updating the repo with the fine-tuned model
Browse files- README.md +104 -0
- added_tokens.json +4 -0
- all_results.json +14 -0
- alphabet.json +1 -0
- config.json +107 -0
- eval_results.json +9 -0
- language_model/attrs.json +1 -0
- language_model/unigrams.txt +0 -0
- language_model/uwb_atcc_4g.binary +0 -0
- log/train_log +0 -0
- preprocessor_config.json +10 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +6 -0
- tokenizer_config.json +11 -0
- train_results.json +8 -0
- trainer_state.json +625 -0
- training_args.bin +3 -0
- vocab.json +31 -0
README.md
ADDED
|
@@ -0,0 +1,104 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- automatic-speech-recognition
|
| 5 |
+
- experiments/data/atcosim_corpus/train
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
metrics:
|
| 8 |
+
- wer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: 0.0ld_0.05ad_0.05attd_0.0fpd_0.03mtp_10mtl_0.0mfp_10mfl
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# 0.0ld_0.05ad_0.05attd_0.0fpd_0.03mtp_10mtl_0.0mfp_10mfl
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the EXPERIMENTS/DATA/ATCOSIM_CORPUS/TRAIN - NA dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.0988
|
| 22 |
+
- Wer: 0.0736
|
| 23 |
+
|
| 24 |
+
## Model description
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Intended uses & limitations
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training and evaluation data
|
| 33 |
+
|
| 34 |
+
More information needed
|
| 35 |
+
|
| 36 |
+
## Training procedure
|
| 37 |
+
|
| 38 |
+
### Training hyperparameters
|
| 39 |
+
|
| 40 |
+
The following hyperparameters were used during training:
|
| 41 |
+
- learning_rate: 0.0005
|
| 42 |
+
- train_batch_size: 24
|
| 43 |
+
- eval_batch_size: 24
|
| 44 |
+
- seed: 42
|
| 45 |
+
- gradient_accumulation_steps: 4
|
| 46 |
+
- total_train_batch_size: 96
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: linear
|
| 49 |
+
- lr_scheduler_warmup_steps: 500
|
| 50 |
+
- training_steps: 20000
|
| 51 |
+
- mixed_precision_training: Native AMP
|
| 52 |
+
|
| 53 |
+
### Training results
|
| 54 |
+
|
| 55 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 56 |
+
|:-------------:|:------:|:-----:|:---------------:|:------:|
|
| 57 |
+
| 1.9105 | 6.41 | 500 | 0.1622 | 0.1531 |
|
| 58 |
+
| 0.1119 | 12.82 | 1000 | 0.0971 | 0.0936 |
|
| 59 |
+
| 0.0614 | 19.23 | 1500 | 0.1002 | 0.0983 |
|
| 60 |
+
| 0.044 | 25.64 | 2000 | 0.1011 | 0.0929 |
|
| 61 |
+
| 0.0366 | 32.05 | 2500 | 0.0932 | 0.0828 |
|
| 62 |
+
| 0.0315 | 38.46 | 3000 | 0.0926 | 0.0880 |
|
| 63 |
+
| 0.0297 | 44.87 | 3500 | 0.0972 | 0.0882 |
|
| 64 |
+
| 0.0216 | 51.28 | 4000 | 0.0911 | 0.0774 |
|
| 65 |
+
| 0.0211 | 57.69 | 4500 | 0.0982 | 0.0891 |
|
| 66 |
+
| 0.0187 | 64.1 | 5000 | 0.1009 | 0.0863 |
|
| 67 |
+
| 0.02 | 70.51 | 5500 | 0.0953 | 0.0852 |
|
| 68 |
+
| 0.0163 | 76.92 | 6000 | 0.1028 | 0.0804 |
|
| 69 |
+
| 0.0128 | 83.33 | 6500 | 0.0930 | 0.0856 |
|
| 70 |
+
| 0.0127 | 89.74 | 7000 | 0.0892 | 0.0676 |
|
| 71 |
+
| 0.0116 | 96.15 | 7500 | 0.0857 | 0.0753 |
|
| 72 |
+
| 0.0139 | 102.56 | 8000 | 0.1078 | 0.0481 |
|
| 73 |
+
| 0.0107 | 108.97 | 8500 | 0.0955 | 0.0683 |
|
| 74 |
+
| 0.0096 | 115.38 | 9000 | 0.0846 | 0.0697 |
|
| 75 |
+
| 0.0089 | 121.79 | 9500 | 0.0854 | 0.0675 |
|
| 76 |
+
| 0.0084 | 128.21 | 10000 | 0.0875 | 0.0779 |
|
| 77 |
+
| 0.0074 | 134.62 | 10500 | 0.0840 | 0.0770 |
|
| 78 |
+
| 0.0061 | 141.03 | 11000 | 0.0903 | 0.0754 |
|
| 79 |
+
| 0.0076 | 147.44 | 11500 | 0.0872 | 0.0769 |
|
| 80 |
+
| 0.0069 | 153.85 | 12000 | 0.0891 | 0.0772 |
|
| 81 |
+
| 0.0061 | 160.26 | 12500 | 0.0971 | 0.0774 |
|
| 82 |
+
| 0.0049 | 166.67 | 13000 | 0.0984 | 0.0726 |
|
| 83 |
+
| 0.0045 | 173.08 | 13500 | 0.0952 | 0.0765 |
|
| 84 |
+
| 0.0039 | 179.49 | 14000 | 0.1015 | 0.0762 |
