Hugging Face's logo

k-nurf
/
whisper_openchs_custom_v5

TensorBoard
Safetensors
whisper
Model card Files
xet
 Metrics Community
k-nurf commited on
Commit
e040c03
·
verified ·
1 Parent(s): 7a44687

Training in progress, step 500

Browse files
Files changed (7) hide show
  1. config.json +60 -0
  2. model-00001-of-00002.safetensors +3 -0
  3. model-00002-of-00002.safetensors +3 -0
  4. model.safetensors.index.json +0 -0
  5. preprocessor_config.json +15 -0
  6. training_args.bin +3 -0
  7. training_metrics.jsonl +61 -0
config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "gelu",
4
+ "apply_spec_augment": false,
5
+ "architectures": [
6
+ "WhisperForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "begin_suppress_tokens": null,
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
+ "dtype": "float32",
20
+ "encoder_attention_heads": 20,
21
+ "encoder_ffn_dim": 5120,
22
+ "encoder_layerdrop": 0.0,
23
+ "encoder_layers": 32,
24
+ "eos_token_id": 50257,
25
+ "forced_decoder_ids": [
26
+ [
27
+ 1,
28
+ 50259
29
+ ],
30
+ [
31
+ 2,
32
+ 50359
33
+ ],
34
+ [
35
+ 3,
36
+ 50363
37
+ ]
38
+ ],
39
+ "init_std": 0.02,
40
+ "is_encoder_decoder": true,
41
+ "mask_feature_length": 10,
42
+ "mask_feature_min_masks": 0,
43
+ "mask_feature_prob": 0.0,
44
+ "mask_time_length": 10,
45
+ "mask_time_min_masks": 2,
46
+ "mask_time_prob": 0.05,
47
+ "max_length": null,
48
+ "max_source_positions": 1500,
49
+ "max_target_positions": 448,
50
+ "median_filter_width": 7,
51
+ "model_type": "whisper",
52
+ "num_hidden_layers": 32,
53
+ "num_mel_bins": 80,
54
+ "pad_token_id": 50257,
55
+ "scale_embedding": false,
56
+ "transformers_version": "4.56.2",
57
+ "use_cache": false,
58
+ "use_weighted_layer_sum": false,
59
+ "vocab_size": 51865
60
+ }
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b5d22a391ee285a1a14e526ec8847bc63f172d49954ddc3d1d3dfa95d0febe7
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:ea0099a501e33c2b6892d3072cc64e8c9f6de49396386fb663f08cdc40f595ac
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 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "dither": 0.0,
4
+ "feature_extractor_type": "WhisperFeatureExtractor",
5
+ "feature_size": 80,
6
+ "hop_length": 160,
7
+ "n_fft": 400,
8
+ "n_samples": 480000,
9
+ "nb_max_frames": 3000,
10
+ "padding_side": "right",
11
+ "padding_value": 0.0,
12
+ "processor_class": "WhisperProcessor",
13
+ "return_attention_mask": false,
14
+ "sampling_rate": 16000
15
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:233d7be66198e12f09c0d5d9a9647b9bebd1ebfb4432e862f571db714160179c
3
+ size 6033
training_metrics.jsonl ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"step": 50, "timestamp": "2025-11-06T13:13:41.797725", "loss": 1.5859, "grad_norm": 10.963883399963379, "learning_rate": 9.800000000000001e-07, "epoch": 0.01}
2
+ {"step": 100, "timestamp": "2025-11-06T13:16:15.220759", "loss": 1.2003, "grad_norm": 6.3657097816467285, "learning_rate": 1.98e-06, "epoch": 0.02}
3
+ {"step": 150, "timestamp": "2025-11-06T13:18:48.330557", "loss": 1.034, "grad_norm": 5.083560466766357, "learning_rate": 2.9800000000000003e-06, "epoch": 0.03}
4
+ {"step": 200, "timestamp": "2025-11-06T13:21:20.730275", "loss": 1.2608, "grad_norm": 10.055756568908691, "learning_rate": 3.96e-06, "epoch": 0.04}
