Upload folder using huggingface_hub
Browse files- checkpoint-326/config.json +59 -0
- checkpoint-326/generation_config.json +118 -0
- checkpoint-326/model-00001-of-00002.safetensors +3 -0
- checkpoint-326/model-00002-of-00002.safetensors +3 -0
- checkpoint-326/model.safetensors.index.json +0 -0
- checkpoint-326/optimizer.pt +3 -0
- checkpoint-326/preprocessor_config.json +14 -0
- checkpoint-326/rng_state.pth +3 -0
- checkpoint-326/scheduler.pt +3 -0
- checkpoint-326/trainer_state.json +151 -0
- checkpoint-326/training_args.bin +3 -0
- checkpoint-652/config.json +59 -0
- checkpoint-652/generation_config.json +118 -0
- checkpoint-652/model-00001-of-00002.safetensors +3 -0
- checkpoint-652/model-00002-of-00002.safetensors +3 -0
- checkpoint-652/model.safetensors.index.json +0 -0
- checkpoint-652/optimizer.pt +3 -0
- checkpoint-652/preprocessor_config.json +14 -0
- checkpoint-652/rng_state.pth +3 -0
- checkpoint-652/scheduler.pt +3 -0
- checkpoint-652/trainer_state.json +281 -0
- checkpoint-652/training_args.bin +3 -0
- checkpoint-978/config.json +59 -0
- checkpoint-978/generation_config.json +118 -0
- checkpoint-978/model-00001-of-00002.safetensors +3 -0
- checkpoint-978/model-00002-of-00002.safetensors +3 -0
- checkpoint-978/model.safetensors.index.json +0 -0
- checkpoint-978/optimizer.pt +3 -0
- checkpoint-978/preprocessor_config.json +14 -0
- checkpoint-978/rng_state.pth +3 -0
- checkpoint-978/scheduler.pt +3 -0
- checkpoint-978/trainer_state.json +417 -0
- checkpoint-978/training_args.bin +3 -0
- config.json +59 -0
- generation_config.json +118 -0
- model-00001-of-00002.safetensors +3 -0
- model-00002-of-00002.safetensors +3 -0
- model.safetensors.index.json +0 -0
- preprocessor_config.json +14 -0
- training_args.bin +3 -0
checkpoint-326/config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-large-v2",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"WhisperForConditionalGeneration"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"bos_token_id": 50257,
|
| 11 |
+
"classifier_proj_size": 256,
|
| 12 |
+
"d_model": 1280,
|
| 13 |
+
"decoder_attention_heads": 20,
|
| 14 |
+
"decoder_ffn_dim": 5120,
|
| 15 |
+
"decoder_layerdrop": 0.0,
|
| 16 |
+
"decoder_layers": 32,
|
| 17 |
+
"decoder_start_token_id": 50258,
|
| 18 |
+
"dropout": 0.0,
|
| 19 |
+
"encoder_attention_heads": 20,
|
| 20 |
+
"encoder_ffn_dim": 5120,
|
| 21 |
+
"encoder_layerdrop": 0.0,
|
| 22 |
+
"encoder_layers": 32,
|
| 23 |
+
"eos_token_id": 50257,
|
| 24 |
+
"forced_decoder_ids": [
|
| 25 |
+
[
|
| 26 |
+
1,
|
| 27 |
+
50259
|
| 28 |
+
],
|
| 29 |
+
[
|
| 30 |
+
2,
|
| 31 |
+
50359
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
3,
|
| 35 |
+
50363
|
| 36 |
+
]
|
| 37 |
+
],
|
| 38 |
+
"init_std": 0.02,
|
| 39 |
+
"is_encoder_decoder": true,
|
| 40 |
+
"mask_feature_length": 10,
|
| 41 |
+
"mask_feature_min_masks": 0,
|
| 42 |
+
"mask_feature_prob": 0.0,
|
| 43 |
+
"mask_time_length": 10,
|
| 44 |
+
"mask_time_min_masks": 2,
|
| 45 |
+
"mask_time_prob": 0.05,
|
| 46 |
+
"max_source_positions": 1500,
|
| 47 |
+
"max_target_positions": 448,
|
| 48 |
+
"median_filter_width": 7,
|
| 49 |
+
"model_type": "whisper",
|
| 50 |
+
"num_hidden_layers": 32,
|
| 51 |
+
"num_mel_bins": 80,
|
| 52 |
+
"pad_token_id": 50257,
|
| 53 |
+
"scale_embedding": false,
|
| 54 |
+
"torch_dtype": "float32",
|
| 55 |
+
"transformers_version": "4.37.2",
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"use_weighted_layer_sum": false,
|
| 58 |
+
"vocab_size": 51865
|
| 59 |
+
}
|
checkpoint-326/generation_config.json
ADDED
|
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"begin_suppress_tokens": [
|
| 3 |
+
220,
|
| 4 |
+
50257
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 50257,
|
| 7 |
+
"decoder_start_token_id": 50258,
|
| 8 |
+
"eos_token_id": 50257,
|
| 9 |
+
"forced_decoder_ids": [
|
| 10 |
+
[
|
| 11 |
+
1,
|
| 12 |
+
50259
|
| 13 |
+
],
|
| 14 |
+
[
|
| 15 |
+
2,
|
| 16 |
+
50359
|
| 17 |
+
],
|
| 18 |
+
[
|
| 19 |
+
3,
|
| 20 |
+
50363
|
| 21 |
+
]
|
| 22 |
+
],
|
| 23 |
+
"language": "english",
|
| 24 |
+
"max_length": 448,
|
| 25 |
+
"pad_token_id": 50257,
|
| 26 |
+
"suppress_tokens": [
|
| 27 |
+
1,
|
| 28 |
+
2,
|
| 29 |
+
7,
|
| 30 |
+
8,
|
| 31 |
+
9,
|
| 32 |
+
10,
|
| 33 |
+
14,
|
| 34 |
+
25,
|
| 35 |
+
26,
|
| 36 |
+
27,
|
| 37 |
+
28,
|
| 38 |
+
29,
|
| 39 |
+
31,
|
| 40 |
+
58,
|
| 41 |
+
59,
|
| 42 |
+
60,
|
| 43 |
+
61,
|
| 44 |
+
62,
|
| 45 |
+
63,
|
| 46 |
+
90,
|
| 47 |
+
91,
|
| 48 |
+
92,
|
| 49 |
+
93,
|
| 50 |
+
359,
|
| 51 |
+
503,
|
| 52 |
+
522,
|
| 53 |
+
542,
|
| 54 |
+
873,
|
| 55 |
+
893,
|
| 56 |
+
902,
|
| 57 |
+
918,
|
| 58 |
+
922,
|
| 59 |
+
931,
|
| 60 |
+
1350,
|
| 61 |
+
1853,
|
| 62 |
+
1982,
|
| 63 |
+
2460,
|
| 64 |
+
2627,
|
| 65 |
+
3246,
|
| 66 |
+
3253,
|
| 67 |
+
3268,
|
| 68 |
+
3536,
|
| 69 |
+
3846,
|
| 70 |
+
3961,
|
| 71 |
+
4183,
|
| 72 |
+
4667,
|
| 73 |
+
6585,
|
| 74 |
+
6647,
|
| 75 |
+
7273,
|
| 76 |
+
9061,
|
| 77 |
+
9383,
|
| 78 |
+
10428,
|
| 79 |
+
10929,
|
| 80 |
+
11938,
|
| 81 |
+
12033,
|
| 82 |
+
12331,
|
| 83 |
+
12562,
|
| 84 |
+
13793,
|
| 85 |
+
14157,
|
| 86 |
+
14635,
|
| 87 |
+
15265,
|
| 88 |
+
15618,
|
| 89 |
+
16553,
|
| 90 |
+
16604,
|
| 91 |
+
18362,
|
| 92 |
+
18956,
|
| 93 |
+
20075,
|
| 94 |
+
21675,
|
| 95 |
+
22520,
|
| 96 |
+
26130,
|
| 97 |
+
26161,
|
| 98 |
+
26435,
|
| 99 |
+
28279,
|
| 100 |
+
29464,
|
| 101 |
+
31650,
|
| 102 |
+
32302,
|
| 103 |
+
32470,
|
| 104 |
+
36865,
|
| 105 |
+
42863,
|
| 106 |
+
47425,
|
| 107 |
+
49870,
|
| 108 |
+
50254,
|
| 109 |
+
50258,
|
| 110 |
+
50358,
|
| 111 |
+
50359,
|
| 112 |
+
50360,
|
| 113 |