|
| 85 |
+
| 0.0031 | 185.9 | 14500 | 0.0937 | 0.0712 |
|
| 86 |
+
| 0.0032 | 192.31 | 15000 | 0.0982 | 0.0635 |
|
| 87 |
+
| 0.0028 | 198.72 | 15500 | 0.0981 | 0.0743 |
|
| 88 |
+
| 0.0024 | 205.13 | 16000 | 0.1019 | 0.0712 |
|
| 89 |
+
| 0.0024 | 211.54 | 16500 | 0.0957 | 0.0732 |
|
| 90 |
+
| 0.002 | 217.95 | 17000 | 0.0941 | 0.0732 |
|
| 91 |
+
| 0.0015 | 224.36 | 17500 | 0.1009 | 0.0717 |
|
| 92 |
+
| 0.0017 | 230.77 | 18000 | 0.0955 | 0.0730 |
|
| 93 |
+
| 0.0013 | 237.18 | 18500 | 0.0989 | 0.0732 |
|
| 94 |
+
| 0.0013 | 243.59 | 19000 | 0.0967 | 0.0738 |
|
| 95 |
+
| 0.0011 | 250.0 | 19500 | 0.0980 | 0.0734 |
|
| 96 |
+
| 0.0008 | 256.41 | 20000 | 0.0988 | 0.0736 |
|
| 97 |
+
|
| 98 |
+
|
| 99 |
+
### Framework versions
|
| 100 |
+
|
| 101 |
+
- Transformers 4.24.0
|
| 102 |
+
- Pytorch 1.13.0+cu117
|
| 103 |
+
- Datasets 2.6.1
|
| 104 |
+
- Tokenizers 0.13.2
|
added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</s>": 30,
|
| 3 |
+
"<s>": 29
|
| 4 |
+
}
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 256.41,
|
| 3 |
+
"eval_loss": 0.09882301837205887,
|
| 4 |
+
"eval_runtime": 66.9757,
|
| 5 |
+
"eval_samples": 1857,
|
| 6 |
+
"eval_samples_per_second": 27.726,
|
| 7 |
+
"eval_steps_per_second": 1.165,
|
| 8 |
+
"eval_wer": 0.07359326539654408,
|
| 9 |
+
"train_loss": 0.06207249406576157,
|
| 10 |
+
"train_runtime": 86281.9422,
|
| 11 |
+
"train_samples": 7480,
|
| 12 |
+
"train_samples_per_second": 22.253,
|
| 13 |
+
"train_steps_per_second": 0.232
|
| 14 |
+
}
|
alphabet.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"labels": [" ", "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z", "\u2047", "", "<s>", "</s>"], "is_bpe": false}
|
config.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-xls-r-300m",
|
| 3 |
+
"activation_dropout": 0.05,
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2ForCTC"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.05,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 256,
|
| 14 |
+
"codevector_dim": 768,
|
| 15 |
+
"contrastive_logits_temperature": 0.1,
|
| 16 |
+
"conv_bias": true,
|
| 17 |
+
"conv_dim": [
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512
|
| 25 |
+
],
|
| 26 |
+
"conv_kernel": [
|
| 27 |
+
10,
|
| 28 |
+
3,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
2,
|
| 33 |
+
2
|
| 34 |
+
],
|
| 35 |
+
"conv_stride": [
|
| 36 |
+
5,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2
|
| 43 |
+
],
|
| 44 |
+
"ctc_loss_reduction": "mean",
|
| 45 |
+
"ctc_zero_infinity": true,
|
| 46 |
+
"diversity_loss_weight": 0.1,
|
| 47 |
+
"do_stable_layer_norm": true,
|
| 48 |
+
"eos_token_id": 2,
|
| 49 |
+
"feat_extract_activation": "gelu",
|
| 50 |
+
"feat_extract_dropout": 0.0,
|
| 51 |
+
"feat_extract_norm": "layer",
|
| 52 |
+
"feat_proj_dropout": 0.0,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"hidden_act": "gelu",
|
| 56 |
+
"hidden_dropout": 0.0,
|
| 57 |
+
"hidden_size": 1024,
|
| 58 |
+
"initializer_range": 0.02,
|
| 59 |
+
"intermediate_size": 4096,
|
| 60 |
+
"layer_norm_eps": 1e-05,
|
| 61 |
+
"layerdrop": 0.0,
|
| 62 |
+
"mask_feature_length": 10,
|
| 63 |
+
"mask_feature_min_masks": 0,
|
| 64 |
+
"mask_feature_prob": 0.0,
|
| 65 |
+
"mask_time_length": 10,
|
| 66 |
+
"mask_time_min_masks": 2,
|
| 67 |
+
"mask_time_prob": 0.03,
|
| 68 |
+
"model_type": "wav2vec2",
|
| 69 |
+
"num_adapter_layers": 3,
|
| 70 |
+
"num_attention_heads": 16,
|
| 71 |
+
"num_codevector_groups": 2,
|
| 72 |
+
"num_codevectors_per_group": 320,
|
| 73 |
+
"num_conv_pos_embedding_groups": 16,
|
| 74 |
+
"num_conv_pos_embeddings": 128,
|
| 75 |
+
"num_feat_extract_layers": 7,
|
| 76 |
+
"num_hidden_layers": 24,
|
| 77 |
+
"num_negatives": 100,
|
| 78 |
+
"output_hidden_size": 1024,
|
| 79 |
+
"pad_token_id": 28,
