Hugging Face's logo

mahwizzzz
/
medwhishper

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
English
whisper
hf-asr-leaderboard
Generated from Trainer
Model card Files
xet
Metrics Community
mahwizzzz commited on
Commit
d7daba0
·
verified ·
1 Parent(s): eab8c53

Training in progress, step 1000

Browse files
Files changed (6) hide show
  1. config.json +47 -47
  2. model.safetensors +1 -1
  3. preprocessor_config.json +15 -15
  4. runs/Jul09_16-32-44_localhost/events.out.tfevents.1752060814.localhost.49149.0 +3 -0
  5. runs/Jul09_16-34-05_localhost/events.out.tfevents.1752060847.localhost.49149.1 +3 -0
  6. training_args.bin +2 -2
config.json CHANGED
@@ -1,47 +1,47 @@
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": 768,
13
- "decoder_attention_heads": 12,
14
- "decoder_ffn_dim": 3072,
15
- "decoder_layerdrop": 0.0,
16
- "decoder_layers": 12,
17
- "decoder_start_token_id": 50258,
18
- "dropout": 0.0,
19
- "encoder_attention_heads": 12,
20
- "encoder_ffn_dim": 3072,
21
- "encoder_layerdrop": 0.0,
22
- "encoder_layers": 12,
23
- "eos_token_id": 50257,
24
- "forced_decoder_ids": null,
25
- "init_std": 0.02,
26
- "is_encoder_decoder": true,
27
- "mask_feature_length": 10,
28
- "mask_feature_min_masks": 0,
29
- "mask_feature_prob": 0.0,
30
- "mask_time_length": 10,
31
- "mask_time_min_masks": 2,
32
- "mask_time_prob": 0.05,
33
- "max_length": null,
34
- "max_source_positions": 1500,
35
- "max_target_positions": 448,
36
- "median_filter_width": 7,
37
- "model_type": "whisper",
38
- "num_hidden_layers": 12,
39
- "num_mel_bins": 80,
40
- "pad_token_id": 50257,
41
- "scale_embedding": false,
42
- "torch_dtype": "float32",
43
- "transformers_version": "4.54.0.dev0",
44
- "use_cache": true,
45
- "use_weighted_layer_sum": false,
46
- "vocab_size": 51865
47
- }
 
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": 768,
13
+ "decoder_attention_heads": 12,
14
+ "decoder_ffn_dim": 3072,
15
+ "decoder_layerdrop": 0.0,
16
+ "decoder_layers": 12,
17
+ "decoder_start_token_id": 50258,
18
+ "dropout": 0.0,
19
+ "encoder_attention_heads": 12,
20
+ "encoder_ffn_dim": 3072,
21
+ "encoder_layerdrop": 0.0,
22
+ "encoder_layers": 12,
23
+ "eos_token_id": 50257,
24
+ "forced_decoder_ids": null,
25
+ "init_std": 0.02,
26
+ "is_encoder_decoder": true,
27
+ "mask_feature_length": 10,
28
+ "mask_feature_min_masks": 0,
29
+ "mask_feature_prob": 0.0,
30
+ "mask_time_length": 10,
31
+ "mask_time_min_masks": 2,
32
+ "mask_time_prob": 0.05,
33
+ "max_length": null,
34
+ "max_source_positions": 1500,
35
+ "max_target_positions": 448,
36
+ "median_filter_width": 7,
37
+ "model_type": "whisper",
38
+ "num_hidden_layers": 12,
39
+ "num_mel_bins": 80,
40
+ "pad_token_id": 50257,
41
+ "scale_embedding": false,
42
+ "torch_dtype": "float32",
43
+ "transformers_version": "4.54.0.dev0",
44
+ "use_cache": true,
45
+ "use_weighted_layer_sum": false,
46
+ "vocab_size": 51865
47
+ }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:41b7464216493b8f2448d5b419bc047622fc0794bf3f1c4ed6feaa2a5baddd70
3
  size 966995080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf3ab318016e7e6140d09eafcb643bc321ed732df24c93e03e5c3fde23e78a3b
3
  size 966995080
preprocessor_config.json CHANGED
@@ -1,15 +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
- }
 
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
+ }
runs/Jul09_16-32-44_localhost/events.out.tfevents.1752060814.localhost.49149.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a808a79610ab6b0e22270d04e111dbeca0da3373fcf19f434689d30ba2632d05
3
+ size 5733
runs/Jul09_16-34-05_localhost/events.out.tfevents.1752060847.localhost.49149.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c52eb549a83a58243af3e5c7947cf15512cda63fd3abcf00832fd9996ae7ecf
3
+ size 10357
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b38cde8986a2da22d1a2328fde52fe5b371c3e9f3bb5c708dd6070953f208fd6
3
- size 5905
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdd3aa6d33848881d626fb1b0a0939804fdb1e533ba74b36903ee583bd4609ef
3
+ size 5841