Upload folder using huggingface_hub
Browse files- config.json +49 -0
- generation_config.json +241 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- preprocessor_config.json +14 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +2345 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-large-v3-turbo",
|
| 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 |
+
50256
|
| 13 |
+
],
|
| 14 |
+
"bos_token_id": 50257,
|
| 15 |
+
"classifier_proj_size": 256,
|
| 16 |
+
"d_model": 1280,
|
| 17 |
+
"decoder_attention_heads": 20,
|
| 18 |
+
"decoder_ffn_dim": 5120,
|
| 19 |
+
"decoder_layerdrop": 0.0,
|
| 20 |
+
"decoder_layers": 4,
|
| 21 |
+
"decoder_start_token_id": 50258,
|
| 22 |
+
"dropout": 0.0,
|
| 23 |
+
"encoder_attention_heads": 20,
|
| 24 |
+
"encoder_ffn_dim": 5120,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 32,
|
| 27 |
+
"eos_token_id": 50257,
|
| 28 |
+
"init_std": 0.02,
|
| 29 |
+
"is_encoder_decoder": true,
|
| 30 |
+
"mask_feature_length": 10,
|
| 31 |
+
"mask_feature_min_masks": 0,
|
| 32 |
+
"mask_feature_prob": 0.0,
|
| 33 |
+
"mask_time_length": 10,
|
| 34 |
+
"mask_time_min_masks": 2,
|
| 35 |
+
"mask_time_prob": 0.05,
|
| 36 |
+
"max_source_positions": 1500,
|
| 37 |
+
"max_target_positions": 448,
|
| 38 |
+
"median_filter_width": 7,
|
| 39 |
+
"model_type": "whisper",
|
| 40 |
+
"num_hidden_layers": 32,
|
| 41 |
+
"num_mel_bins": 128,
|
| 42 |
+
"pad_token_id": 50257,
|
| 43 |
+
"scale_embedding": false,
|
| 44 |
+
"torch_dtype": "float32",
|
| 45 |
+
"transformers_version": "4.45.1",
|
| 46 |
+
"use_cache": true,
|
| 47 |
+
"use_weighted_layer_sum": false,
|
| 48 |
+
"vocab_size": 51866
|
| 49 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,241 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alignment_heads": [
|
| 3 |
+
[
|
| 4 |
+
2,
|
| 5 |
+
4
|
| 6 |
+
],
|
| 7 |
+
[
|
| 8 |
+
2,
|
| 9 |
+
11
|
| 10 |
+
],
|
| 11 |
+
[
|
| 12 |
+
3,
|
| 13 |
+
3
|
| 14 |
+
],
|
| 15 |
+
[
|
| 16 |
+
3,
|
| 17 |
+
6
|
| 18 |
+
],
|
| 19 |
+
[
|
| 20 |
+
3,
|
| 21 |
+
11
|
| 22 |
+
],
|
| 23 |
+
[
|
| 24 |
+
3,
|
| 25 |
+
14
|
| 26 |
+
]
|
| 27 |
+
],
|
| 28 |
+
"begin_suppress_tokens": [
|
| 29 |
+
220,
|
| 30 |
+
50257
|
| 31 |
+
],
|
| 32 |
+
"bos_token_id": 50257,
|
| 33 |
+
"decoder_start_token_id": 50258,
|
| 34 |
+
"eos_token_id": 50257,
|
| 35 |
+
"is_multilingual": true,
|
| 36 |
+
"lang_to_id": {
|
| 37 |
+
"<|af|>": 50327,
|
| 38 |
+
"<|am|>": 50334,
|
| 39 |
+
"<|ar|>": 50272,
|
| 40 |
+
"<|as|>": 50350,
|
| 41 |
+
"<|az|>": 50304,
|
| 42 |
+
"<|ba|>": 50355,
|
| 43 |
+
"<|be|>": 50330,
|
| 44 |
+
"<|bg|>": 50292,
|
| 45 |
+
"<|bn|>": 50302,
|
| 46 |
+
"<|bo|>": 50347,
|
| 47 |
+
"<|br|>": 50309,
|
| 48 |
+
"<|bs|>": 50315,
|
| 49 |
+
"<|ca|>": 50270,
|
| 50 |
+
"<|cs|>": 50283,
|
| 51 |
+
"<|cy|>": 50297,
|
| 52 |
+
"<|da|>": 50285,
|
| 53 |
+
"<|de|>": 50261,
|
| 54 |
+
"<|el|>": 50281,
|
| 55 |
+
"<|en|>": 50259,
|
| 56 |
+
"<|es|>": 50262,
|
| 57 |
+
"<|et|>": 50307,
|
| 58 |
+
"<|eu|>": 50310,
|
| 59 |
+
"<|fa|>": 50300,
|
| 60 |
+
"<|fi|>": 50277,
|
| 61 |
+
"<|fo|>": 50338,
|
| 62 |
+
"<|fr|>": 50265,
|
| 63 |
+
"<|gl|>": 50319,
|
| 64 |
+
"<|gu|>": 50333,
|
| 65 |
+
"<|haw|>": 50352,
|
| 66 |
+
"<|ha|>": 50354,
|
| 67 |
+
"<|he|>": 50279,
|
| 68 |
+
"<|hi|>": 50276,
|
| 69 |
+
"<|hr|>": 50291,
|
| 70 |
+
"<|ht|>": 50339,
|
| 71 |
+
"<|hu|>": 50286,
|
| 72 |
+
"<|hy|>": 50312,
|
| 73 |
+
"<|id|>": 50275,
|
| 74 |
+
"<|is|>": 50311,
|
| 75 |
+
"<|it|>": 50274,
|
| 76 |
+
"<|ja|>": 50266,
|
| 77 |
+
"<|jw|>": 50356,
|
| 78 |
+
"<|ka|>": 50329,
|
| 79 |
+
"<|kk|>": 50316,
|
| 80 |
+
"<|km|>": 50323,
|
| 81 |
+
"<|kn|>": 50306,
|
| 82 |
+
"<|ko|>": 50264,
|
| 83 |
+
"<|la|>": 50294,
|
| 84 |
+
"<|lb|>": 50345,
|
| 85 |
+
"<|ln|>": 50353,
|
| 86 |
+
"<|lo|>": 50336,
|
| 87 |
+
"<|lt|>": 50293,
|
| 88 |
+
"<|lv|>": 50301,
|
| 89 |
+
"<|mg|>": 50349,
|
| 90 |
+
"<|mi|>": 50295,
|
| 91 |
+
"<|mk|>": 50308,
|
| 92 |
+
"<|ml|>": 50296,
|
| 93 |
+
"<|mn|>": 50314,
|
| 94 |
+
"<|mr|>": 50320,
|
| 95 |
+
"<|ms|>": 50282,
|
| 96 |
+
"<|mt|>": 50343,
|
| 97 |
+
"<|my|>": 50346,
|
| 98 |
+
"<|ne|>": 50313,
|
| 99 |
+
"<|nl|>": 50271,
|
| 100 |
+
"<|nn|>": 50342,
|
| 101 |
+
"<|no|>": 50288,
|
| 102 |
+
"<|oc|>": 50328,
|
| 103 |
+
"<|pa|>": 50321,
|
| 104 |
+
"<|pl|>": 50269,
|
| 105 |
+
"<|ps|>": 50340,
|
| 106 |
+
"<|pt|>": 50267,
|
| 107 |
+
"<|ro|>": 50284,
|
| 108 |
+
"<|ru|>": 50263,
|
| 109 |
+
"<|sa|>": 50344,
|
| 110 |
+
"<|sd|>": 50332,
|
| 111 |
+
"<|si|>": 50322,
|
| 112 |
+
"<|sk|>": 50298,
|
| 113 |
+
"<|sl|>": 50305,
|
| 114 |
+
"<|sn|>": 50324,
|
| 115 |
+
"<|so|>": 50326,
|
| 116 |
+
"<|sq|>": 50317,
|
| 117 |
+
"<|sr|>": 50303,
|
| 118 |
+
"<|su|>": 50357,
|
| 119 |
+
"<|sv|>": 50273,
|
| 120 |
+
"<|sw|>": 50318,
|
| 121 |
+
"<|ta|>": 50287,
|
| 122 |
+
"<|te|>": 50299,
|
| 123 |
+
"<|tg|>": 50331,
|
| 124 |
+
"<|th|>": 50289,
|
| 125 |
+
"<|tk|>": 50341,
|
| 126 |
+
"<|tl|>": 50348,
|
| 127 |
+
"<|tr|>": 50268,
|
| 128 |
+
"<|tt|>": 50351,
|
| 129 |
+
"<|uk|>": 50280,
|
| 130 |
+
"<|ur|>": 50290,
|
| 131 |
+
"<|uz|>": 50337,
|
| 132 |
+
"<|vi|>": 50278,
|
| 133 |
+
"<|yi|>": 50335,
|
| 134 |
+
"<|yo|>": 50325,
|
| 135 |
+
"<|yue|>": 50358,
|
| 136 |
+
"<|zh|>": 50260
|
| 137 |
+
},
|
| 138 |
+
"language": "Mongolian",
|
| 139 |
+
"max_initial_timestamp_index": 50,
|
| 140 |
+
"max_length": 448,
|
| 141 |
+
"no_timestamps_token_id": 50364,
|
| 142 |
+
"pad_token_id": 50257,
|
| 143 |
+
"prev_sot_token_id": 50362,
|
| 144 |
+
"return_timestamps": false,
|
| 145 |
+
"suppress_tokens": [
|
| 146 |
+
1,
|
| 147 |
+
2,
|
| 148 |
+
7,
|
| 149 |
+
8,
|
| 150 |
+
9,
|
| 151 |
+
10,
|
| 152 |
+
14,
|
| 153 |
+
25,
|
| 154 |
+
26,
|
| 155 |
+
27,
|
| 156 |
+
28,
|
| 157 |
+
29,
|
| 158 |
+
31,
|
| 159 |
+
58,
|
| 160 |
+
59,
|
| 161 |
+
60,
|
| 162 |
+
61,
|
| 163 |
+
62,
|
| 164 |
+
63,
|
| 165 |
+
90,
|
| 166 |
+
91,
|
| 167 |
+
92,
|
| 168 |
+
93,
|
| 169 |
+
359,
|
| 170 |
+
503,
|
| 171 |
+
522,
|
| 172 |
+
542,
|
| 173 |
+
873,
|
| 174 |
+
893,
|
| 175 |
+
902,
|
| 176 |
+
918,
|
| 177 |
+
922,
|
| 178 |
+
931,
|
| 179 |
+
1350,
|
| 180 |
+
1853,
|
| 181 |
+
1982,
|
| 182 |
+
2460,
|
| 183 |
+
2627,
|
| 184 |
+
3246,
|
| 185 |
+
3253,
|
| 186 |
+
3268,
|
| 187 |
+
3536,
|
| 188 |
+
3846,
|
| 189 |
+
3961,
|
| 190 |
+
4183,
|
| 191 |
+
4667,
|
| 192 |
+
6585,
|
| 193 |
+
6647,
|
| 194 |
+
7273,
|
| 195 |
+
9061,
|
| 196 |
+
9383,
|
| 197 |
+
10428,
|
| 198 |
+
10929,
|
| 199 |
+
11938,
|
| 200 |
+
12033,
|
| 201 |
+
12331,
|
| 202 |
+
12562,
|
| 203 |
+
13793,
|
| 204 |
+
14157,
|
| 205 |
+
14635,
|
| 206 |
+
15265,
|
| 207 |
+
15618,
|
| 208 |
+
16553,
|
| 209 |
+
16604,
|
| 210 |
+
18362,
|
| 211 |
+
18956,
|
| 212 |
+
20075,
|
| 213 |
+
21675,
|
| 214 |
+
22520,
|
| 215 |
+
26130,
|
| 216 |
+
26161,
|
| 217 |
+
26435,
|
| 218 |
+
28279,
|
| 219 |
+
29464,
|
| 220 |
+
31650,
|
| 221 |
+
32302,
|
| 222 |
+
32470,
|
| 223 |
+
36865,
|
| 224 |
+
42863,
|
| 225 |
+
47425,
|
| 226 |
+
49870,
|
| 227 |
+
50254,
|
| 228 |
+
50258,
|
| 229 |
+
50359,
|
| 230 |
+
50360,
|
| 231 |
+
50361,
|
| 232 |
+
50362,
|
| 233 |
+
50363
|
| 234 |
+
],
|
| 235 |
+
"task": "transcribe",
|
| 236 |
+
"task_to_id": {
|
| 237 |
+
"transcribe": 50360,
|
| 238 |
+
"translate": 50359
|
| 239 |
+
},
|
| 240 |
+
"transformers_version": "4.45.1"
|
| 241 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8aee478900f727946e3634b5cf15cc75be09e6d3af95bede3c87172ee207629
|
| 3 |
+
size 3235581408
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9accc831b22227a8e7a217b6f2d51e305e7f1ab095863193ad6bdd0ab594770b
|
| 3 |
+
size 6456162404
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 128,
|
| 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 |
+
}
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4a331edbe622d3be302f3a29820de21988f34c13f3c08b82fe8d4e50768b2dde
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bcd65f5c5ae513a1ca88894657fa81d516edccd9692a6884989dee9560a972a1
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2345 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 15.246076710047603,
|
| 3 |
+
"best_model_checkpoint": "./whisper-turbo/checkpoint-8000",
|
| 4 |
+
"epoch": 3.4057045551298426,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 8000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.010642826734780758,
|
| 13 |
+
"grad_norm": 24.32292938232422,
|
| 14 |
+
"learning_rate": 2.4000000000000003e-07,
|
| 15 |
+
"loss": 2.3553,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.021285653469561516,
|
| 20 |
+
"grad_norm": 13.848426818847656,
|
| 21 |
+
"learning_rate": 4.900000000000001e-07,
|
| 22 |
+
"loss": 1.8027,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.031928480204342274,
|
| 27 |
+
"grad_norm": 12.110209465026855,
|
| 28 |
+
"learning_rate": 7.4e-07,
|
| 29 |
+
"loss": 1.43,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.04257130693912303,
|
| 34 |
+
"grad_norm": 11.48365306854248,
|
| 35 |
+
"learning_rate": 9.9e-07,
|
| 36 |
+
"loss": 1.2826,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.05321413367390379,
|
| 41 |
+
"grad_norm": 12.327783584594727,
|
| 42 |
+
"learning_rate": 1.2400000000000002e-06,
|
| 43 |
+
"loss": 1.0584,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.06385696040868455,
|
| 48 |
+
"grad_norm": 10.366923332214355,
|
| 49 |
+
"learning_rate": 1.4900000000000001e-06,
|
| 50 |
+
"loss": 0.9736,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.07449978714346531,
|
| 55 |
+
"grad_norm": 12.267701148986816,
|
| 56 |
+
"learning_rate": 1.74e-06,
|
| 57 |
+
"loss": 0.9528,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.08514261387824607,
|
| 62 |
+
"grad_norm": 10.672560691833496,
|
| 63 |
+
"learning_rate": 1.9900000000000004e-06,
|
| 64 |
+
"loss": 0.8758,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.09578544061302682,
|
| 69 |
+
"grad_norm": 11.000222206115723,
|
| 70 |
+
"learning_rate": 2.24e-06,
|
| 71 |
+
"loss": 0.7971,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.10642826734780758,
|
| 76 |
+
"grad_norm": 10.404646873474121,
|
| 77 |
+
"learning_rate": 2.4900000000000003e-06,
|
| 78 |
+
"loss": 0.7646,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.11707109408258834,
|
| 83 |
+
"grad_norm": 8.721226692199707,