+
50361,
|
| 114 |
+
50362
|
| 115 |
+
],
|
| 116 |
+
"task": "transcribe",
|
| 117 |
+
"transformers_version": "4.37.2"
|
| 118 |
+
}
|
checkpoint-326/model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2629d2f60a106fc6e439663604eb9b9823b40a3a4ba72f33b951d6c57aa8f2d
|
| 3 |
+
size 4992706480
|
checkpoint-326/model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a6b54d34721e3141bfc364b678f08268deb4d6f19339a6120b0adf797be3730
|
| 3 |
+
size 1180663192
|
checkpoint-326/model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-326/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb22df89a113663fc4207090a9fc8031862b5dc1841d2ec1e5a1202d5130c821
|
| 3 |
+
size 3095074288
|
checkpoint-326/preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
checkpoint-326/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ff4981b7a9823ed2cfb0228d1178fdb9ea2cc63654bc94c24657eed09eae12f
|
| 3 |
+
size 14244
|
checkpoint-326/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ae61b503d1a44593a3163bb6c73b68e71f3332de8bd2aaf1d58d0aafeb80f60
|
| 3 |
+
size 1064
|
checkpoint-326/trainer_state.json
ADDED
|
@@ -0,0 +1,151 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.21568627450980393,
|
| 3 |
+
"best_model_checkpoint": "models/ArtFair/Hank-Green-326-timecoded/checkpoint-326",
|
| 4 |
+
"epoch": 0.3326530612244898,
|
| 5 |
+
"eval_steps": 326,
|
| 6 |
+
"global_step": 326,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 1.1965811965811965e-07,
|
| 14 |
+
"loss": 3.8774,
|
| 15 |
+
"step": 16
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.03,
|
| 19 |
+
"learning_rate": 2.393162393162393e-07,
|
| 20 |
+
"loss": 3.7909,
|
| 21 |
+
"step": 32
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.05,
|
| 25 |
+
"learning_rate": 3.7606837606837604e-07,
|
| 26 |
+
"loss": 3.4601,
|
| 27 |
+
"step": 48
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.07,
|
| 31 |
+
"learning_rate": 5.128205128205127e-07,
|
| 32 |
+
"loss": 3.3684,
|
| 33 |
+
"step": 64
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.08,
|
| 37 |
+
"learning_rate": 6.495726495726495e-07,
|
| 38 |
+
"loss": 2.3532,
|
| 39 |
+
"step": 80
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.1,
|
| 43 |
+
"learning_rate": 7.863247863247862e-07,
|
| 44 |
+
"loss": 2.3592,
|
| 45 |
+
"step": 96
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.11,
|
| 49 |
+
"learning_rate": 9.230769230769231e-07,
|
| 50 |
+
"loss": 1.7341,
|
| 51 |
+
"step": 112
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.13,
|
| 55 |
+
"learning_rate": 9.99837673007631e-07,
|
| 56 |
+
"loss": 1.4398,
|
| 57 |
+
"step": 128
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.15,
|
| 61 |
+
"learning_rate": 9.982484597964214e-07,
|
| 62 |
+
"loss": 1.3764,
|
| 63 |
+
"step": 144
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.16,
|
| 67 |
+
"learning_rate": 9.949694232523914e-07,
|
| 68 |
+
"loss": 1.1601,
|
| 69 |
+
"step": 160
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.18,
|
| 73 |
+
"learning_rate": 9.900116843180192e-07,
|
| 74 |
+
"loss": 1.1587,
|
| 75 |
+
"step": 176
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.2,
|
| 79 |
+
"learning_rate": 9.833920573010362e-07,
|
| 80 |
+
"loss": 1.0174,
|
| 81 |
+
"step": 192
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.21,
|
| 85 |
+
"learning_rate": 9.75132992848238e-07,
|
| 86 |
+
"loss": 0.725,
|
| 87 |
+
"step": 208
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.23,
|
| 91 |
+
"learning_rate": 9.652625018035032e-07,
|
| 92 |
+
"loss": 0.5014,
|
| 93 |
+
"step": 224
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.24,
|
| 97 |
+
"learning_rate": 9.538140602082435e-07,
|
| 98 |
+
"loss": 0.55,
|
| 99 |
+
"step": 240
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.26,
|
| 103 |
+
"learning_rate": 9.408264957664926e-07,
|
| 104 |
+
"loss": 0.4288,
|
| 105 |
+
"step": 256
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.28,
|
| 109 |
+
"learning_rate": 9.263438561596806e-07,
|
| 110 |
+
"loss": 0.467,
|
| 111 |
+
"step": 272
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.29,
|
| 115 |
+
"learning_rate": 9.104152596577132e-07,
|
| 116 |
+
"loss": 0.4463,
|
| 117 |
+
"step": 288
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.31,
|
| 121 |
+
"learning_rate": 8.930947285330108e-07,
|
| 122 |
+
"loss": 0.3361,
|
| 123 |
+
"step": 304
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.33,
|
| 127 |
+
"learning_rate": 8.744410058424878e-07,
|
| 128 |
+
"loss": 0.3554,
|
| 129 |
+
"step": 320
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.33,
|
| 133 |
+
"eval_loss": 0.7186310291290283,
|
| 134 |
+
"eval_raw_wer": 0.2878929349517585,
|
| 135 |
+
"eval_runtime": 477.7988,
|
| 136 |
+
"eval_samples_per_second": 0.82,
|
| 137 |
+
"eval_steps_per_second": 0.82,
|
| 138 |
+
"eval_wer": 0.21568627450980393,
|
| 139 |
+
"step": 326
|
| 140 |
+
}
|
| 141 |
+
],
|
| 142 |
+
"logging_steps": 16,
|
| 143 |
+
"max_steps": 980,
|
| 144 |
+
"num_input_tokens_seen": 0,
|
| 145 |
+
"num_train_epochs": 1,
|
| 146 |
+
"save_steps": 326,
|
| 147 |
+
"total_flos": 2.7686150111232e+18,
|
| 148 |
+
"train_batch_size": 4,
|
| 149 |
+
"trial_name": null,
|
| 150 |
+
"trial_params": null
|
| 151 |
+
}
|
checkpoint-326/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:683240c6785fb1eee21b2d156176395090b713ecd552f2112825d03b27aac81a
|
| 3 |
+
size 4984
|
checkpoint-652/config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-large-v2",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"WhisperForConditionalGeneration"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"bos_token_id": 50257,
