devasheeshG commited on
Commit
5b6ee15
1 Parent(s): 531146d
Files changed (1) hide show
  1. README.md +9 -9
README.md CHANGED
@@ -236,9 +236,9 @@ language:
236
  | Device Name | float32 (Original) | float16 | CudaCores | TensorCores |
237
  | ----------------- | -------------------- | ------- | --------- | ----------- |
238
  | 3060 | 1.7 | 1.1 | 3,584 | 112 |
239
- | 1660 Super | OOM | 3.3 | 1,408 | - |
240
  | Collab (Tesla T4) | 2.8 | 2.2 | 2,560 | 320 |
241
- | Collab (CPU) | 35 | - | - | - |
242
  | M1 (CPU) | - | - | - | - |
243
  | M1 (GPU -> 'mps') | - | - | - | - |
244
 
@@ -255,11 +255,11 @@ language:
255
  - **WIP: Word Information Preserved**
256
  - **CER: Character Error Rate**
257
 
258
- ### Hindi (test.tsv -> 2557 samples used) [Common Voice 14.0](https://commonvoice.mozilla.org/en/datasets)
259
- | | WER | MER | WIL | WIP | CER |
260
- | ----------------- | -------------------- | ------- | --------- | ----------- | ----- |
261
- | Original_Model | - | - | - | - | - |
262
- | This_Model | 54.97 | 47.86 | 66.83 | 33.16 | 30.23 |
263
 
264
  ### English (LibriSpeech -> test-clean -> __ samples used) [LibriSpeech](https://huggingface.co/datasets/librispeech_asr)
265
  | | WER | MER | WIL | WIP | CER |
@@ -275,8 +275,8 @@ language:
275
 
276
  - **'jiwer' library is used for calculations**
277
 
278
- ## Code:
279
- - ### [$\textbf{Will be soon Uploaded on Github}$ ](https://github.com/devasheeshG)
280
 
281
  ## Usage
282
  A file ``__init__.py`` is contained inside this repo which contains all the code to use this model.
 
236
  | Device Name | float32 (Original) | float16 | CudaCores | TensorCores |
237
  | ----------------- | -------------------- | ------- | --------- | ----------- |
238
  | 3060 | 1.7 | 1.1 | 3,584 | 112 |
239
+ | 1660 Super | OOM | 3.3 | 1,408 | N/A |
240
  | Collab (Tesla T4) | 2.8 | 2.2 | 2,560 | 320 |
241
+ | Collab (CPU) | 35 | N/A | N/A | N/A |
242
  | M1 (CPU) | - | - | - | - |
243
  | M1 (GPU -> 'mps') | - | - | - | - |
244
 
 
255
  - **WIP: Word Information Preserved**
256
  - **CER: Character Error Rate**
257
 
258
+ ### Hindi (test.tsv -> 2557 samples used) [Common Voice 14.0](https://huggingface.co/datasets/mozilla-foundation/common_voice_14_0)
259
+ | | WER | MER | WIL | WIP | CER |
260
+ | ----------------------- | -------------------- | ------- | --------- | ----------- | ----- |
261
+ | Original_Model (54 min) | 52.02 | 47.86 | 66.82 | 33.17 | 23.76 |
262
+ | This_Model (38 min) | 54.97 | 47.86 | 66.83 | 33.16 | 30.23 |
263
 
264
  ### English (LibriSpeech -> test-clean -> __ samples used) [LibriSpeech](https://huggingface.co/datasets/librispeech_asr)
265
  | | WER | MER | WIL | WIP | CER |
 
275
 
276
  - **'jiwer' library is used for calculations**
277
 
278
+ ## Code for conversion:
279
+ - ### [Will be soon Uploaded on Github](https://github.com/devasheeshG)
280
 
281
  ## Usage
282
  A file ``__init__.py`` is contained inside this repo which contains all the code to use this model.