|
| 80 |
+
"proj_codevector_dim": 768,
|
| 81 |
+
"tdnn_dilation": [
|
| 82 |
+
1,
|
| 83 |
+
2,
|
| 84 |
+
3,
|
| 85 |
+
1,
|
| 86 |
+
1
|
| 87 |
+
],
|
| 88 |
+
"tdnn_dim": [
|
| 89 |
+
512,
|
| 90 |
+
512,
|
| 91 |
+
512,
|
| 92 |
+
512,
|
| 93 |
+
1500
|
| 94 |
+
],
|
| 95 |
+
"tdnn_kernel": [
|
| 96 |
+
5,
|
| 97 |
+
3,
|
| 98 |
+
3,
|
| 99 |
+
1,
|
| 100 |
+
1
|
| 101 |
+
],
|
| 102 |
+
"torch_dtype": "float32",
|
| 103 |
+
"transformers_version": "4.24.0",
|
| 104 |
+
"use_weighted_layer_sum": false,
|
| 105 |
+
"vocab_size": 31,
|
| 106 |
+
"xvector_output_dim": 512
|
| 107 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 256.41,
|
| 3 |
+
"eval_loss": 0.09882301837205887,
|
| 4 |
+
"eval_runtime": 66.9757,
|
| 5 |
+
"eval_samples": 1857,
|
| 6 |
+
"eval_samples_per_second": 27.726,
|
| 7 |
+
"eval_steps_per_second": 1.165,
|
| 8 |
+
"eval_wer": 0.07359326539654408
|
| 9 |
+
}
|
language_model/attrs.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"alpha": 0.5, "beta": 1.5, "unk_score_offset": -10.0, "score_boundary": true}
|
language_model/unigrams.txt
ADDED
|
File without changes
|
language_model/uwb_atcc_4g.binary
ADDED
|
Binary file (971 kB). View file
|
|
|
log/train_log
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0,
|
| 7 |
+
"processor_class": "Wav2Vec2ProcessorWithLM",
|
| 8 |
+
"return_attention_mask": true,
|
| 9 |
+
"sampling_rate": 16000
|
| 10 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92f00a147a01cfbf9d997364e964c62adb8df2a7f134bc77451664920f2f8367
|
| 3 |
+
size 1262028973
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"eos_token": "</s>",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"unk_token": "[UNK]"
|
| 6 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"do_lower_case": false,
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"pad_token": "[PAD]",
|
| 6 |
+
"processor_class": "Wav2Vec2ProcessorWithLM",
|
| 7 |
+
"replace_word_delimiter_char": " ",
|
| 8 |
+
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
| 9 |
+
"unk_token": "[UNK]",
|
| 10 |
+
"word_delimiter_token": "|"
|
| 11 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 256.41,
|
| 3 |
+
"train_loss": 0.06207249406576157,
|
| 4 |
+
"train_runtime": 86281.9422,
|
| 5 |
+
"train_samples": 7480,
|
| 6 |
+
"train_samples_per_second": 22.253,
|
| 7 |
+
"train_steps_per_second": 0.232
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,625 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 256.4102564102564,
|
| 5 |
+
"global_step": 20000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 6.41,
|
| 12 |
+
"learning_rate": 0.0005,
|
| 13 |
+
"loss": 1.9105,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 6.41,
|
| 18 |
+
"eval_loss": 0.1622474640607834,
|
| 19 |
+
"eval_runtime": 72.1023,
|
| 20 |
+
"eval_samples_per_second": 25.755,
|
| 21 |
+
"eval_steps_per_second": 1.082,
|
| 22 |
+
"eval_wer": 0.15307930881701373,
|
| 23 |
+
"step": 500
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 12.82,
|
| 27 |
+
"learning_rate": 0.0004871794871794872,
|
| 28 |
+
"loss": 0.1119,
|
| 29 |
+
"step": 1000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 12.82,
|
| 33 |
+
"eval_loss": 0.09705421328544617,
|
| 34 |
+
"eval_runtime": 74.0065,
|
| 35 |
+
"eval_samples_per_second": 25.092,
|
| 36 |
+
"eval_steps_per_second": 1.054,
|
| 37 |
+
"eval_wer": 0.09361984935755427,
|
| 38 |
+
"step": 1000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 19.23,
|
| 42 |
+
"learning_rate": 0.00047435897435897434,
|
| 43 |
+
"loss": 0.0614,
|
| 44 |
+
"step": 1500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 19.23,
|
| 48 |
+
"eval_loss": 0.10023126006126404,
|
| 49 |
+
"eval_runtime": 74.336,
|
| 50 |
+
"eval_samples_per_second": 24.981,
|
