wav2vec2-large-xls-r-300m-zhhk

README.md

@@ -1,6 +1,6 @@
 ---
 license: apache-2.0
-base_model: w11wo/wav2vec2-xls-r-300m-zh-HK-lm-v2
+base_model: ctl/wav2vec2-large-xlsr-cantonese
 tags:
 - generated_from_trainer
 model-index:
@@ -13,7 +13,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # wav2vec2-large-xls-r-300m-zhhk
 
-This model is a fine-tuned version of [w11wo/wav2vec2-xls-r-300m-zh-HK-lm-v2](https://huggingface.co/w11wo/wav2vec2-xls-r-300m-zh-HK-lm-v2) on an unknown dataset.
+This model is a fine-tuned version of [ctl/wav2vec2-large-xlsr-cantonese](https://huggingface.co/ctl/wav2vec2-large-xlsr-cantonese) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 4.3414
+- Cer: 0.7094
 
 ## Model description
 
@@ -36,16 +39,22 @@ The following hyperparameters were used during training:
 - train_batch_size: 8
 - eval_batch_size: 8
 - seed: 42
-- gradient_accumulation_steps: 8
-- total_train_batch_size: 64
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
-- num_epochs: 5
+- lr_scheduler_warmup_steps: 100
+- num_epochs: 30
 - mixed_precision_training: Native AMP
 
 ### Training results
 
+| Training Loss | Epoch | Step | Validation Loss | Cer    |
+|:-------------:|:-----:|:----:|:---------------:|:------:|
+| 9.2737        | 6.78  | 400  | 4.0199          | 0.8241 |
+| 1.0554        | 13.56 | 800  | 4.2189          | 0.7531 |
+| 0.5784        | 20.34 | 1200 | 4.3916          | 0.7212 |
+| 0.3313        | 27.12 | 1600 | 4.3414          | 0.7094 |
 
 
 ### Framework versions

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f3a85dc0651b1d782bb018575c16503ee32bc8519c4a1836d50592ee5c13c97
+oid sha256:83586a68f77202dba89826935d83dfa7c81405c04d727dd77a49028b8dacaef6
 size 1276497820

runs/Dec18_06-04-12_b2128b12c385/events.out.tfevents.1702879558.b2128b12c385.18685.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5276d6e7981dadc5b1a78c6c5dc97cd1f6bf153570932fa6101187ae0456ebee
-size 7780
+oid sha256:28ab2b25f7bf8ba27413610669af7e4ba9f12891a3c3ba20e7e1c7f62c2958f4
+size 8134