|
| 84 |
+
"learning_rate": 2.7400000000000004e-06,
|
| 85 |
+
"loss": 0.7578,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1277139208173691,
|
| 90 |
+
"grad_norm": 9.198404312133789,
|
| 91 |
+
"learning_rate": 2.99e-06,
|
| 92 |
+
"loss": 0.6623,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.13835674755214986,
|
| 97 |
+
"grad_norm": 9.394553184509277,
|
| 98 |
+
"learning_rate": 3.2400000000000003e-06,
|
| 99 |
+
"loss": 0.669,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.14899957428693061,
|
| 104 |
+
"grad_norm": 9.5685453414917,
|
| 105 |
+
"learning_rate": 3.49e-06,
|
| 106 |
+
"loss": 0.6431,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.15964240102171137,
|
| 111 |
+
"grad_norm": 10.325225830078125,
|
| 112 |
+
"learning_rate": 3.74e-06,
|
| 113 |
+
"loss": 0.6444,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.17028522775649213,
|
| 118 |
+
"grad_norm": 8.178572654724121,
|
| 119 |
+
"learning_rate": 3.990000000000001e-06,
|
| 120 |
+
"loss": 0.609,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.1809280544912729,
|
| 125 |
+
"grad_norm": 9.746500015258789,
|
| 126 |
+
"learning_rate": 4.24e-06,
|
| 127 |
+
"loss": 0.5755,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.19157088122605365,
|
| 132 |
+
"grad_norm": 8.11845874786377,
|
| 133 |
+
"learning_rate": 4.49e-06,
|
| 134 |
+
"loss": 0.5589,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.2022137079608344,
|
| 139 |
+
"grad_norm": 7.166477203369141,
|
| 140 |
+
"learning_rate": 4.74e-06,
|
| 141 |
+
"loss": 0.5578,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.21285653469561516,
|
| 146 |
+
"grad_norm": 7.681941032409668,
|
| 147 |
+
"learning_rate": 4.9900000000000005e-06,
|
| 148 |
+
"loss": 0.5076,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.22349936143039592,
|
| 153 |
+
"grad_norm": 8.438258171081543,
|
| 154 |
+
"learning_rate": 5.240000000000001e-06,
|
| 155 |
+
"loss": 0.5445,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.23414218816517668,
|
| 160 |
+
"grad_norm": 9.802384376525879,
|
| 161 |
+
"learning_rate": 5.490000000000001e-06,
|
| 162 |
+
"loss": 0.5098,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.24478501489995744,
|
| 167 |
+
"grad_norm": 7.197368144989014,
|
| 168 |
+
"learning_rate": 5.74e-06,
|
| 169 |
+
"loss": 0.4969,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.2554278416347382,
|
| 174 |
+
"grad_norm": 7.666371822357178,
|
| 175 |
+
"learning_rate": 5.99e-06,
|
| 176 |
+
"loss": 0.4861,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.2660706683695189,
|
| 181 |
+
"grad_norm": 7.324782371520996,
|
| 182 |
+
"learning_rate": 6.24e-06,
|
| 183 |
+
"loss": 0.481,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.2767134951042997,
|
| 188 |
+
"grad_norm": 8.649055480957031,
|
| 189 |
+
"learning_rate": 6.4900000000000005e-06,
|
| 190 |
+
"loss": 0.447,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.28735632183908044,
|
| 195 |
+
"grad_norm": 7.136589050292969,
|
| 196 |
+
"learning_rate": 6.740000000000001e-06,
|
| 197 |
+
"loss": 0.4778,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.29799914857386123,
|
| 202 |
+
"grad_norm": 7.153022289276123,
|
| 203 |
+
"learning_rate": 6.99e-06,
|
| 204 |
+
"loss": 0.4674,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.30864197530864196,
|
| 209 |
+
"grad_norm": 6.950058937072754,
|
| 210 |
+
"learning_rate": 7.24e-06,
|
| 211 |
+
"loss": 0.4434,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.31928480204342274,
|
| 216 |
+
"grad_norm": 7.339558124542236,
|
| 217 |
+
"learning_rate": 7.49e-06,
|
| 218 |
+
"loss": 0.4439,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.3299276287782035,
|
| 223 |
+
"grad_norm": 6.6849541664123535,
|
| 224 |
+
"learning_rate": 7.74e-06,
|
| 225 |
+
"loss": 0.442,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.34057045551298426,
|
| 230 |
+
"grad_norm": 7.065944194793701,
|
| 231 |
+
"learning_rate": 7.990000000000001e-06,
|
| 232 |
+
"loss": 0.4147,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.351213282247765,
|
| 237 |
+
"grad_norm": 6.242930889129639,
|
| 238 |
+
"learning_rate": 8.24e-06,
|
| 239 |
+
"loss": 0.3905,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.3618561089825458,
|
| 244 |
+
"grad_norm": 6.885308742523193,
|
| 245 |
+
"learning_rate": 8.48e-06,
|
| 246 |
+
"loss": 0.4077,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.3724989357173265,
|
| 251 |
+
"grad_norm": 5.589861869812012,
|
| 252 |
+
"learning_rate": 8.730000000000001e-06,
|
| 253 |
+
"loss": 0.4074,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.3831417624521073,
|
| 258 |
+
"grad_norm": 6.651442050933838,
|
| 259 |
+
"learning_rate": 8.98e-06,
|
| 260 |
+
"loss": 0.4013,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.393784589186888,
|
| 265 |
+
"grad_norm": 5.705496311187744,
|
| 266 |
+
"learning_rate": 9.230000000000001e-06,
|
| 267 |
+
"loss": 0.3897,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.4044274159216688,
|
| 272 |
+
"grad_norm": 6.6162333488464355,
|
| 273 |
+
"learning_rate": 9.48e-06,
|
| 274 |
+
"loss": 0.365,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.41507024265644954,
|
| 279 |
+
"grad_norm": 7.273537635803223,
|
| 280 |
+
"learning_rate": 9.73e-06,
|
| 281 |
+
"loss": 0.3924,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.4257130693912303,
|
| 286 |
+
"grad_norm": 6.101346969604492,
|
| 287 |
+
"learning_rate": 9.980000000000001e-06,
|
| 288 |
+
"loss": 0.3715,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.4257130693912303,
|
| 293 |
+
"eval_loss": 0.3456858992576599,
|
| 294 |
+
"eval_runtime": 2451.0815,
|
| 295 |
+
"eval_samples_per_second": 1.949,
|
| 296 |
+
"eval_steps_per_second": 0.061,
|
| 297 |
+
"eval_wer": 40.46916574655637,
|
| 298 |
+
"step": 1000
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 0.43635589612601106,
|
| 302 |
+
"grad_norm": 5.119203567504883,
|
| 303 |
+
"learning_rate": 9.967142857142858e-06,
|
| 304 |
+
"loss": 0.3875,
|
| 305 |
+
"step": 1025
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.44699872286079184,
|
| 309 |
+
"grad_norm": 5.488610744476318,
|
| 310 |
+
"learning_rate": 9.931428571428571e-06,
|
| 311 |
+
"loss": 0.3741,
|
| 312 |
+
"step": 1050
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.45764154959557257,
|
| 316 |
+
"grad_norm": 5.710753917694092,
|
| 317 |
+
"learning_rate": 9.895714285714287e-06,
|
| 318 |
+
"loss": 0.3622,
|
| 319 |
+
"step": 1075
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.46828437633035336,
|
| 323 |
+
"grad_norm": 5.7168192863464355,
|
| 324 |
+
"learning_rate": 9.86e-06,
|
| 325 |
+
"loss": 0.3652,
|
| 326 |
+
"step": 1100
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.4789272030651341,
|
| 330 |
+
"grad_norm": 5.204087734222412,
|
| 331 |
+
"learning_rate": 9.824285714285716e-06,
|
| 332 |
+
"loss": 0.3696,
|
| 333 |
+
"step": 1125
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.4895700297999149,
|
| 337 |
+
"grad_norm": 5.014431476593018,
|
| 338 |
+
"learning_rate": 9.78857142857143e-06,
|
| 339 |
+
"loss": 0.347,
|
| 340 |
+
"step": 1150
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 0.5002128565346956,
|
| 344 |
+
"grad_norm": 5.643191814422607,
|
| 345 |
+
"learning_rate": 9.752857142857143e-06,
|
| 346 |
+
"loss": 0.3681,
|
| 347 |
+
"step": 1175
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.5108556832694764,
|
| 351 |
+
"grad_norm": 6.40764045715332,
|
| 352 |
+
"learning_rate": 9.717142857142858e-06,
|
| 353 |
+
"loss": 0.34,
|
| 354 |
+
"step": 1200
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.5214985100042572,
|
| 358 |
+
"grad_norm": 5.89484167098999,
|
| 359 |
+
"learning_rate": 9.681428571428572e-06,
|
| 360 |
+
"loss": 0.3257,
|
| 361 |
+
"step": 1225
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 0.5321413367390379,
|
| 365 |
+
"grad_norm": 6.474817752838135,
|
| 366 |
+
"learning_rate": 9.645714285714286e-06,
|
| 367 |
+
"loss": 0.3398,
|
| 368 |
+
"step": 1250
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.5427841634738186,
|
| 372 |
+
"grad_norm": 6.272877216339111,
|
| 373 |
+
"learning_rate": 9.610000000000001e-06,
|
| 374 |
+
"loss": 0.3315,
|
| 375 |
+
"step": 1275
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.5534269902085994,
|
| 379 |
+
"grad_norm": 4.669580936431885,
|
| 380 |
+
"learning_rate": 9.574285714285715e-06,
|
| 381 |
+
"loss": 0.3296,
|
| 382 |
+
"step": 1300
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 0.5640698169433802,
|
| 386 |
+
"grad_norm": 4.576137065887451,
|
| 387 |
+
"learning_rate": 9.538571428571428e-06,
|
| 388 |
+
"loss": 0.3334,
|
| 389 |
+
"step": 1325
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.5747126436781609,
|
| 393 |
+
"grad_norm": 4.5423665046691895,
|
| 394 |
+
"learning_rate": 9.502857142857144e-06,
|
| 395 |
+
"loss": 0.3087,
|
| 396 |
+
"step": 1350
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 0.5853554704129417,
|
| 400 |
+
"grad_norm": 4.673890113830566,
|
| 401 |
+
"learning_rate": 9.467142857142857e-06,
|
| 402 |
+
"loss": 0.3286,
|
| 403 |
+
"step": 1375
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.5959982971477225,
|
| 407 |
+
"grad_norm": 4.775241374969482,
|
| 408 |
+
"learning_rate": 9.431428571428573e-06,
|
| 409 |
+
"loss": 0.3258,
|
| 410 |
+
"step": 1400
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.6066411238825032,
|
| 414 |
+
"grad_norm": 5.259005069732666,
|
| 415 |
+
"learning_rate": 9.395714285714287e-06,
|
| 416 |
+
"loss": 0.29,
|
| 417 |
+
"step": 1425
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.6172839506172839,
|
| 421 |
+
"grad_norm": 4.7707014083862305,
|
| 422 |
+
"learning_rate": 9.360000000000002e-06,
|
| 423 |
+
"loss": 0.2879,
|
| 424 |
+
"step": 1450
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 0.6279267773520647,
|
| 428 |
+
"grad_norm": 4.998105525970459,
|
| 429 |
+
"learning_rate": 9.324285714285714e-06,
|
| 430 |
+
"loss": 0.2957,
|
| 431 |
+
"step": 1475
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.6385696040868455,
|
| 435 |
+
"grad_norm": 4.137457370758057,
|
| 436 |
+
"learning_rate": 9.28857142857143e-06,
|
| 437 |
+
"loss": 0.2897,
|
| 438 |
+
"step": 1500