|
| 11 |
+
"classifier_proj_size": 256,
|
| 12 |
+
"d_model": 1280,
|
| 13 |
+
"decoder_attention_heads": 20,
|
| 14 |
+
"decoder_ffn_dim": 5120,
|
| 15 |
+
"decoder_layerdrop": 0.0,
|
| 16 |
+
"decoder_layers": 32,
|
| 17 |
+
"decoder_start_token_id": 50258,
|
| 18 |
+
"dropout": 0.0,
|
| 19 |
+
"encoder_attention_heads": 20,
|
| 20 |
+
"encoder_ffn_dim": 5120,
|
| 21 |
+
"encoder_layerdrop": 0.0,
|
| 22 |
+
"encoder_layers": 32,
|
| 23 |
+
"eos_token_id": 50257,
|
| 24 |
+
"forced_decoder_ids": [
|
| 25 |
+
[
|
| 26 |
+
1,
|
| 27 |
+
50259
|
| 28 |
+
],
|
| 29 |
+
[
|
| 30 |
+
2,
|
| 31 |
+
50359
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
3,
|
| 35 |
+
50363
|
| 36 |
+
]
|
| 37 |
+
],
|
| 38 |
+
"init_std": 0.02,
|
| 39 |
+
"is_encoder_decoder": true,
|
| 40 |
+
"mask_feature_length": 10,
|
| 41 |
+
"mask_feature_min_masks": 0,
|
| 42 |
+
"mask_feature_prob": 0.0,
|
| 43 |
+
"mask_time_length": 10,
|
| 44 |
+
"mask_time_min_masks": 2,
|
| 45 |
+
"mask_time_prob": 0.05,
|
| 46 |
+
"max_source_positions": 1500,
|
| 47 |
+
"max_target_positions": 448,
|
| 48 |
+
"median_filter_width": 7,
|
| 49 |
+
"model_type": "whisper",
|
| 50 |
+
"num_hidden_layers": 32,
|
| 51 |
+
"num_mel_bins": 80,
|
| 52 |
+
"pad_token_id": 50257,
|
| 53 |
+
"scale_embedding": false,
|
| 54 |
+
"torch_dtype": "float32",
|
| 55 |
+
"transformers_version": "4.37.2",
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"use_weighted_layer_sum": false,
|
| 58 |
+
"vocab_size": 51865
|
| 59 |
+
}
|
checkpoint-652/generation_config.json
ADDED
|
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"begin_suppress_tokens": [
|
| 3 |
+
220,
|
| 4 |
+
50257
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 50257,
|
| 7 |
+
"decoder_start_token_id": 50258,
|
| 8 |
+
"eos_token_id": 50257,
|
| 9 |
+
"forced_decoder_ids": [
|
| 10 |
+
[
|
| 11 |
+
1,
|
| 12 |
+
50259
|
| 13 |
+
],
|
| 14 |
+
[
|
| 15 |
+
2,
|
| 16 |
+
50359
|
| 17 |
+
],
|
| 18 |
+
[
|
| 19 |
+
3,
|
| 20 |
+
50363
|
| 21 |
+
]
|
| 22 |
+
],
|
| 23 |
+
"language": "english",
|
| 24 |
+
"max_length": 448,
|
| 25 |
+
"pad_token_id": 50257,
|
| 26 |
+
"suppress_tokens": [
|
| 27 |
+
1,
|
| 28 |
+
2,
|
| 29 |
+
7,
|
| 30 |
+
8,
|
| 31 |
+
9,
|
| 32 |
+
10,
|
| 33 |
+
14,
|
| 34 |
+
25,
|
| 35 |
+
26,
|
| 36 |
+
27,
|
| 37 |
+
28,
|
| 38 |
+
29,
|
| 39 |
+
31,
|
| 40 |
+
58,
|
| 41 |
+
59,
|
| 42 |
+
60,
|
| 43 |
+
61,
|
| 44 |
+
62,
|
| 45 |
+
63,
|
| 46 |
+
90,
|
| 47 |
+
91,
|
| 48 |
+
92,
|
| 49 |
+
93,
|
| 50 |
+
359,
|
| 51 |
+
503,
|
| 52 |
+
522,
|
| 53 |
+
542,
|
| 54 |
+
873,
|
| 55 |
+
893,
|
| 56 |
+
902,
|
| 57 |
+
918,
|
| 58 |
+
922,
|
| 59 |
+
931,
|
| 60 |
+
1350,
|
| 61 |
+
1853,
|
| 62 |
+
1982,
|
| 63 |
+
2460,
|
| 64 |
+
2627,
|
| 65 |
+
3246,
|
| 66 |
+
3253,
|
| 67 |
+
3268,
|
| 68 |
+
3536,
|
| 69 |
+
3846,
|
| 70 |
+
3961,
|
| 71 |
+
4183,
|
| 72 |
+
4667,
|
| 73 |
+
6585,
|
| 74 |
+
6647,
|
| 75 |
+
7273,
|
| 76 |
+
9061,
|
| 77 |
+
9383,
|
| 78 |
+
10428,
|
| 79 |
+
10929,
|
| 80 |
+
11938,
|
| 81 |
+
12033,
|
| 82 |
+
12331,
|
| 83 |
+
12562,
|
| 84 |
+
13793,
|
| 85 |
+
14157,
|
| 86 |
+
14635,
|
| 87 |
+
15265,
|
| 88 |
+
15618,
|
| 89 |
+
16553,
|
| 90 |
+
16604,
|
| 91 |
+
18362,
|
| 92 |
+
18956,
|
| 93 |
+
20075,
|
| 94 |
+
21675,
|
| 95 |
+
22520,
|
| 96 |
+
26130,
|
| 97 |
+
26161,
|
| 98 |
+
26435,
|
| 99 |
+
28279,
|
| 100 |
+
29464,
|
| 101 |
+
31650,
|
| 102 |
+
32302,
|
| 103 |
+
32470,
|
| 104 |
+
36865,
|
| 105 |
+
42863,
|
| 106 |
+
47425,
|
| 107 |
+
49870,
|
| 108 |
+
50254,
|
| 109 |
+
50258,
|
| 110 |
+
50358,
|
| 111 |
+
50359,
|
| 112 |
+
50360,
|
| 113 |
+
50361,
|
| 114 |
+
50362
|
| 115 |
+
],
|
| 116 |
+
"task": "transcribe",
|
| 117 |
+
"transformers_version": "4.37.2"
|
| 118 |
+
}
|
checkpoint-652/model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c51c2b1f15816c5b56eeb5aace31923bf2acaa35e031a0f53d1c0e49634f5145
|
| 3 |
+
size 4992706480
|
checkpoint-652/model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:770c17df5051d402bbea62dd0647aa3f922d631327c7013c28f2cc54e5904424
|
| 3 |
+
size 1180663192
|
checkpoint-652/model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-652/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25c0ef18745f8823d0eb7be0960480cf5ccac37de05da4bfc31543f9e4637927
|
| 3 |
+
size 3095074288
|
checkpoint-652/preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
checkpoint-652/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6e252adc41fce4a0a5ceca1c272a400b22b467c438820efe4781bb968f93a787
|
| 3 |
+
size 14244
|
checkpoint-652/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7529b3c880fa0076fcbb749b32b04523426d2180c08f4a4878ae269b08a8ffa2
|
| 3 |
+
size 1064
|
checkpoint-652/trainer_state.json
ADDED
|
@@ -0,0 +1,281 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.20385932150638034,
|
| 3 |
+
"best_model_checkpoint": "models/ArtFair/Hank-Green-326-timecoded/checkpoint-652",
|
| 4 |
+
"epoch": 0.6653061224489796,
|
| 5 |
+
"eval_steps": 326,
|
| 6 |
+
"global_step": 652,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 1.1965811965811965e-07,
|
| 14 |
+
"loss": 3.8774,
|
| 15 |
+
"step": 16
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.03,
|
| 19 |
+
"learning_rate": 2.393162393162393e-07,
|
| 20 |
+
"loss": 3.7909,
|
| 21 |
+
"step": 32
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.05,
|
| 25 |
+
"learning_rate": 3.7606837606837604e-07,
|
| 26 |
+
"loss": 3.4601,
|
| 27 |
+
"step": 48
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.07,
|
| 31 |
+
"learning_rate": 5.128205128205127e-07,
|
| 32 |
