Update PyTorch model

config.json

@@ -3,22 +3,26 @@
     "PerceiverForMaskedLM"
   ],
   "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": 768,
   "hidden_act": "gelu",
   "hidden_dropout_prob": 0.1,
+  "image_size": 56,
   "initializer_range": 0.02,
   "layer_norm_eps": 1e-12,
   "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,
   "position_embedding_init_scale": 0.02,
   "qk_channels": 256,
+  "samples_per_patch": 16,
   "self_attention_widening_factor": 1,
   "seq_len": 2048,
   "torch_dtype": "float32",
@@ -26,7 +30,7 @@
     368,
     496
   ],
-  "transformers_version": "4.11.0.dev0",
+  "transformers_version": "4.12.0.dev0",
   "use_query_residual": true,
   "v_channels": 1280,
   "vocab_size": 262

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:405418b61079daa8780df29d31fa88cc8edb2e36ad2b53543c65fe0675df965b
-size 804615663
+oid sha256:941759e1c7489c5ffc7a922551a29e93581868972007aed48ae40e4a525ecba7
+size 804615599