"<|sk|>": 50298,
|
| 139 |
+
"<|sl|>": 50305,
|
| 140 |
+
"<|sn|>": 50324,
|
| 141 |
+
"<|so|>": 50326,
|
| 142 |
+
"<|sq|>": 50317,
|
| 143 |
+
"<|sr|>": 50303,
|
| 144 |
+
"<|su|>": 50357,
|
| 145 |
+
"<|sv|>": 50273,
|
| 146 |
+
"<|sw|>": 50318,
|
| 147 |
+
"<|ta|>": 50287,
|
| 148 |
+
"<|te|>": 50299,
|
| 149 |
+
"<|tg|>": 50331,
|
| 150 |
+
"<|th|>": 50289,
|
| 151 |
+
"<|tk|>": 50341,
|
| 152 |
+
"<|tl|>": 50348,
|
| 153 |
+
"<|tr|>": 50268,
|
| 154 |
+
"<|tt|>": 50351,
|
| 155 |
+
"<|uk|>": 50280,
|
| 156 |
+
"<|ur|>": 50290,
|
| 157 |
+
"<|uz|>": 50337,
|
| 158 |
+
"<|vi|>": 50278,
|
| 159 |
+
"<|yi|>": 50335,
|
| 160 |
+
"<|yo|>": 50325,
|
| 161 |
+
"<|zh|>": 50260
|
| 162 |
+
},
|
| 163 |
+
"language": "ar",
|
| 164 |
+
"max_initial_timestamp_index": 50,
|
| 165 |
+
"max_length": 448,
|
| 166 |
+
"no_timestamps_token_id": 50363,
|
| 167 |
+
"pad_token_id": 50257,
|
| 168 |
+
"prev_sot_token_id": 50361,
|
| 169 |
+
"return_timestamps": false,
|
| 170 |
+
"suppress_tokens": [
|
| 171 |
+
1,
|
| 172 |
+
2,
|
| 173 |
+
7,
|
| 174 |
+
8,
|
| 175 |
+
9,
|
| 176 |
+
10,
|
| 177 |
+
14,
|
| 178 |
+
25,
|
| 179 |
+
26,
|
| 180 |
+
27,
|
| 181 |
+
28,
|
| 182 |
+
29,
|
| 183 |
+
31,
|
| 184 |
+
58,
|
| 185 |
+
59,
|
| 186 |
+
60,
|
| 187 |
+
61,
|
| 188 |
+
62,
|
| 189 |
+
63,
|
| 190 |
+
90,
|
| 191 |
+
91,
|
| 192 |
+
92,
|
| 193 |
+
93,
|
| 194 |
+
359,
|
| 195 |
+
503,
|
| 196 |
+
522,
|
| 197 |
+
542,
|
| 198 |
+
873,
|
| 199 |
+
893,
|
| 200 |
+
902,
|
| 201 |
+
918,
|
| 202 |
+
922,
|
| 203 |
+
931,
|
| 204 |
+
1350,
|
| 205 |
+
1853,
|
| 206 |
+
1982,
|
| 207 |
+
2460,
|
| 208 |
+
2627,
|
| 209 |
+
3246,
|
| 210 |
+
3253,
|
| 211 |
+
3268,
|
| 212 |
+
3536,
|
| 213 |
+
3846,
|
| 214 |
+
3961,
|
| 215 |
+
4183,
|
| 216 |
+
4667,
|
| 217 |
+
6585,
|
| 218 |
+
6647,
|
| 219 |
+
7273,
|
| 220 |
+
9061,
|
| 221 |
+
9383,
|
| 222 |
+
10428,
|
| 223 |
+
10929,
|
| 224 |
+
11938,
|
| 225 |
+
12033,
|
| 226 |
+
12331,
|
| 227 |
+
12562,
|
| 228 |
+
13793,
|
| 229 |
+
14157,
|
| 230 |
+
14635,
|
| 231 |
+
15265,
|
| 232 |
+
15618,
|
| 233 |
+
16553,
|
| 234 |
+
16604,
|
| 235 |
+
18362,
|
| 236 |
+
18956,
|
| 237 |
+
20075,
|
| 238 |
+
21675,
|
| 239 |
+
22520,
|
| 240 |
+
26130,
|
| 241 |
+
26161,
|
| 242 |
+
26435,
|
| 243 |
+
28279,
|
| 244 |
+
29464,
|
| 245 |
+
31650,
|
| 246 |
+
32302,
|
| 247 |
+
32470,
|
| 248 |
+
36865,
|
| 249 |
+
42863,
|
| 250 |
+
47425,
|
| 251 |
+
49870,
|
| 252 |
+
50254,
|
| 253 |
+
50258,
|
| 254 |
+
50358,
|
| 255 |
+
50359,
|
| 256 |
+
50360,
|
| 257 |
+
50361,
|
| 258 |
+
50362
|
| 259 |
+
],
|
| 260 |
+
"task_to_id": {
|
| 261 |
+
"transcribe": 50359,
|
| 262 |
+
"translate": 50358
|
| 263 |
+
},
|
| 264 |
+
"transformers_version": "4.40.0.dev0"
|
| 265 |
+
}
|
checkpoint-11000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5ed9f4fb251c3f8113ebbc6d540a375af71e626f9e20161425d756a0f2a48ccc
|
| 3 |
+
size 966995080
|
checkpoint-11000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1e098551f9ad5c50c8a5ef6fdd685b08d0aee2ef3a254d45ad88d20a22bed0b1
|
| 3 |
+
size 1925064044
|
checkpoint-11000/preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
checkpoint-11000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d82582023ea966a9634fc21e23be31422ee78df1a62d7de2a14c4c5d1f89543
|
| 3 |
+
size 14244
|
checkpoint-11000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:debad3b978d9b4e201fe71bea07cfbc9668a5e965bb98d0a03f31b8f2144374d
|
| 3 |
+
size 1064
|
checkpoint-11000/trainer_state.json
ADDED
|
@@ -0,0 +1,3299 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 2.9224686264397457,
|
| 3 |
+
"best_model_checkpoint": "./whisper-small-final-v2/checkpoint-5000",
|
| 4 |
+
"epoch": 35.59870550161812,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 11000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.08,
|
| 13 |
+
"grad_norm": 0.0023542253766208887,
|
| 14 |
+
"learning_rate": 5.000000000000001e-07,
|
| 15 |
+
"loss": 0.0001,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.16,
|
| 20 |
+
"grad_norm": 0.003053858410567045,
|
| 21 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 22 |
+
"loss": 0.0001,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.24,
|
| 27 |
+
"grad_norm": 0.007697729859501123,
|
| 28 |
+
"learning_rate": 1.5e-06,
|
| 29 |
+
"loss": 0.0001,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.32,
|
| 34 |
+
"grad_norm": 0.01041096169501543,
|
| 35 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 36 |
+
"loss": 0.0,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.4,
|
| 41 |
+
"grad_norm": 0.0023387186229228973,
|
| 42 |
+
"learning_rate": 2.5e-06,
|
| 43 |
+
"loss": 0.0,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.49,
|
| 48 |
+
"grad_norm": 0.0036297033075243235,
|
| 49 |
+
"learning_rate": 3e-06,
|
| 50 |
+
"loss": 0.0001,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.57,
|
| 55 |
+
"grad_norm": 0.0016004816861823201,
|
| 56 |
+
"learning_rate": 3.5e-06,
|
| 57 |
+
"loss": 0.0,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.65,
|
| 62 |
+
"grad_norm": 0.0035209895577281713,
|
| 63 |
+
"learning_rate": 4.000000000000001e-06,
|
| 64 |
+
"loss": 0.0,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.73,
|
| 69 |
+
"grad_norm": 0.37100252509117126,
|
| 70 |
+
"learning_rate": 4.5e-06,
|
| 71 |
+
"loss": 0.0015,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.81,
|
| 76 |
+
"grad_norm": 0.2593393325805664,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 0.003,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.89,
|
| 83 |
+
"grad_norm": 1.1273244619369507,
|
| 84 |
+
"learning_rate": 5.500000000000001e-06,
|
| 85 |
+
"loss": 0.003,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.97,
|
| 90 |
+
"grad_norm": 0.11678028106689453,
|
| 91 |
+
"learning_rate": 6e-06,
|
| 92 |
+
"loss": 0.0014,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 1.05,
|
| 97 |
+
"grad_norm": 2.602248191833496,
|
| 98 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 99 |
+
"loss": 0.0025,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 1.13,
|
| 104 |
+
"grad_norm": 1.053270697593689,
|
| 105 |
+
"learning_rate": 7e-06,
|
| 106 |
+
"loss": 0.0021,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.21,
|
| 111 |
+
"grad_norm": 0.25527653098106384,
|
| 112 |
+
"learning_rate": 7.500000000000001e-06,
|
| 113 |
+
"loss": 0.0037,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.29,
|
| 118 |
+
"grad_norm": 0.208439901471138,
|
| 119 |
+
"learning_rate": 8.000000000000001e-06,
|
| 120 |
+
"loss": 0.002,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.38,
|
| 125 |
+
"grad_norm": 0.14851850271224976,
|
| 126 |
+
"learning_rate": 8.5e-06,
|
| 127 |
+
"loss": 0.0034,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.46,
|
| 132 |
+
"grad_norm": 1.6750054359436035,
|
| 133 |
+
"learning_rate": 9e-06,
|
| 134 |
+
"loss": 0.0046,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 1.54,
|
| 139 |
+
"grad_norm": 4.503875255584717,
|
| 140 |
+
"learning_rate": 9.48e-06,
|
| 141 |
+
"loss": 0.0093,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 1.62,
|
| 146 |
+
"grad_norm": 0.6791089177131653,
|
| 147 |
+
"learning_rate": 9.980000000000001e-06,
|
| 148 |
+
"loss": 0.0073,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.62,
|
| 153 |
+
"eval_loss": 0.024947550147771835,
|
| 154 |
+
"eval_runtime": 794.9124,
|
| 155 |
+
"eval_samples_per_second": 1.551,
|
| 156 |
+
"eval_steps_per_second": 0.195,
|
| 157 |
+
"eval_wer": 5.002578648788035,
|
| 158 |
+
"step": 500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.7,
|
| 162 |
+
"grad_norm": 0.7141709923744202,
|
| 163 |
+
"learning_rate": 9.931428571428571e-06,
|
| 164 |
+
"loss": 0.0022,
|
| 165 |
+
"step": 525
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.78,
|
| 169 |
+
"grad_norm": 1.1480615139007568,
|
| 170 |
+
"learning_rate": 9.86e-06,
|
| 171 |
+
"loss": 0.0036,
|
| 172 |
+
"step": 550
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 1.86,
|
| 176 |
+
"grad_norm": 1.9711138010025024,
|
| 177 |
+
"learning_rate": 9.78857142857143e-06,
|
| 178 |
+
"loss": 0.0043,
|
| 179 |
+
"step": 575
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.94,
|
| 183 |
+
"grad_norm": 1.0525217056274414,
|
| 184 |
+
"learning_rate": 9.717142857142858e-06,
|
| 185 |
+
"loss": 0.0051,
|
| 186 |
+
"step": 600
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.02,
|
| 190 |
+
"grad_norm": 0.15985222160816193,
|
| 191 |
+
"learning_rate": 9.645714285714286e-06,
|
| 192 |
+
"loss": 0.0057,
|
| 193 |
+
"step": 625
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 2.1,
|
| 197 |
+
"grad_norm": 0.8261871933937073,
|
| 198 |
+
"learning_rate": 9.574285714285715e-06,
|
| 199 |
+
"loss": 0.0023,
|
| 200 |
+
"step": 650
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.18,
|
| 204 |
+
"grad_norm": 0.04620926454663277,
|
| 205 |
+
"learning_rate": 9.502857142857144e-06,
|
| 206 |
+
"loss": 0.0019,
|
| 207 |
+
"step": 675
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 2.27,
|
| 211 |
+
"grad_norm": 0.3023536503314972,
|
| 212 |
+
"learning_rate": 9.431428571428573e-06,
|
| 213 |
+
"loss": 0.0012,
|
| 214 |
+
"step": 700
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 2.35,
|
| 218 |
+
"grad_norm": 0.11112522333860397,
|
| 219 |
+
"learning_rate": 9.360000000000002e-06,
|
| 220 |
+
"loss": 0.0024,
|
| 221 |
+
"step": 725
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 2.43,
|
| 225 |
+
"grad_norm": 1.6971714496612549,
|
| 226 |
+
"learning_rate": 9.28857142857143e-06,
|
| 227 |
+
"loss": 0.0031,
|
| 228 |
+
"step": 750
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 2.51,
|
| 232 |
+
"grad_norm": 0.22176721692085266,
|
| 233 |
+
"learning_rate": 9.217142857142858e-06,
|
| 234 |
+
"loss": 0.0026,
|
| 235 |
+
"step": 775
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 2.59,
|
| 239 |
+
"grad_norm": 0.2839481830596924,
|
| 240 |
+
"learning_rate": 9.145714285714287e-06,
|
| 241 |
+
"loss": 0.0016,
|
| 242 |
+
"step": 800
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 2.67,
|
| 246 |
+
"grad_norm": 0.6732001900672913,
|
| 247 |
+
"learning_rate": 9.074285714285716e-06,
|
| 248 |
+
"loss": 0.0044,
|
| 249 |
+
"step": 825
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.75,
|
| 253 |
+
"grad_norm": 2.15289044380188,
|
| 254 |
+
"learning_rate": 9.002857142857144e-06,
|
| 255 |
+
"loss": 0.0028,
|
| 256 |
+
"step": 850
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.83,
|
| 260 |
+
"grad_norm": 0.4701692759990692,
|
| 261 |
+
"learning_rate": 8.931428571428573e-06,
|
| 262 |
+
"loss": 0.0009,
|
| 263 |
+
"step": 875
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 2.91,
|
| 267 |
+
"grad_norm": 0.38268008828163147,
|
| 268 |
+
"learning_rate": 8.860000000000002e-06,
|
| 269 |
+
"loss": 0.0012,
|
| 270 |
+
"step": 900
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 2.99,
|
| 274 |
+
"grad_norm": 0.057314854115247726,
|
| 275 |
+
"learning_rate": 8.788571428571429e-06,
|
| 276 |
+
"loss": 0.001,
|
| 277 |
+
"step": 925
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 3.07,
|
| 281 |
+
"grad_norm": 0.048205479979515076,
|
| 282 |
+
"learning_rate": 8.717142857142858e-06,
|
| 283 |
+
"loss": 0.0022,
|
| 284 |
+
"step": 950
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 3.16,
|
| 288 |
+
"grad_norm": 0.008282452821731567,
|
| 289 |
+
"learning_rate": 8.645714285714287e-06,
|
| 290 |
+
"loss": 0.0006,
|
| 291 |
+
"step": 975
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 3.24,
|
| 295 |
+
"grad_norm": 1.3108314275741577,
|
| 296 |
+
"learning_rate": 8.574285714285714e-06,
|
| 297 |
+
"loss": 0.0014,
|
| 298 |
+
"step": 1000
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 3.24,
|
| 302 |
+
"eval_loss": 0.02143564447760582,
|
| 303 |
+
"eval_runtime": 802.0669,
|
| 304 |
+
"eval_samples_per_second": 1.537,
|
| 305 |
+
"eval_steps_per_second": 0.193,
|
| 306 |
+
"eval_wer": 4.108647068935878,
|
| 307 |
+
"step": 1000
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 3.32,
|
| 311 |
+
"grad_norm": 0.888531506061554,
|
| 312 |
+
"learning_rate": 8.502857142857143e-06,
|
| 313 |
+
"loss": 0.0025,
|
| 314 |
+
"step": 1025
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 3.4,
|
| 318 |
+
"grad_norm": 0.013467811048030853,
|
| 319 |
+
"learning_rate": 8.431428571428572e-06,
|
| 320 |
+
"loss": 0.0009,
|
| 321 |
+
"step": 1050
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 3.48,
|
| 325 |
+
"grad_norm": 1.272592306137085,
|
| 326 |
+
"learning_rate": 8.36e-06,
|
| 327 |
+
"loss": 0.001,
|
| 328 |
+
"step": 1075
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 3.56,
|
| 332 |
+
"grad_norm": 0.022731788456439972,
|
| 333 |
+
"learning_rate": 8.288571428571429e-06,
|
| 334 |
+
"loss": 0.0002,
|
| 335 |
+
"step": 1100
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 3.64,
|
| 339 |
+
"grad_norm": 0.49200737476348877,
|
| 340 |
+
"learning_rate": 8.217142857142858e-06,
|
| 341 |
+
"loss": 0.0011,
|