| 51 |
+
"eval_steps_per_second": 1.049,
|
| 52 |
+
"eval_wer": 0.09831634913602126,
|
| 53 |
+
"step": 1500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 25.64,
|
| 57 |
+
"learning_rate": 0.0004615384615384616,
|
| 58 |
+
"loss": 0.044,
|
| 59 |
+
"step": 2000
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 25.64,
|
| 63 |
+
"eval_loss": 0.10109349340200424,
|
| 64 |
+
"eval_runtime": 68.958,
|
| 65 |
+
"eval_samples_per_second": 26.929,
|
| 66 |
+
"eval_steps_per_second": 1.131,
|
| 67 |
+
"eval_wer": 0.09286663712893221,
|
| 68 |
+
"step": 2000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 32.05,
|
| 72 |
+
"learning_rate": 0.0004487179487179487,
|
| 73 |
+
"loss": 0.0366,
|
| 74 |
+
"step": 2500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 32.05,
|
| 78 |
+
"eval_loss": 0.09319411963224411,
|
| 79 |
+
"eval_runtime": 67.9099,
|
| 80 |
+
"eval_samples_per_second": 27.345,
|
| 81 |
+
"eval_steps_per_second": 1.149,
|
| 82 |
+
"eval_wer": 0.08280903854674347,
|
| 83 |
+
"step": 2500
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 38.46,
|
| 87 |
+
"learning_rate": 0.0004358974358974359,
|
| 88 |
+
"loss": 0.0315,
|
| 89 |
+
"step": 3000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 38.46,
|
| 93 |
+
"eval_loss": 0.09261373430490494,
|
| 94 |
+
"eval_runtime": 68.2657,
|
| 95 |
+
"eval_samples_per_second": 27.203,
|
| 96 |
+
"eval_steps_per_second": 1.143,
|
| 97 |
+
"eval_wer": 0.08803721754541427,
|
| 98 |
+
"step": 3000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 44.87,
|
| 102 |
+
"learning_rate": 0.0004230769230769231,
|
| 103 |
+
"loss": 0.0297,
|
| 104 |
+
"step": 3500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 44.87,
|
| 108 |
+
"eval_loss": 0.09722220152616501,
|
| 109 |
+
"eval_runtime": 69.3764,
|
| 110 |
+
"eval_samples_per_second": 26.767,
|
| 111 |
+
"eval_steps_per_second": 1.124,
|
| 112 |
+
"eval_wer": 0.08821444395214886,
|
| 113 |
+
"step": 3500
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 51.28,
|
| 117 |
+
"learning_rate": 0.00041025641025641023,
|
| 118 |
+
"loss": 0.0216,
|
| 119 |
+
"step": 4000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 51.28,
|
| 123 |
+
"eval_loss": 0.09113188087940216,
|
| 124 |
+
"eval_runtime": 68.0225,
|
| 125 |
+
"eval_samples_per_second": 27.3,
|
| 126 |
+
"eval_steps_per_second": 1.147,
|
| 127 |
+
"eval_wer": 0.07735932653965441,
|
| 128 |
+
"step": 4000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 57.69,
|
| 132 |
+
"learning_rate": 0.0003974358974358974,
|
| 133 |
+
"loss": 0.0211,
|
| 134 |
+
"step": 4500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 57.69,
|
| 138 |
+
"eval_loss": 0.09818430244922638,
|
| 139 |
+
"eval_runtime": 70.5517,
|
| 140 |
+
"eval_samples_per_second": 26.321,
|
| 141 |
+
"eval_steps_per_second": 1.106,
|
| 142 |
+
"eval_wer": 0.08910057598582188,
|
| 143 |
+
"step": 4500
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 64.1,
|
| 147 |
+
"learning_rate": 0.00038461538461538467,
|
| 148 |
+
"loss": 0.0187,
|
| 149 |
+
"step": 5000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 64.1,
|
| 153 |
+
"eval_loss": 0.10086847096681595,
|
| 154 |
+
"eval_runtime": 68.8313,
|
| 155 |
+
"eval_samples_per_second": 26.979,
|
| 156 |
+
"eval_steps_per_second": 1.133,
|
| 157 |
+
"eval_wer": 0.08626495347806823,
|
| 158 |
+
"step": 5000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 70.51,
|
| 162 |
+
"learning_rate": 0.0003717948717948718,
|
| 163 |
+
"loss": 0.02,
|
| 164 |
+
"step": 5500
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 70.51,
|
| 168 |
+
"eval_loss": 0.09532515704631805,
|
| 169 |
+
"eval_runtime": 65.875,
|
| 170 |
+
"eval_samples_per_second": 28.19,
|
| 171 |
+
"eval_steps_per_second": 1.184,
|
| 172 |