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 0.6492124308216263,
|
| 442 |
+
"grad_norm": 4.685913562774658,
|
| 443 |
+
"learning_rate": 9.252857142857143e-06,
|
| 444 |
+
"loss": 0.3282,
|
| 445 |
+
"step": 1525
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 0.659855257556407,
|
| 449 |
+
"grad_norm": 5.36374044418335,
|
| 450 |
+
"learning_rate": 9.217142857142858e-06,
|
| 451 |
+
"loss": 0.2747,
|
| 452 |
+
"step": 1550
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.6704980842911877,
|
| 456 |
+
"grad_norm": 4.616824150085449,
|
| 457 |
+
"learning_rate": 9.181428571428572e-06,
|
| 458 |
+
"loss": 0.285,
|
| 459 |
+
"step": 1575
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.6811409110259685,
|
| 463 |
+
"grad_norm": 4.848719120025635,
|
| 464 |
+
"learning_rate": 9.145714285714287e-06,
|
| 465 |
+
"loss": 0.2971,
|
| 466 |
+
"step": 1600
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.6917837377607493,
|
| 470 |
+
"grad_norm": 4.435796737670898,
|
| 471 |
+
"learning_rate": 9.110000000000001e-06,
|
| 472 |
+
"loss": 0.2993,
|
| 473 |
+
"step": 1625
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.70242656449553,
|
| 477 |
+
"grad_norm": 4.055502414703369,
|
| 478 |
+
"learning_rate": 9.074285714285716e-06,
|
| 479 |
+
"loss": 0.2713,
|
| 480 |
+
"step": 1650
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.7130693912303108,
|
| 484 |
+
"grad_norm": 5.476015090942383,
|
| 485 |
+
"learning_rate": 9.038571428571428e-06,
|
| 486 |
+
"loss": 0.2553,
|
| 487 |
+
"step": 1675
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.7237122179650916,
|
| 491 |
+
"grad_norm": 4.443753242492676,
|
| 492 |
+
"learning_rate": 9.002857142857144e-06,
|
| 493 |
+
"loss": 0.2772,
|
| 494 |
+
"step": 1700
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.7343550446998723,
|
| 498 |
+
"grad_norm": 4.617072105407715,
|
| 499 |
+
"learning_rate": 8.967142857142857e-06,
|
| 500 |
+
"loss": 0.2745,
|
| 501 |
+
"step": 1725
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.744997871434653,
|
| 505 |
+
"grad_norm": 4.322467803955078,
|
| 506 |
+
"learning_rate": 8.931428571428573e-06,
|
| 507 |
+
"loss": 0.2756,
|
| 508 |
+
"step": 1750
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.7556406981694338,
|
| 512 |
+
"grad_norm": 5.194156169891357,
|
| 513 |
+
"learning_rate": 8.895714285714286e-06,
|
| 514 |
+
"loss": 0.2571,
|
| 515 |
+
"step": 1775
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.7662835249042146,
|
| 519 |
+
"grad_norm": 5.350680828094482,
|
| 520 |
+
"learning_rate": 8.860000000000002e-06,
|
| 521 |
+
"loss": 0.2705,
|
| 522 |
+
"step": 1800
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.7769263516389954,
|
| 526 |
+
"grad_norm": 5.343641757965088,
|
| 527 |
+
"learning_rate": 8.824285714285715e-06,
|
| 528 |
+
"loss": 0.2499,
|
| 529 |
+
"step": 1825
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.787569178373776,
|
| 533 |
+
"grad_norm": 4.356059551239014,
|
| 534 |
+
"learning_rate": 8.788571428571429e-06,
|
| 535 |
+
"loss": 0.2767,
|
| 536 |
+
"step": 1850
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.7982120051085568,
|
| 540 |
+
"grad_norm": 4.316229820251465,
|
| 541 |
+
"learning_rate": 8.752857142857144e-06,
|
| 542 |
+
"loss": 0.2484,
|
| 543 |
+
"step": 1875
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.8088548318433376,
|
| 547 |
+
"grad_norm": 4.627383232116699,
|
| 548 |
+
"learning_rate": 8.717142857142858e-06,
|
| 549 |
+
"loss": 0.2559,
|
| 550 |
+
"step": 1900
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.8194976585781183,
|
| 554 |
+
"grad_norm": 4.916121006011963,
|
| 555 |
+
"learning_rate": 8.681428571428572e-06,
|
| 556 |
+
"loss": 0.2755,
|
| 557 |
+
"step": 1925
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.8301404853128991,
|
| 561 |
+
"grad_norm": 5.244263172149658,
|
| 562 |
+
"learning_rate": 8.645714285714287e-06,
|
| 563 |
+
"loss": 0.2334,
|
| 564 |
+
"step": 1950
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.8407833120476799,
|
| 568 |
+
"grad_norm": 4.568859100341797,
|
| 569 |
+
"learning_rate": 8.61e-06,
|
| 570 |
+
"loss": 0.2542,
|
| 571 |
+
"step": 1975
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.8514261387824607,
|
| 575 |
+
"grad_norm": 3.6536848545074463,
|
| 576 |
+
"learning_rate": 8.574285714285714e-06,
|
| 577 |
+
"loss": 0.251,
|
| 578 |
+
"step": 2000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.8514261387824607,
|
| 582 |
+
"eval_loss": 0.21811740100383759,
|
| 583 |
+
"eval_runtime": 2451.0584,
|
| 584 |
+
"eval_samples_per_second": 1.949,
|
| 585 |
+
"eval_steps_per_second": 0.061,
|
| 586 |
+
"eval_wer": 27.706481799916737,
|
| 587 |
+
"step": 2000
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 0.8620689655172413,
|
| 591 |
+
"grad_norm": 4.318572521209717,
|
| 592 |
+
"learning_rate": 8.53857142857143e-06,
|
| 593 |
+
"loss": 0.2319,
|
| 594 |
+
"step": 2025
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.8727117922520221,
|
| 598 |
+
"grad_norm": 4.489058494567871,
|
| 599 |
+
"learning_rate": 8.502857142857143e-06,
|
| 600 |
+
"loss": 0.2531,
|
| 601 |
+
"step": 2050
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 0.8833546189868029,
|
| 605 |
+
"grad_norm": 4.232712745666504,
|
| 606 |
+
"learning_rate": 8.467142857142859e-06,
|
| 607 |
+
"loss": 0.2491,
|
| 608 |
+
"step": 2075
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.8939974457215837,
|
| 612 |
+
"grad_norm": 4.031393051147461,
|
| 613 |
+
"learning_rate": 8.431428571428572e-06,
|
| 614 |
+
"loss": 0.2485,
|
| 615 |
+
"step": 2100
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.9046402724563644,
|
| 619 |
+
"grad_norm": 3.8136720657348633,
|
| 620 |
+
"learning_rate": 8.395714285714286e-06,
|
| 621 |
+
"loss": 0.2412,
|
| 622 |
+
"step": 2125
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 0.9152830991911451,
|
| 626 |
+
"grad_norm": 4.3343505859375,
|
| 627 |
+
"learning_rate": 8.36e-06,
|
| 628 |
+
"loss": 0.2378,
|
| 629 |
+
"step": 2150
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 0.9259259259259259,
|
| 633 |
+
"grad_norm": 3.6388914585113525,
|
| 634 |
+
"learning_rate": 8.324285714285715e-06,
|
| 635 |
+
"loss": 0.2409,
|
| 636 |
+
"step": 2175
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 0.9365687526607067,
|
| 640 |
+
"grad_norm": 5.596227169036865,
|
| 641 |
+
"learning_rate": 8.288571428571429e-06,
|
| 642 |
+
"loss": 0.2363,
|
| 643 |
+
"step": 2200
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 0.9472115793954874,
|
| 647 |
+
"grad_norm": 4.016772747039795,
|
| 648 |
+
"learning_rate": 8.252857142857144e-06,
|
| 649 |
+
"loss": 0.2281,
|
| 650 |
+
"step": 2225
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.9578544061302682,
|
| 654 |
+
"grad_norm": 5.106402397155762,
|
| 655 |
+
"learning_rate": 8.217142857142858e-06,
|
| 656 |
+
"loss": 0.224,
|
| 657 |
+
"step": 2250
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.968497232865049,
|
| 661 |
+
"grad_norm": 3.714061975479126,
|
| 662 |
+
"learning_rate": 8.181428571428573e-06,
|
| 663 |
+
"loss": 0.2306,
|
| 664 |
+
"step": 2275
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 0.9791400595998297,
|
| 668 |
+
"grad_norm": 4.1780009269714355,
|
| 669 |
+
"learning_rate": 8.145714285714287e-06,
|
| 670 |
+
"loss": 0.2284,
|
| 671 |
+
"step": 2300
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 0.9897828863346104,
|
| 675 |
+
"grad_norm": 4.007058143615723,
|
| 676 |
+
"learning_rate": 8.110000000000002e-06,
|
| 677 |
+
"loss": 0.237,
|
| 678 |
+
"step": 2325
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 1.0004257130693912,
|
| 682 |
+
"grad_norm": 2.91274094581604,
|
| 683 |
+
"learning_rate": 8.074285714285714e-06,
|
| 684 |
+
"loss": 0.227,
|
| 685 |
+
"step": 2350
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.0110685398041719,
|
| 689 |
+
"grad_norm": 4.321012496948242,
|
| 690 |
+
"learning_rate": 8.03857142857143e-06,
|
| 691 |
+
"loss": 0.1917,
|
| 692 |
+
"step": 2375
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.0217113665389528,
|
| 696 |
+
"grad_norm": 4.613705635070801,
|
| 697 |
+
"learning_rate": 8.002857142857143e-06,
|
| 698 |
+
"loss": 0.1861,
|
| 699 |
+
"step": 2400
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.0323541932737335,
|
| 703 |
+
"grad_norm": 3.9575023651123047,
|
| 704 |
+
"learning_rate": 7.967142857142858e-06,
|
| 705 |
+
"loss": 0.1931,
|
| 706 |
+
"step": 2425
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 1.0429970200085144,
|
| 710 |
+
"grad_norm": 4.651571273803711,
|
| 711 |
+
"learning_rate": 7.931428571428572e-06,
|
| 712 |
+
"loss": 0.1784,
|
| 713 |
+
"step": 2450
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 1.053639846743295,
|
| 717 |
+
"grad_norm": 4.0472412109375,
|
| 718 |
+
"learning_rate": 7.895714285714287e-06,
|
| 719 |
+
"loss": 0.2002,
|
| 720 |
+
"step": 2475
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 1.0642826734780757,
|
| 724 |
+
"grad_norm": 3.4641237258911133,
|
| 725 |
+
"learning_rate": 7.860000000000001e-06,
|
| 726 |
+
"loss": 0.1885,
|
| 727 |
+
"step": 2500
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.0749255002128566,
|
| 731 |
+
"grad_norm": 3.260540008544922,
|
| 732 |
+
"learning_rate": 7.824285714285715e-06,
|
| 733 |
+
"loss": 0.1924,
|
| 734 |
+
"step": 2525
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.0855683269476373,
|
| 738 |
+
"grad_norm": 4.416691303253174,
|
| 739 |
+
"learning_rate": 7.788571428571428e-06,
|
| 740 |
+
"loss": 0.181,
|
| 741 |
+
"step": 2550
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 1.096211153682418,
|
| 745 |
+
"grad_norm": 3.7334911823272705,
|
| 746 |
+
"learning_rate": 7.752857142857144e-06,
|
| 747 |
+
"loss": 0.1667,
|
| 748 |
+
"step": 2575
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.1068539804171988,
|
| 752 |
+
"grad_norm": 4.4988555908203125,
|
| 753 |
+
"learning_rate": 7.717142857142857e-06,
|
| 754 |
+
"loss": 0.1864,
|
| 755 |
+
"step": 2600
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.1174968071519795,
|
| 759 |
+
"grad_norm": 4.6382222175598145,
|
| 760 |
+
"learning_rate": 7.681428571428573e-06,
|
| 761 |
+
"loss": 0.1805,
|
| 762 |
+
"step": 2625