| 342 |
+
"step": 1125
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 3.72,
|
| 346 |
+
"grad_norm": 0.011444230563938618,
|
| 347 |
+
"learning_rate": 8.145714285714287e-06,
|
| 348 |
+
"loss": 0.0005,
|
| 349 |
+
"step": 1150
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 3.8,
|
| 353 |
+
"grad_norm": 0.02571061998605728,
|
| 354 |
+
"learning_rate": 8.074285714285714e-06,
|
| 355 |
+
"loss": 0.0002,
|
| 356 |
+
"step": 1175
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 3.88,
|
| 360 |
+
"grad_norm": 0.02670286037027836,
|
| 361 |
+
"learning_rate": 8.002857142857143e-06,
|
| 362 |
+
"loss": 0.0005,
|
| 363 |
+
"step": 1200
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 3.96,
|
| 367 |
+
"grad_norm": 0.04544826224446297,
|
| 368 |
+
"learning_rate": 7.931428571428572e-06,
|
| 369 |
+
"loss": 0.0007,
|
| 370 |
+
"step": 1225
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 4.05,
|
| 374 |
+
"grad_norm": 0.028520189225673676,
|
| 375 |
+
"learning_rate": 7.860000000000001e-06,
|
| 376 |
+
"loss": 0.0003,
|
| 377 |
+
"step": 1250
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 4.13,
|
| 381 |
+
"grad_norm": 0.017675643786787987,
|
| 382 |
+
"learning_rate": 7.788571428571428e-06,
|
| 383 |
+
"loss": 0.0006,
|
| 384 |
+
"step": 1275
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 4.21,
|
| 388 |
+
"grad_norm": 0.16999584436416626,
|
| 389 |
+
"learning_rate": 7.717142857142857e-06,
|
| 390 |
+
"loss": 0.0005,
|
| 391 |
+
"step": 1300
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 4.29,
|
| 395 |
+
"grad_norm": 0.22153036296367645,
|
| 396 |
+
"learning_rate": 7.645714285714286e-06,
|
| 397 |
+
"loss": 0.0004,
|
| 398 |
+
"step": 1325
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 4.37,
|
| 402 |
+
"grad_norm": 0.0049127149395644665,
|
| 403 |
+
"learning_rate": 7.574285714285715e-06,
|
| 404 |
+
"loss": 0.0001,
|
| 405 |
+
"step": 1350
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 4.45,
|
| 409 |
+
"grad_norm": 0.02039247751235962,
|
| 410 |
+
"learning_rate": 7.502857142857144e-06,
|
| 411 |
+
"loss": 0.0003,
|
| 412 |
+
"step": 1375
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 4.53,
|
| 416 |
+
"grad_norm": 0.009562624618411064,
|
| 417 |
+
"learning_rate": 7.431428571428572e-06,
|
| 418 |
+
"loss": 0.0002,
|
| 419 |
+
"step": 1400
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 4.61,
|
| 423 |
+
"grad_norm": 0.019309034571051598,
|
| 424 |
+
"learning_rate": 7.360000000000001e-06,
|
| 425 |
+
"loss": 0.0001,
|
| 426 |
+
"step": 1425
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 4.69,
|
| 430 |
+
"grad_norm": 0.004860865883529186,
|
| 431 |
+
"learning_rate": 7.28857142857143e-06,
|
| 432 |
+
"loss": 0.0008,
|
| 433 |
+
"step": 1450
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 4.77,
|
| 437 |
+
"grad_norm": 0.005680860485881567,
|
| 438 |
+
"learning_rate": 7.217142857142858e-06,
|
| 439 |
+
"loss": 0.0004,
|
| 440 |
+
"step": 1475
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 4.85,
|
| 444 |
+
"grad_norm": 0.02605484426021576,
|
| 445 |
+
"learning_rate": 7.145714285714286e-06,
|
| 446 |
+
"loss": 0.0008,
|
| 447 |
+
"step": 1500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 4.85,
|
| 451 |
+
"eval_loss": 0.022091694176197052,
|
| 452 |
+
"eval_runtime": 791.3534,
|
| 453 |
+
"eval_samples_per_second": 1.558,
|
| 454 |
+
"eval_steps_per_second": 0.196,
|
| 455 |
+
"eval_wer": 3.988310125494241,
|
| 456 |
+
"step": 1500
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 4.94,
|
| 460 |
+
"grad_norm": 1.7420192956924438,
|
| 461 |
+
"learning_rate": 7.074285714285715e-06,
|
| 462 |
+
"loss": 0.0005,
|
| 463 |
+
"step": 1525
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 5.02,
|
| 467 |
+
"grad_norm": 0.15281762182712555,
|
| 468 |
+
"learning_rate": 7.002857142857143e-06,
|
| 469 |
+
"loss": 0.0004,
|
| 470 |
+
"step": 1550
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 5.1,
|
| 474 |
+
"grad_norm": 0.013328706845641136,
|
| 475 |
+
"learning_rate": 6.931428571428572e-06,
|
| 476 |
+
"loss": 0.0002,
|
| 477 |
+
"step": 1575
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 5.18,
|
| 481 |
+
"grad_norm": 0.23917707800865173,
|
| 482 |
+
"learning_rate": 6.860000000000001e-06,
|
| 483 |
+
"loss": 0.0019,
|
| 484 |
+
"step": 1600
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 5.26,
|
| 488 |
+
"grad_norm": 0.02083628997206688,
|
| 489 |
+
"learning_rate": 6.7885714285714286e-06,
|
| 490 |
+
"loss": 0.0003,
|
| 491 |
+
"step": 1625
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 5.34,
|
| 495 |
+
"grad_norm": 0.7623525857925415,
|
| 496 |
+
"learning_rate": 6.7171428571428576e-06,
|
| 497 |
+
"loss": 0.0006,
|
| 498 |
+
"step": 1650
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 5.42,
|
| 502 |
+
"grad_norm": 1.761312484741211,
|
| 503 |
+
"learning_rate": 6.645714285714287e-06,
|
| 504 |
+
"loss": 0.0005,
|
| 505 |
+
"step": 1675
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 5.5,
|
| 509 |
+
"grad_norm": 0.02908872812986374,
|
| 510 |
+
"learning_rate": 6.574285714285716e-06,
|
| 511 |
+
"loss": 0.0003,
|
| 512 |
+
"step": 1700
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 5.58,
|
| 516 |
+
"grad_norm": 0.0062005845829844475,
|
| 517 |
+
"learning_rate": 6.502857142857143e-06,
|
| 518 |
+
"loss": 0.0003,
|
| 519 |
+
"step": 1725
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 5.66,
|
| 523 |
+
"grad_norm": 0.009553766809403896,
|
| 524 |
+
"learning_rate": 6.431428571428572e-06,
|
| 525 |
+
"loss": 0.0001,
|
| 526 |
+
"step": 1750
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 5.74,
|
| 530 |
+
"grad_norm": 0.006461838725954294,
|
| 531 |
+
"learning_rate": 6.360000000000001e-06,
|
| 532 |
+
"loss": 0.0001,
|
| 533 |
+
"step": 1775
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 5.83,
|
| 537 |
+
"grad_norm": 0.006093865260481834,
|
| 538 |
+
"learning_rate": 6.288571428571429e-06,
|
| 539 |
+
"loss": 0.0003,
|
| 540 |
+
"step": 1800
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 5.91,
|
| 544 |
+
"grad_norm": 0.006265025120228529,
|
| 545 |
+
"learning_rate": 6.217142857142857e-06,
|
| 546 |
+
"loss": 0.0005,
|
| 547 |
+
"step": 1825
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 5.99,
|
| 551 |
+
"grad_norm": 0.08442416042089462,
|
| 552 |
+
"learning_rate": 6.145714285714286e-06,
|
| 553 |
+
"loss": 0.0002,
|
| 554 |
+
"step": 1850
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 6.07,
|
| 558 |
+
"grad_norm": 0.005628472194075584,
|
| 559 |
+
"learning_rate": 6.0742857142857145e-06,
|
| 560 |
+
"loss": 0.0001,
|
| 561 |
+
"step": 1875
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 6.15,
|
| 565 |
+
"grad_norm": 0.0012941138120368123,
|
| 566 |
+
"learning_rate": 6.0028571428571435e-06,
|
| 567 |
+
"loss": 0.0001,
|
| 568 |
+
"step": 1900
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 6.23,
|
| 572 |
+
"grad_norm": 0.003825143212452531,
|
| 573 |
+
"learning_rate": 5.9314285714285725e-06,
|
| 574 |
+
"loss": 0.0001,
|
| 575 |
+
"step": 1925
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 6.31,
|
| 579 |
+
"grad_norm": 0.002939299214631319,
|
| 580 |
+
"learning_rate": 5.86e-06,
|
| 581 |
+
"loss": 0.0,
|
| 582 |
+
"step": 1950
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 6.39,
|
| 586 |
+
"grad_norm": 0.0030452311038970947,
|
| 587 |
+
"learning_rate": 5.788571428571429e-06,
|
| 588 |
+
"loss": 0.0001,
|
| 589 |
+
"step": 1975
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 6.47,
|
| 593 |
+
"grad_norm": 0.017602646723389626,
|
| 594 |
+
"learning_rate": 5.717142857142858e-06,
|
| 595 |
+
"loss": 0.0,
|
| 596 |
+
"step": 2000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 6.47,
|
| 600 |
+
"eval_loss": 0.018006248399615288,
|
| 601 |
+
"eval_runtime": 791.4436,
|
| 602 |
+
"eval_samples_per_second": 1.558,
|
| 603 |
+
"eval_steps_per_second": 0.196,
|
| 604 |
+
"eval_wer": 2.974041602200447,
|
| 605 |
+
"step": 2000
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 6.55,
|
| 609 |
+
"grad_norm": 0.09079013019800186,
|
| 610 |
+
"learning_rate": 5.645714285714287e-06,
|
| 611 |
+
"loss": 0.0001,
|
| 612 |
+
"step": 2025
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 6.63,
|
| 616 |
+
"grad_norm": 0.011517921462655067,
|
| 617 |
+
"learning_rate": 5.574285714285714e-06,
|
| 618 |
+
"loss": 0.0001,
|
| 619 |
+
"step": 2050
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 6.72,
|
| 623 |
+
"grad_norm": 0.0036222347989678383,
|
| 624 |
+
"learning_rate": 5.502857142857143e-06,
|
| 625 |
+
"loss": 0.0,
|
| 626 |
+
"step": 2075
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 6.8,
|
| 630 |
+
"grad_norm": 0.09321127831935883,
|
| 631 |
+
"learning_rate": 5.431428571428572e-06,
|
| 632 |
+
"loss": 0.0,
|
| 633 |
+
"step": 2100
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 6.88,
|
| 637 |
+
"grad_norm": 0.019542288035154343,
|
| 638 |
+
"learning_rate": 5.36e-06,
|
| 639 |
+
"loss": 0.0,
|
| 640 |
+
"step": 2125
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 6.96,
|
| 644 |
+
"grad_norm": 0.010700986720621586,
|
| 645 |
+
"learning_rate": 5.2885714285714285e-06,
|
| 646 |
+
"loss": 0.0001,
|
| 647 |
+
"step": 2150
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 7.04,
|
| 651 |
+
"grad_norm": 0.005260708276182413,
|
| 652 |
+
"learning_rate": 5.2171428571428575e-06,
|
| 653 |
+
"loss": 0.0,
|
| 654 |
+
"step": 2175
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 7.12,
|
| 658 |
+
"grad_norm": 0.0018403552239760756,
|
| 659 |
+
"learning_rate": 5.145714285714286e-06,
|
| 660 |
+
"loss": 0.0,
|
| 661 |
+
"step": 2200
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 7.2,
|
| 665 |
+
"grad_norm": 0.00291392276994884,
|
| 666 |
+
"learning_rate": 5.074285714285715e-06,
|
| 667 |
+
"loss": 0.0,
|
| 668 |
+
"step": 2225
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 7.28,
|
| 672 |
+
"grad_norm": 0.001351356622762978,
|
| 673 |
+
"learning_rate": 5.002857142857144e-06,
|
| 674 |
+
"loss": 0.0,
|
| 675 |
+
"step": 2250
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 7.36,
|
| 679 |
+
"grad_norm": 0.002212280174717307,
|
| 680 |
+
"learning_rate": 4.931428571428572e-06,
|
| 681 |
+
"loss": 0.0,
|
| 682 |
+
"step": 2275
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 7.44,
|
| 686 |
+
"grad_norm": 0.0013704759767279029,
|
| 687 |
+
"learning_rate": 4.86e-06,
|
| 688 |
+
"loss": 0.0,
|
| 689 |
+
"step": 2300
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 7.52,
|
| 693 |
+
"grad_norm": 0.004209516104310751,
|
| 694 |
+
"learning_rate": 4.788571428571429e-06,
|
| 695 |
+
"loss": 0.0,
|
| 696 |
+
"step": 2325
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 7.61,
|
| 700 |
+
"grad_norm": 0.004072306212037802,
|
| 701 |
+
"learning_rate": 4.717142857142857e-06,
|
| 702 |
+
"loss": 0.0,
|
| 703 |
+
"step": 2350
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 7.69,
|
| 707 |
+
"grad_norm": 0.0258013978600502,
|
| 708 |
+
"learning_rate": 4.645714285714286e-06,
|
| 709 |
+
"loss": 0.0,
|
| 710 |
+
"step": 2375
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 7.77,
|
| 714 |
+
"grad_norm": 0.0018390474142506719,
|
| 715 |
+
"learning_rate": 4.574285714285714e-06,
|
| 716 |
+
"loss": 0.0,
|
| 717 |
+
"step": 2400
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 7.85,
|
| 721 |
+
"grad_norm": 0.002616502810269594,
|
| 722 |
+
"learning_rate": 4.5028571428571434e-06,
|
| 723 |
+
"loss": 0.0,
|
| 724 |
+
"step": 2425
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 7.93,
|
| 728 |
+
"grad_norm": 0.0016982157249003649,
|
| 729 |
+
"learning_rate": 4.431428571428572e-06,
|
| 730 |
+
"loss": 0.0,
|
| 731 |
+
"step": 2450
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 8.01,
|
| 735 |
+
"grad_norm": 0.0023189985658973455,
|
| 736 |
+
"learning_rate": 4.360000000000001e-06,
|
| 737 |
+
"loss": 0.0,
|
| 738 |
+
"step": 2475
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 8.09,
|
| 742 |
+
"grad_norm": 0.003353666979819536,
|
| 743 |
+
"learning_rate": 4.288571428571429e-06,
|
| 744 |
+
"loss": 0.0,
|
| 745 |
+
"step": 2500
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 8.09,
|
| 749 |
+
"eval_loss": 0.017685044556856155,
|
| 750 |
+
"eval_runtime": 798.3523,
|
| 751 |
+
"eval_samples_per_second": 1.544,
|
| 752 |
+
"eval_steps_per_second": 0.194,
|
| 753 |
+
"eval_wer": 3.094378545642084,
|
| 754 |
+
"step": 2500
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 8.17,
|
| 758 |
+
"grad_norm": 0.0012944688787683845,
|
| 759 |
+
"learning_rate": 4.217142857142858e-06,
|
| 760 |
+
"loss": 0.0,
|
| 761 |
+
"step": 2525
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 8.25,
|
| 765 |
+
"grad_norm": 0.002697590971365571,
|
| 766 |
+
"learning_rate": 4.145714285714286e-06,
|
| 767 |
+
"loss": 0.0,
|
| 768 |
+
"step": 2550
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 8.33,
|
| 772 |
+
"grad_norm": 0.0015665855025872588,
