Training in progress, step 500

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "abletobetable/spec_soul_ast",
+  "_name_or_path": "MIT/ast-finetuned-audioset-10-10-0.4593",
   "architectures": [
     "ASTForAudioClassification"
   ],
@@ -29,7 +29,7 @@
     "sadness": 6
   },
   "layer_norm_eps": 1e-12,
-  "max_length": 1024,
+  "max_length": 2048,
   "model_type": "audio-spectrogram-transformer",
   "num_attention_heads": 12,
   "num_hidden_layers": 12,

preprocessor_config.json

@@ -2,7 +2,7 @@
   "do_normalize": true,
   "feature_extractor_type": "ASTFeatureExtractor",
   "feature_size": 1,
-  "max_length": 1024,
+  "max_length": 2048,
   "mean": -4.2677393,
   "num_mel_bins": 128,
   "padding_side": "right",

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fcc046380045f49f5ede5648187a4b4cffad264d9127931cd1817c29c77f3f2b
-size 344850809
+oid sha256:ccf4439733a49dd9e5b1111c67996b55ddcd140743caa7c7b864513514804c6d
+size 348647801

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ff80f475e5501058925745c38f3106d432f30979e70dd1ff6f5ce6a3361d5464
+oid sha256:228403db0d84d4bb8a54b1e41d7d13df513837db8312ac0d36e1ce88e662c599
 size 3579