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.1281396338867604,
|
| 766 |
+
"grad_norm": 4.512842178344727,
|
| 767 |
+
"learning_rate": 7.645714285714286e-06,
|
| 768 |
+
"loss": 0.1848,
|
| 769 |
+
"step": 2650
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.138782460621541,
|
| 773 |
+
"grad_norm": 3.889390468597412,
|
| 774 |
+
"learning_rate": 7.610000000000001e-06,
|
| 775 |
+
"loss": 0.1846,
|
| 776 |
+
"step": 2675
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.1494252873563218,
|
| 780 |
+
"grad_norm": 4.247312068939209,
|
| 781 |
+
"learning_rate": 7.574285714285715e-06,
|
| 782 |
+
"loss": 0.1799,
|
| 783 |
+
"step": 2700
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 1.1600681140911027,
|
| 787 |
+
"grad_norm": 4.321536540985107,
|
| 788 |
+
"learning_rate": 7.53857142857143e-06,
|
| 789 |
+
"loss": 0.1764,
|
| 790 |
+
"step": 2725
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.1707109408258833,
|
| 794 |
+
"grad_norm": 4.06414794921875,
|
| 795 |
+
"learning_rate": 7.502857142857144e-06,
|
| 796 |
+
"loss": 0.1903,
|
| 797 |
+
"step": 2750
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 1.181353767560664,
|
| 801 |
+
"grad_norm": 3.314551591873169,
|
| 802 |
+
"learning_rate": 7.467142857142857e-06,
|
| 803 |
+
"loss": 0.1614,
|
| 804 |
+
"step": 2775
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 1.191996594295445,
|
| 808 |
+
"grad_norm": 4.245212078094482,
|
| 809 |
+
"learning_rate": 7.431428571428572e-06,
|
| 810 |
+
"loss": 0.1867,
|
| 811 |
+
"step": 2800
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.2026394210302256,
|
| 815 |
+
"grad_norm": 3.1465117931365967,
|
| 816 |
+
"learning_rate": 7.395714285714286e-06,
|
| 817 |
+
"loss": 0.1584,
|
| 818 |
+
"step": 2825
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.2132822477650063,
|
| 822 |
+
"grad_norm": 4.1284637451171875,
|
| 823 |
+
"learning_rate": 7.360000000000001e-06,
|
| 824 |
+
"loss": 0.1856,
|
| 825 |
+
"step": 2850
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 1.2239250744997872,
|
| 829 |
+
"grad_norm": 3.685889720916748,
|
| 830 |
+
"learning_rate": 7.324285714285715e-06,
|
| 831 |
+
"loss": 0.171,
|
| 832 |
+
"step": 2875
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.2345679012345678,
|
| 836 |
+
"grad_norm": 4.70512580871582,
|
| 837 |
+
"learning_rate": 7.28857142857143e-06,
|
| 838 |
+
"loss": 0.1752,
|
| 839 |
+
"step": 2900
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 1.2452107279693487,
|
| 843 |
+
"grad_norm": 3.846862316131592,
|
| 844 |
+
"learning_rate": 7.252857142857143e-06,
|
| 845 |
+
"loss": 0.1697,
|
| 846 |
+
"step": 2925
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 1.2558535547041294,
|
| 850 |
+
"grad_norm": 3.7466206550598145,
|
| 851 |
+
"learning_rate": 7.217142857142858e-06,
|
| 852 |
+
"loss": 0.1796,
|
| 853 |
+
"step": 2950
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.2664963814389103,
|
| 857 |
+
"grad_norm": 3.8162903785705566,
|
| 858 |
+
"learning_rate": 7.182857142857144e-06,
|
| 859 |
+
"loss": 0.1591,
|
| 860 |
+
"step": 2975
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.277139208173691,
|
| 864 |
+
"grad_norm": 3.880910873413086,
|
| 865 |
+
"learning_rate": 7.147142857142858e-06,
|
| 866 |
+
"loss": 0.1569,
|
| 867 |
+
"step": 3000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 1.277139208173691,
|
| 871 |
+
"eval_loss": 0.1813717633485794,
|
| 872 |
+
"eval_runtime": 2447.146,
|
| 873 |
+
"eval_samples_per_second": 1.952,
|
| 874 |
+
"eval_steps_per_second": 0.061,
|
| 875 |
+
"eval_wer": 24.153347693087408,
|
| 876 |
+
"step": 3000
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 1.2877820349084717,
|
| 880 |
+
"grad_norm": 3.98262882232666,
|
| 881 |
+
"learning_rate": 7.111428571428572e-06,
|
| 882 |
+
"loss": 0.1804,
|
| 883 |
+
"step": 3025
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.2984248616432525,
|
| 887 |
+
"grad_norm": 3.6790521144866943,
|
| 888 |
+
"learning_rate": 7.075714285714286e-06,
|
| 889 |
+
"loss": 0.1647,
|
| 890 |
+
"step": 3050
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.3090676883780332,
|
| 894 |
+
"grad_norm": 3.431762456893921,
|
| 895 |
+
"learning_rate": 7.04e-06,
|
| 896 |
+
"loss": 0.1662,
|
| 897 |
+
"step": 3075
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 1.319710515112814,
|
| 901 |
+
"grad_norm": 4.0635247230529785,
|
| 902 |
+
"learning_rate": 7.004285714285715e-06,
|
| 903 |
+
"loss": 0.1726,
|
| 904 |
+
"step": 3100
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.3303533418475948,
|
| 908 |
+
"grad_norm": 3.1607766151428223,
|
| 909 |
+
"learning_rate": 6.968571428571429e-06,
|
| 910 |
+
"loss": 0.1544,
|
| 911 |
+
"step": 3125
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 1.3409961685823755,
|
| 915 |
+
"grad_norm": 4.5737385749816895,
|
| 916 |
+
"learning_rate": 6.932857142857143e-06,
|
| 917 |
+
"loss": 0.1644,
|
| 918 |
+
"step": 3150
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 1.3516389953171561,
|
| 922 |
+
"grad_norm": 4.182763576507568,
|
| 923 |
+
"learning_rate": 6.8971428571428575e-06,
|
| 924 |
+
"loss": 0.167,
|
| 925 |
+
"step": 3175
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.362281822051937,
|
| 929 |
+
"grad_norm": 3.3566439151763916,
|
| 930 |
+
"learning_rate": 6.861428571428572e-06,
|
| 931 |
+
"loss": 0.1631,
|
| 932 |
+
"step": 3200
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.3729246487867177,
|
| 936 |
+
"grad_norm": 3.771667718887329,
|
| 937 |
+
"learning_rate": 6.8257142857142866e-06,
|
| 938 |
+
"loss": 0.1675,
|
| 939 |
+
"step": 3225
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.3835674755214984,
|
| 943 |
+
"grad_norm": 4.14226770401001,
|
| 944 |
+
"learning_rate": 6.790000000000001e-06,
|
| 945 |
+
"loss": 0.1785,
|
| 946 |
+
"step": 3250
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.3942103022562793,
|
| 950 |
+
"grad_norm": 4.599484443664551,
|
| 951 |
+
"learning_rate": 6.754285714285715e-06,
|
| 952 |
+
"loss": 0.1832,
|
| 953 |
+
"step": 3275
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.40485312899106,
|
| 957 |
+
"grad_norm": 3.78108286857605,
|
| 958 |
+
"learning_rate": 6.718571428571428e-06,
|
| 959 |
+
"loss": 0.1765,
|
| 960 |
+
"step": 3300
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 1.4154959557258409,
|
| 964 |
+
"grad_norm": 3.3249051570892334,
|
| 965 |
+
"learning_rate": 6.682857142857143e-06,
|
| 966 |
+
"loss": 0.1517,
|
| 967 |
+
"step": 3325
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.4261387824606215,
|
| 971 |
+
"grad_norm": 3.299750804901123,
|
| 972 |
+
"learning_rate": 6.647142857142857e-06,
|
| 973 |
+
"loss": 0.1632,
|
| 974 |
+
"step": 3350
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 1.4367816091954024,
|
| 978 |
+
"grad_norm": 4.0860066413879395,
|
| 979 |
+
"learning_rate": 6.611428571428572e-06,
|
| 980 |
+
"loss": 0.1457,
|
| 981 |
+
"step": 3375
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 1.447424435930183,
|
| 985 |
+
"grad_norm": 4.305485725402832,
|
| 986 |
+
"learning_rate": 6.575714285714286e-06,
|
| 987 |
+
"loss": 0.1638,
|
| 988 |
+
"step": 3400
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 1.4580672626649638,
|
| 992 |
+
"grad_norm": 3.656642436981201,
|
| 993 |
+
"learning_rate": 6.540000000000001e-06,
|
| 994 |
+
"loss": 0.1681,
|
| 995 |
+
"step": 3425
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.4687100893997447,
|
| 999 |
+
"grad_norm": 3.596554756164551,
|
| 1000 |
+
"learning_rate": 6.504285714285715e-06,
|
| 1001 |
+
"loss": 0.1473,
|
| 1002 |
+
"step": 3450
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 1.4793529161345254,
|
| 1006 |
+
"grad_norm": 3.35798716545105,
|
| 1007 |
+
"learning_rate": 6.46857142857143e-06,
|
| 1008 |
+
"loss": 0.1443,
|
| 1009 |
+
"step": 3475
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.489995742869306,
|
| 1013 |
+
"grad_norm": 3.782789468765259,
|
| 1014 |
+
"learning_rate": 6.432857142857143e-06,
|
| 1015 |
+
"loss": 0.1399,
|
| 1016 |
+
"step": 3500
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 1.500638569604087,
|
| 1020 |
+
"grad_norm": 3.556546926498413,
|
| 1021 |
+
"learning_rate": 6.397142857142857e-06,
|
| 1022 |
+
"loss": 0.1657,
|
| 1023 |
+
"step": 3525
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 1.5112813963388676,
|
| 1027 |
+
"grad_norm": 4.0330657958984375,
|
| 1028 |
+
"learning_rate": 6.361428571428572e-06,
|
| 1029 |
+
"loss": 0.1455,
|
| 1030 |
+
"step": 3550
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 1.5219242230736483,
|
| 1034 |
+
"grad_norm": 3.4194424152374268,
|
| 1035 |
+
"learning_rate": 6.325714285714286e-06,
|
| 1036 |
+
"loss": 0.1558,
|
| 1037 |
+
"step": 3575
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 1.5325670498084292,
|
| 1041 |
+
"grad_norm": 3.4053897857666016,
|
| 1042 |
+
"learning_rate": 6.290000000000001e-06,
|
| 1043 |
+
"loss": 0.1667,
|
| 1044 |
+
"step": 3600
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 1.5432098765432098,
|
| 1048 |
+
"grad_norm": 3.4398772716522217,
|
| 1049 |
+
"learning_rate": 6.254285714285715e-06,
|
| 1050 |
+
"loss": 0.1704,
|
| 1051 |
+
"step": 3625
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.5538527032779905,
|
| 1055 |
+
"grad_norm": 3.950698137283325,
|
| 1056 |
+
"learning_rate": 6.21857142857143e-06,
|
| 1057 |
+
"loss": 0.1587,
|
| 1058 |
+
"step": 3650
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 1.5644955300127714,
|
| 1062 |
+
"grad_norm": 3.5105514526367188,
|
| 1063 |
+
"learning_rate": 6.1828571428571434e-06,
|
| 1064 |
+
"loss": 0.1662,
|
| 1065 |
+
"step": 3675
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 1.5751383567475523,
|
| 1069 |
+
"grad_norm": 3.1570792198181152,
|
| 1070 |
+
"learning_rate": 6.147142857142858e-06,
|
| 1071 |
+
"loss": 0.1542,
|
| 1072 |
+
"step": 3700
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 1.5857811834823328,
|
| 1076 |
+
"grad_norm": 3.395730495452881,
|
| 1077 |
+
"learning_rate": 6.111428571428572e-06,
|
| 1078 |
+
"loss": 0.1419,
|
| 1079 |
+
"step": 3725
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.5964240102171137,
|
| 1083 |
+
"grad_norm": 3.692760944366455,
|
| 1084 |
+