|
| 773 |
+
"learning_rate": 4.074285714285714e-06,
|
| 774 |
+
"loss": 0.0,
|
| 775 |
+
"step": 2575
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 8.41,
|
| 779 |
+
"grad_norm": 0.0023723498452454805,
|
| 780 |
+
"learning_rate": 4.002857142857143e-06,
|
| 781 |
+
"loss": 0.0,
|
| 782 |
+
"step": 2600
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 8.5,
|
| 786 |
+
"grad_norm": 0.0024561272002756596,
|
| 787 |
+
"learning_rate": 3.931428571428571e-06,
|
| 788 |
+
"loss": 0.0,
|
| 789 |
+
"step": 2625
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 8.58,
|
| 793 |
+
"grad_norm": 0.0012809446780011058,
|
| 794 |
+
"learning_rate": 3.86e-06,
|
| 795 |
+
"loss": 0.0,
|
| 796 |
+
"step": 2650
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 8.66,
|
| 800 |
+
"grad_norm": 0.002741657430306077,
|
| 801 |
+
"learning_rate": 3.7885714285714285e-06,
|
| 802 |
+
"loss": 0.0,
|
| 803 |
+
"step": 2675
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 8.74,
|
| 807 |
+
"grad_norm": 0.0022798278369009495,
|
| 808 |
+
"learning_rate": 3.7171428571428575e-06,
|
| 809 |
+
"loss": 0.0,
|
| 810 |
+
"step": 2700
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 8.82,
|
| 814 |
+
"grad_norm": 0.0008324767695739865,
|
| 815 |
+
"learning_rate": 3.6457142857142857e-06,
|
| 816 |
+
"loss": 0.0,
|
| 817 |
+
"step": 2725
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 8.9,
|
| 821 |
+
"grad_norm": 0.0021829677280038595,
|
| 822 |
+
"learning_rate": 3.5742857142857147e-06,
|
| 823 |
+
"loss": 0.0,
|
| 824 |
+
"step": 2750
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 8.98,
|
| 828 |
+
"grad_norm": 0.0010186375584453344,
|
| 829 |
+
"learning_rate": 3.5028571428571433e-06,
|
| 830 |
+
"loss": 0.0,
|
| 831 |
+
"step": 2775
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 9.06,
|
| 835 |
+
"grad_norm": 0.0021447737235575914,
|
| 836 |
+
"learning_rate": 3.431428571428572e-06,
|
| 837 |
+
"loss": 0.0,
|
| 838 |
+
"step": 2800
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 9.14,
|
| 842 |
+
"grad_norm": 0.0014905674615874887,
|
| 843 |
+
"learning_rate": 3.3600000000000004e-06,
|
| 844 |
+
"loss": 0.0,
|
| 845 |
+
"step": 2825
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 9.22,
|
| 849 |
+
"grad_norm": 0.0023917625658214092,
|
| 850 |
+
"learning_rate": 3.2885714285714286e-06,
|
| 851 |
+
"loss": 0.0,
|
| 852 |
+
"step": 2850
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 9.3,
|
| 856 |
+
"grad_norm": 0.0014455171767622232,
|
| 857 |
+
"learning_rate": 3.2171428571428576e-06,
|
| 858 |
+
"loss": 0.0,
|
| 859 |
+
"step": 2875
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 9.39,
|
| 863 |
+
"grad_norm": 0.009605071507394314,
|
| 864 |
+
"learning_rate": 3.1457142857142858e-06,
|
| 865 |
+
"loss": 0.0,
|
| 866 |
+
"step": 2900
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 9.47,
|
| 870 |
+
"grad_norm": 0.0006564524956047535,
|
| 871 |
+
"learning_rate": 3.074285714285715e-06,
|
| 872 |
+
"loss": 0.0,
|
| 873 |
+
"step": 2925
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 9.55,
|
| 877 |
+
"grad_norm": 0.0008893849444575608,
|
| 878 |
+
"learning_rate": 3.002857142857143e-06,
|
| 879 |
+
"loss": 0.0,
|
| 880 |
+
"step": 2950
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 9.63,
|
| 884 |
+
"grad_norm": 0.001537854433991015,
|
| 885 |
+
"learning_rate": 2.9314285714285716e-06,
|
| 886 |
+
"loss": 0.0,
|
| 887 |
+
"step": 2975
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 9.71,
|
| 891 |
+
"grad_norm": 0.0020573874935507774,
|
| 892 |
+
"learning_rate": 2.86e-06,
|
| 893 |
+
"loss": 0.0,
|
| 894 |
+
"step": 3000
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 9.71,
|
| 898 |
+
"eval_loss": 0.017799125984311104,
|
| 899 |
+
"eval_runtime": 794.7434,
|
| 900 |
+
"eval_samples_per_second": 1.551,
|
| 901 |
+
"eval_steps_per_second": 0.195,
|
| 902 |
+
"eval_wer": 3.094378545642084,
|
| 903 |
+
"step": 3000
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 9.79,
|
| 907 |
+
"grad_norm": 0.001788494409993291,
|
| 908 |
+
"learning_rate": 2.7885714285714287e-06,
|
| 909 |
+
"loss": 0.0,
|
| 910 |
+
"step": 3025
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 9.87,
|
| 914 |
+
"grad_norm": 0.0009725405252538621,
|
| 915 |
+
"learning_rate": 2.7171428571428577e-06,
|
| 916 |
+
"loss": 0.0,
|
| 917 |
+
"step": 3050
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 9.95,
|
| 921 |
+
"grad_norm": 0.001284493599087,
|
| 922 |
+
"learning_rate": 2.645714285714286e-06,
|
| 923 |
+
"loss": 0.0,
|
| 924 |
+
"step": 3075
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 10.03,
|
| 928 |
+
"grad_norm": 0.0017106755403801799,
|
| 929 |
+
"learning_rate": 2.574285714285715e-06,
|
| 930 |
+
"loss": 0.0,
|
| 931 |
+
"step": 3100
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 10.11,
|
| 935 |
+
"grad_norm": 0.0007921723299659789,
|
| 936 |
+
"learning_rate": 2.502857142857143e-06,
|
| 937 |
+
"loss": 0.0,
|
| 938 |
+
"step": 3125
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 10.19,
|
| 942 |
+
"grad_norm": 0.0018981642788276076,
|
| 943 |
+
"learning_rate": 2.4314285714285717e-06,
|
| 944 |
+
"loss": 0.0,
|
| 945 |
+
"step": 3150
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 10.28,
|
| 949 |
+
"grad_norm": 0.0014530618209391832,
|
| 950 |
+
"learning_rate": 2.3600000000000003e-06,
|
| 951 |
+
"loss": 0.0,
|
| 952 |
+
"step": 3175
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 10.36,
|
| 956 |
+
"grad_norm": 0.0011344915255904198,
|
| 957 |
+
"learning_rate": 2.288571428571429e-06,
|
| 958 |
+
"loss": 0.0,
|
| 959 |
+
"step": 3200
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 10.44,
|
| 963 |
+
"grad_norm": 0.0006109612877480686,
|
| 964 |
+
"learning_rate": 2.2171428571428575e-06,
|
| 965 |
+
"loss": 0.0,
|
| 966 |
+
"step": 3225
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 10.52,
|
| 970 |
+
"grad_norm": 0.0011850837618112564,
|
| 971 |
+
"learning_rate": 2.145714285714286e-06,
|
| 972 |
+
"loss": 0.0,
|
| 973 |
+
"step": 3250
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 10.6,
|
| 977 |
+
"grad_norm": 0.0005283011123538017,
|
| 978 |
+
"learning_rate": 2.0742857142857146e-06,
|
| 979 |
+
"loss": 0.0,
|
| 980 |
+
"step": 3275
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 10.68,
|
| 984 |
+
"grad_norm": 0.0010328389471396804,
|
| 985 |
+
"learning_rate": 2.0028571428571432e-06,
|
| 986 |
+
"loss": 0.0,
|
| 987 |
+
"step": 3300
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 10.76,
|
| 991 |
+
"grad_norm": 0.0009599952027201653,
|
| 992 |
+
"learning_rate": 1.9314285714285714e-06,
|
| 993 |
+
"loss": 0.0,
|
| 994 |
+
"step": 3325
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 10.84,
|
| 998 |
+
"grad_norm": 0.0011755761224776506,
|
| 999 |
+
"learning_rate": 1.8600000000000002e-06,
|
| 1000 |
+
"loss": 0.0,
|
| 1001 |
+
"step": 3350
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 10.92,
|
| 1005 |
+
"grad_norm": 0.0018132536206394434,
|
| 1006 |
+
"learning_rate": 1.7885714285714288e-06,
|
| 1007 |
+
"loss": 0.0,
|
| 1008 |
+
"step": 3375
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 11.0,
|
| 1012 |
+
"grad_norm": 0.0018807642627507448,
|
| 1013 |
+
"learning_rate": 1.7171428571428572e-06,
|
| 1014 |
+
"loss": 0.0,
|
| 1015 |
+
"step": 3400
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 11.08,
|
| 1019 |
+
"grad_norm": 0.003156110644340515,
|
| 1020 |
+
"learning_rate": 1.6457142857142857e-06,
|
| 1021 |
+
"loss": 0.0,
|
| 1022 |
+
"step": 3425
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 11.17,
|
| 1026 |
+
"grad_norm": 0.0010045066010206938,
|
| 1027 |
+
"learning_rate": 1.5742857142857143e-06,
|
| 1028 |
+
"loss": 0.0,
|
| 1029 |
+
"step": 3450
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 11.25,
|
| 1033 |
+
"grad_norm": 0.0008904847782105207,
|
| 1034 |
+
"learning_rate": 1.502857142857143e-06,
|
| 1035 |
+
"loss": 0.0,
|
| 1036 |
+
"step": 3475
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 11.33,
|
| 1040 |
+
"grad_norm": 0.0013107324484735727,
|
| 1041 |
+
"learning_rate": 1.4314285714285717e-06,
|
| 1042 |
+
"loss": 0.0,
|
| 1043 |
+
"step": 3500
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 11.33,
|
| 1047 |
+
"eval_loss": 0.017873059958219528,
|
| 1048 |
+
"eval_runtime": 799.0063,
|
| 1049 |
+
"eval_samples_per_second": 1.543,
|
| 1050 |
+
"eval_steps_per_second": 0.194,
|
| 1051 |
+
"eval_wer": 3.1287605294825513,
|
| 1052 |
+
"step": 3500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 11.41,
|
| 1056 |
+
"grad_norm": 0.0013505747774615884,
|
| 1057 |
+
"learning_rate": 1.3600000000000001e-06,
|
| 1058 |
+
"loss": 0.0,
|
| 1059 |
+
"step": 3525
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 11.49,
|
| 1063 |
+
"grad_norm": 0.0013679265975952148,
|
| 1064 |
+
"learning_rate": 1.2885714285714287e-06,
|
| 1065 |
+
"loss": 0.0,
|
| 1066 |
+
"step": 3550
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 11.57,
|
| 1070 |
+
"grad_norm": 0.0021877065300941467,
|
| 1071 |
+
"learning_rate": 1.2171428571428573e-06,
|
| 1072 |
+
"loss": 0.0,
|
| 1073 |
+
"step": 3575
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 11.65,
|
| 1077 |
+
"grad_norm": 0.0011414324399083853,
|
| 1078 |
+
"learning_rate": 1.1457142857142859e-06,
|
| 1079 |
+
"loss": 0.0,
|
| 1080 |
+
"step": 3600
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 11.73,
|
| 1084 |
+
"grad_norm": 0.0008932927739806473,
|
| 1085 |
+
"learning_rate": 1.0742857142857145e-06,
|
| 1086 |
+
"loss": 0.0,
|
| 1087 |
+
"step": 3625
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 11.81,
|
| 1091 |
+
"grad_norm": 0.0016671326011419296,
|
| 1092 |
+
"learning_rate": 1.0028571428571428e-06,
|
| 1093 |
+
"loss": 0.0,
|
| 1094 |
+
"step": 3650
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 11.89,
|
| 1098 |
+
"grad_norm": 0.0011023151455447078,
|
| 1099 |
+
"learning_rate": 9.314285714285714e-07,
|
| 1100 |
+
"loss": 0.0,
|
| 1101 |
+
"step": 3675
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 11.97,
|
| 1105 |
+
"grad_norm": 0.0008648216025903821,
|
| 1106 |
+
"learning_rate": 8.6e-07,
|
| 1107 |
+
"loss": 0.0,
|
| 1108 |
+
"step": 3700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 12.06,
|
| 1112 |
+
"grad_norm": 0.003709783311933279,
|
| 1113 |
+
"learning_rate": 7.885714285714287e-07,
|
| 1114 |
+
"loss": 0.0,
|
| 1115 |
+
"step": 3725
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 12.14,
|
| 1119 |
+
"grad_norm": 0.0005944801960140467,
|
| 1120 |
+
"learning_rate": 7.171428571428572e-07,
|
| 1121 |
+
"loss": 0.0,
|
| 1122 |
+
"step": 3750
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 12.22,
|
| 1126 |
+
"grad_norm": 0.0010830622632056475,
|
| 1127 |
+
"learning_rate": 6.457142857142858e-07,
|
| 1128 |
+
"loss": 0.0,
|
| 1129 |
+
"step": 3775
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 12.3,
|
| 1133 |
+
"grad_norm": 0.0018471087096258998,
|
| 1134 |
+
"learning_rate": 5.742857142857143e-07,
|
| 1135 |
+
"loss": 0.0,
|
| 1136 |
+
"step": 3800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 12.38,
|
| 1140 |
+
"grad_norm": 0.0018315908964723349,
|
| 1141 |
+
"learning_rate": 5.028571428571429e-07,
|
| 1142 |
+
"loss": 0.0,
|
| 1143 |
+
"step": 3825
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 12.46,
|
| 1147 |
+
"grad_norm": 0.0010556795168668032,
|
| 1148 |
+
"learning_rate": 4.3142857142857146e-07,
|
| 1149 |
+
"loss": 0.0,
|
| 1150 |
+
"step": 3850
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 12.54,
|
| 1154 |
+
"grad_norm": 0.001286807470023632,
|
| 1155 |
+
"learning_rate": 3.6e-07,
|
| 1156 |
+
"loss": 0.0,
|
| 1157 |
+
"step": 3875
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 12.62,
|
| 1161 |
+
"grad_norm": 0.0025280534755438566,
|
| 1162 |
+
"learning_rate": 2.885714285714286e-07,
|
| 1163 |
+
"loss": 0.0,
|
| 1164 |
+
"step": 3900
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 12.7,
|
| 1168 |
+
"grad_norm": 0.0008814557804726064,
|
| 1169 |
+
"learning_rate": 2.1714285714285715e-07,
|
| 1170 |
+
"loss": 0.0,
|
| 1171 |
+
"step": 3925
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 12.78,
|
| 1175 |
+
"grad_norm": 0.0007184511632658541,
|
| 1176 |
+
"learning_rate": 1.4571428571428574e-07,
|
| 1177 |
+
"loss": 0.0,
|
| 1178 |
+
"step": 3950
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 12.86,
|
| 1182 |
+
"grad_norm": 0.0024200354237109423,
|
| 1183 |
+
"learning_rate": 7.428571428571429e-08,
|
| 1184 |
+
"loss": 0.0,
|
| 1185 |
+
"step": 3975
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 12.94,
|
| 1189 |
+
"grad_norm": 0.0013352702371776104,
|
| 1190 |
+
"learning_rate": 2.8571428571428576e-09,
|
| 1191 |
+
"loss": 0.0,
|
| 1192 |
+
"step": 4000
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 12.94,
|
| 1196 |
+
"eval_loss": 0.017900340259075165,
|
| 1197 |
+
"eval_runtime": 798.968,
|
| 1198 |
+
"eval_samples_per_second": 1.543,
|
| 1199 |
+