5
+ {"step": 250, "timestamp": "2025-11-06T13:23:52.470726", "loss": 2.1807, "grad_norm": 8.506660461425781, "learning_rate": 4.960000000000001e-06, "epoch": 0.05}
6
+ {"step": 300, "timestamp": "2025-11-06T13:26:20.788603", "loss": 2.161, "grad_norm": 7.902853488922119, "learning_rate": 5.9600000000000005e-06, "epoch": 0.06}
7
+ {"step": 350, "timestamp": "2025-11-06T13:28:54.747758", "loss": 2.0955, "grad_norm": 9.336529731750488, "learning_rate": 6.96e-06, "epoch": 0.07}
8
+ {"step": 400, "timestamp": "2025-11-06T13:31:28.177015", "loss": 2.0366, "grad_norm": 8.473003387451172, "learning_rate": 7.960000000000002e-06, "epoch": 0.08}
9
+ {"step": 450, "timestamp": "2025-11-06T13:34:01.690044", "loss": 2.0117, "grad_norm": 6.365055561065674, "learning_rate": 8.96e-06, "epoch": 0.09}
10
+ {"step": 500, "timestamp": "2025-11-06T13:36:34.800499", "loss": 2.0294, "grad_norm": 7.956766128540039, "learning_rate": 9.960000000000001e-06, "epoch": 0.1}
11
+ {"step": 500, "timestamp": "2025-11-06T13:49:37.917470", "eval_loss": 0.9187508821487427, "eval_wer": 57.42849457545055, "eval_runtime": 776.9084, "eval_samples_per_second": 1.12, "eval_steps_per_second": 0.071, "epoch": 0.1}
12
+ {"step": 550, "timestamp": "2025-11-06T13:52:57.833879", "loss": 1.9729, "grad_norm": 7.512238025665283, "learning_rate": 9.893333333333334e-06, "epoch": 0.11}
13
+ {"step": 600, "timestamp": "2025-11-06T13:55:31.909836", "loss": 2.0109, "grad_norm": 7.466618537902832, "learning_rate": 9.782222222222222e-06, "epoch": 0.12}
14
+ {"step": 650, "timestamp": "2025-11-06T13:58:06.657330", "loss": 1.8975, "grad_norm": 7.688978672027588, "learning_rate": 9.673333333333334e-06, "epoch": 0.13}
15
+ {"step": 700, "timestamp": "2025-11-06T14:00:36.024078", "loss": 1.9269, "grad_norm": 9.745697975158691, "learning_rate": 9.562222222222223e-06, "epoch": 0.14}
16
+ {"step": 750, "timestamp": "2025-11-06T14:03:08.629739", "loss": 1.8819, "grad_norm": 9.41437816619873, "learning_rate": 9.451111111111112e-06, "epoch": 0.15}
17
+ {"step": 800, "timestamp": "2025-11-06T14:05:41.212845", "loss": 1.8635, "grad_norm": 7.229548931121826, "learning_rate": 9.340000000000002e-06, "epoch": 0.16}
18
+ {"step": 850, "timestamp": "2025-11-06T14:08:13.788452", "loss": 1.8425, "grad_norm": 7.563614845275879, "learning_rate": 9.22888888888889e-06, "epoch": 0.17}
19
+ {"step": 900, "timestamp": "2025-11-06T14:10:42.912767", "loss": 1.7859, "grad_norm": 7.745034694671631, "learning_rate": 9.117777777777778e-06, "epoch": 0.18}
20
+ {"step": 950, "timestamp": "2025-11-06T14:13:17.557805", "loss": 1.86, "grad_norm": 7.088816165924072, "learning_rate": 9.006666666666666e-06, "epoch": 0.19}
21
+ {"step": 1000, "timestamp": "2025-11-06T14:15:50.359088", "loss": 0.6576, "grad_norm": 3.6605756282806396, "learning_rate": 8.895555555555556e-06, "epoch": 1.0098}
22
+ {"step": 1000, "timestamp": "2025-11-06T14:31:53.847138", "eval_loss": 0.8343904614448547, "eval_wer": 58.280283331838966, "eval_runtime": 959.4902, "eval_samples_per_second": 0.907, "eval_steps_per_second": 0.057, "epoch": 1.0098}
23
+ {"step": 1050, "timestamp": "2025-11-06T14:35:14.288671", "loss": 0.5495, "grad_norm": 3.4421613216400146, "learning_rate": 8.784444444444446e-06, "epoch": 1.0198}