+
"loss": 3.3684,
|
| 33 |
+
"step": 64
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.08,
|
| 37 |
+
"learning_rate": 6.495726495726495e-07,
|
| 38 |
+
"loss": 2.3532,
|
| 39 |
+
"step": 80
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.1,
|
| 43 |
+
"learning_rate": 7.863247863247862e-07,
|
| 44 |
+
"loss": 2.3592,
|
| 45 |
+
"step": 96
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.11,
|
| 49 |
+
"learning_rate": 9.230769230769231e-07,
|
| 50 |
+
"loss": 1.7341,
|
| 51 |
+
"step": 112
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.13,
|
| 55 |
+
"learning_rate": 9.99837673007631e-07,
|
| 56 |
+
"loss": 1.4398,
|
| 57 |
+
"step": 128
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.15,
|
| 61 |
+
"learning_rate": 9.982484597964214e-07,
|
| 62 |
+
"loss": 1.3764,
|
| 63 |
+
"step": 144
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.16,
|
| 67 |
+
"learning_rate": 9.949694232523914e-07,
|
| 68 |
+
"loss": 1.1601,
|
| 69 |
+
"step": 160
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.18,
|
| 73 |
+
"learning_rate": 9.900116843180192e-07,
|
| 74 |
+
"loss": 1.1587,
|
| 75 |
+
"step": 176
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.2,
|
| 79 |
+
"learning_rate": 9.833920573010362e-07,
|
| 80 |
+
"loss": 1.0174,
|
| 81 |
+
"step": 192
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.21,
|
| 85 |
+
"learning_rate": 9.75132992848238e-07,
|
| 86 |
+
"loss": 0.725,
|
| 87 |
+
"step": 208
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.23,
|
| 91 |
+
"learning_rate": 9.652625018035032e-07,
|
| 92 |
+
"loss": 0.5014,
|
| 93 |
+
"step": 224
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.24,
|
| 97 |
+
"learning_rate": 9.538140602082435e-07,
|
| 98 |
+
"loss": 0.55,
|
| 99 |
+
"step": 240
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.26,
|
| 103 |
+
"learning_rate": 9.408264957664926e-07,
|
| 104 |
+
"loss": 0.4288,
|
| 105 |
+
"step": 256
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.28,
|
| 109 |
+
"learning_rate": 9.263438561596806e-07,
|
| 110 |
+
"loss": 0.467,
|
| 111 |
+
"step": 272
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.29,
|
| 115 |
+
"learning_rate": 9.104152596577132e-07,
|
| 116 |
+
"loss": 0.4463,
|
| 117 |
+
"step": 288
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.31,
|
| 121 |
+
"learning_rate": 8.930947285330108e-07,
|
| 122 |
+
"loss": 0.3361,
|
| 123 |
+
"step": 304
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.33,
|
| 127 |
+
"learning_rate": 8.744410058424878e-07,
|
| 128 |
+
"loss": 0.3554,
|
| 129 |
+
"step": 320
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.33,
|
| 133 |
+
"eval_loss": 0.7186310291290283,
|
| 134 |
+
"eval_raw_wer": 0.2878929349517585,
|
| 135 |
+
"eval_runtime": 477.7988,
|
| 136 |
+
"eval_samples_per_second": 0.82,
|
| 137 |
+
"eval_steps_per_second": 0.82,
|
| 138 |
+
"eval_wer": 0.21568627450980393,
|
| 139 |
+
"step": 326
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 0.34,
|
| 143 |
+
"learning_rate": 8.545173561988624e-07,
|
| 144 |
+
"loss": 0.3713,
|
| 145 |
+
"step": 336
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.36,
|
| 149 |
+
"learning_rate": 8.333913512069848e-07,
|
| 150 |
+
"loss": 0.4327,
|
| 151 |
+
"step": 352
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 0.38,
|
| 155 |
+
"learning_rate": 8.111346402928848e-07,
|
| 156 |
+
"loss": 0.4631,
|
| 157 |
+
"step": 368
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.39,
|
| 161 |
+
"learning_rate": 7.878227077027753e-07,
|
| 162 |
+
"loss": 0.3439,
|
| 163 |
+
"step": 384
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.41,
|
| 167 |
+
"learning_rate": 7.635346164961587e-07,
|
| 168 |
+
"loss": 0.4046,
|
| 169 |
+
"step": 400
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.42,
|
| 173 |
+
"learning_rate": 7.383527404012899e-07,
|
| 174 |
+
"loss": 0.2902,
|
| 175 |
+
"step": 416
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.44,
|
| 179 |
+
"learning_rate": 7.123624844424145e-07,
|
| 180 |
+
"loss": 0.4497,
|
| 181 |
+
"step": 432
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 0.46,
|
| 185 |
+
"learning_rate": 6.856519952862844e-07,
|
| 186 |
+
"loss": 0.2758,
|
| 187 |
+
"step": 448
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 0.47,
|
| 191 |
+
"learning_rate": 6.583118622903166e-07,
|
| 192 |
+
"loss": 0.4701,
|
| 193 |
+
"step": 464
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.49,
|
| 197 |
+
"learning_rate": 6.304348102663004e-07,
|
| 198 |
+
"loss": 0.3829,
|
| 199 |
+
"step": 480
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.51,
|
| 203 |
+
"learning_rate": 6.021153850016527e-07,
|
| 204 |
+
"loss": 0.3568,
|
| 205 |
+
"step": 496
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.52,
|
| 209 |
+
"learning_rate": 5.734496326047821e-07,
|
| 210 |
+
"loss": 0.4111,
|
| 211 |
+
"step": 512
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 0.54,
|
| 215 |
+
"learning_rate": 5.445347737620766e-07,
|
| 216 |
+
"loss": 0.5163,
|
| 217 |
+
"step": 528
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.56,
|
| 221 |
+
"learning_rate": 5.154688740112749e-07,
|
| 222 |
+
"loss": 0.502,
|
| 223 |
+
"step": 544
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 0.57,
|
| 227 |
+
"learning_rate": 4.86350511149504e-07,
|
| 228 |
+
"loss": 0.325,
|
| 229 |
+
"step": 560
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.59,
|
| 233 |
+
"learning_rate": 4.5727844090397133e-07,
|
| 234 |
+
"loss": 0.3893,
|
| 235 |
+
"step": 576
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.6,
|
| 239 |
+