"learning_rate": 6.075714285714286e-06,
|
| 1085 |
+
"loss": 0.1515,
|
| 1086 |
+
"step": 3750
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 1.6070668369518946,
|
| 1090 |
+
"grad_norm": 4.292817115783691,
|
| 1091 |
+
"learning_rate": 6.040000000000001e-06,
|
| 1092 |
+
"loss": 0.1558,
|
| 1093 |
+
"step": 3775
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.617709663686675,
|
| 1097 |
+
"grad_norm": 2.7795393466949463,
|
| 1098 |
+
"learning_rate": 6.004285714285715e-06,
|
| 1099 |
+
"loss": 0.1603,
|
| 1100 |
+
"step": 3800
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 1.628352490421456,
|
| 1104 |
+
"grad_norm": 3.6494193077087402,
|
| 1105 |
+
"learning_rate": 5.968571428571429e-06,
|
| 1106 |
+
"loss": 0.1527,
|
| 1107 |
+
"step": 3825
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 1.6389953171562368,
|
| 1111 |
+
"grad_norm": 3.185007333755493,
|
| 1112 |
+
"learning_rate": 5.932857142857143e-06,
|
| 1113 |
+
"loss": 0.1415,
|
| 1114 |
+
"step": 3850
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 1.6496381438910175,
|
| 1118 |
+
"grad_norm": 4.0278143882751465,
|
| 1119 |
+
"learning_rate": 5.897142857142858e-06,
|
| 1120 |
+
"loss": 0.1595,
|
| 1121 |
+
"step": 3875
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.6602809706257982,
|
| 1125 |
+
"grad_norm": 3.8083670139312744,
|
| 1126 |
+
"learning_rate": 5.861428571428572e-06,
|
| 1127 |
+
"loss": 0.1596,
|
| 1128 |
+
"step": 3900
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 1.670923797360579,
|
| 1132 |
+
"grad_norm": 5.412234783172607,
|
| 1133 |
+
"learning_rate": 5.825714285714286e-06,
|
| 1134 |
+
"loss": 0.1418,
|
| 1135 |
+
"step": 3925
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.6815666240953597,
|
| 1139 |
+
"grad_norm": 3.8275325298309326,
|
| 1140 |
+
"learning_rate": 5.7900000000000005e-06,
|
| 1141 |
+
"loss": 0.1725,
|
| 1142 |
+
"step": 3950
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 1.6922094508301404,
|
| 1146 |
+
"grad_norm": 3.4874017238616943,
|
| 1147 |
+
"learning_rate": 5.754285714285714e-06,
|
| 1148 |
+
"loss": 0.1334,
|
| 1149 |
+
"step": 3975
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 1.7028522775649213,
|
| 1153 |
+
"grad_norm": 2.9034647941589355,
|
| 1154 |
+
"learning_rate": 5.718571428571429e-06,
|
| 1155 |
+
"loss": 0.1436,
|
| 1156 |
+
"step": 4000
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 1.7028522775649213,
|
| 1160 |
+
"eval_loss": 0.1530725359916687,
|
| 1161 |
+
"eval_runtime": 2470.8785,
|
| 1162 |
+
"eval_samples_per_second": 1.933,
|
| 1163 |
+
"eval_steps_per_second": 0.061,
|
| 1164 |
+
"eval_wer": 20.381197169077055,
|
| 1165 |
+
"step": 4000
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.713495104299702,
|
| 1169 |
+
"grad_norm": 3.192444086074829,
|
| 1170 |
+
"learning_rate": 5.682857142857143e-06,
|
| 1171 |
+
"loss": 0.1391,
|
| 1172 |
+
"step": 4025
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.7241379310344827,
|
| 1176 |
+
"grad_norm": 3.376185655593872,
|
| 1177 |
+
"learning_rate": 5.647142857142858e-06,
|
| 1178 |
+
"loss": 0.1447,
|
| 1179 |
+
"step": 4050
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.7347807577692635,
|
| 1183 |
+
"grad_norm": 3.2235193252563477,
|
| 1184 |
+
"learning_rate": 5.611428571428572e-06,
|
| 1185 |
+
"loss": 0.1473,
|
| 1186 |
+
"step": 4075
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.7454235845040442,
|
| 1190 |
+
"grad_norm": 3.4376378059387207,
|
| 1191 |
+
"learning_rate": 5.575714285714287e-06,
|
| 1192 |
+
"loss": 0.1526,
|
| 1193 |
+
"step": 4100
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.756066411238825,
|
| 1197 |
+
"grad_norm": 3.4150240421295166,
|
| 1198 |
+
"learning_rate": 5.540000000000001e-06,
|
| 1199 |
+
"loss": 0.1503,
|
| 1200 |
+
"step": 4125
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.7667092379736058,
|
| 1204 |
+
"grad_norm": 3.757262706756592,
|
| 1205 |
+
"learning_rate": 5.504285714285714e-06,
|
| 1206 |
+
"loss": 0.1311,
|
| 1207 |
+
"step": 4150
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.7773520647083867,
|
| 1211 |
+
"grad_norm": 3.725192070007324,
|
| 1212 |
+
"learning_rate": 5.4685714285714285e-06,
|
| 1213 |
+
"loss": 0.1506,
|
| 1214 |
+
"step": 4175
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.7879948914431671,
|
| 1218 |
+
"grad_norm": 3.243486166000366,
|
| 1219 |
+
"learning_rate": 5.432857142857143e-06,
|
| 1220 |
+
"loss": 0.1529,
|
| 1221 |
+
"step": 4200
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.798637718177948,
|
| 1225 |
+
"grad_norm": 3.1005189418792725,
|
| 1226 |
+
"learning_rate": 5.3971428571428575e-06,
|
| 1227 |
+
"loss": 0.1592,
|
| 1228 |
+
"step": 4225
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.809280544912729,
|
| 1232 |
+
"grad_norm": 2.6923441886901855,
|
| 1233 |
+
"learning_rate": 5.361428571428572e-06,
|
| 1234 |
+
"loss": 0.1373,
|
| 1235 |
+
"step": 4250
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.8199233716475096,
|
| 1239 |
+
"grad_norm": 3.4601283073425293,
|
| 1240 |
+
"learning_rate": 5.3257142857142865e-06,
|
| 1241 |
+
"loss": 0.1358,
|
| 1242 |
+
"step": 4275
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.8305661983822903,
|
| 1246 |
+
"grad_norm": 4.46110200881958,
|
| 1247 |
+
"learning_rate": 5.290000000000001e-06,
|
| 1248 |
+
"loss": 0.1406,
|
| 1249 |
+
"step": 4300
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.8412090251170712,
|
| 1253 |
+
"grad_norm": 3.4556360244750977,
|
| 1254 |
+
"learning_rate": 5.254285714285715e-06,
|
| 1255 |
+
"loss": 0.1314,
|
| 1256 |
+
"step": 4325
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.8518518518518519,
|
| 1260 |
+
"grad_norm": 2.851836919784546,
|
| 1261 |
+
"learning_rate": 5.218571428571429e-06,
|
| 1262 |
+
"loss": 0.1263,
|
| 1263 |
+
"step": 4350
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.8624946785866325,
|
| 1267 |
+
"grad_norm": 3.1507768630981445,
|
| 1268 |
+
"learning_rate": 5.182857142857143e-06,
|
| 1269 |
+
"loss": 0.1263,
|
| 1270 |
+
"step": 4375
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.8731375053214134,
|
| 1274 |
+
"grad_norm": 3.7861220836639404,
|
| 1275 |
+
"learning_rate": 5.147142857142857e-06,
|
| 1276 |
+
"loss": 0.1423,
|
| 1277 |
+
"step": 4400
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.883780332056194,
|
| 1281 |
+
"grad_norm": 2.670792818069458,
|
| 1282 |
+
"learning_rate": 5.111428571428572e-06,
|
| 1283 |
+
"loss": 0.1378,
|
| 1284 |
+
"step": 4425
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.8944231587909748,
|
| 1288 |
+
"grad_norm": 3.21482515335083,
|
| 1289 |
+
"learning_rate": 5.075714285714286e-06,
|
| 1290 |
+
"loss": 0.1305,
|
| 1291 |
+
"step": 4450
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.9050659855257557,
|
| 1295 |
+
"grad_norm": 3.0958456993103027,
|
| 1296 |
+
"learning_rate": 5.04e-06,
|
| 1297 |
+
"loss": 0.1312,
|
| 1298 |
+
"step": 4475
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.9157088122605364,
|
| 1302 |
+
"grad_norm": 3.2010111808776855,
|
| 1303 |
+
"learning_rate": 5.0042857142857145e-06,
|
| 1304 |
+
"loss": 0.1358,
|
| 1305 |
+
"step": 4500
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.926351638995317,
|
| 1309 |
+
"grad_norm": 4.211108684539795,
|
| 1310 |
+
"learning_rate": 4.968571428571429e-06,
|
| 1311 |
+
"loss": 0.149,
|
| 1312 |
+
"step": 4525
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.936994465730098,
|
| 1316 |
+
"grad_norm": 3.6158218383789062,
|
| 1317 |
+
"learning_rate": 4.932857142857143e-06,
|
| 1318 |
+
"loss": 0.1456,
|
| 1319 |
+
"step": 4550
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.9476372924648788,
|
| 1323 |
+
"grad_norm": 3.1304032802581787,
|
| 1324 |
+
"learning_rate": 4.897142857142857e-06,
|
| 1325 |
+
"loss": 0.1474,
|
| 1326 |
+
"step": 4575
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.9582801191996593,
|
| 1330 |
+
"grad_norm": 3.7992565631866455,
|
| 1331 |
+
"learning_rate": 4.861428571428572e-06,
|
| 1332 |
+
"loss": 0.1252,
|
| 1333 |
+
"step": 4600
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.9689229459344402,
|
| 1337 |
+
"grad_norm": 3.0859761238098145,
|
| 1338 |
+
"learning_rate": 4.825714285714286e-06,
|
| 1339 |
+
"loss": 0.1459,
|
| 1340 |
+
"step": 4625
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.979565772669221,
|
| 1344 |
+
"grad_norm": 4.332040309906006,
|
| 1345 |
+
"learning_rate": 4.79e-06,
|
| 1346 |
+
"loss": 0.129,
|
| 1347 |
+
"step": 4650
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.9902085994040017,
|
| 1351 |
+
"grad_norm": 4.2954816818237305,
|
| 1352 |
+
"learning_rate": 4.754285714285714e-06,
|
| 1353 |
+
"loss": 0.1566,
|
| 1354 |
+
"step": 4675
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 2.0008514261387824,
|
| 1358 |
+
"grad_norm": 2.788947105407715,
|
| 1359 |
+
"learning_rate": 4.718571428571429e-06,
|
| 1360 |
+
"loss": 0.145,
|
| 1361 |
+
"step": 4700
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 2.0114942528735633,
|
| 1365 |
+
"grad_norm": 3.2599875926971436,
|
| 1366 |
+
"learning_rate": 4.682857142857143e-06,
|
| 1367 |
+
"loss": 0.1063,
|
| 1368 |
+
"step": 4725
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 2.0221370796083438,
|
| 1372 |
+
"grad_norm": 3.0225577354431152,
|
| 1373 |
+
"learning_rate": 4.647142857142857e-06,
|
| 1374 |
+
"loss": 0.0877,
|
| 1375 |
+
"step": 4750
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 2.0327799063431247,
|
| 1379 |
+
"grad_norm": 3.564682960510254,
|
| 1380 |
+
"learning_rate": 4.6114285714285716e-06,
|
| 1381 |
+
"loss": 0.1014,
|
| 1382 |
+
"step": 4775
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 2.0434227330779056,
|
| 1386 |
+
"grad_norm": 2.5339510440826416,
|
| 1387 |
+
"learning_rate": 4.575714285714286e-06,
|
| 1388 |
+
"loss": 0.0906,
|
| 1389 |
+
"step": 4800
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 2.0540655598126865,
|
| 1393 |
+
"grad_norm": 2.7343597412109375,
|
| 1394 |
+
"learning_rate": 4.540000000000001e-06,
|
| 1395 |
+
"loss": 0.0994,
|
| 1396 |
+
"step": 4825
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 2.064708386547467,
|
| 1400 |
+
"grad_norm": 2.6490981578826904,
|
| 1401 |
+
