gglabs/stt-turbo-multilingual-test

README.md

@@ -20,8 +20,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the custom_multilingual dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3980
-- Wer: 15.3116
+- Loss: 0.3434
+- Wer: 25.3086
 
 ## Model description
 
@@ -40,32 +40,22 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 0.0005
-- train_batch_size: 16
-- eval_batch_size: 32
+- learning_rate: 0.001
+- train_batch_size: 4
+- eval_batch_size: 4
 - seed: 42
-- gradient_accumulation_steps: 4
-- total_train_batch_size: 64
 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 200
-- training_steps: 5000
+- lr_scheduler_warmup_steps: 5
+- training_steps: 20
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch   | Step | Validation Loss | Wer     |
-|:-------------:|:-------:|:----:|:---------------:|:-------:|
-| 0.3199        | 1.0971  | 500  | 0.3644          | 18.0913 |
-| 0.2659        | 2.1942  | 1000 | 0.3579          | 17.1415 |
-| 0.2321        | 3.2913  | 1500 | 0.3579          | 16.6319 |
-| 0.2101        | 4.3884  | 2000 | 0.3635          | 16.3539 |
-| 0.1901        | 5.4855  | 2500 | 0.3677          | 15.7470 |
-| 0.1673        | 6.5826  | 3000 | 0.3748          | 15.4413 |
-| 0.146         | 7.6796  | 3500 | 0.3813          | 15.4691 |
-| 0.13          | 8.7767  | 4000 | 0.3879          | 15.6266 |
-| 0.1206        | 9.8738  | 4500 | 0.3950          | 15.4459 |
-| 0.1091        | 10.9709 | 5000 | 0.3980          | 15.3116 |
+| Training Loss | Epoch  | Step | Validation Loss | Wer     |
+|:-------------:|:------:|:----:|:---------------:|:-------:|
+| 0.6956        | 0.1754 | 10   | 0.4175          | 32.7160 |
+| 0.2372        | 0.3509 | 20   | 0.3434          | 25.3086 |
 
 
 ### Framework versions

adapter_config.json

@@ -28,8 +28,8 @@
   "revision": null,
   "target_modules": [
     "v_proj",
-    "o_proj",
     "k_proj",
+    "o_proj",
     "q_proj"
   ],
   "task_type": null,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:30c0d6d7a16b36b1af07c24d22195c702e8f39cb3baf65c0cdde8080ffa2f2d3
+oid sha256:795b992857e478e64e53623aa7a99b91a8d9b5f8100961a67d679b74f05d8cdb
 size 78677456

runs/May25_00-40-08_ubuntu/events.out.tfevents.1748101208.ubuntu.32776.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c372f0490278e277a2d0116e95f28360a2d44f2f1ca20e95abdb5a4b35760361
+size 9414

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8373683a2465563929a973e2ec6b94a5075a0414524cfa447c6b53e234a32892
+oid sha256:b9b26113d4187becce7aef3bd62b47862e056d6587add7f8094e7ba04fc0a4b8
 size 5496