update model

README.md

@@ -24,10 +24,10 @@ model-index:
     metrics:
        - name: Test WER
          type: wer
-         value: 21.16
+         value: 21.53
        - name: Test CER
          type: cer
-         value: 9.53
+         value: 9.66
 ---
 
 # Wav2vec2-Large-English
@@ -81,16 +81,16 @@ for i, predicted_sentence in enumerate(predicted_sentences):
 
 | Reference  | Prediction |
 | ------------- | ------------- |
-| "SHE'LL BE ALL RIGHT." | SHE'D BE AL RIGHT |
+| "SHE'LL BE ALL RIGHT." | SHELL BE ALL RIGHT |
 | SIX | SIX |
-| "ALL'S WELL THAT ENDS WELL." | ALL IS WELL THAT ENDS WELL |
-| DO YOU MEAN IT? | DO YOU MEAN IT |
-| THE NEW PATCH IS LESS INVASIVE THAN THE OLD ONE, BUT STILL CAUSES REGRESSIONS. | THE NEW PATCH IS LESS INVASIVE THAN THE OLD ONE BUT STILL CAUSES REGRESSION |
-| HOW IS MOZILLA GOING TO HANDLE AMBIGUITIES LIKE QUEUE AND CUE? | HOW IS MOSTYURLA GOING TO BANDO AMBIHOTIS LIKE YU AND Q |
-| "I GUESS YOU MUST THINK I'M KINDA BATTY." | QUESTIONS IN CANTON TE PARC |
+| "ALL'S WELL THAT ENDS WELL." | ALLAS WELL THAT ENDS WELL |
+| DO YOU MEAN IT? | W MEAN IT |
+| THE NEW PATCH IS LESS INVASIVE THAN THE OLD ONE, BUT STILL CAUSES REGRESSIONS. | THE NEW PATCH IS LESS INVASIVE THAN THE OLD ONE BUT STILL CAUSES REGRESTION |
+| HOW IS MOZILLA GOING TO HANDLE AMBIGUITIES LIKE QUEUE AND CUE? | HOW IS MOSILLA GOING TO BANDL AND BE WHIT IS LIKE QU AND QU |
+| "I GUESS YOU MUST THINK I'M KINDA BATTY." | RUSTION AS HAME AK AN THE POT |
 | NO ONE NEAR THE REMOTE MACHINE YOU COULD RING? | NO ONE NEAR THE REMOTE MACHINE YOU COULD RING |
-| SAUCE FOR THE GOOSE IS SAUCE FOR THE GANDER. | SAUCE FOR THE GOICE IS SAUCE FOR THE GONDER |
-| GROVES STARTED WRITING SONGS WHEN SHE WAS FOUR YEARS OLD. | GRAFFES STORTED WRITING SONGS WHEN SHE WAS FOUR YEARS OLD |
+| SAUCE FOR THE GOOSE IS SAUCE FOR THE GANDER. | SAUCE FOR THE GUCE IS SAUCE FOR THE GONDER |
+| GROVES STARTED WRITING SONGS WHEN SHE WAS FOUR YEARS OLD. | GRAFS STARTED WRITING SONGS WHEN SHE WAS FOUR YEARS OLD |
 
 ## Evaluation
 
@@ -159,76 +159,18 @@ print(f"CER: {cer.compute(predictions=predictions, references=references, chunk_
 
 **Test Result**:
 
-In the table below I report the Word Error Rate (WER) and the Character Error Rate (CER) of the model. I ran the evaluation script described above on other models as well (on 2021-05-20). Note that the table below may show different results from those already reported, this may have been caused due to some specificity of the other evaluation scripts used. Initially, I've tested the model only using the Common Voice dataset. Later I've also tested the model using the LibriSpeech and TIMIT datasets, which are better-behaved datasets than the Common Voice, containing only examples in US English extracted from audiobooks.
-
----
-
-**Common Voice**
+In the table below I report the Word Error Rate (WER) and the Character Error Rate (CER) of the model. I ran the evaluation script described above on other models as well (on 2021-06-17). Note that the table below may show different results from those already reported, this may have been caused due to some specificity of the other evaluation scripts used.
 
