Training in progress, epoch 1

config.json

@@ -3,10 +3,10 @@
   "architectures": [
     "ASTForAudioClassification"
   ],
-  "attention_probs_dropout_prob": 0.3,
+  "attention_probs_dropout_prob": 0.1,
   "frequency_stride": 10,
   "hidden_act": "gelu",
-  "hidden_dropout_prob": 0.5,
+  "hidden_dropout_prob": 0.3,
   "hidden_size": 768,
   "id2label": {
     "0": "anger",
@@ -29,7 +29,7 @@
     "sadness": 6
   },
   "layer_norm_eps": 1e-12,
-  "max_length": 2048,
+  "max_length": 1024,
   "model_type": "audio-spectrogram-transformer",
   "num_attention_heads": 12,
   "num_hidden_layers": 12,
@@ -39,5 +39,5 @@
   "qkv_bias": true,
   "time_stride": 10,
   "torch_dtype": "float32",
-  "transformers_version": "4.28.1"
+  "transformers_version": "4.28.0"
 }

preprocessor_config.json

@@ -2,7 +2,7 @@
   "do_normalize": true,
   "feature_extractor_type": "ASTFeatureExtractor",
   "feature_size": 1,
-  "max_length": 2048,
+  "max_length": 1024,
   "mean": -2.30462729231313e-05,
   "num_mel_bins": 128,
   "padding_side": "right",

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b8fcb21b65f827e91f9767e457819026c84136ba89a0929c36bdd324c20454e
-size 348647801
+oid sha256:8d0940c80eac265e0b370480ab0d3d9fbcc38ef12fbe550828131bea2bb1a471
+size 344850809

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c9e46db368eab91a184a012bbed1809ca8e4fc5aa71311392fb2fdcaa066326
+oid sha256:8604bc2b04577b7139cfb448f3e31b3da694cc27aa17bc80441411ed84766419
 size 3579