"learning_rate": 4.2835126199920545e-07,
|
| 240 |
+
"loss": 0.4843,
|
| 241 |
+
"step": 592
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.62,
|
| 245 |
+
"learning_rate": 3.996670817567741e-07,
|
| 246 |
+
"loss": 0.4143,
|
| 247 |
+
"step": 608
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.64,
|
| 251 |
+
"learning_rate": 3.713231833616096e-07,
|
| 252 |
+
"loss": 0.3176,
|
| 253 |
+
"step": 624
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 0.65,
|
| 257 |
+
"learning_rate": 3.434156959234152e-07,
|
| 258 |
+
"loss": 0.4304,
|
| 259 |
+
"step": 640
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 0.67,
|
| 263 |
+
"eval_loss": 0.6797897219657898,
|
| 264 |
+
"eval_raw_wer": 0.2773109243697479,
|
| 265 |
+
"eval_runtime": 485.642,
|
| 266 |
+
"eval_samples_per_second": 0.807,
|
| 267 |
+
"eval_steps_per_second": 0.807,
|
| 268 |
+
"eval_wer": 0.20385932150638034,
|
| 269 |
+
"step": 652
|
| 270 |
+
}
|
| 271 |
+
],
|
| 272 |
+
"logging_steps": 16,
|
| 273 |
+
"max_steps": 980,
|
| 274 |
+
"num_input_tokens_seen": 0,
|
| 275 |
+
"num_train_epochs": 1,
|
| 276 |
+
"save_steps": 326,
|
| 277 |
+
"total_flos": 5.5372300222464e+18,
|
| 278 |
+
"train_batch_size": 4,
|
| 279 |
+
"trial_name": null,
|
| 280 |
+
"trial_params": null
|
| 281 |
+
}
|
checkpoint-652/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:683240c6785fb1eee21b2d156176395090b713ecd552f2112825d03b27aac81a
|
| 3 |
+
size 4984
|
checkpoint-978/config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-large-v2",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"WhisperForConditionalGeneration"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"bos_token_id": 50257,
|
| 11 |
+
"classifier_proj_size": 256,
|
| 12 |
+
"d_model": 1280,
|
| 13 |
+
"decoder_attention_heads": 20,
|
| 14 |
+
"decoder_ffn_dim": 5120,
|
| 15 |
+
"decoder_layerdrop": 0.0,
|
| 16 |
+
"decoder_layers": 32,
|
| 17 |
+
"decoder_start_token_id": 50258,
|
| 18 |
+
"dropout": 0.0,
|
| 19 |
+
"encoder_attention_heads": 20,
|
| 20 |
+
"encoder_ffn_dim": 5120,
|
| 21 |
+
"encoder_layerdrop": 0.0,
|
| 22 |
+
"encoder_layers": 32,
|
| 23 |
+
"eos_token_id": 50257,
|
| 24 |
+
"forced_decoder_ids": [
|
| 25 |
+
[
|
| 26 |
+
1,
|
| 27 |
+
50259
|
| 28 |
+
],
|
| 29 |
+
[
|
| 30 |
+
2,
|
| 31 |
+
50359
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
3,
|
| 35 |
+
50363
|
| 36 |
+
]
|
| 37 |
+
],
|
| 38 |
+
"init_std": 0.02,
|
| 39 |
+
"is_encoder_decoder": true,
|
| 40 |
+
"mask_feature_length": 10,
|
| 41 |
+
"mask_feature_min_masks": 0,
|
| 42 |
+
"mask_feature_prob": 0.0,
|
| 43 |
+
"mask_time_length": 10,
|
| 44 |
+
"mask_time_min_masks": 2,
|
| 45 |
+
"mask_time_prob": 0.05,
|
| 46 |
+
"max_source_positions": 1500,
|
| 47 |
+
"max_target_positions": 448,
|
| 48 |
+
"median_filter_width": 7,
|
| 49 |
+
"model_type": "whisper",
|
| 50 |
+
"num_hidden_layers": 32,
|
| 51 |
+
"num_mel_bins": 80,
|
| 52 |
+
"pad_token_id": 50257,
|
| 53 |
+
"scale_embedding": false,
|
| 54 |
+
"torch_dtype": "float32",
|
| 55 |
+
"transformers_version": "4.37.2",
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"use_weighted_layer_sum": false,
|
| 58 |
+
"vocab_size": 51865
|
| 59 |
+
}
|
checkpoint-978/generation_config.json
ADDED
|
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"begin_suppress_tokens": [
|
| 3 |
+
220,
|
| 4 |
+
50257
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 50257,
|
| 7 |
+
"decoder_start_token_id": 50258,
|
| 8 |
+
"eos_token_id": 50257,
|
| 9 |
+
"forced_decoder_ids": [
|
| 10 |
+
[
|
| 11 |
+
1,
|
| 12 |
+
50259
|
| 13 |
+
],
|
| 14 |
+
[
|
| 15 |
+
2,
|
| 16 |
+
50359
|
| 17 |
+
],
|
| 18 |
+
[
|
| 19 |
+
3,
|
| 20 |
+
50363
|
| 21 |
+
]
|
| 22 |
+
],
|
| 23 |
+
"language": "english",
|
| 24 |
+
"max_length": 448,
|
| 25 |
+
"pad_token_id": 50257,
|
| 26 |
+
"suppress_tokens": [
|
| 27 |
+
1,
|
| 28 |
+
2,
|
| 29 |
+
7,
|
| 30 |
+
8,
|
| 31 |
+
9,
|
| 32 |
+
10,
|
| 33 |
+
14,
|
| 34 |
+
25,
|
| 35 |
+
26,
|
| 36 |
+
27,
|
| 37 |
+
28,
|
| 38 |
+
29,
|
| 39 |
+
31,
|
| 40 |
+
58,
|
| 41 |
+
59,
|
| 42 |
+
60,
|
| 43 |
+
61,
|
| 44 |
+
62,
|
| 45 |
+
63,
|
| 46 |
+
90,
|
| 47 |
+
91,
|
| 48 |
+
92,
|
| 49 |
+
93,
|
| 50 |
+
359,
|
| 51 |
+
503,
|
| 52 |
+
522,
|
| 53 |
+
542,
|
| 54 |
+
873,
|
| 55 |
+
893,
|
| 56 |
+
902,
|
| 57 |
+
918,
|
| 58 |
+
922,
|
| 59 |
+
931,
|
| 60 |
+
1350,
|
| 61 |
+
1853,
|
| 62 |
+
1982,
|
| 63 |
+
2460,
|
| 64 |
+
2627,
|
| 65 |
+
3246,
|
| 66 |
+
3253,
|
| 67 |
+
3268,
|
| 68 |
+
3536,
|
| 69 |
+
3846,
|
| 70 |
+
3961,
|
| 71 |
+
4183,
|
| 72 |
+
4667,
|
| 73 |
+
6585,
|
| 74 |
+
6647,
|
| 75 |
+
7273,
|
| 76 |
+
9061,
|
| 77 |
+
9383,
|
| 78 |
+
10428,
|
| 79 |
+
10929,
|
| 80 |
+
11938,
|
| 81 |
+
12033,
|
| 82 |
+
12331,
|
| 83 |
+
12562,
|
| 84 |
+
13793,
|
| 85 |
+
14157,
|
| 86 |
+
14635,
|
| 87 |
+
15265,
|
| 88 |
+
15618,
|
| 89 |
+
16553,
|
| 90 |
+
16604,
|
| 91 |
+
18362,
|
| 92 |
+
18956,
|
| 93 |
+
20075,
|
| 94 |
+
21675,
|
| 95 |
+
22520,
|
| 96 |
+
26130,
|
| 97 |
+
26161,
|
| 98 |
+
26435,
|
| 99 |
+
28279,
|
| 100 |
+
29464,
|
| 101 |
+
31650,
|
| 102 |
+
32302,
|
| 103 |
+
32470,
|
| 104 |
+
36865,
|
| 105 |
+
42863,
|
| 106 |
+
47425,
|
| 107 |
+
49870,
|
| 108 |
+
50254,
|
| 109 |
+
50258,
|
| 110 |
+
50358,
|
| 111 |
+
50359,
|
| 112 |
+
50360,
|
| 113 |
+
50361,
|
| 114 |
+
50362
|
| 115 |
+
],
|
| 116 |
+
"task": "transcribe",
|
| 117 |
+
"transformers_version": "4.37.2"
|
| 118 |
+
}
|
checkpoint-978/model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2eb4054a883d623038a477dad3a5bf67cafa64ef99f5b96d19b6ed319ad99c26
|
| 3 |
+
size 4992706480
|