"learning_rate": 4.504285714285715e-06,
|
| 1402 |
+
"loss": 0.0979,
|
| 1403 |
+
"step": 4850
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 2.075351213282248,
|
| 1407 |
+
"grad_norm": 1.9775068759918213,
|
| 1408 |
+
"learning_rate": 4.468571428571429e-06,
|
| 1409 |
+
"loss": 0.1023,
|
| 1410 |
+
"step": 4875
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 2.0859940400170287,
|
| 1414 |
+
"grad_norm": 2.2302167415618896,
|
| 1415 |
+
"learning_rate": 4.432857142857143e-06,
|
| 1416 |
+
"loss": 0.101,
|
| 1417 |
+
"step": 4900
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 2.096636866751809,
|
| 1421 |
+
"grad_norm": 2.7685494422912598,
|
| 1422 |
+
"learning_rate": 4.397142857142858e-06,
|
| 1423 |
+
"loss": 0.0934,
|
| 1424 |
+
"step": 4925
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 2.10727969348659,
|
| 1428 |
+
"grad_norm": 2.8027827739715576,
|
| 1429 |
+
"learning_rate": 4.361428571428572e-06,
|
| 1430 |
+
"loss": 0.0962,
|
| 1431 |
+
"step": 4950
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 2.117922520221371,
|
| 1435 |
+
"grad_norm": 2.9173505306243896,
|
| 1436 |
+
"learning_rate": 4.325714285714286e-06,
|
| 1437 |
+
"loss": 0.0934,
|
| 1438 |
+
"step": 4975
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 2.1285653469561514,
|
| 1442 |
+
"grad_norm": 2.7315633296966553,
|
| 1443 |
+
"learning_rate": 4.2900000000000004e-06,
|
| 1444 |
+
"loss": 0.0931,
|
| 1445 |
+
"step": 5000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 2.1285653469561514,
|
| 1449 |
+
"eval_loss": 0.13744878768920898,
|
| 1450 |
+
"eval_runtime": 2441.1543,
|
| 1451 |
+
"eval_samples_per_second": 1.957,
|
| 1452 |
+
"eval_steps_per_second": 0.061,
|
| 1453 |
+
"eval_wer": 18.466161058519013,
|
| 1454 |
+
"step": 5000
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 2.1392081736909323,
|
| 1458 |
+
"grad_norm": 2.411224126815796,
|
| 1459 |
+
"learning_rate": 4.254285714285715e-06,
|
| 1460 |
+
"loss": 0.1058,
|
| 1461 |
+
"step": 5025
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 2.149851000425713,
|
| 1465 |
+
"grad_norm": 2.7599411010742188,
|
| 1466 |
+
"learning_rate": 4.2185714285714294e-06,
|
| 1467 |
+
"loss": 0.105,
|
| 1468 |
+
"step": 5050
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 2.1604938271604937,
|
| 1472 |
+
"grad_norm": 2.873077392578125,
|
| 1473 |
+
"learning_rate": 4.182857142857143e-06,
|
| 1474 |
+
"loss": 0.1122,
|
| 1475 |
+
"step": 5075
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 2.1711366538952745,
|
| 1479 |
+
"grad_norm": 2.4859185218811035,
|
| 1480 |
+
"learning_rate": 4.147142857142858e-06,
|
| 1481 |
+
"loss": 0.0956,
|
| 1482 |
+
"step": 5100
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 2.1817794806300554,
|
| 1486 |
+
"grad_norm": 2.307053565979004,
|
| 1487 |
+
"learning_rate": 4.111428571428572e-06,
|
| 1488 |
+
"loss": 0.0936,
|
| 1489 |
+
"step": 5125
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 2.192422307364836,
|
| 1493 |
+
"grad_norm": 2.692552328109741,
|
| 1494 |
+
"learning_rate": 4.075714285714286e-06,
|
| 1495 |
+
"loss": 0.0814,
|
| 1496 |
+
"step": 5150
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 2.203065134099617,
|
| 1500 |
+
"grad_norm": 2.640380382537842,
|
| 1501 |
+
"learning_rate": 4.04e-06,
|
| 1502 |
+
"loss": 0.0961,
|
| 1503 |
+
"step": 5175
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 2.2137079608343977,
|
| 1507 |
+
"grad_norm": 1.9715120792388916,
|
| 1508 |
+
"learning_rate": 4.004285714285715e-06,
|
| 1509 |
+
"loss": 0.0911,
|
| 1510 |
+
"step": 5200
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 2.224350787569178,
|
| 1514 |
+
"grad_norm": 2.4855728149414062,
|
| 1515 |
+
"learning_rate": 3.9685714285714284e-06,
|
| 1516 |
+
"loss": 0.0871,
|
| 1517 |
+
"step": 5225
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 2.234993614303959,
|
| 1521 |
+
"grad_norm": 2.190443992614746,
|
| 1522 |
+
"learning_rate": 3.932857142857143e-06,
|
| 1523 |
+
"loss": 0.0923,
|
| 1524 |
+
"step": 5250
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 2.24563644103874,
|
| 1528 |
+
"grad_norm": 2.5768940448760986,
|
| 1529 |
+
"learning_rate": 3.8971428571428575e-06,
|
| 1530 |
+
"loss": 0.1033,
|
| 1531 |
+
"step": 5275
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 2.256279267773521,
|
| 1535 |
+
"grad_norm": 2.527087926864624,
|
| 1536 |
+
"learning_rate": 3.861428571428571e-06,
|
| 1537 |
+
"loss": 0.081,
|
| 1538 |
+
"step": 5300
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 2.2669220945083013,
|
| 1542 |
+
"grad_norm": 3.3411247730255127,
|
| 1543 |
+
"learning_rate": 3.825714285714286e-06,
|
| 1544 |
+
"loss": 0.0931,
|
| 1545 |
+
"step": 5325
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 2.277564921243082,
|
| 1549 |
+
"grad_norm": 2.852933645248413,
|
| 1550 |
+
"learning_rate": 3.79e-06,
|
| 1551 |
+
"loss": 0.0736,
|
| 1552 |
+
"step": 5350
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 2.288207747977863,
|
| 1556 |
+
"grad_norm": 3.596585512161255,
|
| 1557 |
+
"learning_rate": 3.7542857142857146e-06,
|
| 1558 |
+
"loss": 0.0892,
|
| 1559 |
+
"step": 5375
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 2.2988505747126435,
|
| 1563 |
+
"grad_norm": 3.5326387882232666,
|
| 1564 |
+
"learning_rate": 3.7185714285714287e-06,
|
| 1565 |
+
"loss": 0.0827,
|
| 1566 |
+
"step": 5400
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 2.3094934014474244,
|
| 1570 |
+
"grad_norm": 2.5857245922088623,
|
| 1571 |
+
"learning_rate": 3.682857142857143e-06,
|
| 1572 |
+
"loss": 0.0957,
|
| 1573 |
+
"step": 5425
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 2.3201362281822053,
|
| 1577 |
+
"grad_norm": 2.7961575984954834,
|
| 1578 |
+
"learning_rate": 3.6471428571428573e-06,
|
| 1579 |
+
"loss": 0.0968,
|
| 1580 |
+
"step": 5450
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 2.330779054916986,
|
| 1584 |
+
"grad_norm": 2.5830881595611572,
|
| 1585 |
+
"learning_rate": 3.611428571428572e-06,
|
| 1586 |
+
"loss": 0.086,
|
| 1587 |
+
"step": 5475
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 2.3414218816517667,
|
| 1591 |
+
"grad_norm": 3.009079694747925,
|
| 1592 |
+
"learning_rate": 3.5757142857142863e-06,
|
| 1593 |
+
"loss": 0.0821,
|
| 1594 |
+
"step": 5500
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 2.3520647083865476,
|
| 1598 |
+
"grad_norm": 3.2206666469573975,
|
| 1599 |
+
"learning_rate": 3.54e-06,
|
| 1600 |
+
"loss": 0.1,
|
| 1601 |
+
"step": 5525
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 2.362707535121328,
|
| 1605 |
+
"grad_norm": 2.6536972522735596,
|
| 1606 |
+
"learning_rate": 3.5042857142857145e-06,
|
| 1607 |
+
"loss": 0.0911,
|
| 1608 |
+
"step": 5550
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 2.373350361856109,
|
| 1612 |
+
"grad_norm": 2.0286781787872314,
|
| 1613 |
+
"learning_rate": 3.468571428571429e-06,
|
| 1614 |
+
"loss": 0.083,
|
| 1615 |
+
"step": 5575
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 2.38399318859089,
|
| 1619 |
+
"grad_norm": 3.5354936122894287,
|
| 1620 |
+
"learning_rate": 3.4328571428571435e-06,
|
| 1621 |
+
"loss": 0.0994,
|
| 1622 |
+
"step": 5600
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 2.3946360153256707,
|
| 1626 |
+
"grad_norm": 2.823812246322632,
|
| 1627 |
+
"learning_rate": 3.397142857142857e-06,
|
| 1628 |
+
"loss": 0.0921,
|
| 1629 |
+
"step": 5625
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 2.405278842060451,
|
| 1633 |
+
"grad_norm": 3.5603067874908447,
|
| 1634 |
+
"learning_rate": 3.3614285714285717e-06,
|
| 1635 |
+
"loss": 0.1015,
|
| 1636 |
+
"step": 5650
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 2.415921668795232,
|
| 1640 |
+
"grad_norm": 2.4219422340393066,
|
| 1641 |
+
"learning_rate": 3.325714285714286e-06,
|
| 1642 |
+
"loss": 0.098,
|
| 1643 |
+
"step": 5675
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 2.4265644955300125,
|
| 1647 |
+
"grad_norm": 3.9650704860687256,
|
| 1648 |
+
"learning_rate": 3.2900000000000003e-06,
|
| 1649 |
+
"loss": 0.0914,
|
| 1650 |
+
"step": 5700
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 2.4372073222647934,
|
| 1654 |
+
"grad_norm": 2.7661550045013428,
|
| 1655 |
+
"learning_rate": 3.2542857142857148e-06,
|
| 1656 |
+
"loss": 0.0733,
|
| 1657 |
+
"step": 5725
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 2.4478501489995743,
|
| 1661 |
+
"grad_norm": 2.8396358489990234,
|
| 1662 |
+
"learning_rate": 3.218571428571429e-06,
|
| 1663 |
+
"loss": 0.0954,
|
| 1664 |
+
"step": 5750
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 2.458492975734355,
|
| 1668 |
+
"grad_norm": 2.8353986740112305,
|
| 1669 |
+
"learning_rate": 3.182857142857143e-06,
|
| 1670 |
+
"loss": 0.0848,
|
| 1671 |
+
"step": 5775
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 2.4691358024691357,
|
| 1675 |
+
"grad_norm": 2.9679837226867676,
|
| 1676 |
+
"learning_rate": 3.1471428571428574e-06,
|
| 1677 |
+
"loss": 0.084,
|
| 1678 |
+
"step": 5800
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 2.4797786292039166,
|
| 1682 |
+
"grad_norm": 2.0554795265197754,
|
| 1683 |
+
"learning_rate": 3.111428571428572e-06,
|
| 1684 |
+
"loss": 0.0894,
|
| 1685 |
+
"step": 5825
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 2.4904214559386975,
|
| 1689 |
+
"grad_norm": 2.5439860820770264,
|
| 1690 |
+
"learning_rate": 3.0757142857142856e-06,
|
| 1691 |
+
"loss": 0.0836,
|
| 1692 |
+
"step": 5850
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 2.501064282673478,
|
| 1696 |
+
"grad_norm": 2.93955135345459,
|
| 1697 |
+
"learning_rate": 3.04e-06,
|
| 1698 |
+
"loss": 0.0915,
|
| 1699 |
+
"step": 5875
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 2.511707109408259,
|
| 1703 |
+
"grad_norm": 2.3502097129821777,
|
| 1704 |
+
"learning_rate": 3.0042857142857146e-06,
|
| 1705 |
+
"loss": 0.0963,
|
| 1706 |
+
"step": 5900
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 2.5223499361430397,
|
| 1710 |
+
"grad_norm": 2.289599895477295,
|
| 1711 |
+
"learning_rate": 2.968571428571429e-06,
|
| 1712 |
+
"loss": 0.0892,
|
| 1713 |
+
"step": 5925
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 2.5329927628778206,
|
| 1717 |
+
"grad_norm": 4.718634128570557,
|
| 1718 |
