Hugging Face's logo

skit-ai
/
speechllm-2B

Feature Extraction
Transformers
Safetensors
English
custom_model
multi-modal
speech-language
custom_code
Eval Results
Model card Files Community
shangeth commited on
Commit
1e24fb8
•
1 Parent(s): 187d4d3

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -34
README.md CHANGED
@@ -56,33 +56,7 @@ model-index:
56
  language: en
57
  metrics:
58
  - type: wer
59
- value: 24.47
60
- name: Test WER
61
- - task:
62
- type: automatic-speech-recognition
63
- name: Automatic Speech Recognition
64
- dataset:
65
- name: ML Spoken Words
66
- type: MLCommons/ml_spoken_words
67
- split: test
68
- args:
69
- language: en
70
- metrics:
71
- - type: wer
72
- value: 36.12
73
- name: Test WER
74
- - task:
75
- type: automatic-speech-recognition
76
- name: Automatic Speech Recognition
77
- dataset:
78
- name: IEMOCAP
79
- type: Ar4ikov/iemocap_audio_text_splitted
80
- split: test
81
- args:
82
- language: en
83
- metrics:
84
- - type: wer
85
- value: 44.15
86
  name: Test WER
87
  - task:
88
  type: audio-classification
@@ -95,10 +69,10 @@ model-index:
95
  language: en
96
  metrics:
97
  - type: accuracy
98
- value: 62.51
99
  name: Test Age Accuracy
100
  - type: accuracy
101
- value: 64.57
102
  name: Test Accent Accuracy
103
  ---
104
 
@@ -160,8 +134,6 @@ Try the model in [Google Colab Notebook](https://colab.research.google.com/drive
160
 
161
  | **Dataset** | **Type** | **Word Error Rate** | **Gender Acc** | **Age Acc** | **Accent Acc** |
162
  |:--------------------------:|:-------------------:|:-------------------:|:--------------:|:-----------:|:--------------:|
163
- | **librispeech-test-clean** | Read Speech | 6.73 | 0.9536 | | |
164
- | **librispeech-test-other** | Read Speech | 9.13 | 0.9099 | | |
165
- | **CommonVoice test** | Diverse Accent, Age | 24.27 | 0.8680 | 0.6251 | 0.6457 |
166
- | **ML Spoken Words test** | Short Utterance | 36.12 | 0.6587 | | |
167
- | **IEMOCAP test** | Emotional Speech | 44.15 | 0.7557 | | |
 
56
  language: en
57
  metrics:
58
  - type: wer
59
+ value: 26.02
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60
  name: Test WER
61
  - task:
62
  type: audio-classification
 
69
  language: en
70
  metrics:
71
  - type: accuracy
72
+ value: 60.41
73
  name: Test Age Accuracy
74
  - type: accuracy
75
+ value: 69.59
76
  name: Test Accent Accuracy
77
  ---
78
 
 
134
 
135
  | **Dataset** | **Type** | **Word Error Rate** | **Gender Acc** | **Age Acc** | **Accent Acc** |
136
  |:--------------------------:|:-------------------:|:-------------------:|:--------------:|:-----------:|:--------------:|
137
+ | **librispeech-test-clean** | Read Speech | 6.73 | 0.9496 | | |
138
+ | **librispeech-test-other** | Read Speech | 9.13 | 0.9217 | | |
139
+ | **CommonVoice test** | Diverse Accent, Age | 26.02 | 0.8680 | 0.6041 | 0.6959 |