First commit
Browse files- config.json +37 -0
- pytorch_model.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"PerceiverForOpticalFlow"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"audio_samples_per_frame": 1920,
|
| 7 |
+
"cross_attention_shape_for_attention": "kv",
|
| 8 |
+
"cross_attention_widening_factor": 1,
|
| 9 |
+
"d_latents": 512,
|
| 10 |
+
"d_model": 322,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"image_size": 56,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"layer_norm_eps": 1e-12,
|
| 16 |
+
"model_type": "perceiver",
|
| 17 |
+
"num_blocks": 1,
|
| 18 |
+
"num_cross_attention_heads": 1,
|
| 19 |
+
"num_frames": 16,
|
| 20 |
+
"num_latents": 2048,
|
| 21 |
+
"num_self_attends_per_block": 24,
|
| 22 |
+
"num_self_attention_heads": 16,
|
| 23 |
+
"position_embedding_init_scale": 0.02,
|
| 24 |
+
"qk_channels": null,
|
| 25 |
+
"samples_per_patch": 16,
|
| 26 |
+
"self_attention_widening_factor": 1,
|
| 27 |
+
"seq_len": 2048,
|
| 28 |
+
"torch_dtype": "float32",
|
| 29 |
+
"train_size": [
|
| 30 |
+
368,
|
| 31 |
+
496
|
| 32 |
+
],
|
| 33 |
+
"transformers_version": "4.11.0.dev0",
|
| 34 |
+
"use_query_residual": true,
|
| 35 |
+
"v_channels": null,
|
| 36 |
+
"vocab_size": 262
|
| 37 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3633a73cec625d7d5fcf77a087893d8035da26e74f4083a99b45db35734455dd
|
| 3 |
+
size 164398575
|