End of training

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: wetdog/TUT-urban-acoustic-scenes-2018-development
+base_model: wetdog/TUT-urban-acoustic-scenes-2018-development-16bit
 tags:
 - audio-classification
 - generated_from_trainer
@@ -14,13 +14,13 @@ model-index:
       name: Audio Classification
       type: audio-classification
     dataset:
-      name: TUT-urban-acoustic-scenes-2018-development
+      name: TUT-urban-acoustic-scenes-2018-development-16bit
       type: acoustic-scenes
       args: 'split: train'
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.52
+      value: 0.715647339158062
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -28,10 +28,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # ast-finetuned-audioset-10-10-0.4593-TUT-acoustic-scenes
 
-This model is a fine-tuned version of [wetdog/TUT-urban-acoustic-scenes-2018-development](https://huggingface.co/wetdog/TUT-urban-acoustic-scenes-2018-development) on the TUT-urban-acoustic-scenes-2018-development dataset.
+This model is a fine-tuned version of [wetdog/TUT-urban-acoustic-scenes-2018-development-16bit](https://huggingface.co/wetdog/TUT-urban-acoustic-scenes-2018-development-16bit) on the TUT-urban-acoustic-scenes-2018-development-16bit dataset.
 It achieves the following results on the evaluation set:
-- Loss: 4.4942
-- Accuracy: 0.52
+- Loss: 0.8055
+- Accuracy: 0.7156
 
 ## Model description
 
@@ -50,32 +50,26 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 3e-05
-- train_batch_size: 4
-- eval_batch_size: 4
+- learning_rate: 3e-06
+- train_batch_size: 16
+- eval_batch_size: 16
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_ratio: 0.1
-- training_steps: 4000
+- lr_scheduler_warmup_steps: 50
+- training_steps: 1000
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 1.212         | 0.12  | 500  | 2.5780          | 0.375    |
-| 0.5031        | 1.12  | 1000 | 3.6818          | 0.435    |
-| 0.2583        | 2.12  | 1500 | 3.8194          | 0.465    |
-| 0.0828        | 3.12  | 2000 | 4.3022          | 0.455    |
-| 0.0367        | 4.12  | 2500 | 4.6687          | 0.45     |
-| 0.0054        | 5.12  | 3000 | 4.8838          | 0.465    |
-| 0.0014        | 6.12  | 3500 | 4.6808          | 0.495    |
-| 0.0029        | 7.12  | 4000 | 4.4942          | 0.52     |
+| 0.4664        | 1.12  | 500  | 1.3147          | 0.6136   |
+| 0.6605        | 2.23  | 1000 | 0.8055          | 0.7156   |
 
 
 ### Framework versions
 
-- Transformers 4.33.1
+- Transformers 4.33.2
 - Pytorch 2.0.1+cu118
 - Datasets 2.14.5
 - Tokenizers 0.13.3

config.json

@@ -45,5 +45,5 @@
   "qkv_bias": true,
   "time_stride": 10,
   "torch_dtype": "float32",
-  "transformers_version": "4.33.1"
+  "transformers_version": "4.33.2"
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cbd64872244b5fa19120711014adbe6b1c89a36816c02ecea27b4574560c8ca7
+oid sha256:d5e2bbd1229f24536c2f52f207c9c014d1fec067d66c23ce9927116f5fc4ce82
 size 344860025

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1535226bd3b8383179146ce3812049edfe2b4e320360a0a101a8f835651c4290
+oid sha256:168549945405aa47173c2f0b3ffcc3ac1a0d1d5fb0ee021f4eda1767874b16b4
 size 4091