add model and config file

Add config.json and model file for Perceiver-based emotion recognition model trained on RAVDESS data

config.json

@@ -0,0 +1,61 @@
+{
+  "architectures": [
+    "PerceiverModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "audio_samples_per_frame": 1920,
+  "cross_attention_shape_for_attention": "kv",
+  "cross_attention_widening_factor": 1,
+  "d_latents": 1280,
+  "d_model": 2048,
+  "hidden_act": "gelu",
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2",
+    "3": "LABEL_3",
+    "4": "LABEL_4",
+    "5": "LABEL_5",
+    "6": "LABEL_6",
+    "7": "LABEL_7"
+  },
+  "image_size": 56,
+  "initializer_range": 0.02,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_2": 2,
+    "LABEL_3": 3,
+    "LABEL_4": 4,
+    "LABEL_5": 5,
+    "LABEL_6": 6,
+    "LABEL_7": 7
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 2048,
+  "model_type": "perceiver",
+  "num_blocks": 1,
+  "num_cross_attention_heads": 8,
+  "num_frames": 16,
+  "num_latents": 256,
+  "num_self_attends_per_block": 26,
+  "num_self_attention_heads": 8,
+  "output_shape": [
+    1,
+    16,
+    224,
+    224
+  ],
+  "qk_channels": null,
+  "samples_per_patch": 16,
+  "self_attention_widening_factor": 1,
+  "torch_dtype": "float32",
+  "train_size": [
+    368,
+    496
+  ],
+  "transformers_version": "4.20.1",
+  "use_query_residual": true,
+  "v_channels": null,
+  "vocab_size": 262
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fadacdf305c81fa5a0ec391697835875f931e00b602f04d157cd39d6ed532e6
+size 1132295665