Model save

README.md

@@ -1,15 +1,10 @@
 ---
-language:
-- tw
 license: apache-2.0
 base_model: facebook/wav2vec2-xls-r-300m
 tags:
-- automatic-speech-recognition
-- mozilla-foundation/common_voice_16_1
-- mms
 - generated_from_trainer
 datasets:
-- common_voice_16_1
+- common_voice_17_0
 metrics:
 - wer
 model-index:
@@ -19,11 +14,11 @@ model-index:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
     dataset:
-      name: MOZILLA-FOUNDATION/COMMON_VOICE_16_1 - TW
-      type: common_voice_16_1
+      name: common_voice_17_0
+      type: common_voice_17_0
       config: tw
       split: None
-      args: 'Config: tw, Training split: train, Eval split: validation+test'
+      args: tw
     metrics:
     - name: Wer
       type: wer
@@ -35,11 +30,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 # wav2vec2-twi-adapter
 
-This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_16_1 - TW dataset.
+This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice_17_0 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 35.6947
+- Loss: 2.4281
 - Wer: 1.0
-- Cer: 5.1304
+- Cer: 1.0
 
 ## Model description
 
@@ -70,11 +65,15 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 100
-- num_epochs: 4.0
+- num_epochs: 30.0
 - mixed_precision_training: Native AMP
 
 ### Training results
 
+| Training Loss | Epoch   | Step | Validation Loss | Wer | Cer |
+|:-------------:|:-------:|:----:|:---------------:|:---:|:---:|
+| No log        | 11.1111 | 50   | 5.8930          | 1.0 | 1.0 |
+| No log        | 22.2222 | 100  | 2.4281          | 1.0 | 1.0 |
 
 
 ### Framework versions

config.json

@@ -103,6 +103,6 @@
   "torch_dtype": "float32",
   "transformers_version": "4.41.1",
   "use_weighted_layer_sum": false,
-  "vocab_size": 32,
+  "vocab_size": 38,
   "xvector_output_dim": 512
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a8642a139276c58649ec3962b6c38c21adb554c0ee7e666d71b9e975a3ae3f9d
-size 1265398792
+oid sha256:45e47801a0c10552648c6368e667f1ed98845f5e54174df8f8970feb7816a7c0
+size 1265423392

runs/May24_03-32-47_cc57b0143c57/events.out.tfevents.1716521591.cc57b0143c57.549.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a236f05672d60cc3b45a221cd78f74212a61781b26eb4790f5c273961d55261
+size 4136

runs/May24_03-34-21_cc57b0143c57/events.out.tfevents.1716521688.cc57b0143c57.714.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84a152e8574f56a4bea0699260f83faec2a449b2e46fac82234a702755ac5a4b
+size 7271

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2b832995a5266615b0e3e4086f70522aa34d00f91dc165597fd1d810432779ba
+oid sha256:a02b74a29f3bd2fb75ea22c3890bcb31ee009c72ec0e942299216385158a9151
 size 4667

vocab.json

@@ -1,7 +1,7 @@
 {
   "aka": {
-    "[PAD]": 29,
-    "[UNK]": 28,
+    "[PAD]": 35,
+    "[UNK]": 34,
     "a": 1,
     "b": 2,
     "c": 3,
@@ -23,12 +23,18 @@
     "u": 19,
     "w": 20,
     "y": 21,
+    "z": 22,
     "|": 0,
-    "á": 22,
-    "ɔ": 23,
-    "ɛ": 24,
-    "—": 25,
-    "’": 26,
-    "•": 27
+    "·": 23,
+    "á": 24,
+    "ã": 25,
+    "ĩ": 26,
+    "ũ": 27,
+    "ɔ": 28,
+    "ɛ": 29,
+    "ʹ": 30,
+    "—": 31,
+    "’": 32,
+    "•": 33
   }
 }