"eval_steps_per_second": 0.194,
|
| 1200 |
+
"eval_wer": 3.1287605294825513,
|
| 1201 |
+
"step": 4000
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 13.03,
|
| 1205 |
+
"grad_norm": 0.0006061609601601958,
|
| 1206 |
+
"learning_rate": 5.301333333333334e-06,
|
| 1207 |
+
"loss": 0.0,
|
| 1208 |
+
"step": 4025
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 13.11,
|
| 1212 |
+
"grad_norm": 0.000903214851859957,
|
| 1213 |
+
"learning_rate": 5.268000000000001e-06,
|
| 1214 |
+
"loss": 0.0,
|
| 1215 |
+
"step": 4050
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 13.19,
|
| 1219 |
+
"grad_norm": 0.0009095696732401848,
|
| 1220 |
+
"learning_rate": 5.234666666666667e-06,
|
| 1221 |
+
"loss": 0.0,
|
| 1222 |
+
"step": 4075
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 13.27,
|
| 1226 |
+
"grad_norm": 0.0009569237590767443,
|
| 1227 |
+
"learning_rate": 5.201333333333334e-06,
|
| 1228 |
+
"loss": 0.0,
|
| 1229 |
+
"step": 4100
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 13.35,
|
| 1233 |
+
"grad_norm": 0.0013346378691494465,
|
| 1234 |
+
"learning_rate": 5.168000000000001e-06,
|
| 1235 |
+
"loss": 0.0,
|
| 1236 |
+
"step": 4125
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 13.43,
|
| 1240 |
+
"grad_norm": 0.0007105122786015272,
|
| 1241 |
+
"learning_rate": 5.134666666666667e-06,
|
| 1242 |
+
"loss": 0.0,
|
| 1243 |
+
"step": 4150
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 13.51,
|
| 1247 |
+
"grad_norm": 0.001324965967796743,
|
| 1248 |
+
"learning_rate": 5.101333333333334e-06,
|
| 1249 |
+
"loss": 0.0,
|
| 1250 |
+
"step": 4175
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 13.59,
|
| 1254 |
+
"grad_norm": 0.002936706179752946,
|
| 1255 |
+
"learning_rate": 5.0680000000000004e-06,
|
| 1256 |
+
"loss": 0.0,
|
| 1257 |
+
"step": 4200
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 13.67,
|
| 1261 |
+
"grad_norm": 0.0006480725132860243,
|
| 1262 |
+
"learning_rate": 5.034666666666667e-06,
|
| 1263 |
+
"loss": 0.0,
|
| 1264 |
+
"step": 4225
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 13.75,
|
| 1268 |
+
"grad_norm": 0.0008248516824096441,
|
| 1269 |
+
"learning_rate": 5.001333333333333e-06,
|
| 1270 |
+
"loss": 0.0,
|
| 1271 |
+
"step": 4250
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 13.83,
|
| 1275 |
+
"grad_norm": 0.000976664014160633,
|
| 1276 |
+
"learning_rate": 4.9680000000000005e-06,
|
| 1277 |
+
"loss": 0.0,
|
| 1278 |
+
"step": 4275
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 13.92,
|
| 1282 |
+
"grad_norm": 0.0007901455392129719,
|
| 1283 |
+
"learning_rate": 4.934666666666667e-06,
|
| 1284 |
+
"loss": 0.0,
|
| 1285 |
+
"step": 4300
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 14.0,
|
| 1289 |
+
"grad_norm": 0.0008691786788403988,
|
| 1290 |
+
"learning_rate": 4.901333333333333e-06,
|
| 1291 |
+
"loss": 0.0,
|
| 1292 |
+
"step": 4325
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 14.08,
|
| 1296 |
+
"grad_norm": 0.00039108938653953373,
|
| 1297 |
+
"learning_rate": 4.868000000000001e-06,
|
| 1298 |
+
"loss": 0.0,
|
| 1299 |
+
"step": 4350
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 14.16,
|
| 1303 |
+
"grad_norm": 0.000672497262712568,
|
| 1304 |
+
"learning_rate": 4.834666666666667e-06,
|
| 1305 |
+
"loss": 0.0,
|
| 1306 |
+
"step": 4375
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 14.24,
|
| 1310 |
+
"grad_norm": 0.0006290368619374931,
|
| 1311 |
+
"learning_rate": 4.801333333333334e-06,
|
| 1312 |
+
"loss": 0.0,
|
| 1313 |
+
"step": 4400
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 14.32,
|
| 1317 |
+
"grad_norm": 0.0012541558826342225,
|
| 1318 |
+
"learning_rate": 4.768000000000001e-06,
|
| 1319 |
+
"loss": 0.0,
|
| 1320 |
+
"step": 4425
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 14.4,
|
| 1324 |
+
"grad_norm": 0.0015149698592722416,
|
| 1325 |
+
"learning_rate": 4.734666666666667e-06,
|
| 1326 |
+
"loss": 0.0,
|
| 1327 |
+
"step": 4450
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 14.48,
|
| 1331 |
+
"grad_norm": 0.001397727057337761,
|
| 1332 |
+
"learning_rate": 4.701333333333334e-06,
|
| 1333 |
+
"loss": 0.0,
|
| 1334 |
+
"step": 4475
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 14.56,
|
| 1338 |
+
"grad_norm": 0.001000353367999196,
|
| 1339 |
+
"learning_rate": 4.668e-06,
|
| 1340 |
+
"loss": 0.0,
|
| 1341 |
+
"step": 4500
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 14.56,
|
| 1345 |
+
"eval_loss": 0.018149884417653084,
|
| 1346 |
+
"eval_runtime": 806.023,
|
| 1347 |
+
"eval_samples_per_second": 1.53,
|
| 1348 |
+
"eval_steps_per_second": 0.192,
|
| 1349 |
+
"eval_wer": 2.888086642599278,
|
| 1350 |
+
"step": 4500
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 14.64,
|
| 1354 |
+
"grad_norm": 0.0004112539754714817,
|
| 1355 |
+
"learning_rate": 4.634666666666667e-06,
|
| 1356 |
+
"loss": 0.0,
|
| 1357 |
+
"step": 4525
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 14.72,
|
| 1361 |
+
"grad_norm": 0.00042669521644711494,
|
| 1362 |
+
"learning_rate": 4.601333333333334e-06,
|
| 1363 |
+
"loss": 0.0,
|
| 1364 |
+
"step": 4550
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 14.81,
|
| 1368 |
+
"grad_norm": 0.0005972330691292882,
|
| 1369 |
+
"learning_rate": 4.568e-06,
|
| 1370 |
+
"loss": 0.0,
|
| 1371 |
+
"step": 4575
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 14.89,
|
| 1375 |
+
"grad_norm": 0.000703622295986861,
|
| 1376 |
+
"learning_rate": 4.534666666666667e-06,
|
| 1377 |
+
"loss": 0.0,
|
| 1378 |
+
"step": 4600
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 14.97,
|
| 1382 |
+
"grad_norm": 0.0008054905338212848,
|
| 1383 |
+
"learning_rate": 4.501333333333334e-06,
|
| 1384 |
+
"loss": 0.0,
|
| 1385 |
+
"step": 4625
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 15.05,
|
| 1389 |
+
"grad_norm": 0.0006107372464612126,
|
| 1390 |
+
"learning_rate": 4.468e-06,
|
| 1391 |
+
"loss": 0.0,
|
| 1392 |
+
"step": 4650
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 15.13,
|
| 1396 |
+
"grad_norm": 0.0011443241965025663,
|
| 1397 |
+
"learning_rate": 4.4346666666666675e-06,
|
| 1398 |
+
"loss": 0.0,
|
| 1399 |
+
"step": 4675
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 15.21,
|
| 1403 |
+
"grad_norm": 0.0007337784627452493,
|
| 1404 |
+
"learning_rate": 4.401333333333334e-06,
|
| 1405 |
+
"loss": 0.0,
|
| 1406 |
+
"step": 4700
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 15.29,
|
| 1410 |
+
"grad_norm": 0.0005867818254046142,
|
| 1411 |
+
"learning_rate": 4.368e-06,
|
| 1412 |
+
"loss": 0.0,
|
| 1413 |
+
"step": 4725
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 15.37,
|
| 1417 |
+
"grad_norm": 0.00047605938743799925,
|
| 1418 |
+
"learning_rate": 4.334666666666667e-06,
|
| 1419 |
+
"loss": 0.0,
|
| 1420 |
+
"step": 4750
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 15.45,
|
| 1424 |
+
"grad_norm": 0.0007535215700045228,
|
| 1425 |
+
"learning_rate": 4.301333333333333e-06,
|
| 1426 |
+
"loss": 0.0,
|
| 1427 |
+
"step": 4775
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 15.53,
|
| 1431 |
+
"grad_norm": 0.0004895786987617612,
|
| 1432 |
+
"learning_rate": 4.2680000000000005e-06,
|
| 1433 |
+
"loss": 0.0,
|
| 1434 |
+
"step": 4800
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 15.61,
|
| 1438 |
+
"grad_norm": 0.0005502207204699516,
|
| 1439 |
+
"learning_rate": 4.234666666666667e-06,
|
| 1440 |
+
"loss": 0.0,
|
| 1441 |
+
"step": 4825
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 15.7,
|
| 1445 |
+
"grad_norm": 0.0008273226558230817,
|
| 1446 |
+
"learning_rate": 4.201333333333334e-06,
|
| 1447 |
+
"loss": 0.0,
|
| 1448 |
+
"step": 4850
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 15.78,
|
| 1452 |
+
"grad_norm": 0.0008853155304677784,
|
| 1453 |
+
"learning_rate": 4.168000000000001e-06,
|
| 1454 |
+
"loss": 0.0,
|
| 1455 |
+
"step": 4875
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 15.86,
|
| 1459 |
+
"grad_norm": 0.0008551743812859058,
|
| 1460 |
+
"learning_rate": 4.134666666666667e-06,
|
| 1461 |
+
"loss": 0.0,
|
| 1462 |
+
"step": 4900
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 15.94,
|
| 1466 |
+
"grad_norm": 0.0012048580683767796,
|
| 1467 |
+
"learning_rate": 4.1013333333333334e-06,
|
| 1468 |
+
"loss": 0.0,
|
| 1469 |
+
"step": 4925
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 16.02,
|
| 1473 |
+
"grad_norm": 0.00043612357694655657,
|
| 1474 |
+
"learning_rate": 4.068000000000001e-06,
|
| 1475 |
+
"loss": 0.0,
|
| 1476 |
+
"step": 4950
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 16.1,
|
| 1480 |
+
"grad_norm": 0.0005169506184756756,
|
| 1481 |
+
"learning_rate": 4.034666666666667e-06,
|
| 1482 |
+
"loss": 0.0,
|
| 1483 |
+
"step": 4975
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 16.18,
|
| 1487 |
+
"grad_norm": 0.000585487752687186,
|
| 1488 |
+
"learning_rate": 4.0013333333333335e-06,
|
| 1489 |
+
"loss": 0.0,
|
| 1490 |
+
"step": 5000
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 16.18,
|
| 1494 |
+
"eval_loss": 0.018417511135339737,
|
| 1495 |
+
"eval_runtime": 808.2166,
|
| 1496 |
+
"eval_samples_per_second": 1.526,
|
| 1497 |
+
"eval_steps_per_second": 0.192,
|
| 1498 |
+
"eval_wer": 2.9224686264397457,
|
| 1499 |
+
"step": 5000
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 16.26,
|
| 1503 |
+
"grad_norm": 0.000627236848231405,
|
| 1504 |
+
"learning_rate": 3.968e-06,
|
| 1505 |
+
"loss": 0.0,
|
| 1506 |
+
"step": 5025
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 16.34,
|
| 1510 |
+
"grad_norm": 0.0008906592265702784,
|
| 1511 |
+
"learning_rate": 3.934666666666667e-06,
|
| 1512 |
+
"loss": 0.0,
|
| 1513 |
+
"step": 5050
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 16.42,
|
| 1517 |
+
"grad_norm": 0.0005441014654934406,
|
| 1518 |
+
"learning_rate": 3.901333333333334e-06,
|
| 1519 |
+
"loss": 0.0,
|
| 1520 |
+
"step": 5075
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 16.5,
|
| 1524 |
+
"grad_norm": 0.001127504394389689,
|
| 1525 |
+
"learning_rate": 3.868e-06,
|
| 1526 |
+
"loss": 0.0,
|
| 1527 |
+
"step": 5100
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 16.59,
|
| 1531 |
+
"grad_norm": 0.0006261191447265446,
|
| 1532 |
+
"learning_rate": 3.834666666666667e-06,
|
| 1533 |
+
"loss": 0.0,
|
| 1534 |
+
"step": 5125
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 16.67,
|
| 1538 |
+
"grad_norm": 0.0007385109784081578,
|
| 1539 |
+
"learning_rate": 3.8013333333333333e-06,
|
| 1540 |
+
"loss": 0.0,
|
| 1541 |
+
"step": 5150
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 16.75,
|
| 1545 |
+
"grad_norm": 0.0010095132747665048,
|
| 1546 |
+
"learning_rate": 3.7680000000000006e-06,
|
| 1547 |
+
"loss": 0.0,
|
| 1548 |
+
"step": 5175
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 16.83,
|
| 1552 |
+
"grad_norm": 0.0008644785266369581,
|
| 1553 |
+
"learning_rate": 3.734666666666667e-06,
|
| 1554 |
+
"loss": 0.0,
|
| 1555 |
+
"step": 5200
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 16.91,
|
| 1559 |
+
"grad_norm": 0.0005047526792623103,
|
| 1560 |
+
"learning_rate": 3.7013333333333334e-06,
|
| 1561 |
+
"loss": 0.0,
|
| 1562 |
+
"step": 5225
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 16.99,
|
| 1566 |
+
"grad_norm": 0.0005285277147777379,
|
| 1567 |
+
"learning_rate": 3.6680000000000003e-06,
|
| 1568 |
+
"loss": 0.0,
|
| 1569 |
+
"step": 5250
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 17.07,
|
| 1573 |
+
"grad_norm": 0.0005454490310512483,
|
| 1574 |
+
"learning_rate": 3.6346666666666667e-06,
|
| 1575 |
+
"loss": 0.0,
|
| 1576 |
+
"step": 5275
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 17.15,
|
| 1580 |
+
"grad_norm": 0.0005612596869468689,
|
| 1581 |
+
"learning_rate": 3.601333333333334e-06,
|
| 1582 |
+
"loss": 0.0,
|
| 1583 |
+
"step": 5300
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 17.23,
|
| 1587 |
+
"grad_norm": 0.0006952615804038942,
|
| 1588 |
+
"learning_rate": 3.5680000000000004e-06,
|
| 1589 |
+
"loss": 0.0,
|
| 1590 |
+
"step": 5325
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 17.31,
|
| 1594 |
+
"grad_norm": 0.0005712429410777986,
|
| 1595 |
+
"learning_rate": 3.534666666666667e-06,
|
| 1596 |
+
"loss": 0.0,
|
| 1597 |
+
"step": 5350
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 17.39,
|
| 1601 |
+
"grad_norm": 0.00041945508564822376,
|
| 1602 |
+
"learning_rate": 3.5013333333333337e-06,
|
| 1603 |
+
"loss": 0.0,
|
| 1604 |
+
"step": 5375
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 17.48,
|
| 1608 |
+
"grad_norm": 0.00044482393423095345,
|
| 1609 |
+
"learning_rate": 3.468e-06,
|
| 1610 |
+
"loss": 0.0,
|
| 1611 |
+
"step": 5400
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 17.56,
|
| 1615 |
+
"grad_norm": 0.0004090140573680401,
|
| 1616 |
+
"learning_rate": 3.4346666666666665e-06,
|
| 1617 |
+
"loss": 0.0,
|
| 1618 |
+
"step": 5425
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 17.64,
|
| 1622 |
+
"grad_norm": 0.00047689909115433693,
|
| 1623 |