24
+ {"step": 1100, "timestamp": "2025-11-06T14:37:43.667858", "loss": 0.4864, "grad_norm": 3.0287463665008545, "learning_rate": 8.673333333333334e-06, "epoch": 1.0298}
25
+ {"step": 1150, "timestamp": "2025-11-06T14:40:17.548568", "loss": 0.6504, "grad_norm": 7.032417297363281, "learning_rate": 8.562222222222224e-06, "epoch": 1.0398}
26
+ {"step": 1200, "timestamp": "2025-11-06T14:42:51.662196", "loss": 1.5265, "grad_norm": 7.1791558265686035, "learning_rate": 8.451111111111112e-06, "epoch": 1.0498}
27
+ {"step": 1250, "timestamp": "2025-11-06T14:45:25.447785", "loss": 1.5199, "grad_norm": 6.664968490600586, "learning_rate": 8.34e-06, "epoch": 1.0598}
28
+ {"step": 1300, "timestamp": "2025-11-06T14:47:59.540006", "loss": 1.4571, "grad_norm": 7.519531726837158, "learning_rate": 8.22888888888889e-06, "epoch": 1.0698}
29
+ {"step": 1350, "timestamp": "2025-11-06T14:50:34.242414", "loss": 1.3813, "grad_norm": 6.71161413192749, "learning_rate": 8.11777777777778e-06, "epoch": 1.0798}
30
+ {"step": 1400, "timestamp": "2025-11-06T14:53:08.411514", "loss": 1.3521, "grad_norm": 6.759510517120361, "learning_rate": 8.006666666666667e-06, "epoch": 1.0898}
31
+ {"step": 1450, "timestamp": "2025-11-06T14:55:41.837120", "loss": 1.3298, "grad_norm": 8.092727661132812, "learning_rate": 7.895555555555557e-06, "epoch": 1.0998}
32
+ {"step": 1500, "timestamp": "2025-11-06T14:58:15.877947", "loss": 1.2793, "grad_norm": 6.280128002166748, "learning_rate": 7.784444444444445e-06, "epoch": 1.1098}
33
+ {"step": 1500, "timestamp": "2025-11-06T15:09:05.039518", "eval_loss": 0.9000957012176514, "eval_wer": 49.40374787052811, "eval_runtime": 643.291, "eval_samples_per_second": 1.352, "eval_steps_per_second": 0.085, "epoch": 1.1098}
34
+ {"step": 1550, "timestamp": "2025-11-06T15:12:22.369075", "loss": 1.3244, "grad_norm": 8.540104866027832, "learning_rate": 7.673333333333333e-06, "epoch": 1.1198}
35
+ {"step": 1600, "timestamp": "2025-11-06T15:14:53.984714", "loss": 1.2325, "grad_norm": 8.547293663024902, "learning_rate": 7.562222222222223e-06, "epoch": 1.1298}
36
+ {"step": 1650, "timestamp": "2025-11-06T15:17:25.217062", "loss": 1.2571, "grad_norm": 10.56464672088623, "learning_rate": 7.451111111111111e-06, "epoch": 1.1398}
37
+ {"step": 1700, "timestamp": "2025-11-06T15:19:56.582393", "loss": 1.2757, "grad_norm": 8.089674949645996, "learning_rate": 7.340000000000001e-06, "epoch": 1.1498}
38
+ {"step": 1750, "timestamp": "2025-11-06T15:22:32.457393", "loss": 1.2688, "grad_norm": 6.506724834442139, "learning_rate": 7.22888888888889e-06, "epoch": 1.1598}
39
+ {"step": 1800, "timestamp": "2025-11-06T15:25:04.227338", "loss": 1.2138, "grad_norm": 7.030559539794922, "learning_rate": 7.117777777777778e-06, "epoch": 1.1698}
40
+ {"step": 1850, "timestamp": "2025-11-06T15:27:35.780317", "loss": 1.1783, "grad_norm": 6.814513683319092, "learning_rate": 7.006666666666667e-06, "epoch": 1.1798}
41
+ {"step": 1900, "timestamp": "2025-11-06T15:30:07.515654", "loss": 1.2874, "grad_norm": 7.095033645629883, "learning_rate": 6.8955555555555565e-06, "epoch": 1.1898}
42
+ {"step": 1950, "timestamp": "2025-11-06T15:32:38.118758", "loss": 0.2454, "grad_norm": 3.2571322917938232, "learning_rate": 6.784444444444445e-06, "epoch": 2.0096}
