bigmorning commited on
Commit
012a4d2
1 Parent(s): 07b52a6

Upload TFWhisperForConditionalGeneration

Browse files
Files changed (2) hide show
  1. README.md +28 -26
  2. tf_model.h5 +1 -1
README.md CHANGED
@@ -14,10 +14,12 @@ probably proofread and complete it, then remove this comment. -->
14
 
15
  This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on an unknown dataset.
16
  It achieves the following results on the evaluation set:
17
- - Train Loss: 0.1785
18
- - Train Accuracy: 0.0334
19
- - Validation Loss: 0.5522
20
- - Validation Accuracy: 0.0314
 
 
21
  - Epoch: 19
22
 
23
  ## Model description
@@ -42,28 +44,28 @@ The following hyperparameters were used during training:
42
 
43
  ### Training results
44
 
45
- | Train Loss | Train Accuracy | Validation Loss | Validation Accuracy | Epoch |
46
- |:----------:|:--------------:|:---------------:|:-------------------:|:-----:|
47
- | 5.0949 | 0.0116 | 4.4444 | 0.0124 | 0 |
48
- | 4.3242 | 0.0130 | 4.0648 | 0.0140 | 1 |
49
- | 3.9308 | 0.0145 | 3.6837 | 0.0157 | 2 |
50
- | 3.5552 | 0.0159 | 3.3410 | 0.0171 | 3 |
51
- | 3.1591 | 0.0175 | 2.8089 | 0.0198 | 4 |
52
- | 2.2408 | 0.0221 | 1.7104 | 0.0255 | 5 |
53
- | 1.4220 | 0.0261 | 1.2181 | 0.0279 | 6 |
54
- | 1.0460 | 0.0280 | 0.9912 | 0.0290 | 7 |
55
- | 0.8363 | 0.0291 | 0.8645 | 0.0296 | 8 |
56
- | 0.6967 | 0.0299 | 0.7748 | 0.0301 | 9 |
57
- | 0.5942 | 0.0305 | 0.7201 | 0.0304 | 10 |
58
- | 0.5151 | 0.0309 | 0.6675 | 0.0307 | 11 |
59
- | 0.4496 | 0.0314 | 0.6382 | 0.0308 | 12 |
60
- | 0.3951 | 0.0318 | 0.6060 | 0.0310 | 13 |
61
- | 0.3473 | 0.0321 | 0.5945 | 0.0311 | 14 |
62
- | 0.3053 | 0.0324 | 0.5752 | 0.0312 | 15 |
63
- | 0.2684 | 0.0327 | 0.5700 | 0.0313 | 16 |
64
- | 0.2355 | 0.0330 | 0.5651 | 0.0313 | 17 |
65
- | 0.2065 | 0.0332 | 0.5619 | 0.0313 | 18 |
66
- | 0.1785 | 0.0334 | 0.5522 | 0.0314 | 19 |
67
 
68
 
69
  ### Framework versions
 
14
 
15
  This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on an unknown dataset.
16
  It achieves the following results on the evaluation set:
17
+ - Train Loss: 4.2860
18
+ - Train Accuracy: 0.0128
19
+ - Train Do Wer: 1.0
20
+ - Validation Loss: 4.6401
21
+ - Validation Accuracy: 0.0125
22
+ - Validation Do Wer: 1.0
23
  - Epoch: 19
24
 
25
  ## Model description
 
44
 
45
  ### Training results
46
 
47
+ | Train Loss | Train Accuracy | Train Do Wer | Validation Loss | Validation Accuracy | Validation Do Wer | Epoch |
48
+ |:----------:|:--------------:|:------------:|:---------------:|:-------------------:|:-----------------:|:-----:|
49
+ | 9.9191 | 0.0046 | 1.0 | 8.5836 | 0.0067 | 1.0 | 0 |
50
+ | 8.0709 | 0.0083 | 1.0 | 7.4667 | 0.0089 | 1.0 | 1 |
51
+ | 7.1652 | 0.0100 | 1.0 | 6.8204 | 0.0112 | 1.0 | 2 |
52
+ | 6.7196 | 0.0114 | 1.0 | 6.5192 | 0.0114 | 1.0 | 3 |
53
+ | 6.4115 | 0.0115 | 1.0 | 6.2357 | 0.0115 | 1.0 | 4 |
54
+ | 6.1085 | 0.0115 | 1.0 | 5.9657 | 0.0115 | 1.0 | 5 |
55
+ | 5.8206 | 0.0115 | 1.0 | 5.7162 | 0.0115 | 1.0 | 6 |
56
+ | 5.5567 | 0.0115 | 1.0 | 5.4963 | 0.0115 | 1.0 | 7 |
57
+ | 5.3223 | 0.0116 | 1.0 | 5.3096 | 0.0116 | 1.0 | 8 |
58
+ | 5.1222 | 0.0117 | 1.0 | 5.1600 | 0.0117 | 1.0 | 9 |
59
+ | 4.9580 | 0.0117 | 1.0 | 5.0391 | 0.0118 | 1.0 | 10 |
60
+ | 4.8251 | 0.0119 | 1.0 | 4.9427 | 0.0118 | 1.0 | 11 |
61
+ | 4.7171 | 0.0119 | 1.0 | 4.8691 | 0.0119 | 1.0 | 12 |
62
+ | 4.6284 | 0.0121 | 1.0 | 4.8123 | 0.0120 | 1.0 | 13 |
63
+ | 4.5508 | 0.0121 | 1.0 | 4.7620 | 0.0121 | 1.0 | 14 |
64
+ | 4.4855 | 0.0123 | 1.0 | 4.7260 | 0.0121 | 1.0 | 15 |
65
+ | 4.4305 | 0.0124 | 1.0 | 4.7018 | 0.0123 | 1.0 | 16 |
66
+ | 4.3788 | 0.0125 | 1.0 | 4.6738 | 0.0123 | 1.0 | 17 |
67
+ | 4.3305 | 0.0127 | 1.0 | 4.6525 | 0.0124 | 1.0 | 18 |
68
+ | 4.2860 | 0.0128 | 1.0 | 4.6401 | 0.0125 | 1.0 | 19 |
69
 
70
 
71
  ### Framework versions
tf_model.h5 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:21f6bddb95988b793e797dcd66caa04fa053fac47c50591ba6908e745f5e2972
3
  size 151253984
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07a0dec50cf8d54e4bf9bdb709705179d8c23574897735da9fe15098c010dcc0
3
  size 151253984