+
"learning_rate": 3.4013333333333338e-06,
|
| 1624 |
+
"loss": 0.0,
|
| 1625 |
+
"step": 5450
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 17.72,
|
| 1629 |
+
"grad_norm": 0.0006317782681435347,
|
| 1630 |
+
"learning_rate": 3.368e-06,
|
| 1631 |
+
"loss": 0.0,
|
| 1632 |
+
"step": 5475
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 17.8,
|
| 1636 |
+
"grad_norm": 0.0004752794047817588,
|
| 1637 |
+
"learning_rate": 3.334666666666667e-06,
|
| 1638 |
+
"loss": 0.0,
|
| 1639 |
+
"step": 5500
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 17.8,
|
| 1643 |
+
"eval_loss": 0.018600720912218094,
|
| 1644 |
+
"eval_runtime": 803.0485,
|
| 1645 |
+
"eval_samples_per_second": 1.535,
|
| 1646 |
+
"eval_steps_per_second": 0.193,
|
| 1647 |
+
"eval_wer": 3.0256145779611483,
|
| 1648 |
+
"step": 5500
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 17.88,
|
| 1652 |
+
"grad_norm": 0.0007763198809698224,
|
| 1653 |
+
"learning_rate": 3.3013333333333335e-06,
|
| 1654 |
+
"loss": 0.0,
|
| 1655 |
+
"step": 5525
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 17.96,
|
| 1659 |
+
"grad_norm": 0.0006742349942214787,
|
| 1660 |
+
"learning_rate": 3.268e-06,
|
| 1661 |
+
"loss": 0.0,
|
| 1662 |
+
"step": 5550
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 18.04,
|
| 1666 |
+
"grad_norm": 0.00037475177668966353,
|
| 1667 |
+
"learning_rate": 3.234666666666667e-06,
|
| 1668 |
+
"loss": 0.0,
|
| 1669 |
+
"step": 5575
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 18.12,
|
| 1673 |
+
"grad_norm": 0.0003356238012202084,
|
| 1674 |
+
"learning_rate": 3.2013333333333336e-06,
|
| 1675 |
+
"loss": 0.0,
|
| 1676 |
+
"step": 5600
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 18.2,
|
| 1680 |
+
"grad_norm": 0.0004451247223187238,
|
| 1681 |
+
"learning_rate": 3.1680000000000004e-06,
|
| 1682 |
+
"loss": 0.0,
|
| 1683 |
+
"step": 5625
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 18.28,
|
| 1687 |
+
"grad_norm": 0.0006252903258427978,
|
| 1688 |
+
"learning_rate": 3.134666666666667e-06,
|
| 1689 |
+
"loss": 0.0,
|
| 1690 |
+
"step": 5650
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 18.37,
|
| 1694 |
+
"grad_norm": 0.00027708217385225,
|
| 1695 |
+
"learning_rate": 3.1013333333333333e-06,
|
| 1696 |
+
"loss": 0.0,
|
| 1697 |
+
"step": 5675
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 18.45,
|
| 1701 |
+
"grad_norm": 0.00046747917076572776,
|
| 1702 |
+
"learning_rate": 3.0680000000000005e-06,
|
| 1703 |
+
"loss": 0.0,
|
| 1704 |
+
"step": 5700
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 18.53,
|
| 1708 |
+
"grad_norm": 0.0009358280221931636,
|
| 1709 |
+
"learning_rate": 3.034666666666667e-06,
|
| 1710 |
+
"loss": 0.0,
|
| 1711 |
+
"step": 5725
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 18.61,
|
| 1715 |
+
"grad_norm": 0.0007056727772578597,
|
| 1716 |
+
"learning_rate": 3.0013333333333334e-06,
|
| 1717 |
+
"loss": 0.0,
|
| 1718 |
+
"step": 5750
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 18.69,
|
| 1722 |
+
"grad_norm": 0.00043415065738372505,
|
| 1723 |
+
"learning_rate": 2.9680000000000002e-06,
|
| 1724 |
+
"loss": 0.0,
|
| 1725 |
+
"step": 5775
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 18.77,
|
| 1729 |
+
"grad_norm": 0.0005403661634773016,
|
| 1730 |
+
"learning_rate": 2.9346666666666666e-06,
|
| 1731 |
+
"loss": 0.0,
|
| 1732 |
+
"step": 5800
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 18.85,
|
| 1736 |
+
"grad_norm": 0.0002807588898576796,
|
| 1737 |
+
"learning_rate": 2.901333333333334e-06,
|
| 1738 |
+
"loss": 0.0,
|
| 1739 |
+
"step": 5825
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 18.93,
|
| 1743 |
+
"grad_norm": 0.00064972008112818,
|
| 1744 |
+
"learning_rate": 2.8680000000000003e-06,
|
| 1745 |
+
"loss": 0.0,
|
| 1746 |
+
"step": 5850
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 19.01,
|
| 1750 |
+
"grad_norm": 0.00047511656885035336,
|
| 1751 |
+
"learning_rate": 2.8346666666666667e-06,
|
| 1752 |
+
"loss": 0.0,
|
| 1753 |
+
"step": 5875
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 19.09,
|
| 1757 |
+
"grad_norm": 0.0008176405099220574,
|
| 1758 |
+
"learning_rate": 2.8013333333333336e-06,
|
| 1759 |
+
"loss": 0.0,
|
| 1760 |
+
"step": 5900
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 19.17,
|
| 1764 |
+
"grad_norm": 0.0006749618332833052,
|
| 1765 |
+
"learning_rate": 2.768e-06,
|
| 1766 |
+
"loss": 0.0,
|
| 1767 |
+
"step": 5925
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 19.26,
|
| 1771 |
+
"grad_norm": 0.00030714174499735236,
|
| 1772 |
+
"learning_rate": 2.7346666666666673e-06,
|
| 1773 |
+
"loss": 0.0,
|
| 1774 |
+
"step": 5950
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 19.34,
|
| 1778 |
+
"grad_norm": 0.0002531746868044138,
|
| 1779 |
+
"learning_rate": 2.7013333333333337e-06,
|
| 1780 |
+
"loss": 0.0,
|
| 1781 |
+
"step": 5975
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 19.42,
|
| 1785 |
+
"grad_norm": 0.0002513741492293775,
|
| 1786 |
+
"learning_rate": 2.668e-06,
|
| 1787 |
+
"loss": 0.0,
|
| 1788 |
+
"step": 6000
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 19.42,
|
| 1792 |
+
"eval_loss": 0.01877419464290142,
|
| 1793 |
+
"eval_runtime": 802.3536,
|
| 1794 |
+
"eval_samples_per_second": 1.537,
|
| 1795 |
+
"eval_steps_per_second": 0.193,
|
| 1796 |
+
"eval_wer": 3.1803335052432526,
|
| 1797 |
+
"step": 6000
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 19.5,
|
| 1801 |
+
"grad_norm": 0.00022407456708606333,
|
| 1802 |
+
"learning_rate": 2.634666666666667e-06,
|
| 1803 |
+
"loss": 0.0,
|
| 1804 |
+
"step": 6025
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 19.58,
|
| 1808 |
+
"grad_norm": 0.0006584114744327962,
|
| 1809 |
+
"learning_rate": 2.6013333333333334e-06,
|
| 1810 |
+
"loss": 0.0,
|
| 1811 |
+
"step": 6050
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 19.66,
|
| 1815 |
+
"grad_norm": 0.0003251735179219395,
|
| 1816 |
+
"learning_rate": 2.568e-06,
|
| 1817 |
+
"loss": 0.0,
|
| 1818 |
+
"step": 6075
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 19.74,
|
| 1822 |
+
"grad_norm": 0.00032441489747725427,
|
| 1823 |
+
"learning_rate": 2.534666666666667e-06,
|
| 1824 |
+
"loss": 0.0,
|
| 1825 |
+
"step": 6100
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 19.82,
|
| 1829 |
+
"grad_norm": 0.000556399580091238,
|
| 1830 |
+
"learning_rate": 2.5013333333333335e-06,
|
| 1831 |
+
"loss": 0.0,
|
| 1832 |
+
"step": 6125
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 19.9,
|
| 1836 |
+
"grad_norm": 0.0003794110380113125,
|
| 1837 |
+
"learning_rate": 2.468e-06,
|
| 1838 |
+
"loss": 0.0,
|
| 1839 |
+
"step": 6150
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 19.98,
|
| 1843 |
+
"grad_norm": 0.0008817287161946297,
|
| 1844 |
+
"learning_rate": 2.4346666666666668e-06,
|
| 1845 |
+
"loss": 0.0,
|
| 1846 |
+
"step": 6175
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 20.06,
|
| 1850 |
+
"grad_norm": 0.0003605467500165105,
|
| 1851 |
+
"learning_rate": 2.4013333333333336e-06,
|
| 1852 |
+
"loss": 0.0,
|
| 1853 |
+
"step": 6200
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 20.15,
|
| 1857 |
+
"grad_norm": 0.00034257510560564697,
|
| 1858 |
+
"learning_rate": 2.3680000000000005e-06,
|
| 1859 |
+
"loss": 0.0,
|
| 1860 |
+
"step": 6225
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 20.23,
|
| 1864 |
+
"grad_norm": 0.0006304741837084293,
|
| 1865 |
+
"learning_rate": 2.334666666666667e-06,
|
| 1866 |
+
"loss": 0.0,
|
| 1867 |
+
"step": 6250
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 20.31,
|
| 1871 |
+
"grad_norm": 0.0003804029547609389,
|
| 1872 |
+
"learning_rate": 2.3013333333333333e-06,
|
| 1873 |
+
"loss": 0.0,
|
| 1874 |
+
"step": 6275
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 20.39,
|
| 1878 |
+
"grad_norm": 0.0003537291777320206,
|
| 1879 |
+
"learning_rate": 2.268e-06,
|
| 1880 |
+
"loss": 0.0,
|
| 1881 |
+
"step": 6300
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 20.47,
|
| 1885 |
+
"grad_norm": 0.0005464838468469679,
|
| 1886 |
+
"learning_rate": 2.234666666666667e-06,
|
| 1887 |
+
"loss": 0.0,
|
| 1888 |
+
"step": 6325
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 20.55,
|
| 1892 |
+
"grad_norm": 0.00035579383256845176,
|
| 1893 |
+
"learning_rate": 2.2013333333333334e-06,
|
| 1894 |
+
"loss": 0.0,
|
| 1895 |
+
"step": 6350
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 20.63,
|
| 1899 |
+
"grad_norm": 0.00029211092623881996,
|
| 1900 |
+
"learning_rate": 2.1680000000000002e-06,
|
| 1901 |
+
"loss": 0.0,
|
| 1902 |
+
"step": 6375
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 20.71,
|
| 1906 |
+
"grad_norm": 0.00046534286229871213,
|
| 1907 |
+
"learning_rate": 2.1346666666666667e-06,
|
| 1908 |
+
"loss": 0.0,
|
| 1909 |
+
"step": 6400
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 20.79,
|
| 1913 |
+
"grad_norm": 0.0003821933933068067,
|
| 1914 |
+
"learning_rate": 2.1013333333333335e-06,
|
| 1915 |
+
"loss": 0.0,
|
| 1916 |
+
"step": 6425
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 20.87,
|
| 1920 |
+
"grad_norm": 0.00034173333551734686,
|
| 1921 |
+
"learning_rate": 2.0680000000000004e-06,
|
| 1922 |
+
"loss": 0.0,
|
| 1923 |
+
"step": 6450
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 20.95,
|
| 1927 |
+
"grad_norm": 0.00022973580053076148,
|
| 1928 |
+
"learning_rate": 2.0346666666666668e-06,
|
| 1929 |
+
"loss": 0.0,
|
| 1930 |
+
"step": 6475
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 21.04,
|
| 1934 |
+
"grad_norm": 0.00030270396382547915,
|
| 1935 |
+
"learning_rate": 2.0013333333333336e-06,
|
| 1936 |
+
"loss": 0.0,
|
| 1937 |
+
"step": 6500
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 21.04,
|
| 1941 |
+
"eval_loss": 0.018963992595672607,
|
| 1942 |
+
"eval_runtime": 809.4746,
|
| 1943 |
+
"eval_samples_per_second": 1.523,
|
| 1944 |
+
"eval_steps_per_second": 0.191,
|
| 1945 |
+
"eval_wer": 3.163142513323019,
|
| 1946 |
+
"step": 6500
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 21.12,
|
| 1950 |
+
"grad_norm": 0.00032841338543221354,
|
| 1951 |
+
"learning_rate": 1.968e-06,
|
| 1952 |
+
"loss": 0.0,
|
| 1953 |
+
"step": 6525
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 21.2,
|
| 1957 |
+
"grad_norm": 0.0006685039843432605,
|
| 1958 |
+
"learning_rate": 1.934666666666667e-06,
|
| 1959 |
+
"loss": 0.0,
|
| 1960 |
+
"step": 6550
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 21.28,
|
| 1964 |
+
"grad_norm": 0.0002688696840777993,
|
| 1965 |
+
"learning_rate": 1.9013333333333333e-06,
|
| 1966 |
+
"loss": 0.0,
|
| 1967 |
+
"step": 6575
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 21.36,
|
| 1971 |
+
"grad_norm": 0.00033872510539367795,
|
| 1972 |
+
"learning_rate": 1.8680000000000002e-06,
|
| 1973 |
+
"loss": 0.0,
|
| 1974 |
+
"step": 6600
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 21.44,
|
| 1978 |
+
"grad_norm": 0.0004930770955979824,
|
| 1979 |
+
"learning_rate": 1.8346666666666668e-06,
|
| 1980 |
+
"loss": 0.0,
|
| 1981 |
+
"step": 6625
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 21.52,
|
| 1985 |
+
"grad_norm": 0.00019405547936912626,
|
| 1986 |
+
"learning_rate": 1.8013333333333336e-06,
|
| 1987 |
+
"loss": 0.0,
|
| 1988 |
+
"step": 6650
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 21.6,
|
| 1992 |
+
"grad_norm": 0.00027933728415519,
|
| 1993 |
+
"learning_rate": 1.7680000000000003e-06,
|
| 1994 |
+
"loss": 0.0,
|
| 1995 |
+
"step": 6675
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 21.68,
|
| 1999 |
+
"grad_norm": 0.0006270716548897326,
|
| 2000 |
+
"learning_rate": 1.7346666666666667e-06,
|
| 2001 |
+
"loss": 0.0,
|
| 2002 |
+
"step": 6700
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 21.76,
|
| 2006 |
+
"grad_norm": 0.00039901587297208607,
|
| 2007 |
+
"learning_rate": 1.7013333333333335e-06,
|
| 2008 |
+
"loss": 0.0,
|
| 2009 |
+
"step": 6725
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 21.84,
|
| 2013 |
+
"grad_norm": 0.0003480367304291576,
|
| 2014 |
+
"learning_rate": 1.6680000000000002e-06,
|
| 2015 |
+
"loss": 0.0,
|
| 2016 |
+
"step": 6750
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 21.93,
|
| 2020 |
+
"grad_norm": 0.00031261061667464674,
|
| 2021 |
+
"learning_rate": 1.6346666666666668e-06,
|
| 2022 |
+
"loss": 0.0,
|
| 2023 |
+
"step": 6775
|
| 2024 |
+
},
|
| 2025 |
+
{
|
| 2026 |
+
"epoch": 22.01,
|
| 2027 |
+
"grad_norm": 0.0002611794334370643,
|
| 2028 |
+
"learning_rate": 1.6013333333333334e-06,
|
| 2029 |
+
"loss": 0.0,
|
| 2030 |
+
"step": 6800
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 22.09,
|
| 2034 |
+
"grad_norm": 0.00025896457373164594,
|
| 2035 |
+
"learning_rate": 1.568e-06,
|
| 2036 |
+
"loss": 0.0,
|
| 2037 |
+
"step": 6825
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 22.17,
|
| 2041 |
+
"grad_norm": 0.00022639252711087465,
|
| 2042 |
+
"learning_rate": 1.534666666666667e-06,
|
| 2043 |
+
"loss": 0.0,
|
| 2044 |
+
"step": 6850