checkpoint-978/model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e119a93c59b470e69c4ccdb5e4278405f8a36c1464f7352e919b6e425d35e56d
|
| 3 |
+
size 1180663192
|
checkpoint-978/model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-978/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:787584cbc3bbc62877d53a7ba7a3b6ec15709b5d2abefc197af5298aec4e17bc
|
| 3 |
+
size 3095074288
|
checkpoint-978/preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
checkpoint-978/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51e8b08cfa8fd5aa6f83287b7f008824a765c81db175ab746f43508379b5388b
|
| 3 |
+
size 14244
|
checkpoint-978/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0124c5ebdf26e1c8c695f5185fd13e81520935f4de77d6eff87ecd6dc9c8e99c
|
| 3 |
+
size 1064
|
checkpoint-978/trainer_state.json
ADDED
|
@@ -0,0 +1,417 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.20292561469032058,
|
| 3 |
+
"best_model_checkpoint": "models/ArtFair/Hank-Green-326-timecoded/checkpoint-978",
|
| 4 |
+
"epoch": 0.9979591836734694,
|
| 5 |
+
"eval_steps": 326,
|
| 6 |
+
"global_step": 978,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 1.1965811965811965e-07,
|
| 14 |
+
"loss": 3.8774,
|
| 15 |
+
"step": 16
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.03,
|
| 19 |
+
"learning_rate": 2.393162393162393e-07,
|
| 20 |
+
"loss": 3.7909,
|
| 21 |
+
"step": 32
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.05,
|
| 25 |
+
"learning_rate": 3.7606837606837604e-07,
|
| 26 |
+
"loss": 3.4601,
|
| 27 |
+
"step": 48
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.07,
|
| 31 |
+
"learning_rate": 5.128205128205127e-07,
|
| 32 |
+
"loss": 3.3684,
|
| 33 |
+
"step": 64
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.08,
|
| 37 |
+
"learning_rate": 6.495726495726495e-07,
|
| 38 |
+
"loss": 2.3532,
|
| 39 |
+
"step": 80
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.1,
|
| 43 |
+
"learning_rate": 7.863247863247862e-07,
|
| 44 |
+
"loss": 2.3592,
|
| 45 |
+
"step": 96
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.11,
|
| 49 |
+
"learning_rate": 9.230769230769231e-07,
|
| 50 |
+
"loss": 1.7341,
|
| 51 |
+
"step": 112
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.13,
|
| 55 |
+
"learning_rate": 9.99837673007631e-07,
|
| 56 |
+
"loss": 1.4398,
|
| 57 |
+
"step": 128
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.15,
|
| 61 |
+
"learning_rate": 9.982484597964214e-07,
|
| 62 |
+
"loss": 1.3764,
|
| 63 |
+
"step": 144
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.16,
|
| 67 |
+
"learning_rate": 9.949694232523914e-07,
|
| 68 |
+
"loss": 1.1601,
|
| 69 |
+
"step": 160
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.18,
|
| 73 |
+
"learning_rate": 9.900116843180192e-07,
|
| 74 |
+
"loss": 1.1587,
|
| 75 |
+
"step": 176
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.2,
|
| 79 |
+
"learning_rate": 9.833920573010362e-07,
|
| 80 |
+
"loss": 1.0174,
|
| 81 |
+
"step": 192
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.21,
|
| 85 |
+
"learning_rate": 9.75132992848238e-07,
|
| 86 |
+
"loss": 0.725,
|
| 87 |
+
"step": 208
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.23,
|
| 91 |
+
"learning_rate": 9.652625018035032e-07,
|
| 92 |
+
"loss": 0.5014,
|
| 93 |
+
"step": 224
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.24,
|
| 97 |
+
"learning_rate": 9.538140602082435e-07,
|
| 98 |
+
"loss": 0.55,
|
| 99 |
+
"step": 240
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.26,
|
| 103 |
+
"learning_rate": 9.408264957664926e-07,
|
| 104 |
+
"loss": 0.4288,
|
| 105 |
+
"step": 256
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.28,
|
| 109 |
+
"learning_rate": 9.263438561596806e-07,
|
| 110 |
+
"loss": 0.467,
|
| 111 |
+
"step": 272
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.29,
|
| 115 |
+
"learning_rate": 9.104152596577132e-07,
|
| 116 |
+
"loss": 0.4463,
|
| 117 |
+
"step": 288
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.31,
|
| 121 |
+
"learning_rate": 8.930947285330108e-07,
|
| 122 |
+
"loss": 0.3361,
|
| 123 |
+
"step": 304
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.33,
|
| 127 |
+
"learning_rate": 8.744410058424878e-07,
|
| 128 |
+
"loss": 0.3554,
|
| 129 |
+
"step": 320
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.33,
|
| 133 |
+
"eval_loss": 0.7186310291290283,
|
| 134 |
+
"eval_raw_wer": 0.2878929349517585,
|
| 135 |
+
"eval_runtime": 477.7988,
|
| 136 |
+
"eval_samples_per_second": 0.82,
|
| 137 |
+
"eval_steps_per_second": 0.82,
|
| 138 |
+
"eval_wer": 0.21568627450980393,
|
| 139 |
+
"step": 326
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 0.34,
|
| 143 |
+
"learning_rate": 8.545173561988624e-07,
|
| 144 |
+
"loss": 0.3713,
|
| 145 |
+
"step": 336
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.36,
|
| 149 |
+
"learning_rate": 8.333913512069848e-07,
|
| 150 |
+
"loss": 0.4327,
|
| 151 |
+
"step": 352
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 0.38,
|
| 155 |
+
"learning_rate": 8.111346402928848e-07,
|
| 156 |
+
"loss": 0.4631,
|
| 157 |
+
"step": 368
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.39,
|
| 161 |
+
"learning_rate": 7.878227077027753e-07,
|
| 162 |
+
"loss": 0.3439,
|
| 163 |
+
"step": 384
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.41,
|
| 167 |
+
"learning_rate": 7.635346164961587e-07,
|
| 168 |
+
"loss": 0.4046,
|
| 169 |
+
"step": 400
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.42,
|
| 173 |
+
"learning_rate": 7.383527404012899e-07,
|
| 174 |
+
"loss": 0.2902,
|
| 175 |
+
"step": 416
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.44,