+
"eval_wer": 0.08520159503766062,
|
| 173 |
+
"step": 5500
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 76.92,
|
| 177 |
+
"learning_rate": 0.000358974358974359,
|
| 178 |
+
"loss": 0.0163,
|
| 179 |
+
"step": 6000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 76.92,
|
| 183 |
+
"eval_loss": 0.1028498187661171,
|
| 184 |
+
"eval_runtime": 62.6651,
|
| 185 |
+
"eval_samples_per_second": 29.634,
|
| 186 |
+
"eval_steps_per_second": 1.245,
|
| 187 |
+
"eval_wer": 0.08041648205582631,
|
| 188 |
+
"step": 6000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 83.33,
|
| 192 |
+
"learning_rate": 0.00034615384615384613,
|
| 193 |
+
"loss": 0.0128,
|
| 194 |
+
"step": 6500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 83.33,
|
| 198 |
+
"eval_loss": 0.09295257925987244,
|
| 199 |
+
"eval_runtime": 68.7425,
|
| 200 |
+
"eval_samples_per_second": 27.014,
|
| 201 |
+
"eval_steps_per_second": 1.135,
|
| 202 |
+
"eval_wer": 0.08555604785112982,
|
| 203 |
+
"step": 6500
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 89.74,
|
| 207 |
+
"learning_rate": 0.0003333333333333333,
|
| 208 |
+
"loss": 0.0127,
|
| 209 |
+
"step": 7000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 89.74,
|
| 213 |
+
"eval_loss": 0.08917286247015,
|
| 214 |
+
"eval_runtime": 65.0895,
|
| 215 |
+
"eval_samples_per_second": 28.53,
|
| 216 |
+
"eval_steps_per_second": 1.198,
|
| 217 |
+
"eval_wer": 0.06756756756756757,
|
| 218 |
+
"step": 7000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 96.15,
|
| 222 |
+
"learning_rate": 0.00032051282051282057,
|
| 223 |
+
"loss": 0.0116,
|
| 224 |
+
"step": 7500
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 96.15,
|
| 228 |
+
"eval_loss": 0.08566667139530182,
|
| 229 |
+
"eval_runtime": 62.4843,
|
| 230 |
+
"eval_samples_per_second": 29.719,
|
| 231 |
+
"eval_steps_per_second": 1.248,
|
| 232 |
+
"eval_wer": 0.07527691626052282,
|
| 233 |
+
"step": 7500
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 102.56,
|
| 237 |
+
"learning_rate": 0.0003076923076923077,
|
| 238 |
+
"loss": 0.0139,
|
| 239 |
+
"step": 8000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 102.56,
|
| 243 |
+
"eval_loss": 0.10782884061336517,
|
| 244 |
+
"eval_runtime": 62.9655,
|
| 245 |
+
"eval_samples_per_second": 29.492,
|
| 246 |
+
"eval_steps_per_second": 1.239,
|
| 247 |
+
"eval_wer": 0.04811696942844484,
|
| 248 |
+
"step": 8000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 108.97,
|
| 252 |
+
"learning_rate": 0.0002948717948717949,
|
| 253 |
+
"loss": 0.0107,
|
| 254 |
+
"step": 8500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 108.97,
|
| 258 |
+
"eval_loss": 0.09546608477830887,
|
| 259 |
+
"eval_runtime": 64.1669,
|
| 260 |
+
"eval_samples_per_second": 28.94,
|
| 261 |
+
"eval_steps_per_second": 1.216,
|
| 262 |
+
"eval_wer": 0.06827647319450598,
|
| 263 |
+
"step": 8500
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 115.38,
|
| 267 |
+
"learning_rate": 0.0002820769230769231,
|
| 268 |
+
"loss": 0.0096,
|
| 269 |
+
"step": 9000
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 115.38,
|
| 273 |
+
"eval_loss": 0.08463115245103836,
|
| 274 |
+
"eval_runtime": 63.761,
|
| 275 |
+
"eval_samples_per_second": 29.124,
|
| 276 |
+
"eval_steps_per_second": 1.223,
|
| 277 |
+
"eval_wer": 0.06973859105006645,
|
| 278 |
+
"step": 9000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 121.79,
|
| 282 |
+
"learning_rate": 0.0002692564102564103,
|
| 283 |
+
"loss": 0.0089,
|
| 284 |
+
"step": 9500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 121.79,
|
| 288 |
+
"eval_loss": 0.08535169810056686,
|
| 289 |
+
"eval_runtime": 63.7326,
|
| 290 |
+
"eval_samples_per_second": 29.137,
|
| 291 |
+
"eval_steps_per_second": 1.224,
|
| 292 |
+