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 22.25,
|
| 2048 |
+
"grad_norm": 0.0004359305021353066,
|
| 2049 |
+
"learning_rate": 1.5013333333333335e-06,
|
| 2050 |
+
"loss": 0.0,
|
| 2051 |
+
"step": 6875
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 22.33,
|
| 2055 |
+
"grad_norm": 0.00036377765354700387,
|
| 2056 |
+
"learning_rate": 1.4680000000000002e-06,
|
| 2057 |
+
"loss": 0.0,
|
| 2058 |
+
"step": 6900
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 22.41,
|
| 2062 |
+
"grad_norm": 0.0004344619228504598,
|
| 2063 |
+
"learning_rate": 1.4346666666666668e-06,
|
| 2064 |
+
"loss": 0.0,
|
| 2065 |
+
"step": 6925
|
| 2066 |
+
},
|
| 2067 |
+
{
|
| 2068 |
+
"epoch": 22.49,
|
| 2069 |
+
"grad_norm": 0.00028026316431351006,
|
| 2070 |
+
"learning_rate": 1.4013333333333334e-06,
|
| 2071 |
+
"loss": 0.0,
|
| 2072 |
+
"step": 6950
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 22.57,
|
| 2076 |
+
"grad_norm": 0.00035817205207422376,
|
| 2077 |
+
"learning_rate": 1.368e-06,
|
| 2078 |
+
"loss": 0.0,
|
| 2079 |
+
"step": 6975
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 22.65,
|
| 2083 |
+
"grad_norm": 0.0002789015998132527,
|
| 2084 |
+
"learning_rate": 1.334666666666667e-06,
|
| 2085 |
+
"loss": 0.0,
|
| 2086 |
+
"step": 7000
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 22.65,
|
| 2090 |
+
"eval_loss": 0.019092679023742676,
|
| 2091 |
+
"eval_runtime": 801.9043,
|
| 2092 |
+
"eval_samples_per_second": 1.538,
|
| 2093 |
+
"eval_steps_per_second": 0.193,
|
| 2094 |
+
"eval_wer": 3.163142513323019,
|
| 2095 |
+
"step": 7000
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 22.73,
|
| 2099 |
+
"grad_norm": 0.0006544535863213241,
|
| 2100 |
+
"learning_rate": 1.3013333333333333e-06,
|
| 2101 |
+
"loss": 0.0,
|
| 2102 |
+
"step": 7025
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 22.82,
|
| 2106 |
+
"grad_norm": 0.00028150764410384,
|
| 2107 |
+
"learning_rate": 1.268e-06,
|
| 2108 |
+
"loss": 0.0,
|
| 2109 |
+
"step": 7050
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 22.9,
|
| 2113 |
+
"grad_norm": 0.0002222236798843369,
|
| 2114 |
+
"learning_rate": 1.2346666666666668e-06,
|
| 2115 |
+
"loss": 0.0,
|
| 2116 |
+
"step": 7075
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 22.98,
|
| 2120 |
+
"grad_norm": 0.00017937009397428483,
|
| 2121 |
+
"learning_rate": 1.2013333333333334e-06,
|
| 2122 |
+
"loss": 0.0,
|
| 2123 |
+
"step": 7100
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 23.06,
|
| 2127 |
+
"grad_norm": 0.00022395767155103385,
|
| 2128 |
+
"learning_rate": 1.168e-06,
|
| 2129 |
+
"loss": 0.0,
|
| 2130 |
+
"step": 7125
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 23.14,
|
| 2134 |
+
"grad_norm": 0.0005528877954930067,
|
| 2135 |
+
"learning_rate": 1.1346666666666667e-06,
|
| 2136 |
+
"loss": 0.0,
|
| 2137 |
+
"step": 7150
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 23.22,
|
| 2141 |
+
"grad_norm": 0.00019384044571779668,
|
| 2142 |
+
"learning_rate": 1.1013333333333333e-06,
|
| 2143 |
+
"loss": 0.0,
|
| 2144 |
+
"step": 7175
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 23.3,
|
| 2148 |
+
"grad_norm": 0.0004787093785125762,
|
| 2149 |
+
"learning_rate": 1.0680000000000002e-06,
|
| 2150 |
+
"loss": 0.0,
|
| 2151 |
+
"step": 7200
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 23.38,
|
| 2155 |
+
"grad_norm": 0.0002817475178744644,
|
| 2156 |
+
"learning_rate": 1.0346666666666668e-06,
|
| 2157 |
+
"loss": 0.0,
|
| 2158 |
+
"step": 7225
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 23.46,
|
| 2162 |
+
"grad_norm": 0.000205877106054686,
|
| 2163 |
+
"learning_rate": 1.0013333333333335e-06,
|
| 2164 |
+
"loss": 0.0,
|
| 2165 |
+
"step": 7250
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 23.54,
|
| 2169 |
+
"grad_norm": 0.0002100019046338275,
|
| 2170 |
+
"learning_rate": 9.68e-07,
|
| 2171 |
+
"loss": 0.0,
|
| 2172 |
+
"step": 7275
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 23.62,
|
| 2176 |
+
"grad_norm": 0.00047292860108427703,
|
| 2177 |
+
"learning_rate": 9.346666666666668e-07,
|
| 2178 |
+
"loss": 0.0,
|
| 2179 |
+
"step": 7300
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 23.71,
|
| 2183 |
+
"grad_norm": 0.00027352923643775284,
|
| 2184 |
+
"learning_rate": 9.013333333333334e-07,
|
| 2185 |
+
"loss": 0.0,
|
| 2186 |
+
"step": 7325
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 23.79,
|
| 2190 |
+
"grad_norm": 0.00024422837304882705,
|
| 2191 |
+
"learning_rate": 8.680000000000001e-07,
|
| 2192 |
+
"loss": 0.0,
|
| 2193 |
+
"step": 7350
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 23.87,
|
| 2197 |
+
"grad_norm": 0.0003593063447624445,
|
| 2198 |
+
"learning_rate": 8.346666666666667e-07,
|
| 2199 |
+
"loss": 0.0,
|
| 2200 |
+
"step": 7375
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 23.95,
|
| 2204 |
+
"grad_norm": 0.00028577508055604994,
|
| 2205 |
+
"learning_rate": 8.013333333333334e-07,
|
| 2206 |
+
"loss": 0.0,
|
| 2207 |
+
"step": 7400
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 24.03,
|
| 2211 |
+
"grad_norm": 0.00022404512856155634,
|
| 2212 |
+
"learning_rate": 7.68e-07,
|
| 2213 |
+
"loss": 0.0,
|
| 2214 |
+
"step": 7425
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 24.11,
|
| 2218 |
+
"grad_norm": 0.000312491727527231,
|
| 2219 |
+
"learning_rate": 7.346666666666667e-07,
|
| 2220 |
+
"loss": 0.0,
|
| 2221 |
+
"step": 7450
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 24.19,
|
| 2225 |
+
"grad_norm": 0.0001654025836614892,
|
| 2226 |
+
"learning_rate": 7.013333333333335e-07,
|
| 2227 |
+
"loss": 0.0,
|
| 2228 |
+
"step": 7475
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 24.27,
|
| 2232 |
+
"grad_norm": 0.0004445198574103415,
|
| 2233 |
+
"learning_rate": 6.68e-07,
|
| 2234 |
+
"loss": 0.0,
|
| 2235 |
+
"step": 7500
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 24.27,
|
| 2239 |
+
"eval_loss": 0.019191982224583626,
|
| 2240 |
+
"eval_runtime": 797.2943,
|
| 2241 |
+
"eval_samples_per_second": 1.546,
|
| 2242 |
+
"eval_steps_per_second": 0.194,
|
| 2243 |
+
"eval_wer": 3.1803335052432526,
|
| 2244 |
+
"step": 7500
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 24.35,
|
| 2248 |
+
"grad_norm": 0.00019284736481495202,
|
| 2249 |
+
"learning_rate": 6.346666666666667e-07,
|
| 2250 |
+
"loss": 0.0,
|
| 2251 |
+
"step": 7525
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 24.43,
|
| 2255 |
+
"grad_norm": 0.00020476471399888396,
|
| 2256 |
+
"learning_rate": 6.013333333333334e-07,
|
| 2257 |
+
"loss": 0.0,
|
| 2258 |
+
"step": 7550
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 24.51,
|
| 2262 |
+
"grad_norm": 0.0003657681227196008,
|
| 2263 |
+
"learning_rate": 5.680000000000001e-07,
|
| 2264 |
+
"loss": 0.0,
|
| 2265 |
+
"step": 7575
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 24.6,
|
| 2269 |
+
"grad_norm": 0.00033989950316026807,
|
| 2270 |
+
"learning_rate": 5.346666666666667e-07,
|
| 2271 |
+
"loss": 0.0,
|
| 2272 |
+
"step": 7600
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 24.68,
|
| 2276 |
+
"grad_norm": 0.0003244269755668938,
|
| 2277 |
+
"learning_rate": 5.013333333333334e-07,
|
| 2278 |
+
"loss": 0.0,
|
| 2279 |
+
"step": 7625
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 24.76,
|
| 2283 |
+
"grad_norm": 0.0003897367278113961,
|
| 2284 |
+
"learning_rate": 4.6800000000000006e-07,
|
| 2285 |
+
"loss": 0.0,
|
| 2286 |
+
"step": 7650
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 24.84,
|
| 2290 |
+
"grad_norm": 0.0003248370485380292,
|
| 2291 |
+
"learning_rate": 4.346666666666667e-07,
|
| 2292 |
+
"loss": 0.0,
|
| 2293 |
+
"step": 7675
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 24.92,
|
| 2297 |
+
"grad_norm": 0.00026839631027542055,
|
| 2298 |
+
"learning_rate": 4.013333333333334e-07,
|
| 2299 |
+
"loss": 0.0,
|
| 2300 |
+
"step": 7700
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 25.0,
|
| 2304 |
+
"grad_norm": 0.0012187513057142496,
|
| 2305 |
+
"learning_rate": 3.68e-07,
|
| 2306 |
+
"loss": 0.0,
|
| 2307 |
+
"step": 7725
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 25.08,
|
| 2311 |
+
"grad_norm": 0.0006300898385234177,
|
| 2312 |
+
"learning_rate": 3.346666666666667e-07,
|
| 2313 |
+
"loss": 0.0,
|
| 2314 |
+
"step": 7750
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 25.16,
|
| 2318 |
+
"grad_norm": 0.0002933947544079274,
|
| 2319 |
+
"learning_rate": 3.013333333333334e-07,
|
| 2320 |
+
"loss": 0.0,
|
| 2321 |
+
"step": 7775
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 25.24,
|
| 2325 |
+
"grad_norm": 0.00026157114189118147,
|
| 2326 |
+
"learning_rate": 2.68e-07,
|
| 2327 |
+
"loss": 0.0,
|
| 2328 |
+
"step": 7800
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 25.32,
|
| 2332 |
+
"grad_norm": 0.0002829213335644454,
|
| 2333 |
+
"learning_rate": 2.3466666666666668e-07,
|
| 2334 |
+
"loss": 0.0,
|
| 2335 |
+
"step": 7825
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 25.4,
|
| 2339 |
+
"grad_norm": 0.00021175014262553304,
|
| 2340 |
+
"learning_rate": 2.0133333333333334e-07,
|
| 2341 |
+
"loss": 0.0,
|
| 2342 |
+
"step": 7850
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 25.49,
|
| 2346 |
+
"grad_norm": 0.00024275723262690008,
|
| 2347 |
+
"learning_rate": 1.68e-07,
|
| 2348 |
+
"loss": 0.0,
|
| 2349 |
+
"step": 7875
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 25.57,
|
| 2353 |
+
"grad_norm": 0.0002537575492169708,
|
| 2354 |
+
"learning_rate": 1.3466666666666668e-07,
|
| 2355 |
+
"loss": 0.0,
|
| 2356 |
+
"step": 7900
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 25.65,
|
| 2360 |
+
"grad_norm": 0.0002950854832306504,
|
| 2361 |
+
"learning_rate": 1.0133333333333334e-07,
|
| 2362 |
+
"loss": 0.0,
|
| 2363 |
+
"step": 7925
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 25.73,
|
| 2367 |
+
"grad_norm": 0.0002355340839130804,
|
| 2368 |
+
"learning_rate": 6.8e-08,
|
| 2369 |
+
"loss": 0.0,
|
| 2370 |
+
"step": 7950
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 25.81,
|
| 2374 |
+
"grad_norm": 0.00031494017457589507,
|
| 2375 |
+
"learning_rate": 3.4666666666666666e-08,
|
| 2376 |
+
"loss": 0.0,
|
| 2377 |
+
"step": 7975
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 25.89,
|
| 2381 |
+
"grad_norm": 0.00018703413661569357,
|
| 2382 |
+
"learning_rate": 1.3333333333333335e-09,
|
| 2383 |
+
"loss": 0.0,
|
| 2384 |
+
"step": 8000
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 25.89,
|
| 2388 |
+
"eval_loss": 0.019220387563109398,
|
| 2389 |
+
"eval_runtime": 799.5677,
|
| 2390 |
+
"eval_samples_per_second": 1.542,
|
| 2391 |
+
"eval_steps_per_second": 0.194,
|
| 2392 |
+
"eval_wer": 3.163142513323019,
|
| 2393 |
+
"step": 8000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 25.97,
|
| 2397 |
+
"grad_norm": 0.0001846036611823365,
|
| 2398 |
+
"learning_rate": 3.4573913043478264e-06,
|
| 2399 |
+
"loss": 0.0,
|
| 2400 |
+
"step": 8025
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 26.05,
|
| 2404 |
+
"grad_norm": 0.0002695720177143812,
|
| 2405 |
+
"learning_rate": 3.435652173913044e-06,
|
| 2406 |
+
"loss": 0.0,
|
| 2407 |
+
"step": 8050
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 26.13,
|
| 2411 |
+
"grad_norm": 0.0002683591446839273,
|
| 2412 |
+
"learning_rate": 3.4139130434782615e-06,
|
| 2413 |
+
"loss": 0.0,
|
| 2414 |
+
"step": 8075
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 26.21,
|
| 2418 |
+
"grad_norm": 0.00035550855682231486,
|
| 2419 |
+
"learning_rate": 3.3921739130434782e-06,
|
| 2420 |
+
"loss": 0.0,
|
| 2421 |
+
"step": 8100
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 26.29,
|
| 2425 |
+
"grad_norm": 0.00032984025892801583,
|
| 2426 |
+
"learning_rate": 3.370434782608696e-06,
|
| 2427 |
+
"loss": 0.0,
|
| 2428 |
+
"step": 8125
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 26.38,
|
| 2432 |
+
"grad_norm": 0.00020441650121938437,
|
| 2433 |
+
"learning_rate": 3.3486956521739134e-06,
|
| 2434 |
+
"loss": 0.0,
|
| 2435 |
+
"step": 8150
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 26.46,
|
| 2439 |
+
"grad_norm": 0.0003888191422447562,
|
| 2440 |
+
"learning_rate": 3.326956521739131e-06,
|
| 2441 |
+
"loss": 0.0,
|
| 2442 |
+
"step": 8175
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 26.54,
|
| 2446 |
+
"grad_norm": 0.00029555149376392365,
|
| 2447 |
+
"learning_rate": 3.3052173913043476e-06,
|
| 2448 |
+
"loss": 0.0,
|
| 2449 |
+
"step": 8200
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 26.62,
|
| 2453 |
+
"grad_norm": 0.0003228304849471897,
|
| 2454 |
+
"learning_rate": 3.283478260869565e-06,
|
| 2455 |
+
"loss": 0.0,
|
| 2456 |
+
"step": 8225
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 26.7,
|
| 2460 |
+
"grad_norm": 0.00024845023290254176,
|
| 2461 |
+
"learning_rate": 3.2617391304347828e-06,
|
| 2462 |
+
"loss": 0.0,
|
| 2463 |
+
"step": 8250
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 26.78,
|
| 2467 |
+
"grad_norm": 0.00021859569824300706,
|
| 2468 |
+