|
| 179 |
+
"learning_rate": 7.123624844424145e-07,
|
| 180 |
+
"loss": 0.4497,
|
| 181 |
+
"step": 432
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 0.46,
|
| 185 |
+
"learning_rate": 6.856519952862844e-07,
|
| 186 |
+
"loss": 0.2758,
|
| 187 |
+
"step": 448
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 0.47,
|
| 191 |
+
"learning_rate": 6.583118622903166e-07,
|
| 192 |
+
"loss": 0.4701,
|
| 193 |
+
"step": 464
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.49,
|
| 197 |
+
"learning_rate": 6.304348102663004e-07,
|
| 198 |
+
"loss": 0.3829,
|
| 199 |
+
"step": 480
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.51,
|
| 203 |
+
"learning_rate": 6.021153850016527e-07,
|
| 204 |
+
"loss": 0.3568,
|
| 205 |
+
"step": 496
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.52,
|
| 209 |
+
"learning_rate": 5.734496326047821e-07,
|
| 210 |
+
"loss": 0.4111,
|
| 211 |
+
"step": 512
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 0.54,
|
| 215 |
+
"learning_rate": 5.445347737620766e-07,
|
| 216 |
+
"loss": 0.5163,
|
| 217 |
+
"step": 528
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.56,
|
| 221 |
+
"learning_rate": 5.154688740112749e-07,
|
| 222 |
+
"loss": 0.502,
|
| 223 |
+
"step": 544
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 0.57,
|
| 227 |
+
"learning_rate": 4.86350511149504e-07,
|
| 228 |
+
"loss": 0.325,
|
| 229 |
+
"step": 560
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.59,
|
| 233 |
+
"learning_rate": 4.5727844090397133e-07,
|
| 234 |
+
"loss": 0.3893,
|
| 235 |
+
"step": 576
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.6,
|
| 239 |
+
"learning_rate": 4.2835126199920545e-07,
|
| 240 |
+
"loss": 0.4843,
|
| 241 |
+
"step": 592
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.62,
|
| 245 |
+
"learning_rate": 3.996670817567741e-07,
|
| 246 |
+
"loss": 0.4143,
|
| 247 |
+
"step": 608
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.64,
|
| 251 |
+
"learning_rate": 3.713231833616096e-07,
|
| 252 |
+
"loss": 0.3176,
|
| 253 |
+
"step": 624
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 0.65,
|
| 257 |
+
"learning_rate": 3.434156959234152e-07,
|
| 258 |
+
"loss": 0.4304,
|
| 259 |
+
"step": 640
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 0.67,
|
| 263 |
+
"eval_loss": 0.6797897219657898,
|
| 264 |
+
"eval_raw_wer": 0.2773109243697479,
|
| 265 |
+
"eval_runtime": 485.642,
|
| 266 |
+
"eval_samples_per_second": 0.807,
|
| 267 |
+
"eval_steps_per_second": 0.807,
|
| 268 |
+
"eval_wer": 0.20385932150638034,
|
| 269 |
+
"step": 652
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.67,
|
| 273 |
+
"learning_rate": 3.1603926845215e-07,
|
| 274 |
+
"loss": 0.4579,
|
| 275 |
+
"step": 656
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.69,
|
| 279 |
+
"learning_rate": 2.8928674885331293e-07,
|
| 280 |
+
"loss": 0.4704,
|
| 281 |
+
"step": 672
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.7,
|
| 285 |
+
"learning_rate": 2.632488690317219e-07,
|
| 286 |
+
"loss": 0.4061,
|
| 287 |
+
"step": 688
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.72,
|
| 291 |
+
"learning_rate": 2.380139371717667e-07,
|
| 292 |
+
"loss": 0.3334,
|
| 293 |
+
"step": 704
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 0.73,
|
| 297 |
+
"learning_rate": 2.136675382377774e-07,
|
| 298 |
+
"loss": 0.3986,
|
| 299 |
+
"step": 720
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 0.75,
|
| 303 |
+
"learning_rate": 1.9029224371026699e-07,
|
| 304 |
+
"loss": 0.4019,
|
| 305 |
+
"step": 736
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.77,
|
| 309 |
+
"learning_rate": 1.6796733154248855e-07,
|
| 310 |
+
"loss": 0.4271,
|
| 311 |
+
"step": 752
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.78,
|
| 315 |
+
"learning_rate": 1.467685172870794e-07,
|
| 316 |
+
"loss": 0.4094,
|
| 317 |
+
"step": 768
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.8,
|
| 321 |
+
"learning_rate": 1.26767697304686e-07,
|
| 322 |
+
"loss": 0.3653,
|
| 323 |
+
"step": 784
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.82,
|
| 327 |
+
"learning_rate": 1.0803270492548155e-07,
|
| 328 |
+
"loss": 0.3873,
|
| 329 |
+
"step": 800
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.83,
|
| 333 |
+
"learning_rate": 9.062708039056515e-08,
|
| 334 |
+
"loss": 0.4225,
|
| 335 |
+
"step": 816
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.85,
|
| 339 |
+
"learning_rate": 7.46098553534884e-08,
|
| 340 |
+
"loss": 0.4128,
|
| 341 |
+
"step": 832
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 0.87,
|
| 345 |
+
"learning_rate": 6.00353526727821e-08,
|
| 346 |
+
"loss": 0.3909,
|
| 347 |
+
"step": 848
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.88,
|
| 351 |
+
"learning_rate": 4.695300217448944e-08,
|
| 352 |
+
"loss": 0.3881,
|
| 353 |
+
"step": 864
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.9,
|
| 357 |
+
"learning_rate": 3.540717300955431e-08,
|
| 358 |
+
"loss": 0.3468,
|
| 359 |
+
"step": 880
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.91,
|
| 363 |
+
"learning_rate": 2.543702317462676e-08,
|
| 364 |
+
"loss": 0.3516,
|
| 365 |
+
"step": 896
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 0.93,
|
| 369 |
+
"learning_rate": 1.7076366706642587e-08,
|
| 370 |
+
"loss": 0.3635,
|
| 371 |
+
"step": 912
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.95,
|
| 375 |
+
"learning_rate": 1.035355900158863e-08,
|
| 376 |
+
"loss": 0.3995,
|
| 377 |
+
"step": 928
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.96,
|
| 381 |
+
"learning_rate": 5.29140064639827e-09,
|