+
"learning_rate": 2.932857142857143e-06,
|
| 1719 |
+
"loss": 0.089,
|
| 1720 |
+
"step": 5950
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 2.543635589612601,
|
| 1724 |
+
"grad_norm": 2.9124553203582764,
|
| 1725 |
+
"learning_rate": 2.8971428571428573e-06,
|
| 1726 |
+
"loss": 0.0943,
|
| 1727 |
+
"step": 5975
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 2.554278416347382,
|
| 1731 |
+
"grad_norm": 3.2406508922576904,
|
| 1732 |
+
"learning_rate": 2.861428571428572e-06,
|
| 1733 |
+
"loss": 0.0891,
|
| 1734 |
+
"step": 6000
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 2.554278416347382,
|
| 1738 |
+
"eval_loss": 0.1251918077468872,
|
| 1739 |
+
"eval_runtime": 2435.0048,
|
| 1740 |
+
"eval_samples_per_second": 1.962,
|
| 1741 |
+
"eval_steps_per_second": 0.062,
|
| 1742 |
+
"eval_wer": 16.934856191286404,
|
| 1743 |
+
"step": 6000
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 2.5649212430821624,
|
| 1747 |
+
"grad_norm": 2.5758533477783203,
|
| 1748 |
+
"learning_rate": 2.825714285714286e-06,
|
| 1749 |
+
"loss": 0.0909,
|
| 1750 |
+
"step": 6025
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 2.5755640698169433,
|
| 1754 |
+
"grad_norm": 2.308535575866699,
|
| 1755 |
+
"learning_rate": 2.7900000000000004e-06,
|
| 1756 |
+
"loss": 0.0903,
|
| 1757 |
+
"step": 6050
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 2.586206896551724,
|
| 1761 |
+
"grad_norm": 3.0140132904052734,
|
| 1762 |
+
"learning_rate": 2.7542857142857145e-06,
|
| 1763 |
+
"loss": 0.1005,
|
| 1764 |
+
"step": 6075
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 2.596849723286505,
|
| 1768 |
+
"grad_norm": 3.0237767696380615,
|
| 1769 |
+
"learning_rate": 2.7185714285714286e-06,
|
| 1770 |
+
"loss": 0.1032,
|
| 1771 |
+
"step": 6100
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 2.6074925500212855,
|
| 1775 |
+
"grad_norm": 2.413677930831909,
|
| 1776 |
+
"learning_rate": 2.682857142857143e-06,
|
| 1777 |
+
"loss": 0.0753,
|
| 1778 |
+
"step": 6125
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 2.6181353767560664,
|
| 1782 |
+
"grad_norm": 2.406214475631714,
|
| 1783 |
+
"learning_rate": 2.6471428571428576e-06,
|
| 1784 |
+
"loss": 0.0744,
|
| 1785 |
+
"step": 6150
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 2.628778203490847,
|
| 1789 |
+
"grad_norm": 2.9371650218963623,
|
| 1790 |
+
"learning_rate": 2.6114285714285712e-06,
|
| 1791 |
+
"loss": 0.0795,
|
| 1792 |
+
"step": 6175
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 2.639421030225628,
|
| 1796 |
+
"grad_norm": 3.0647592544555664,
|
| 1797 |
+
"learning_rate": 2.5757142857142857e-06,
|
| 1798 |
+
"loss": 0.0885,
|
| 1799 |
+
"step": 6200
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.6500638569604087,
|
| 1803 |
+
"grad_norm": 2.245195150375366,
|
| 1804 |
+
"learning_rate": 2.5400000000000002e-06,
|
| 1805 |
+
"loss": 0.0951,
|
| 1806 |
+
"step": 6225
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 2.6607066836951896,
|
| 1810 |
+
"grad_norm": 3.212939977645874,
|
| 1811 |
+
"learning_rate": 2.5042857142857148e-06,
|
| 1812 |
+
"loss": 0.1081,
|
| 1813 |
+
"step": 6250
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 2.67134951042997,
|
| 1817 |
+
"grad_norm": 2.987602949142456,
|
| 1818 |
+
"learning_rate": 2.468571428571429e-06,
|
| 1819 |
+
"loss": 0.0694,
|
| 1820 |
+
"step": 6275
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 2.681992337164751,
|
| 1824 |
+
"grad_norm": 2.6746339797973633,
|
| 1825 |
+
"learning_rate": 2.4328571428571433e-06,
|
| 1826 |
+
"loss": 0.0879,
|
| 1827 |
+
"step": 6300
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 2.692635163899532,
|
| 1831 |
+
"grad_norm": 2.3074121475219727,
|
| 1832 |
+
"learning_rate": 2.3971428571428574e-06,
|
| 1833 |
+
"loss": 0.0771,
|
| 1834 |
+
"step": 6325
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 2.7032779906343123,
|
| 1838 |
+
"grad_norm": 2.62947940826416,
|
| 1839 |
+
"learning_rate": 2.361428571428572e-06,
|
| 1840 |
+
"loss": 0.0882,
|
| 1841 |
+
"step": 6350
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 2.713920817369093,
|
| 1845 |
+
"grad_norm": 2.5452988147735596,
|
| 1846 |
+
"learning_rate": 2.325714285714286e-06,
|
| 1847 |
+
"loss": 0.081,
|
| 1848 |
+
"step": 6375
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 2.724563644103874,
|
| 1852 |
+
"grad_norm": 1.9240838289260864,
|
| 1853 |
+
"learning_rate": 2.29e-06,
|
| 1854 |
+
"loss": 0.0672,
|
| 1855 |
+
"step": 6400
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 2.735206470838655,
|
| 1859 |
+
"grad_norm": 2.3632349967956543,
|
| 1860 |
+
"learning_rate": 2.2542857142857146e-06,
|
| 1861 |
+
"loss": 0.0716,
|
| 1862 |
+
"step": 6425
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 2.7458492975734354,
|
| 1866 |
+
"grad_norm": 1.9626713991165161,
|
| 1867 |
+
"learning_rate": 2.2185714285714287e-06,
|
| 1868 |
+
"loss": 0.0857,
|
| 1869 |
+
"step": 6450
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 2.7564921243082163,
|
| 1873 |
+
"grad_norm": 1.8497956991195679,
|
| 1874 |
+
"learning_rate": 2.1828571428571428e-06,
|
| 1875 |
+
"loss": 0.0774,
|
| 1876 |
+
"step": 6475
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 2.767134951042997,
|
| 1880 |
+
"grad_norm": 1.9737045764923096,
|
| 1881 |
+
"learning_rate": 2.1471428571428573e-06,
|
| 1882 |
+
"loss": 0.0884,
|
| 1883 |
+
"step": 6500
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 2.7777777777777777,
|
| 1887 |
+
"grad_norm": 3.017702102661133,
|
| 1888 |
+
"learning_rate": 2.1114285714285714e-06,
|
| 1889 |
+
"loss": 0.0894,
|
| 1890 |
+
"step": 6525
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 2.7884206045125586,
|
| 1894 |
+
"grad_norm": 2.41921067237854,
|
| 1895 |
+
"learning_rate": 2.075714285714286e-06,
|
| 1896 |
+
"loss": 0.0855,
|
| 1897 |
+
"step": 6550
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 2.7990634312473395,
|
| 1901 |
+
"grad_norm": 2.0304954051971436,
|
| 1902 |
+
"learning_rate": 2.04e-06,
|
| 1903 |
+
"loss": 0.0802,
|
| 1904 |
+
"step": 6575
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 2.80970625798212,
|
| 1908 |
+
"grad_norm": 2.724147319793701,
|
| 1909 |
+
"learning_rate": 2.0042857142857145e-06,
|
| 1910 |
+
"loss": 0.0892,
|
| 1911 |
+
"step": 6600
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 2.820349084716901,
|
| 1915 |
+
"grad_norm": 1.7320371866226196,
|
| 1916 |
+
"learning_rate": 1.968571428571429e-06,
|
| 1917 |
+
"loss": 0.1036,
|
| 1918 |
+
"step": 6625
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 2.8309919114516817,
|
| 1922 |
+
"grad_norm": 2.932657241821289,
|
| 1923 |
+
"learning_rate": 1.932857142857143e-06,
|
| 1924 |
+
"loss": 0.0902,
|
| 1925 |
+
"step": 6650
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 2.841634738186462,
|
| 1929 |
+
"grad_norm": 2.653630256652832,
|
| 1930 |
+
"learning_rate": 1.8971428571428573e-06,
|
| 1931 |
+
"loss": 0.0807,
|
| 1932 |
+
"step": 6675
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 2.852277564921243,
|
| 1936 |
+
"grad_norm": 2.851041078567505,
|
| 1937 |
+
"learning_rate": 1.8614285714285714e-06,
|
| 1938 |
+
"loss": 0.0908,
|
| 1939 |
+
"step": 6700
|
| 1940 |
+
},
|
| 1941 |
+
{
|
| 1942 |
+
"epoch": 2.862920391656024,
|
| 1943 |
+
"grad_norm": 3.30446720123291,
|
| 1944 |
+
"learning_rate": 1.825714285714286e-06,
|
| 1945 |
+
"loss": 0.1001,
|
| 1946 |
+
"step": 6725
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 2.873563218390805,
|
| 1950 |
+
"grad_norm": 3.250701427459717,
|
| 1951 |
+
"learning_rate": 1.79e-06,
|
| 1952 |
+
"loss": 0.0825,
|
| 1953 |
+
"step": 6750
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 2.8842060451255853,
|
| 1957 |
+
"grad_norm": 2.4845850467681885,
|
| 1958 |
+
"learning_rate": 1.7542857142857145e-06,
|
| 1959 |
+
"loss": 0.0705,
|
| 1960 |
+
"step": 6775
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 2.894848871860366,
|
| 1964 |
+
"grad_norm": 2.6934683322906494,
|
| 1965 |
+
"learning_rate": 1.7185714285714286e-06,
|
| 1966 |
+
"loss": 0.0667,
|
| 1967 |
+
"step": 6800
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 2.9054916985951467,
|
| 1971 |
+
"grad_norm": 2.785459518432617,
|
| 1972 |
+
"learning_rate": 1.6828571428571431e-06,
|
| 1973 |
+
"loss": 0.0746,
|
| 1974 |
+
"step": 6825
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 2.9161345253299276,
|
| 1978 |
+
"grad_norm": 2.5107369422912598,
|
| 1979 |
+
"learning_rate": 1.6471428571428572e-06,
|
| 1980 |
+
"loss": 0.077,
|
| 1981 |
+
"step": 6850
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 2.9267773520647085,
|
| 1985 |
+
"grad_norm": 3.4977328777313232,
|
| 1986 |
+
"learning_rate": 1.6114285714285715e-06,
|
| 1987 |
+
"loss": 0.0857,
|
| 1988 |
+
"step": 6875
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 2.9374201787994894,
|
| 1992 |
+
"grad_norm": 2.6151537895202637,
|
| 1993 |
+
"learning_rate": 1.575714285714286e-06,
|
| 1994 |
+
"loss": 0.0729,
|
| 1995 |
+
"step": 6900
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 2.94806300553427,
|
| 1999 |
+
"grad_norm": 2.975446939468384,
|
| 2000 |
+
"learning_rate": 1.54e-06,
|
| 2001 |
+
"loss": 0.0581,
|
| 2002 |
+
"step": 6925
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 2.9587058322690507,
|
| 2006 |
+
"grad_norm": 2.03027606010437,
|
| 2007 |
+
"learning_rate": 1.5042857142857146e-06,
|
| 2008 |
+
"loss": 0.1017,
|
| 2009 |
+
"step": 6950
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 2.969348659003831,
|
| 2013 |
+
"grad_norm": 3.547647476196289,
|
| 2014 |
+
"learning_rate": 1.4685714285714287e-06,
|
| 2015 |
+
"loss": 0.0603,
|
| 2016 |
+
"step": 6975
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 2.979991485738612,
|
| 2020 |
+
"grad_norm": 1.8231449127197266,
|
| 2021 |
+
"learning_rate": 1.432857142857143e-06,
|
| 2022 |
+
"loss": 0.0738,
|
| 2023 |
+
"step": 7000
|
| 2024 |
+
},
|
| 2025 |
+
{
|
| 2026 |
+
"epoch": 2.979991485738612,
|
| 2027 |
+
"eval_loss": 0.11986401677131653,
|
| 2028 |
+
"eval_runtime": 2460.4728,
|
| 2029 |
+
"eval_samples_per_second": 1.941,
|
| 2030 |
+
"eval_steps_per_second": 0.061,
|
| 2031 |
+
"eval_wer": 15.561025938059986,
|
| 2032 |
+
"step": 7000