"learning_rate": 3.2400000000000003e-06,
|
| 2469 |
+
"loss": 0.0,
|
| 2470 |
+
"step": 8275
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 26.86,
|
| 2474 |
+
"grad_norm": 0.0005411595338955522,
|
| 2475 |
+
"learning_rate": 3.218260869565218e-06,
|
| 2476 |
+
"loss": 0.0,
|
| 2477 |
+
"step": 8300
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 26.94,
|
| 2481 |
+
"grad_norm": 0.00022630894090980291,
|
| 2482 |
+
"learning_rate": 3.196521739130435e-06,
|
| 2483 |
+
"loss": 0.0,
|
| 2484 |
+
"step": 8325
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 27.02,
|
| 2488 |
+
"grad_norm": 0.00018071234808303416,
|
| 2489 |
+
"learning_rate": 3.174782608695652e-06,
|
| 2490 |
+
"loss": 0.0,
|
| 2491 |
+
"step": 8350
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 27.1,
|
| 2495 |
+
"grad_norm": 0.0002543312730267644,
|
| 2496 |
+
"learning_rate": 3.1530434782608697e-06,
|
| 2497 |
+
"loss": 0.0,
|
| 2498 |
+
"step": 8375
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 27.18,
|
| 2502 |
+
"grad_norm": 0.00024747196584939957,
|
| 2503 |
+
"learning_rate": 3.1313043478260873e-06,
|
| 2504 |
+
"loss": 0.0,
|
| 2505 |
+
"step": 8400
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 27.27,
|
| 2509 |
+
"grad_norm": 0.0004030639829579741,
|
| 2510 |
+
"learning_rate": 3.1095652173913044e-06,
|
| 2511 |
+
"loss": 0.0,
|
| 2512 |
+
"step": 8425
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 27.35,
|
| 2516 |
+
"grad_norm": 0.0003001134900841862,
|
| 2517 |
+
"learning_rate": 3.087826086956522e-06,
|
| 2518 |
+
"loss": 0.0,
|
| 2519 |
+
"step": 8450
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 27.43,
|
| 2523 |
+
"grad_norm": 0.00024000635312404484,
|
| 2524 |
+
"learning_rate": 3.0660869565217395e-06,
|
| 2525 |
+
"loss": 0.0,
|
| 2526 |
+
"step": 8475
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 27.51,
|
| 2530 |
+
"grad_norm": 0.0002641078317537904,
|
| 2531 |
+
"learning_rate": 3.0443478260869567e-06,
|
| 2532 |
+
"loss": 0.0,
|
| 2533 |
+
"step": 8500
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 27.51,
|
| 2537 |
+
"eval_loss": 0.019592655822634697,
|
| 2538 |
+
"eval_runtime": 796.0106,
|
| 2539 |
+
"eval_samples_per_second": 1.549,
|
| 2540 |
+
"eval_steps_per_second": 0.195,
|
| 2541 |
+
"eval_wer": 3.2490974729241873,
|
| 2542 |
+
"step": 8500
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 27.59,
|
| 2546 |
+
"grad_norm": 0.00046909687807783484,
|
| 2547 |
+
"learning_rate": 3.0226086956521743e-06,
|
| 2548 |
+
"loss": 0.0,
|
| 2549 |
+
"step": 8525
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 27.67,
|
| 2553 |
+
"grad_norm": 0.00029794371221214533,
|
| 2554 |
+
"learning_rate": 3.0008695652173914e-06,
|
| 2555 |
+
"loss": 0.0,
|
| 2556 |
+
"step": 8550
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 27.75,
|
| 2560 |
+
"grad_norm": 0.00040284113492816687,
|
| 2561 |
+
"learning_rate": 2.979130434782609e-06,
|
| 2562 |
+
"loss": 0.0,
|
| 2563 |
+
"step": 8575
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 27.83,
|
| 2567 |
+
"grad_norm": 0.00018848094623535872,
|
| 2568 |
+
"learning_rate": 2.9573913043478265e-06,
|
| 2569 |
+
"loss": 0.0,
|
| 2570 |
+
"step": 8600
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 27.91,
|
| 2574 |
+
"grad_norm": 0.00014023661788087338,
|
| 2575 |
+
"learning_rate": 2.935652173913044e-06,
|
| 2576 |
+
"loss": 0.0,
|
| 2577 |
+
"step": 8625
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 27.99,
|
| 2581 |
+
"grad_norm": 0.0002997580450028181,
|
| 2582 |
+
"learning_rate": 2.913913043478261e-06,
|
| 2583 |
+
"loss": 0.0,
|
| 2584 |
+
"step": 8650
|
| 2585 |
+
},
|
| 2586 |
+
{
|
| 2587 |
+
"epoch": 28.07,
|
| 2588 |
+
"grad_norm": 0.0002090928319375962,
|
| 2589 |
+
"learning_rate": 2.8921739130434784e-06,
|
| 2590 |
+
"loss": 0.0,
|
| 2591 |
+
"step": 8675
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 28.16,
|
| 2595 |
+
"grad_norm": 0.0002271802513860166,
|
| 2596 |
+
"learning_rate": 2.870434782608696e-06,
|
| 2597 |
+
"loss": 0.0,
|
| 2598 |
+
"step": 8700
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 28.24,
|
| 2602 |
+
"grad_norm": 0.00020665275224018842,
|
| 2603 |
+
"learning_rate": 2.8486956521739135e-06,
|
| 2604 |
+
"loss": 0.0,
|
| 2605 |
+
"step": 8725
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 28.32,
|
| 2609 |
+
"grad_norm": 0.000151711908983998,
|
| 2610 |
+
"learning_rate": 2.826956521739131e-06,
|
| 2611 |
+
"loss": 0.0,
|
| 2612 |
+
"step": 8750
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 28.4,
|
| 2616 |
+
"grad_norm": 0.00018002108845394105,
|
| 2617 |
+
"learning_rate": 2.8052173913043478e-06,
|
| 2618 |
+
"loss": 0.0,
|
| 2619 |
+
"step": 8775
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 28.48,
|
| 2623 |
+
"grad_norm": 0.00017080060206353664,
|
| 2624 |
+
"learning_rate": 2.7834782608695653e-06,
|
| 2625 |
+
"loss": 0.0,
|
| 2626 |
+
"step": 8800
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 28.56,
|
| 2630 |
+
"grad_norm": 0.00011185290350113064,
|
| 2631 |
+
"learning_rate": 2.761739130434783e-06,
|
| 2632 |
+
"loss": 0.0,
|
| 2633 |
+
"step": 8825
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 28.64,
|
| 2637 |
+
"grad_norm": 0.00017247968935407698,
|
| 2638 |
+
"learning_rate": 2.7400000000000004e-06,
|
| 2639 |
+
"loss": 0.0,
|
| 2640 |
+
"step": 8850
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 28.72,
|
| 2644 |
+
"grad_norm": 0.0001423845678800717,
|
| 2645 |
+
"learning_rate": 2.7182608695652176e-06,
|
| 2646 |
+
"loss": 0.0,
|
| 2647 |
+
"step": 8875
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 28.8,
|
| 2651 |
+
"grad_norm": 0.00012165143562015146,
|
| 2652 |
+
"learning_rate": 2.6965217391304347e-06,
|
| 2653 |
+
"loss": 0.0,
|
| 2654 |
+
"step": 8900
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 28.88,
|
| 2658 |
+
"grad_norm": 0.00017706782091408968,
|
| 2659 |
+
"learning_rate": 2.6747826086956523e-06,
|
| 2660 |
+
"loss": 0.0,
|
| 2661 |
+
"step": 8925
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 28.96,
|
| 2665 |
+
"grad_norm": 0.00015399801486637443,
|
| 2666 |
+
"learning_rate": 2.65304347826087e-06,
|
| 2667 |
+
"loss": 0.0,
|
| 2668 |
+
"step": 8950
|
| 2669 |
+
},
|
| 2670 |
+
{
|
| 2671 |
+
"epoch": 29.05,
|
| 2672 |
+
"grad_norm": 0.00011916056246263906,
|
| 2673 |
+
"learning_rate": 2.631304347826087e-06,
|
| 2674 |
+
"loss": 0.0,
|
| 2675 |
+
"step": 8975
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 29.13,
|
| 2679 |
+
"grad_norm": 0.00012513797264546156,
|
| 2680 |
+
"learning_rate": 2.6095652173913046e-06,
|
| 2681 |
+
"loss": 0.0,
|
| 2682 |
+
"step": 9000
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 29.13,
|
| 2686 |
+
"eval_loss": 0.019877396523952484,
|
| 2687 |
+
"eval_runtime": 789.2205,
|
| 2688 |
+
"eval_samples_per_second": 1.562,
|
| 2689 |
+
"eval_steps_per_second": 0.196,
|
| 2690 |
+
"eval_wer": 3.2490974729241873,
|
| 2691 |
+
"step": 9000
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 29.21,
|
| 2695 |
+
"grad_norm": 0.0001386833901051432,
|
| 2696 |
+
"learning_rate": 2.587826086956522e-06,
|
| 2697 |
+
"loss": 0.0,
|
| 2698 |
+
"step": 9025
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 29.29,
|
| 2702 |
+
"grad_norm": 0.0002729986154008657,
|
| 2703 |
+
"learning_rate": 2.5660869565217393e-06,
|
| 2704 |
+
"loss": 0.0,
|
| 2705 |
+
"step": 9050
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 29.37,
|
| 2709 |
+
"grad_norm": 0.00016692836652509868,
|
| 2710 |
+
"learning_rate": 2.544347826086957e-06,
|
| 2711 |
+
"loss": 0.0,
|
| 2712 |
+
"step": 9075
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 29.45,
|
| 2716 |
+
"grad_norm": 0.00012008147314190865,
|
| 2717 |
+
"learning_rate": 2.522608695652174e-06,
|
| 2718 |
+
"loss": 0.0,
|
| 2719 |
+
"step": 9100
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 29.53,
|
| 2723 |
+
"grad_norm": 0.0001401961490046233,
|
| 2724 |
+
"learning_rate": 2.5008695652173915e-06,
|
| 2725 |
+
"loss": 0.0,
|
| 2726 |
+
"step": 9125
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 29.61,
|
| 2730 |
+
"grad_norm": 0.0001520535151939839,
|
| 2731 |
+
"learning_rate": 2.479130434782609e-06,
|
| 2732 |
+
"loss": 0.0,
|
| 2733 |
+
"step": 9150
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 29.69,
|
| 2737 |
+
"grad_norm": 0.00014652697427663952,
|
| 2738 |
+
"learning_rate": 2.4573913043478262e-06,
|
| 2739 |
+
"loss": 0.0,
|
| 2740 |
+
"step": 9175
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 29.77,
|
| 2744 |
+
"grad_norm": 0.00019798542780335993,
|
| 2745 |
+
"learning_rate": 2.4356521739130438e-06,
|
| 2746 |
+
"loss": 0.0,
|
| 2747 |
+
"step": 9200
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 29.85,
|
| 2751 |
+
"grad_norm": 0.00010751246736617759,
|
| 2752 |
+
"learning_rate": 2.413913043478261e-06,
|
| 2753 |
+
"loss": 0.0,
|
| 2754 |
+
"step": 9225
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 29.94,
|
| 2758 |
+
"grad_norm": 0.0002326727844774723,
|
| 2759 |
+
"learning_rate": 2.3921739130434785e-06,
|
| 2760 |
+
"loss": 0.0,
|
| 2761 |
+
"step": 9250
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 30.02,
|
| 2765 |
+
"grad_norm": 0.0002426872670184821,
|
| 2766 |
+
"learning_rate": 2.370434782608696e-06,
|
| 2767 |
+
"loss": 0.0,
|
| 2768 |
+
"step": 9275
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 30.1,
|
| 2772 |
+
"grad_norm": 0.0001499813370173797,
|
| 2773 |
+
"learning_rate": 2.348695652173913e-06,
|
| 2774 |
+
"loss": 0.0,
|
| 2775 |
+
"step": 9300
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 30.18,
|
| 2779 |
+
"grad_norm": 0.00013380887685343623,
|
| 2780 |
+
"learning_rate": 2.3269565217391307e-06,
|
| 2781 |
+
"loss": 0.0,
|
| 2782 |
+
"step": 9325
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 30.26,
|
| 2786 |
+
"grad_norm": 0.00020537307136692107,
|
| 2787 |
+
"learning_rate": 2.305217391304348e-06,
|
| 2788 |
+
"loss": 0.0,
|
| 2789 |
+
"step": 9350
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 30.34,
|
| 2793 |
+
"grad_norm": 0.00012789785978384316,
|
| 2794 |
+
"learning_rate": 2.2834782608695655e-06,
|
| 2795 |
+
"loss": 0.0,
|
| 2796 |
+
"step": 9375
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 30.42,
|
| 2800 |
+
"grad_norm": 9.495256381342188e-05,
|
| 2801 |
+
"learning_rate": 2.2617391304347826e-06,
|
| 2802 |
+
"loss": 0.0,
|
| 2803 |
+
"step": 9400
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 30.5,
|
| 2807 |
+
"grad_norm": 0.00011715501750586554,
|
| 2808 |
+
"learning_rate": 2.24e-06,
|
| 2809 |
+
"loss": 0.0,
|
| 2810 |
+
"step": 9425
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 30.58,
|
| 2814 |
+
"grad_norm": 0.0001109652075683698,
|
| 2815 |
+
"learning_rate": 2.2182608695652173e-06,
|
| 2816 |
+
"loss": 0.0,
|
| 2817 |
+
"step": 9450
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 30.66,
|
| 2821 |
+
"grad_norm": 0.0001521283556940034,
|
| 2822 |
+
"learning_rate": 2.196521739130435e-06,
|
| 2823 |
+
"loss": 0.0,
|
| 2824 |
+
"step": 9475
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 30.74,
|
| 2828 |
+
"grad_norm": 8.459954551653937e-05,
|
| 2829 |
+
"learning_rate": 2.1747826086956524e-06,
|
| 2830 |
+
"loss": 0.0,
|
| 2831 |
+
"step": 9500
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 30.74,
|
| 2835 |
+
"eval_loss": 0.020166443660855293,
|
| 2836 |
+
"eval_runtime": 789.9916,
|
| 2837 |
+
"eval_samples_per_second": 1.561,
|
| 2838 |
+
"eval_steps_per_second": 0.196,
|
| 2839 |
+
"eval_wer": 3.2834794567646557,
|
| 2840 |
+
"step": 9500
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 30.83,
|
| 2844 |
+
"grad_norm": 0.00013170181773602962,
|
| 2845 |
+
"learning_rate": 2.1530434782608696e-06,
|
| 2846 |
+
"loss": 0.0,
|
| 2847 |
+
"step": 9525
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 30.91,
|
| 2851 |
+
"grad_norm": 0.00011664188787108287,
|
| 2852 |
+
"learning_rate": 2.131304347826087e-06,
|
| 2853 |
+
"loss": 0.0,
|
| 2854 |
+
"step": 9550
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 30.99,
|
| 2858 |
+
"grad_norm": 0.0001086597767425701,
|
| 2859 |
+
"learning_rate": 2.1095652173913047e-06,
|
| 2860 |
+
"loss": 0.0,
|
| 2861 |
+
"step": 9575
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 31.07,
|
| 2865 |
+
"grad_norm": 9.175323066301644e-05,
|
| 2866 |
+
"learning_rate": 2.087826086956522e-06,
|
| 2867 |
+
"loss": 0.0,
|
| 2868 |
+
"step": 9600
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 31.15,
|
| 2872 |
+
"grad_norm": 0.0001475626340834424,
|
| 2873 |
+
"learning_rate": 2.0660869565217394e-06,
|
| 2874 |
+
"loss": 0.0,
|
| 2875 |
+
"step": 9625
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 31.23,
|
| 2879 |
+
"grad_norm": 0.00022788842034060508,
|
| 2880 |
+
"learning_rate": 2.044347826086957e-06,
|
| 2881 |
+
"loss": 0.0,
|
| 2882 |
+
"step": 9650
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 31.31,
|
| 2886 |
+
"grad_norm": 0.00014438826474361122,
|
| 2887 |
+
