ai-team-ori commited on
Commit
768c866
·
verified ·
1 Parent(s): 172d34d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +41 -34
README.md CHANGED
@@ -1,10 +1,14 @@
1
  ---
2
- language:
3
  - en
 
4
  tags:
5
  - audio
6
  - automatic-speech-recognition
7
  - whisper-event
 
 
 
8
  model-index:
9
  - name: Whisper-Hindi2Hinglish-Prime
10
  results:
@@ -45,38 +49,41 @@ model-index:
45
  value: 60.8224
46
  name: WER
47
  widget:
48
- - src: audios/c0637211-7384-4abc-af69-5aacf7549824_1_2629072_2656224.wav
49
- output:
50
- text: Mehnat to poora karte hain.
51
- - src: audios/c0faba11-27ba-4837-a2eb-ccd67be07f40_1_3185088_3227568.wav
52
- output:
53
- text: Haan vahi ek aapko bataaya na.
54
- - src: audios/663eb653-d6b5-4fda-b5f2-9ef98adc0a61_0_1098400_1118688.wav
55
- output:
56
- text: Aap pandrah log hain.
57
- - src: audios/f5e0178c-354c-40c9-b3a7-687c86240a77_1_2613728_2630112.wav
58
- output:
59
- text: Kitne saal ki?
60
- - src: audios/f5e0178c-354c-40c9-b3a7-687c86240a77_1_1152496_1175488.wav
61
- output:
62
- text: Lander cycle chaahie.
63
- - src: audios/c0637211-7384-4abc-af69-5aacf7549824_1_2417088_2444224.wav
64
- output:
65
- text: Haan haan, dekhe hain.
66
-
67
- - src: audios/common_voice_hi_23796065.mp3
68
- example_title: Speech Example 1
69
- - src: audios/common_voice_hi_41666099.mp3
70
- example_title: Speech Example 2
71
- - src: audios/common_voice_hi_41429198.mp3
72
- example_title: Speech Example 3
73
- - src: audios/common_voice_hi_41429259.mp3
74
- example_title: Speech Example 4
75
- - src: audios/common_voice_hi_40904697.mp3
76
- example_title: Speech Example 5
77
-
78
  pipeline_tag: automatic-speech-recognition
79
  license: apache-2.0
 
 
 
 
 
80
  ---
81
 
82
  ## Whisper-Hindi2Hinglish-Prime:
@@ -219,7 +226,7 @@ from collections import OrderedDict
219
  import json
220
 
221
  # Load parameter name mapping from HF to OpenAI format
222
- with open('convert_hf2openai.json.json', 'r') as f:
223
  reverse_translation = json.load(f)
224
 
225
  reverse_translation = OrderedDict(reverse_translation)
@@ -272,7 +279,7 @@ model = AutoModelForSpeechSeq2Seq.from_pretrained(
272
 
273
  # Convert and save model
274
  model_save_path = "Whisper-Hindi2Hinglish-Prime.pt"
275
- save_model(model_save_path)
276
  ```
277
 
278
  - Transcribe
@@ -287,4 +294,4 @@ print(result["text"])
287
 
288
 
289
  ### Miscellaneous
290
- This model is from a family of transformers-based ASR models trained by Oriserve. To compare this model against other models from the same family or other SOTA models please head to our [Speech-To-Text Arena](https://huggingface.co/spaces/Oriserve/ASR_arena). To learn more about our other models, and other queries regarding AI voice agents you can reach out to us at our email [email]([email protected])
 
1
  ---
2
+ language:
3
  - en
4
+ - hi
5
  tags:
6
  - audio
7
  - automatic-speech-recognition
8
  - whisper-event
9
+ - pytorch
10
+ - hinglish
11
+ inference: true
12
  model-index:
13
  - name: Whisper-Hindi2Hinglish-Prime
14
  results:
 
49
  value: 60.8224
50
  name: WER
51
  widget:
52
+ - src: audios/c0637211-7384-4abc-af69-5aacf7549824_1_2629072_2656224.wav
53
+ output:
54
+ text: Mehnat to poora karte hain.
55
+ - src: audios/c0faba11-27ba-4837-a2eb-ccd67be07f40_1_3185088_3227568.wav
56
+ output:
57
+ text: Haan vahi ek aapko bataaya na.
58
+ - src: audios/663eb653-d6b5-4fda-b5f2-9ef98adc0a61_0_1098400_1118688.wav
59
+ output:
60
+ text: Aap pandrah log hain.
61
+ - src: audios/f5e0178c-354c-40c9-b3a7-687c86240a77_1_2613728_2630112.wav
62
+ output:
63
+ text: Kitne saal ki?
64
+ - src: audios/f5e0178c-354c-40c9-b3a7-687c86240a77_1_1152496_1175488.wav
65
+ output:
66
+ text: Lander cycle chaahie.
67
+ - src: audios/c0637211-7384-4abc-af69-5aacf7549824_1_2417088_2444224.wav
68
+ output:
69
+ text: Haan haan, dekhe hain.
70
+ - src: audios/common_voice_hi_23796065.mp3
71
+ example_title: Speech Example 1
72
+ - src: audios/common_voice_hi_41666099.mp3
73
+ example_title: Speech Example 2
74
+ - src: audios/common_voice_hi_41429198.mp3
75
+ example_title: Speech Example 3
76
+ - src: audios/common_voice_hi_41429259.mp3
77
+ example_title: Speech Example 4
78
+ - src: audios/common_voice_hi_40904697.mp3
79
+ example_title: Speech Example 5
 
 
80
  pipeline_tag: automatic-speech-recognition
81
  license: apache-2.0
82
+ metrics:
83
+ - wer
84
+ base_model:
85
+ - openai/whisper-large-v3
86
+ library_name: transformers
87
  ---
88
 
89
  ## Whisper-Hindi2Hinglish-Prime:
 
226
  import json
227
 
228
  # Load parameter name mapping from HF to OpenAI format
229
+ with open('convert_hf2openai.json', 'r') as f:
230
  reverse_translation = json.load(f)
231
 
232
  reverse_translation = OrderedDict(reverse_translation)
 
279
 
280
  # Convert and save model
281
  model_save_path = "Whisper-Hindi2Hinglish-Prime.pt"
282
+ save_model(model,model_save_path)
283
  ```
284
 
285
  - Transcribe
 
294
 
295
 
296
  ### Miscellaneous
297
+ This model is from a family of transformers-based ASR models trained by Oriserve. To compare this model against other models from the same family or other SOTA models please head to our [Speech-To-Text Arena](https://huggingface.co/spaces/Oriserve/ASR_arena). To learn more about our other models, and other queries regarding AI voice agents you can reach out to us at our email [[email protected]]([email protected])