"eval_wer": 0.06752326096588392,
|
| 293 |
+
"step": 9500
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 128.21,
|
| 297 |
+
"learning_rate": 0.00025643589743589747,
|
| 298 |
+
"loss": 0.0084,
|
| 299 |
+
"step": 10000
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 128.21,
|
| 303 |
+
"eval_loss": 0.08750651776790619,
|
| 304 |
+
"eval_runtime": 61.61,
|
| 305 |
+
"eval_samples_per_second": 30.141,
|
| 306 |
+
"eval_steps_per_second": 1.266,
|
| 307 |
+
"eval_wer": 0.07793531236154187,
|
| 308 |
+
"step": 10000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 134.62,
|
| 312 |
+
"learning_rate": 0.00024361538461538463,
|
| 313 |
+
"loss": 0.0074,
|
| 314 |
+
"step": 10500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 134.62,
|
| 318 |
+
"eval_loss": 0.08403545618057251,
|
| 319 |
+
"eval_runtime": 60.7541,
|
| 320 |
+
"eval_samples_per_second": 30.566,
|
| 321 |
+
"eval_steps_per_second": 1.284,
|
| 322 |
+
"eval_wer": 0.0770048737261852,
|
| 323 |
+
"step": 10500
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 141.03,
|
| 327 |
+
"learning_rate": 0.00023082051282051282,
|
| 328 |
+
"loss": 0.0061,
|
| 329 |
+
"step": 11000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 141.03,
|
| 333 |
+
"eval_loss": 0.09034867584705353,
|
| 334 |
+
"eval_runtime": 63.3565,
|
| 335 |
+
"eval_samples_per_second": 29.31,
|
| 336 |
+
"eval_steps_per_second": 1.231,
|
| 337 |
+
"eval_wer": 0.07540983606557378,
|
| 338 |
+
"step": 11000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 147.44,
|
| 342 |
+
"learning_rate": 0.000218,
|
| 343 |
+
"loss": 0.0076,
|
| 344 |
+
"step": 11500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 147.44,
|
| 348 |
+
"eval_loss": 0.08722745627164841,
|
| 349 |
+
"eval_runtime": 63.711,
|
| 350 |
+
"eval_samples_per_second": 29.147,
|
| 351 |
+
"eval_steps_per_second": 1.224,
|
| 352 |
+
"eval_wer": 0.0769162605228179,
|
| 353 |
+
"step": 11500
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 153.85,
|
| 357 |
+
"learning_rate": 0.00020517948717948718,
|
| 358 |
+
"loss": 0.0069,
|
| 359 |
+
"step": 12000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 153.85,
|
| 363 |
+
"eval_loss": 0.08911967277526855,
|
| 364 |
+
"eval_runtime": 61.1855,
|
| 365 |
+
"eval_samples_per_second": 30.35,
|
| 366 |
+
"eval_steps_per_second": 1.275,
|
| 367 |
+
"eval_wer": 0.07722640673460346,
|
| 368 |
+
"step": 12000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 160.26,
|
| 372 |
+
"learning_rate": 0.00019235897435897437,
|
| 373 |
+
"loss": 0.0061,
|
| 374 |
+
"step": 12500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 160.26,
|
| 378 |
+
"eval_loss": 0.09709189832210541,
|
| 379 |
+
"eval_runtime": 63.0238,
|
| 380 |
+
"eval_samples_per_second": 29.465,
|
| 381 |
+
"eval_steps_per_second": 1.238,
|
| 382 |
+
"eval_wer": 0.07735932653965441,
|
| 383 |
+
"step": 12500
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 166.67,
|
| 387 |
+
"learning_rate": 0.00017953846153846153,
|
| 388 |
+
"loss": 0.0049,
|
| 389 |
+
"step": 13000
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 166.67,
|
| 393 |
+
"eval_loss": 0.09844444692134857,
|
| 394 |
+
"eval_runtime": 61.9554,
|
| 395 |
+
"eval_samples_per_second": 29.973,
|
| 396 |
+
"eval_steps_per_second": 1.259,
|
| 397 |
+
"eval_wer": 0.07261852015950376,
|
| 398 |
+
"step": 13000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 173.08,
|
| 402 |
+
"learning_rate": 0.00016671794871794872,
|
| 403 |
+
"loss": 0.0045,
|
| 404 |
+
"step": 13500
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 173.08,
|
| 408 |
+
"eval_loss": 0.09517823159694672,
|
| 409 |
+
"eval_runtime": 61.492,
|
| 410 |
+
"eval_samples_per_second": 30.199,
|
| 411 |
+
"eval_steps_per_second": 1.268,
