Push model using huggingface_hub.
Browse files- config.json +39 -0
config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cfg": {
|
| 3 |
+
"decoder": {
|
| 4 |
+
"d_model": 768,
|
| 5 |
+
"freeze_text_encoder": true,
|
| 6 |
+
"input_feat_dim": 256,
|
| 7 |
+
"num_decoder_layers": 8,
|
| 8 |
+
"num_queries": 256,
|
| 9 |
+
"text_encoder": "clip",
|
| 10 |
+
"transformer_dim_feedforward": 3072,
|
| 11 |
+
"transformer_dropout": 0.1,
|
| 12 |
+
"transformer_max_drop_path": 0.0,
|
| 13 |
+
"transformer_n_heads": 12,
|
| 14 |
+
"transformer_use_checkpointing": true
|
| 15 |
+
},
|
| 16 |
+
"encoder": {
|
| 17 |
+
"embed_dim": 256,
|
| 18 |
+
"input_feat_dim": 1536,
|
| 19 |
+
"num_rgb_harmonic_functions": 16,
|
| 20 |
+
"ptv3_args": {
|
| 21 |
+
"dec_channels": [
|
| 22 |
+
256,
|
| 23 |
+
256,
|
| 24 |
+
512,
|
| 25 |
+
512
|
| 26 |
+
],
|
| 27 |
+
"enc_channels": [
|
| 28 |
+
256,
|
| 29 |
+
256,
|
| 30 |
+
512,
|
| 31 |
+
512,
|
| 32 |
+
1024
|
| 33 |
+
],
|
| 34 |
+
"in_channels": 256
|
| 35 |
+
},
|
| 36 |
+
"rgb_proj_dim": 256
|
| 37 |
+
}
|
| 38 |
+
}
|
| 39 |
+
}
|