43
+ {"step": 2000, "timestamp": "2025-11-06T15:35:10.154590", "loss": 0.1645, "grad_norm": 2.6791586875915527, "learning_rate": 6.6733333333333335e-06, "epoch": 2.0196}
44
+ {"step": 2000, "timestamp": "2025-11-06T15:46:18.422127", "eval_loss": 0.9856981635093689, "eval_wer": 53.40267192683583, "eval_runtime": 664.7593, "eval_samples_per_second": 1.309, "eval_steps_per_second": 0.083, "epoch": 2.0196}
45
+ {"step": 2050, "timestamp": "2025-11-06T15:49:24.019624", "loss": 0.1297, "grad_norm": 2.694838285446167, "learning_rate": 6.562222222222223e-06, "epoch": 2.0296}
46
+ {"step": 2100, "timestamp": "2025-11-06T15:51:52.181576", "loss": 0.2727, "grad_norm": 16.655744552612305, "learning_rate": 6.451111111111111e-06, "epoch": 2.0396}
47
+ {"step": 2150, "timestamp": "2025-11-06T15:54:19.957993", "loss": 1.0179, "grad_norm": 7.122961521148682, "learning_rate": 6.34e-06, "epoch": 2.0496}
48
+ {"step": 2200, "timestamp": "2025-11-06T15:56:47.984982", "loss": 1.0174, "grad_norm": 5.766362190246582, "learning_rate": 6.22888888888889e-06, "epoch": 2.0596}
49
+ {"step": 2250, "timestamp": "2025-11-06T15:59:15.808027", "loss": 0.9404, "grad_norm": 4.847174644470215, "learning_rate": 6.117777777777778e-06, "epoch": 2.0696}
50
+ {"step": 2300, "timestamp": "2025-11-06T16:01:43.574192", "loss": 0.8876, "grad_norm": 6.817285537719727, "learning_rate": 6.006666666666667e-06, "epoch": 2.0796}
51
+ {"step": 50, "timestamp": "2025-11-06T16:39:30.021523", "loss": 2.0204, "grad_norm": 9.757033348083496, "learning_rate": 9.600000000000001e-07, "epoch": 0.005}
52
+ {"step": 100, "timestamp": "2025-11-06T16:42:02.771616", "loss": 1.5336, "grad_norm": 5.912972450256348, "learning_rate": 1.9600000000000003e-06, "epoch": 0.01}
53
+ {"step": 150, "timestamp": "2025-11-06T16:44:35.512065", "loss": 1.2863, "grad_norm": 6.189149379730225, "learning_rate": 2.96e-06, "epoch": 0.015}
54
+ {"step": 200, "timestamp": "2025-11-06T16:47:07.474568", "loss": 1.2046, "grad_norm": 4.846399307250977, "learning_rate": 3.94e-06, "epoch": 0.02}
55
+ {"step": 250, "timestamp": "2025-11-06T16:49:39.578675", "loss": 1.1293, "grad_norm": 6.970186710357666, "learning_rate": 4.94e-06, "epoch": 0.025}
56
+ {"step": 300, "timestamp": "2025-11-06T16:52:11.674263", "loss": 1.028, "grad_norm": 5.767773151397705, "learning_rate": 5.94e-06, "epoch": 0.03}
57
+ {"step": 350, "timestamp": "2025-11-06T16:54:43.962907", "loss": 0.988, "grad_norm": 7.702676296234131, "learning_rate": 6.9400000000000005e-06, "epoch": 0.035}
58
+ {"step": 400, "timestamp": "2025-11-06T16:57:15.862306", "loss": 0.9465, "grad_norm": 5.122243404388428, "learning_rate": 7.94e-06, "epoch": 0.04}
59
+ {"step": 450, "timestamp": "2025-11-06T16:59:48.237965", "loss": 0.8703, "grad_norm": 4.715704441070557, "learning_rate": 8.94e-06, "epoch": 0.045}
60
+ {"step": 500, "timestamp": "2025-11-06T17:02:20.580332", "loss": 0.8817, "grad_norm": 5.4225687980651855, "learning_rate": 9.940000000000001e-06, "epoch": 0.05}
61
+ {"step": 500, "timestamp": "2025-11-06T17:20:18.990705", "eval_loss": 0.8744064569473267, "eval_wer": 54.04749055585537, "eval_runtime": 1074.2837, "eval_samples_per_second": 1.089, "eval_steps_per_second": 0.069, "epoch": 0.05}