"learning_rate": 2.022608695652174e-06,
|
| 2888 |
+
"loss": 0.0,
|
| 2889 |
+
"step": 9675
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 31.39,
|
| 2893 |
+
"grad_norm": 0.0001402303169015795,
|
| 2894 |
+
"learning_rate": 2.0008695652173916e-06,
|
| 2895 |
+
"loss": 0.0,
|
| 2896 |
+
"step": 9700
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 31.47,
|
| 2900 |
+
"grad_norm": 0.00013189137098379433,
|
| 2901 |
+
"learning_rate": 1.9791304347826088e-06,
|
| 2902 |
+
"loss": 0.0,
|
| 2903 |
+
"step": 9725
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 31.55,
|
| 2907 |
+
"grad_norm": 0.0002211190148955211,
|
| 2908 |
+
"learning_rate": 1.9573913043478263e-06,
|
| 2909 |
+
"loss": 0.0,
|
| 2910 |
+
"step": 9750
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 31.63,
|
| 2914 |
+
"grad_norm": 7.398930029012263e-05,
|
| 2915 |
+
"learning_rate": 1.935652173913044e-06,
|
| 2916 |
+
"loss": 0.0,
|
| 2917 |
+
"step": 9775
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 31.72,
|
| 2921 |
+
"grad_norm": 0.00014985835878178477,
|
| 2922 |
+
"learning_rate": 1.913913043478261e-06,
|
| 2923 |
+
"loss": 0.0,
|
| 2924 |
+
"step": 9800
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 31.8,
|
| 2928 |
+
"grad_norm": 0.00015712910681031644,
|
| 2929 |
+
"learning_rate": 1.8921739130434786e-06,
|
| 2930 |
+
"loss": 0.0,
|
| 2931 |
+
"step": 9825
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 31.88,
|
| 2935 |
+
"grad_norm": 0.00012111233081668615,
|
| 2936 |
+
"learning_rate": 1.8704347826086958e-06,
|
| 2937 |
+
"loss": 0.0,
|
| 2938 |
+
"step": 9850
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 31.96,
|
| 2942 |
+
"grad_norm": 0.0001239510893356055,
|
| 2943 |
+
"learning_rate": 1.8486956521739133e-06,
|
| 2944 |
+
"loss": 0.0,
|
| 2945 |
+
"step": 9875
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 32.04,
|
| 2949 |
+
"grad_norm": 0.00010567443678155541,
|
| 2950 |
+
"learning_rate": 1.8269565217391305e-06,
|
| 2951 |
+
"loss": 0.0,
|
| 2952 |
+
"step": 9900
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 32.12,
|
| 2956 |
+
"grad_norm": 9.18906953302212e-05,
|
| 2957 |
+
"learning_rate": 1.805217391304348e-06,
|
| 2958 |
+
"loss": 0.0,
|
| 2959 |
+
"step": 9925
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 32.2,
|
| 2963 |
+
"grad_norm": 7.413190905936062e-05,
|
| 2964 |
+
"learning_rate": 1.7834782608695654e-06,
|
| 2965 |
+
"loss": 0.0,
|
| 2966 |
+
"step": 9950
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 32.28,
|
| 2970 |
+
"grad_norm": 7.43623822927475e-05,
|
| 2971 |
+
"learning_rate": 1.7617391304347827e-06,
|
| 2972 |
+
"loss": 0.0,
|
| 2973 |
+
"step": 9975
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 32.36,
|
| 2977 |
+
"grad_norm": 0.00015115566202439368,
|
| 2978 |
+
"learning_rate": 1.74e-06,
|
| 2979 |
+
"loss": 0.0,
|
| 2980 |
+
"step": 10000
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 32.36,
|
| 2984 |
+
"eval_loss": 0.02041783183813095,
|
| 2985 |
+
"eval_runtime": 788.1335,
|
| 2986 |
+
"eval_samples_per_second": 1.564,
|
| 2987 |
+
"eval_steps_per_second": 0.197,
|
| 2988 |
+
"eval_wer": 3.2319064810039535,
|
| 2989 |
+
"step": 10000
|
| 2990 |
+
},
|
| 2991 |
+
{
|
| 2992 |
+
"epoch": 32.44,
|
| 2993 |
+
"grad_norm": 0.00013316261174622923,
|
| 2994 |
+
"learning_rate": 1.7182608695652176e-06,
|
| 2995 |
+
"loss": 0.0,
|
| 2996 |
+
"step": 10025
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 32.52,
|
| 3000 |
+
"grad_norm": 0.000142704215249978,
|
| 3001 |
+
"learning_rate": 1.696521739130435e-06,
|
| 3002 |
+
"loss": 0.0,
|
| 3003 |
+
"step": 10050
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 32.61,
|
| 3007 |
+
"grad_norm": 0.0001261725410586223,
|
| 3008 |
+
"learning_rate": 1.6747826086956523e-06,
|
| 3009 |
+
"loss": 0.0,
|
| 3010 |
+
"step": 10075
|
| 3011 |
+
},
|
| 3012 |
+
{
|
| 3013 |
+
"epoch": 32.69,
|
| 3014 |
+
"grad_norm": 0.00016059931658674031,
|
| 3015 |
+
"learning_rate": 1.65304347826087e-06,
|
| 3016 |
+
"loss": 0.0,
|
| 3017 |
+
"step": 10100
|
| 3018 |
+
},
|
| 3019 |
+
{
|
| 3020 |
+
"epoch": 32.77,
|
| 3021 |
+
"grad_norm": 0.00013550084258895367,
|
| 3022 |
+
"learning_rate": 1.631304347826087e-06,
|
| 3023 |
+
"loss": 0.0,
|
| 3024 |
+
"step": 10125
|
| 3025 |
+
},
|
| 3026 |
+
{
|
| 3027 |
+
"epoch": 32.85,
|
| 3028 |
+
"grad_norm": 8.947417518356815e-05,
|
| 3029 |
+
"learning_rate": 1.6095652173913046e-06,
|
| 3030 |
+
"loss": 0.0,
|
| 3031 |
+
"step": 10150
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 32.93,
|
| 3035 |
+
"grad_norm": 8.361903019249439e-05,
|
| 3036 |
+
"learning_rate": 1.5878260869565217e-06,
|
| 3037 |
+
"loss": 0.0,
|
| 3038 |
+
"step": 10175
|
| 3039 |
+
},
|
| 3040 |
+
{
|
| 3041 |
+
"epoch": 33.01,
|
| 3042 |
+
"grad_norm": 0.00011841193190775812,
|
| 3043 |
+
"learning_rate": 1.5660869565217393e-06,
|
| 3044 |
+
"loss": 0.0,
|
| 3045 |
+
"step": 10200
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 33.09,
|
| 3049 |
+
"grad_norm": 0.00011437163630034775,
|
| 3050 |
+
"learning_rate": 1.5443478260869567e-06,
|
| 3051 |
+
"loss": 0.0,
|
| 3052 |
+
"step": 10225
|
| 3053 |
+
},
|
| 3054 |
+
{
|
| 3055 |
+
"epoch": 33.17,
|
| 3056 |
+
"grad_norm": 0.0001065267642843537,
|
| 3057 |
+
"learning_rate": 1.522608695652174e-06,
|
| 3058 |
+
"loss": 0.0,
|
| 3059 |
+
"step": 10250
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 33.25,
|
| 3063 |
+
"grad_norm": 0.00010684959124773741,
|
| 3064 |
+
"learning_rate": 1.5008695652173914e-06,
|
| 3065 |
+
"loss": 0.0,
|
| 3066 |
+
"step": 10275
|
| 3067 |
+
},
|
| 3068 |
+
{
|
| 3069 |
+
"epoch": 33.33,
|
| 3070 |
+
"grad_norm": 7.87422395660542e-05,
|
| 3071 |
+
"learning_rate": 1.479130434782609e-06,
|
| 3072 |
+
"loss": 0.0,
|
| 3073 |
+
"step": 10300
|
| 3074 |
+
},
|
| 3075 |
+
{
|
| 3076 |
+
"epoch": 33.41,
|
| 3077 |
+
"grad_norm": 0.00013346740161068738,
|
| 3078 |
+
"learning_rate": 1.4573913043478263e-06,
|
| 3079 |
+
"loss": 0.0,
|
| 3080 |
+
"step": 10325
|
| 3081 |
+
},
|
| 3082 |
+
{
|
| 3083 |
+
"epoch": 33.5,
|
| 3084 |
+
"grad_norm": 8.954344230005518e-05,
|
| 3085 |
+
"learning_rate": 1.4356521739130436e-06,
|
| 3086 |
+
"loss": 0.0,
|
| 3087 |
+
"step": 10350
|
| 3088 |
+
},
|
| 3089 |
+
{
|
| 3090 |
+
"epoch": 33.58,
|
| 3091 |
+
"grad_norm": 0.00015138478192966431,
|
| 3092 |
+
"learning_rate": 1.4139130434782612e-06,
|
| 3093 |
+
"loss": 0.0,
|
| 3094 |
+
"step": 10375
|
| 3095 |
+
},
|
| 3096 |
+
{
|
| 3097 |
+
"epoch": 33.66,
|
| 3098 |
+
"grad_norm": 0.00010763735917862505,
|
| 3099 |
+
"learning_rate": 1.3921739130434783e-06,
|
| 3100 |
+
"loss": 0.0,
|
| 3101 |
+
"step": 10400
|
| 3102 |
+
},
|
| 3103 |
+
{
|
| 3104 |
+
"epoch": 33.74,
|
| 3105 |
+
"grad_norm": 0.00014248549996409565,
|
| 3106 |
+
"learning_rate": 1.3704347826086959e-06,
|
| 3107 |
+
"loss": 0.0,
|
| 3108 |
+
"step": 10425
|
| 3109 |
+
},
|
| 3110 |
+
{
|
| 3111 |
+
"epoch": 33.82,
|
| 3112 |
+
"grad_norm": 9.735339699545875e-05,
|
| 3113 |
+
"learning_rate": 1.348695652173913e-06,
|
| 3114 |
+
"loss": 0.0,
|
| 3115 |
+
"step": 10450
|
| 3116 |
+
},
|
| 3117 |
+
{
|
| 3118 |
+
"epoch": 33.9,
|
| 3119 |
+
"grad_norm": 0.0001395919971400872,
|
| 3120 |
+
"learning_rate": 1.3269565217391306e-06,
|
| 3121 |
+
"loss": 0.0,
|
| 3122 |
+
"step": 10475
|
| 3123 |
+
},
|
| 3124 |
+
{
|
| 3125 |
+
"epoch": 33.98,
|
| 3126 |
+
"grad_norm": 7.965522672748193e-05,
|
| 3127 |
+
"learning_rate": 1.3052173913043477e-06,
|
| 3128 |
+
"loss": 0.0,
|
| 3129 |
+
"step": 10500
|
| 3130 |
+
},
|
| 3131 |
+
{
|
| 3132 |
+
"epoch": 33.98,
|
| 3133 |
+
"eval_loss": 0.020673181861639023,
|
| 3134 |
+
"eval_runtime": 787.8771,
|
| 3135 |
+
"eval_samples_per_second": 1.565,
|
| 3136 |
+
"eval_steps_per_second": 0.197,
|
| 3137 |
+
"eval_wer": 3.2834794567646557,
|
| 3138 |
+
"step": 10500
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 34.06,
|
| 3142 |
+
"grad_norm": 5.796532423119061e-05,
|
| 3143 |
+
"learning_rate": 1.2834782608695653e-06,
|
| 3144 |
+
"loss": 0.0,
|
| 3145 |
+
"step": 10525
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 34.14,
|
| 3149 |
+
"grad_norm": 0.00011847128916997463,
|
| 3150 |
+
"learning_rate": 1.2617391304347826e-06,
|
| 3151 |
+
"loss": 0.0,
|
| 3152 |
+
"step": 10550
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 34.22,
|
| 3156 |
+
"grad_norm": 0.00010759083670563996,
|
| 3157 |
+
"learning_rate": 1.2400000000000002e-06,
|
| 3158 |
+
"loss": 0.0,
|
| 3159 |
+
"step": 10575
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 34.3,
|
| 3163 |
+
"grad_norm": 8.88080830918625e-05,
|
| 3164 |
+
"learning_rate": 1.2182608695652175e-06,
|
| 3165 |
+
"loss": 0.0,
|
| 3166 |
+
"step": 10600
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 34.39,
|
| 3170 |
+
"grad_norm": 0.0001434662553947419,
|
| 3171 |
+
"learning_rate": 1.196521739130435e-06,
|
| 3172 |
+
"loss": 0.0,
|
| 3173 |
+
"step": 10625
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 34.47,
|
| 3177 |
+
"grad_norm": 0.00010111245501320809,
|
| 3178 |
+
"learning_rate": 1.1747826086956523e-06,
|
| 3179 |
+
"loss": 0.0,
|
| 3180 |
+
"step": 10650
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 34.55,
|
| 3184 |
+
"grad_norm": 0.00010944180394290015,
|
| 3185 |
+
"learning_rate": 1.1530434782608696e-06,
|
| 3186 |
+
"loss": 0.0,
|
| 3187 |
+
"step": 10675
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 34.63,
|
| 3191 |
+
"grad_norm": 9.52091722865589e-05,
|
| 3192 |
+
"learning_rate": 1.1313043478260872e-06,
|
| 3193 |
+
"loss": 0.0,
|
| 3194 |
+
"step": 10700
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 34.71,
|
| 3198 |
+
"grad_norm": 9.008622146211565e-05,
|
| 3199 |
+
"learning_rate": 1.1095652173913045e-06,
|
| 3200 |
+
"loss": 0.0,
|
| 3201 |
+
"step": 10725
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 34.79,
|
| 3205 |
+
"grad_norm": 8.872839680407196e-05,
|
| 3206 |
+
"learning_rate": 1.0878260869565219e-06,
|
| 3207 |
+
"loss": 0.0,
|
| 3208 |
+
"step": 10750
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 34.87,
|
| 3212 |
+
"grad_norm": 9.502990724286065e-05,
|
| 3213 |
+
"learning_rate": 1.0660869565217392e-06,
|
| 3214 |
+
"loss": 0.0,
|
| 3215 |
+
"step": 10775
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 34.95,
|
| 3219 |
+
"grad_norm": 0.00011394196189939976,
|
| 3220 |
+
"learning_rate": 1.0443478260869566e-06,
|
| 3221 |
+
"loss": 0.0,
|
| 3222 |
+
"step": 10800
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 35.03,
|
| 3226 |
+
"grad_norm": 9.975407010642812e-05,
|
| 3227 |
+
"learning_rate": 1.022608695652174e-06,
|
| 3228 |
+
"loss": 0.0,
|
| 3229 |
+
"step": 10825
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 35.11,
|
| 3233 |
+
"grad_norm": 8.709837857168168e-05,
|
| 3234 |
+
"learning_rate": 1.0008695652173913e-06,
|
| 3235 |
+
"loss": 0.0,
|
| 3236 |
+
"step": 10850
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 35.19,
|
| 3240 |
+
"grad_norm": 7.526788249379024e-05,
|
| 3241 |
+
"learning_rate": 9.791304347826088e-07,
|
| 3242 |
+
"loss": 0.0,
|
| 3243 |
+
"step": 10875
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 35.28,
|
| 3247 |
+
"grad_norm": 7.429657125612721e-05,
|
| 3248 |
+
"learning_rate": 9.573913043478262e-07,
|
| 3249 |
+
"loss": 0.0,
|
| 3250 |
+
"step": 10900
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 35.36,
|
| 3254 |
+
"grad_norm": 8.827085548546165e-05,
|
| 3255 |
+
"learning_rate": 9.356521739130435e-07,
|
| 3256 |
+
"loss": 0.0,
|
| 3257 |
+
"step": 10925
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 35.44,
|
| 3261 |
+
"grad_norm": 9.55950454226695e-05,
|
| 3262 |
+
"learning_rate": 9.139130434782609e-07,
|
| 3263 |
+
"loss": 0.0,
|
| 3264 |
+
"step": 10950
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 35.52,
|
| 3268 |
+
"grad_norm": 7.405767246382311e-05,
|
| 3269 |
+
"learning_rate": 8.921739130434783e-07,
|
| 3270 |
+
"loss": 0.0,
|
| 3271 |
+
"step": 10975
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 35.6,
|
| 3275 |
+
"grad_norm": 0.0001504940155427903,
|
| 3276 |
+
"learning_rate": 8.704347826086958e-07,
|
| 3277 |
+
"loss": 0.0,
|
| 3278 |
+
"step": 11000
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 35.6,
|
| 3282 |
+
"eval_loss": 0.020881911739706993,
|
| 3283 |
+
"eval_runtime": 786.2062,
|
| 3284 |
+
"eval_samples_per_second": 1.568,
|
| 3285 |
+
"eval_steps_per_second": 0.197,
|
| 3286 |
+
"eval_wer": 3.266288464844422,
|
| 3287 |
+
"step": 11000
|
| 3288 |
+
}
|
| 3289 |
+
],
|
| 3290 |
+
"logging_steps": 25,
|
| 3291 |
+
"max_steps": 12000,
|
| 3292 |
+
"num_input_tokens_seen": 0,
|
| 3293 |
+
"num_train_epochs": 39,
|
| 3294 |
+
"save_steps": 1000,
|
| 3295 |
+
"total_flos": 5.06698245881856e+19,
|
| 3296 |
+
"train_batch_size": 16,
|
| 3297 |
+
"trial_name": null,
|
| 3298 |
+
"trial_params": null
|
| 3299 |
+
}
|
checkpoint-11000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25f43d4dd7a8551c57662ca9a88d8424f4510fd4145fbdde0c72911176d8013d
|
| 3 |
+
size 5112
|