 | Model | WER | CER |
 | ------------- | ------------- | ------------- |
-| jonatasgrosman/wav2vec2-large-xlsr-53-english | **19.76%** | **8.60%** |
-| jonatasgrosman/wav2vec2-large-english | 21.16% | 9.53% |
+| jonatasgrosman/wav2vec2-large-xlsr-53-english | **18.98%** | **8.29%** |
+| jonatasgrosman/wav2vec2-large-english | 21.53% | 9.66% |
 | facebook/wav2vec2-large-960h-lv60-self | 22.03% | 10.39% |
 | facebook/wav2vec2-large-960h-lv60 | 23.97% | 11.14% |
+| boris/xlsr-en-punctuation | 29.10% | 10.75% |
 | facebook/wav2vec2-large-960h | 32.79% | 16.03% |
-| boris/xlsr-en-punctuation | 34.81% | 15.51% |
 | facebook/wav2vec2-base-960h | 39.86% | 19.89% |
 | facebook/wav2vec2-base-100h | 51.06% | 25.06% |
 | elgeish/wav2vec2-large-lv60-timit-asr | 59.96% | 34.28% |
 | facebook/wav2vec2-base-10k-voxpopuli-ft-en | 66.41% | 36.76% |
 | elgeish/wav2vec2-base-timit-asr | 68.78% | 36.81% |
-
----
-
-**LibriSpeech (clean)**
-
-| Model | WER | CER |
-| ------------- | ------------- | ------------- |
-| facebook/wav2vec2-large-960h-lv60-self | **1.86%** | **0.54%** |
-| facebook/wav2vec2-large-960h-lv60 | 2.15% | 0.61% |
-| facebook/wav2vec2-large-960h | 2.82% | 0.84% |
-| facebook/wav2vec2-base-960h | 3.44% | 1.06% |
-| jonatasgrosman/wav2vec2-large-xlsr-53-english | 4.16% | 1.28% |
-| facebook/wav2vec2-base-100h | 6.26% | 2.00% |
-| jonatasgrosman/wav2vec2-large-english | 8.00% | 2.55% |
-| elgeish/wav2vec2-large-lv60-timit-asr | 15.53% | 4.93% |
-| boris/xlsr-en-punctuation | 19.28% | 6.45% |
-| elgeish/wav2vec2-base-timit-asr | 29.19% | 8.38% |
-| facebook/wav2vec2-base-10k-voxpopuli-ft-en | 31.82% | 12.41% |
-
----
-
-**LibriSpeech (other)**
-
-| Model | WER | CER |
-| ------------- | ------------- | ------------- |
-| facebook/wav2vec2-large-960h-lv60-self | **3.89%** | **1.40%** |
-| facebook/wav2vec2-large-960h-lv60 | 4.45% | 1.56% |
-| facebook/wav2vec2-large-960h | 6.49% | 2.52% |
-| jonatasgrosman/wav2vec2-large-xlsr-53-english | 8.82% | 3.42% |
-| facebook/wav2vec2-base-960h | 8.90% | 3.55% |
-| jonatasgrosman/wav2vec2-large-english | 13.62% | 5.24% |
-| facebook/wav2vec2-base-100h | 13.97% | 5.51% |
-| boris/xlsr-en-punctuation | 26.40% | 10.11% |
-| elgeish/wav2vec2-large-lv60-timit-asr | 28.39% | 12.08% |
-| elgeish/wav2vec2-base-timit-asr | 42.04% | 15.57% |
-| facebook/wav2vec2-base-10k-voxpopuli-ft-en | 45.19% | 20.32% |
-
----
-
-**TIMIT**
-
-| Model | WER | CER |
-| ------------- | ------------- | ------------- |
-| facebook/wav2vec2-large-960h-lv60-self | **5.17%** | **1.33%** |
-| facebook/wav2vec2-large-960h-lv60 | 6.24% | 1.54% |
-| jonatasgrosman/wav2vec2-large-xlsr-53-english | 6.81% | 2.02% |
-| facebook/wav2vec2-large-960h | 9.63% | 2.19% |
-| facebook/wav2vec2-base-960h | 11.48% | 2.76% |
-| elgeish/wav2vec2-large-lv60-timit-asr | 13.83% | 4.36% |
-| jonatasgrosman/wav2vec2-large-english | 13.91% | 4.01% |
-| facebook/wav2vec2-base-100h | 16.75% | 4.79% |
-| elgeish/wav2vec2-base-timit-asr | 25.40% | 8.16% |
-| boris/xlsr-en-punctuation | 25.93% | 9.99% |
-| facebook/wav2vec2-base-10k-voxpopuli-ft-en | 51.08% | 19.84% |

config.json

@@ -64,6 +64,6 @@
   "num_feat_extract_layers": 7,
   "num_hidden_layers": 24,
   "pad_token_id": 0,
-  "transformers_version": "4.5.0.dev0",
+  "transformers_version": "4.7.0.dev0",
   "vocab_size": 33
 }

preprocessor_config.json

@@ -1,5 +1,6 @@
 {
   "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
   "feature_size": 1,
   "padding_side": "right",
   "padding_value": 0.0,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:53a299003920a2ebacc857bf3f0c2706c06cc88e6c45924e4aa4cb5b9fb0c04d
+oid sha256:ef32d54a3ebc911d64e7a8b8d04896f0957bbab4e3da4c6c0ae42ab901d6e4e7
 size 1262022892