| 382 |
+
"loss": 0.4308,
|
| 383 |
+
"step": 944
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.98,
|
| 387 |
+
"learning_rate": 1.907060090135082e-09,
|
| 388 |
+
"loss": 0.3883,
|
| 389 |
+
"step": 960
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.0,
|
| 393 |
+
"learning_rate": 2.1201541672644097e-10,
|
| 394 |
+
"loss": 0.4248,
|
| 395 |
+
"step": 976
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.0,
|
| 399 |
+
"eval_loss": 0.6699328422546387,
|
| 400 |
+
"eval_raw_wer": 0.27388733271086213,
|
| 401 |
+
"eval_runtime": 482.3136,
|
| 402 |
+
"eval_samples_per_second": 0.813,
|
| 403 |
+
"eval_steps_per_second": 0.813,
|
| 404 |
+
"eval_wer": 0.20292561469032058,
|
| 405 |
+
"step": 978
|
| 406 |
+
}
|
| 407 |
+
],
|
| 408 |
+
"logging_steps": 16,
|
| 409 |
+
"max_steps": 980,
|
| 410 |
+
"num_input_tokens_seen": 0,
|
| 411 |
+
"num_train_epochs": 1,
|
| 412 |
+
"save_steps": 326,
|
| 413 |
+
"total_flos": 8.3058450333696e+18,
|
| 414 |
+
"train_batch_size": 4,
|
| 415 |
+
"trial_name": null,
|
| 416 |
+
"trial_params": null
|
| 417 |
+
}
|
checkpoint-978/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:683240c6785fb1eee21b2d156176395090b713ecd552f2112825d03b27aac81a
|
| 3 |
+
size 4984
|
config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-large-v2",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"WhisperForConditionalGeneration"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"bos_token_id": 50257,
|
| 11 |
+
"classifier_proj_size": 256,
|
| 12 |
+
"d_model": 1280,
|
| 13 |
+
"decoder_attention_heads": 20,
|
| 14 |
+
"decoder_ffn_dim": 5120,
|
| 15 |
+
"decoder_layerdrop": 0.0,
|
| 16 |
+
"decoder_layers": 32,
|
| 17 |
+
"decoder_start_token_id": 50258,
|
| 18 |
+
"dropout": 0.0,
|
| 19 |
+
"encoder_attention_heads": 20,
|
| 20 |
+
"encoder_ffn_dim": 5120,
|
| 21 |
+
"encoder_layerdrop": 0.0,
|
| 22 |
+
"encoder_layers": 32,
|
| 23 |
+
"eos_token_id": 50257,
|
| 24 |
+
"forced_decoder_ids": [
|
| 25 |
+
[
|
| 26 |
+
1,
|
| 27 |
+
50259
|
| 28 |
+
],
|
| 29 |
+
[
|
| 30 |
+
2,
|
| 31 |
+
50359
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
3,
|
| 35 |
+
50363
|
| 36 |
+
]
|
| 37 |
+
],
|
| 38 |
+
"init_std": 0.02,
|
| 39 |
+
"is_encoder_decoder": true,
|
| 40 |
+
"mask_feature_length": 10,
|
| 41 |
+
"mask_feature_min_masks": 0,
|
| 42 |
+
"mask_feature_prob": 0.0,
|
| 43 |
+
"mask_time_length": 10,
|
| 44 |
+
"mask_time_min_masks": 2,
|
| 45 |
+
"mask_time_prob": 0.05,
|
| 46 |
+
"max_source_positions": 1500,
|
| 47 |
+
"max_target_positions": 448,
|
| 48 |
+
"median_filter_width": 7,
|
| 49 |
+
"model_type": "whisper",
|
| 50 |
+
"num_hidden_layers": 32,
|
| 51 |
+
"num_mel_bins": 80,
|
| 52 |
+
"pad_token_id": 50257,
|
| 53 |
+
"scale_embedding": false,
|
| 54 |
+
"torch_dtype": "float32",
|
| 55 |
+
"transformers_version": "4.37.2",
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"use_weighted_layer_sum": false,
|
| 58 |
+
"vocab_size": 51865
|
| 59 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"begin_suppress_tokens": [
|
| 3 |
+
220,
|
| 4 |
+
50257
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 50257,
|
| 7 |
+
"decoder_start_token_id": 50258,
|
| 8 |
+
"eos_token_id": 50257,
|
| 9 |
+
"forced_decoder_ids": [
|
| 10 |
+
[
|
| 11 |
+
1,
|
| 12 |
+
50259
|
| 13 |
+
],
|
| 14 |
+
[
|
| 15 |
+
2,
|
| 16 |
+
50359
|
| 17 |
+
],
|
| 18 |
+
[
|
| 19 |
+
3,
|
| 20 |
+
50363
|
| 21 |
+
]
|
| 22 |
+
],
|
| 23 |
+
"language": "english",
|
| 24 |
+
"max_length": 448,
|
| 25 |
+
"pad_token_id": 50257,
|
| 26 |
+
"suppress_tokens": [
|
| 27 |
+
1,
|
| 28 |
+
2,
|
| 29 |
+
7,
|
| 30 |
+
8,
|
| 31 |
+
9,
|
| 32 |
+
10,
|
| 33 |
+
14,
|
| 34 |
+
25,
|
| 35 |
+
26,
|
| 36 |
+
27,
|
| 37 |
+
28,
|
| 38 |
+
29,
|
| 39 |
+
31,
|
| 40 |
+
58,
|
| 41 |
+
59,
|
| 42 |
+
60,
|
| 43 |
+
61,
|
| 44 |
+
62,
|
| 45 |
+
63,
|
| 46 |
+
90,
|
| 47 |
+
91,
|
| 48 |
+
92,
|
| 49 |
+
93,
|
| 50 |
+
359,
|
| 51 |
+
503,
|
| 52 |
+
522,
|
| 53 |
+
542,
|
| 54 |
+
873,
|
| 55 |
+
893,
|
| 56 |
+
902,
|
| 57 |
+
918,
|
| 58 |
+
922,
|
| 59 |
+
931,
|
| 60 |
+
1350,
|
| 61 |
+
1853,
|
| 62 |
+
1982,
|
| 63 |
+
2460,
|
| 64 |
+
2627,
|
| 65 |
+
3246,
|
| 66 |
+
3253,
|
| 67 |
+
3268,
|
| 68 |
+
3536,
|
| 69 |
+
3846,
|
| 70 |
+
3961,
|
| 71 |
+
4183,
|
| 72 |
+
4667,
|
| 73 |
+
6585,
|
| 74 |
+
6647,
|
| 75 |
+
7273,
|
| 76 |
+
9061,
|
| 77 |
+
9383,
|
| 78 |
+
10428,
|
| 79 |
+
10929,
|
| 80 |
+
11938,
|
| 81 |
+
12033,
|
| 82 |
+
12331,
|
| 83 |
+
12562,
|
| 84 |
+
13793,
|
| 85 |
+
14157,
|
| 86 |
+
14635,
|
| 87 |
+
15265,
|
| 88 |
+
15618,
|
| 89 |
+
16553,
|
| 90 |
+
16604,
|
| 91 |
+
18362,
|
| 92 |
+
18956,
|
| 93 |
+
20075,
|
| 94 |
+
21675,
|
| 95 |
+
22520,
|
| 96 |
+
26130,
|
| 97 |
+
26161,
|
| 98 |
+
26435,
|
| 99 |
+
28279,
|
| 100 |
+
29464,
|
| 101 |
+
31650,
|
| 102 |
+
32302,
|
| 103 |
+
32470,
|
| 104 |
+
36865,
|
| 105 |
+
42863,
|
| 106 |
+
47425,
|
| 107 |
+
49870,
|
| 108 |
+
50254,
|
| 109 |
+
50258,
|
| 110 |
+
50358,
|
| 111 |
+
50359,
|
| 112 |
+
50360,
|
| 113 |
+
50361,
|
| 114 |
+
50362
|
| 115 |
+
],
|
| 116 |
+
"task": "transcribe",
|
| 117 |
+
"transformers_version": "4.37.2"
|
| 118 |
+
}
|
model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ecdfeddba97debdbee5a19b40aaed7f786a38c4964571abb6bb45ead1c970a1
|
| 3 |
+
size 4992706480
|
model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:30aa1ee2143a8df5cf2f84537f9ae6584819a63a52359c9c2d0549e24fa1a6ae
|
| 3 |
+
size 1180663192
|
model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:683240c6785fb1eee21b2d156176395090b713ecd552f2112825d03b27aac81a
|
| 3 |
+
size 4984
|