|
| 412 |
+
"eval_wer": 0.07651750110766505,
|
| 413 |
+
"step": 13500
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 179.49,
|
| 417 |
+
"learning_rate": 0.0001538974358974359,
|
| 418 |
+
"loss": 0.0039,
|
| 419 |
+
"step": 14000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 179.49,
|
| 423 |
+
"eval_loss": 0.10154110193252563,
|
| 424 |
+
"eval_runtime": 61.18,
|
| 425 |
+
"eval_samples_per_second": 30.353,
|
| 426 |
+
"eval_steps_per_second": 1.275,
|
| 427 |
+
"eval_wer": 0.07620735489587949,
|
| 428 |
+
"step": 14000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 185.9,
|
| 432 |
+
"learning_rate": 0.00014107692307692307,
|
| 433 |
+
"loss": 0.0031,
|
| 434 |
+
"step": 14500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 185.9,
|
| 438 |
+
"eval_loss": 0.09374968707561493,
|
| 439 |
+
"eval_runtime": 61.1162,
|
| 440 |
+
"eval_samples_per_second": 30.385,
|
| 441 |
+
"eval_steps_per_second": 1.276,
|
| 442 |
+
"eval_wer": 0.07124501550731059,
|
| 443 |
+
"step": 14500
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 192.31,
|
| 447 |
+
"learning_rate": 0.00012825641025641026,
|
| 448 |
+
"loss": 0.0032,
|
| 449 |
+
"step": 15000
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 192.31,
|
| 453 |
+
"eval_loss": 0.09821684658527374,
|
| 454 |
+
"eval_runtime": 60.073,
|
| 455 |
+
"eval_samples_per_second": 30.912,
|
| 456 |
+
"eval_steps_per_second": 1.298,
|
| 457 |
+
"eval_wer": 0.06353566681435534,
|
| 458 |
+
"step": 15000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 198.72,
|
| 462 |
+
"learning_rate": 0.00011546153846153847,
|
| 463 |
+
"loss": 0.0028,
|
| 464 |
+
"step": 15500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 198.72,
|
| 468 |
+
"eval_loss": 0.098084457218647,
|
| 469 |
+
"eval_runtime": 61.9773,
|
| 470 |
+
"eval_samples_per_second": 29.963,
|
| 471 |
+
"eval_steps_per_second": 1.259,
|
| 472 |
+
"eval_wer": 0.07434647762516615,
|
| 473 |
+
"step": 15500
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 205.13,
|
| 477 |
+
"learning_rate": 0.00010266666666666668,
|
| 478 |
+
"loss": 0.0024,
|
| 479 |
+
"step": 16000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 205.13,
|
| 483 |
+
"eval_loss": 0.10191462188959122,
|
| 484 |
+
"eval_runtime": 65.8946,
|
| 485 |
+
"eval_samples_per_second": 28.181,
|
| 486 |
+
"eval_steps_per_second": 1.184,
|
| 487 |
+
"eval_wer": 0.07120070890562694,
|
| 488 |
+
"step": 16000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 211.54,
|
| 492 |
+
"learning_rate": 8.987179487179488e-05,
|
| 493 |
+
"loss": 0.0024,
|
| 494 |
+
"step": 16500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 211.54,
|
| 498 |
+
"eval_loss": 0.09566177427768707,
|
| 499 |
+
"eval_runtime": 64.1104,
|
| 500 |
+
"eval_samples_per_second": 28.966,
|
| 501 |
+
"eval_steps_per_second": 1.217,
|
| 502 |
+
"eval_wer": 0.07323881258307488,
|
| 503 |
+
"step": 16500
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 217.95,
|
| 507 |
+
"learning_rate": 7.705128205128205e-05,
|
| 508 |
+
"loss": 0.002,
|
| 509 |
+
"step": 17000
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 217.95,
|
| 513 |
+
"eval_loss": 0.09413377195596695,
|
| 514 |
+
"eval_runtime": 61.0383,
|
| 515 |
+
"eval_samples_per_second": 30.424,
|
| 516 |
+
"eval_steps_per_second": 1.278,
|
| 517 |
+
"eval_wer": 0.07315019937970757,
|
| 518 |
+
"step": 17000
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 224.36,
|
| 522 |
+
"learning_rate": 6.423076923076924e-05,
|
| 523 |
+
"loss": 0.0015,
|
| 524 |
+
"step": 17500
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 224.36,
|
| 528 |
+
"eval_loss": 0.10090558975934982,
|
| 529 |
+
"eval_runtime": 61.85,
|
| 530 |
+
"eval_samples_per_second": 30.024,
|
| 531 |
+