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 2.990634312473393,
|
| 2036 |
+
"grad_norm": 2.170557737350464,
|
| 2037 |
+
"learning_rate": 1.3971428571428573e-06,
|
| 2038 |
+
"loss": 0.0847,
|
| 2039 |
+
"step": 7025
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 3.001277139208174,
|
| 2043 |
+
"grad_norm": 1.225490689277649,
|
| 2044 |
+
"learning_rate": 1.3614285714285716e-06,
|
| 2045 |
+
"loss": 0.0737,
|
| 2046 |
+
"step": 7050
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 3.0119199659429543,
|
| 2050 |
+
"grad_norm": 2.1241679191589355,
|
| 2051 |
+
"learning_rate": 1.3257142857142856e-06,
|
| 2052 |
+
"loss": 0.0595,
|
| 2053 |
+
"step": 7075
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 3.022562792677735,
|
| 2057 |
+
"grad_norm": 2.3180058002471924,
|
| 2058 |
+
"learning_rate": 1.2900000000000001e-06,
|
| 2059 |
+
"loss": 0.0615,
|
| 2060 |
+
"step": 7100
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 3.033205619412516,
|
| 2064 |
+
"grad_norm": 2.4434351921081543,
|
| 2065 |
+
"learning_rate": 1.2542857142857142e-06,
|
| 2066 |
+
"loss": 0.0536,
|
| 2067 |
+
"step": 7125
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 3.0438484461472965,
|
| 2071 |
+
"grad_norm": 2.712207317352295,
|
| 2072 |
+
"learning_rate": 1.2185714285714287e-06,
|
| 2073 |
+
"loss": 0.0558,
|
| 2074 |
+
"step": 7150
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 3.0544912728820774,
|
| 2078 |
+
"grad_norm": 2.7258520126342773,
|
| 2079 |
+
"learning_rate": 1.182857142857143e-06,
|
| 2080 |
+
"loss": 0.0727,
|
| 2081 |
+
"step": 7175
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 3.0651340996168583,
|
| 2085 |
+
"grad_norm": 2.103072166442871,
|
| 2086 |
+
"learning_rate": 1.1471428571428573e-06,
|
| 2087 |
+
"loss": 0.0479,
|
| 2088 |
+
"step": 7200
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 3.075776926351639,
|
| 2092 |
+
"grad_norm": 1.9003605842590332,
|
| 2093 |
+
"learning_rate": 1.1114285714285714e-06,
|
| 2094 |
+
"loss": 0.0554,
|
| 2095 |
+
"step": 7225
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 3.0864197530864197,
|
| 2099 |
+
"grad_norm": 1.4967641830444336,
|
| 2100 |
+
"learning_rate": 1.0757142857142857e-06,
|
| 2101 |
+
"loss": 0.0409,
|
| 2102 |
+
"step": 7250
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 3.0970625798212006,
|
| 2106 |
+
"grad_norm": 1.389493703842163,
|
| 2107 |
+
"learning_rate": 1.04e-06,
|
| 2108 |
+
"loss": 0.0474,
|
| 2109 |
+
"step": 7275
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 3.107705406555981,
|
| 2113 |
+
"grad_norm": 1.4253233671188354,
|
| 2114 |
+
"learning_rate": 1.0042857142857143e-06,
|
| 2115 |
+
"loss": 0.0445,
|
| 2116 |
+
"step": 7300
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 3.118348233290762,
|
| 2120 |
+
"grad_norm": 2.6737582683563232,
|
| 2121 |
+
"learning_rate": 9.685714285714288e-07,
|
| 2122 |
+
"loss": 0.0584,
|
| 2123 |
+
"step": 7325
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 3.128991060025543,
|
| 2127 |
+
"grad_norm": 2.5511069297790527,
|
| 2128 |
+
"learning_rate": 9.32857142857143e-07,
|
| 2129 |
+
"loss": 0.0557,
|
| 2130 |
+
"step": 7350
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 3.1396338867603237,
|
| 2134 |
+
"grad_norm": 2.139846086502075,
|
| 2135 |
+
"learning_rate": 8.971428571428573e-07,
|
| 2136 |
+
"loss": 0.0467,
|
| 2137 |
+
"step": 7375
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 3.150276713495104,
|
| 2141 |
+
"grad_norm": 1.826206088066101,
|
| 2142 |
+
"learning_rate": 8.614285714285716e-07,
|
| 2143 |
+
"loss": 0.054,
|
| 2144 |
+
"step": 7400
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 3.160919540229885,
|
| 2148 |
+
"grad_norm": 2.8576643466949463,
|
| 2149 |
+
"learning_rate": 8.257142857142858e-07,
|
| 2150 |
+
"loss": 0.0517,
|
| 2151 |
+
"step": 7425
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 3.171562366964666,
|
| 2155 |
+
"grad_norm": 2.1208717823028564,
|
| 2156 |
+
"learning_rate": 7.900000000000001e-07,
|
| 2157 |
+
"loss": 0.0667,
|
| 2158 |
+
"step": 7450
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 3.1822051936994464,
|
| 2162 |
+
"grad_norm": 1.534239649772644,
|
| 2163 |
+
"learning_rate": 7.542857142857144e-07,
|
| 2164 |
+
"loss": 0.0592,
|
| 2165 |
+
"step": 7475
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 3.1928480204342273,
|
| 2169 |
+
"grad_norm": 2.3605740070343018,
|
| 2170 |
+
"learning_rate": 7.185714285714286e-07,
|
| 2171 |
+
"loss": 0.063,
|
| 2172 |
+
"step": 7500
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 3.2034908471690082,
|
| 2176 |
+
"grad_norm": 2.1266493797302246,
|
| 2177 |
+
"learning_rate": 6.842857142857143e-07,
|
| 2178 |
+
"loss": 0.0567,
|
| 2179 |
+
"step": 7525
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 3.2141336739037887,
|
| 2183 |
+
"grad_norm": 1.5303648710250854,
|
| 2184 |
+
"learning_rate": 6.485714285714287e-07,
|
| 2185 |
+
"loss": 0.0619,
|
| 2186 |
+
"step": 7550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 3.2247765006385696,
|
| 2190 |
+
"grad_norm": 2.740006446838379,
|
| 2191 |
+
"learning_rate": 6.128571428571429e-07,
|
| 2192 |
+
"loss": 0.0773,
|
| 2193 |
+
"step": 7575
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 3.2354193273733505,
|
| 2197 |
+
"grad_norm": 1.5786134004592896,
|
| 2198 |
+
"learning_rate": 5.771428571428572e-07,
|
| 2199 |
+
"loss": 0.0629,
|
| 2200 |
+
"step": 7600
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 3.246062154108131,
|
| 2204 |
+
"grad_norm": 1.3754280805587769,
|
| 2205 |
+
"learning_rate": 5.414285714285715e-07,
|
| 2206 |
+
"loss": 0.0614,
|
| 2207 |
+
"step": 7625
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 3.256704980842912,
|
| 2211 |
+
"grad_norm": 0.8814867734909058,
|
| 2212 |
+
"learning_rate": 5.057142857142858e-07,
|
| 2213 |
+
"loss": 0.0574,
|
| 2214 |
+
"step": 7650
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 3.2673478075776927,
|
| 2218 |
+
"grad_norm": 2.909646511077881,
|
| 2219 |
+
"learning_rate": 4.7000000000000005e-07,
|
| 2220 |
+
"loss": 0.0408,
|
| 2221 |
+
"step": 7675
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 3.2779906343124736,
|
| 2225 |
+
"grad_norm": 2.272367238998413,
|
| 2226 |
+
"learning_rate": 4.342857142857143e-07,
|
| 2227 |
+
"loss": 0.0539,
|
| 2228 |
+
"step": 7700
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 3.288633461047254,
|
| 2232 |
+
"grad_norm": 2.039271831512451,
|
| 2233 |
+
"learning_rate": 3.985714285714286e-07,
|
| 2234 |
+
"loss": 0.0688,
|
| 2235 |
+
"step": 7725
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 3.299276287782035,
|
| 2239 |
+
"grad_norm": 2.0516164302825928,
|
| 2240 |
+
"learning_rate": 3.6285714285714283e-07,
|
| 2241 |
+
"loss": 0.0546,
|
| 2242 |
+
"step": 7750
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 3.3099191145168154,
|
| 2246 |
+
"grad_norm": 1.9131453037261963,
|
| 2247 |
+
"learning_rate": 3.271428571428572e-07,
|
| 2248 |
+
"loss": 0.0532,
|
| 2249 |
+
"step": 7775
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 3.3205619412515963,
|
| 2253 |
+
"grad_norm": 1.66374933719635,
|
| 2254 |
+
"learning_rate": 2.914285714285715e-07,
|
| 2255 |
+
"loss": 0.0536,
|
| 2256 |
+
"step": 7800
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 3.331204767986377,
|
| 2260 |
+
"grad_norm": 1.596907615661621,
|
| 2261 |
+
"learning_rate": 2.557142857142857e-07,
|
| 2262 |
+
"loss": 0.0456,
|
| 2263 |
+
"step": 7825
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 3.341847594721158,
|
| 2267 |
+
"grad_norm": 2.430992603302002,
|
| 2268 |
+
"learning_rate": 2.2e-07,
|
| 2269 |
+
"loss": 0.0635,
|
| 2270 |
+
"step": 7850
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 3.3524904214559386,
|
| 2274 |
+
"grad_norm": 2.4150683879852295,
|
| 2275 |
+
"learning_rate": 1.842857142857143e-07,
|
| 2276 |
+
"loss": 0.0638,
|
| 2277 |
+
"step": 7875
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 3.3631332481907195,
|
| 2281 |
+
"grad_norm": 1.5517698526382446,
|
| 2282 |
+
"learning_rate": 1.4857142857142857e-07,
|
| 2283 |
+
"loss": 0.0615,
|
| 2284 |
+
"step": 7900
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 3.3737760749255004,
|
| 2288 |
+
"grad_norm": 1.8786826133728027,
|
| 2289 |
+
"learning_rate": 1.1285714285714287e-07,
|
| 2290 |
+
"loss": 0.0607,
|
| 2291 |
+
"step": 7925
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 3.384418901660281,
|
| 2295 |
+
"grad_norm": 2.849170446395874,
|
| 2296 |
+
"learning_rate": 7.714285714285715e-08,
|
| 2297 |
+
"loss": 0.0567,
|
| 2298 |
+
"step": 7950
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 3.3950617283950617,
|
| 2302 |
+
"grad_norm": 2.8803513050079346,
|
| 2303 |
+
"learning_rate": 4.1428571428571426e-08,
|
| 2304 |
+
"loss": 0.0507,
|
| 2305 |
+
"step": 7975
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 3.4057045551298426,
|
| 2309 |
+
"grad_norm": 2.7549145221710205,
|
| 2310 |
+
"learning_rate": 5.714285714285715e-09,
|
| 2311 |
+
"loss": 0.0544,
|
| 2312 |
+
"step": 8000
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 3.4057045551298426,
|
| 2316 |
+
"eval_loss": 0.1155887097120285,
|
| 2317 |
+
"eval_runtime": 2444.7471,
|
| 2318 |
+
"eval_samples_per_second": 1.954,
|
| 2319 |
+
"eval_steps_per_second": 0.061,
|
| 2320 |
+
"eval_wer": 15.246076710047603,
|
| 2321 |
+
"step": 8000
|
| 2322 |
+
}
|
| 2323 |
+
],
|
| 2324 |
+
"logging_steps": 25,
|
| 2325 |
+
"max_steps": 8000,
|
| 2326 |
+
"num_input_tokens_seen": 0,
|
| 2327 |
+
"num_train_epochs": 4,
|
| 2328 |
+
"save_steps": 1000,
|
| 2329 |
+
"stateful_callbacks": {
|
| 2330 |
+
"TrainerControl": {
|
| 2331 |
+
"args": {
|
| 2332 |
+
"should_epoch_stop": false,
|
| 2333 |
+
"should_evaluate": false,
|
| 2334 |
+
"should_log": false,
|
| 2335 |
+
"should_save": true,
|
| 2336 |
+
"should_training_stop": true
|
| 2337 |
+
},
|
| 2338 |
+
"attributes": {}
|
| 2339 |
+
}
|
| 2340 |
+
},
|
| 2341 |
+
"total_flos": 4.364112316878029e+20,
|
| 2342 |
+
"train_batch_size": 32,
|
| 2343 |
+
"trial_name": null,
|
| 2344 |
+
"trial_params": null
|
| 2345 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:df4c3f6cdbdc7d71c42768190b989231c9ffbea37fb13bfc17d552f5f4c2732d
|
| 3 |
+
size 5368
|