Upload 4 files

audiocaps_base.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1c435b1cf05a2b0058dae6f096c4eb4e71c685a19754ed84ea1ee812257434b
+size 55293225

audiocaps_config.yaml

@@ -0,0 +1,30 @@
+tokenizer:
+    type: text_tokenizer.DictTokenizer
+    args:
+        max_length: 20
+
+target_sr: 16000
+
+model:
+    args:
+        shared_dim: 1024
+        tchr_dim: 768
+    model:
+        args: {}
+        decoder:
+            args:
+                attn_emb_dim: 1408
+                dropout: 0.2
+                emb_dim: 256
+                fc_emb_dim: 1408
+                nlayers: 2
+                tie_weights: true
+                vocab_size: 4981
+            type: models.transformer_decoder.TransformerDecoder
+        encoder:
+            args:
+                freeze: false
+                pretrained: true
+            type: models.cnn_encoder.EfficientNetB2
+        type: models.transformer_model.TransformerModel
+    type: models.kd_wrapper.ContraEncoderKdWrapper

clotho_base.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:694c9e7139be7ec5aff2153d1af980d6bc305403a76be0d8940481579ea51483
+size 54651005

clotho_config.yaml

@@ -0,0 +1,30 @@
+tokenizer:
+    type: text_tokenizer.DictTokenizer
+    args:
+        max_length: 20
+
+target_sr: 16000
+
+model:
+    args:
+        shared_dim: 1024
+        tchr_dim: 768
+    model:
+        args: {}
+        decoder:
+            args:
+                attn_emb_dim: 1408
+                dropout: 0.2
+                emb_dim: 256
+                fc_emb_dim: 1408
+                nlayers: 2
+                tie_weights: true
+                vocab_size: 4368
+            type: models.transformer_decoder.TransformerDecoder
+        encoder:
+            args:
+                freeze: false
+                pretrained: true
+            type: models.cnn_encoder.EfficientNetB2
+        type: models.transformer_model.TransformerModel
+    type: models.kd_wrapper.ContraEncoderKdWrapper