"eval_steps_per_second": 1.261,
|
| 532 |
+
"eval_wer": 0.07173238812583074,
|
| 533 |
+
"step": 17500
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 230.77,
|
| 537 |
+
"learning_rate": 5.1410256410256414e-05,
|
| 538 |
+
"loss": 0.0017,
|
| 539 |
+
"step": 18000
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 230.77,
|
| 543 |
+
"eval_loss": 0.09554142504930496,
|
| 544 |
+
"eval_runtime": 60.313,
|
| 545 |
+
"eval_samples_per_second": 30.789,
|
| 546 |
+
"eval_steps_per_second": 1.293,
|
| 547 |
+
"eval_wer": 0.07301727957465662,
|
| 548 |
+
"step": 18000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 237.18,
|
| 552 |
+
"learning_rate": 3.858974358974359e-05,
|
| 553 |
+
"loss": 0.0013,
|
| 554 |
+
"step": 18500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 237.18,
|
| 558 |
+
"eval_loss": 0.09889042377471924,
|
| 559 |
+
"eval_runtime": 67.9718,
|
| 560 |
+
"eval_samples_per_second": 27.32,
|
| 561 |
+
"eval_steps_per_second": 1.148,
|
| 562 |
+
"eval_wer": 0.07315019937970757,
|
| 563 |
+
"step": 18500
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 243.59,
|
| 567 |
+
"learning_rate": 2.576923076923077e-05,
|
| 568 |
+
"loss": 0.0013,
|
| 569 |
+
"step": 19000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 243.59,
|
| 573 |
+
"eval_loss": 0.09670563787221909,
|
| 574 |
+
"eval_runtime": 66.758,
|
| 575 |
+
"eval_samples_per_second": 27.817,
|
| 576 |
+
"eval_steps_per_second": 1.168,
|
| 577 |
+
"eval_wer": 0.07377049180327869,
|
| 578 |
+
"step": 19000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 250.0,
|
| 582 |
+
"learning_rate": 1.2948717948717948e-05,
|
| 583 |
+
"loss": 0.0011,
|
| 584 |
+
"step": 19500
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 250.0,
|
| 588 |
+
"eval_loss": 0.0980001762509346,
|
| 589 |
+
"eval_runtime": 66.2994,
|
| 590 |
+
"eval_samples_per_second": 28.009,
|
| 591 |
+
"eval_steps_per_second": 1.176,
|
| 592 |
+
"eval_wer": 0.07341603898980949,
|
| 593 |
+
"step": 19500
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 256.41,
|
| 597 |
+
"learning_rate": 1.282051282051282e-07,
|
| 598 |
+
"loss": 0.0008,
|
| 599 |
+
"step": 20000
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 256.41,
|
| 603 |
+
"eval_loss": 0.09882301837205887,
|
| 604 |
+
"eval_runtime": 65.0891,
|
| 605 |
+
"eval_samples_per_second": 28.53,
|
| 606 |
+
"eval_steps_per_second": 1.198,
|
| 607 |
+
"eval_wer": 0.07359326539654408,
|
| 608 |
+
"step": 20000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 256.41,
|
| 612 |
+
"step": 20000,
|
| 613 |
+
"total_flos": 2.406420737737408e+20,
|
| 614 |
+
"train_loss": 0.06207249406576157,
|
| 615 |
+
"train_runtime": 86281.9422,
|
| 616 |
+
"train_samples_per_second": 22.253,
|
| 617 |
+
"train_steps_per_second": 0.232
|
| 618 |
+
}
|
| 619 |
+
],
|
| 620 |
+
"max_steps": 20000,
|
| 621 |
+
"num_train_epochs": 257,
|
| 622 |
+
"total_flos": 2.406420737737408e+20,
|
| 623 |
+
"trial_name": null,
|
| 624 |
+
"trial_params": null
|
| 625 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b87b5a7cec3775ec06abf0b23d9291d9d82a3d6220cd662b2c0e96c8cdf1bebc
|
| 3 |
+
size 3707
|
vocab.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[PAD]": 28,
|
| 3 |
+
"[UNK]": 27,
|
| 4 |
+
"a": 1,
|
| 5 |
+
"b": 2,
|
| 6 |
+
"c": 3,
|
| 7 |
+
"d": 4,
|
| 8 |
+
"e": 5,
|
| 9 |
+
"f": 6,
|
| 10 |
+
"g": 7,
|
| 11 |
+
"h": 8,
|
| 12 |
+
"i": 9,
|
| 13 |
+
"j": 10,
|
| 14 |
+
"k": 11,
|
| 15 |
+
"l": 12,
|
| 16 |
+
"m": 13,
|
| 17 |
+
"n": 14,
|
| 18 |
+
"o": 15,
|
| 19 |
+
"p": 16,
|
| 20 |
+
"q": 17,
|
| 21 |
+
"r": 18,
|
| 22 |
+
"s": 19,
|
| 23 |
+
"t": 20,
|
| 24 |
+
"u": 21,
|
| 25 |
+
"v": 22,
|
| 26 |
+
"w": 23,
|
| 27 |
+
"x": 24,
|
| 28 |
+
"y": 25,
|
| 29 |
+
"z": 26,
|
| 30 |
+
"|": 0
|
| 31 |
+
}
|