Update tiny models for VJEPA2Model
Browse files- config.json +31 -0
- model.safetensors +3 -0
- video_preprocessor_config.json +48 -0
config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"VJEPA2Model"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"crop_size": 16,
|
| 8 |
+
"drop_path_rate": 0.0,
|
| 9 |
+
"dtype": "float32",
|
| 10 |
+
"frames_per_clip": 2,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_size": 32,
|
| 13 |
+
"in_chans": 3,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"layer_norm_eps": 1e-06,
|
| 16 |
+
"mlp_ratio": 1.0,
|
| 17 |
+
"model_type": "vjepa2",
|
| 18 |
+
"num_attention_heads": 2,
|
| 19 |
+
"num_hidden_layers": 2,
|
| 20 |
+
"num_pooler_layers": 3,
|
| 21 |
+
"patch_size": 16,
|
| 22 |
+
"pred_hidden_size": 32,
|
| 23 |
+
"pred_mlp_ratio": 4.0,
|
| 24 |
+
"pred_num_attention_heads": 2,
|
| 25 |
+
"pred_num_hidden_layers": 2,
|
| 26 |
+
"pred_num_mask_tokens": 10,
|
| 27 |
+
"pred_zero_init_mask_tokens": true,
|
| 28 |
+
"qkv_bias": true,
|
| 29 |
+
"transformers_version": "5.16.0.dev0",
|
| 30 |
+
"tubelet_size": 2
|
| 31 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f635f52afe9a21eddc328e53545d0aa72495a8541c64e50858b3f44c5ec0261b
|
| 3 |
+
size 367656
|
video_preprocessor_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_size": {
|
| 3 |
+
"height": 256,
|
| 4 |
+
"width": 256
|
| 5 |
+
},
|
| 6 |
+
"data_format": "channels_first",
|
| 7 |
+
"default_to_square": true,
|
| 8 |
+
"do_center_crop": true,
|
| 9 |
+
"do_normalize": true,
|
| 10 |
+
"do_rescale": true,
|
| 11 |
+
"do_resize": true,
|
| 12 |
+
"image_mean": [
|
| 13 |
+
0.485,
|
| 14 |
+
0.456,
|
| 15 |
+
0.406
|
| 16 |
+
],
|
| 17 |
+
"image_std": [
|
| 18 |
+
0.229,
|
| 19 |
+
0.224,
|
| 20 |
+
0.225
|
| 21 |
+
],
|
| 22 |
+
"model_valid_processing_keys": [
|
| 23 |
+
"do_convert_rgb",
|
| 24 |
+
"do_resize",
|
| 25 |
+
"size",
|
| 26 |
+
"size_divisor",
|
| 27 |
+
"default_to_square",
|
| 28 |
+
"resample",
|
| 29 |
+
"do_rescale",
|
| 30 |
+
"rescale_factor",
|
| 31 |
+
"do_normalize",
|
| 32 |
+
"image_mean",
|
| 33 |
+
"image_std",
|
| 34 |
+
"do_pad",
|
| 35 |
+
"do_center_crop",
|
| 36 |
+
"crop_size",
|
| 37 |
+
"data_format",
|
| 38 |
+
"input_data_format",
|
| 39 |
+
"device"
|
| 40 |
+
],
|
| 41 |
+
"resample": 2,
|
| 42 |
+
"rescale_factor": 0.00392156862745098,
|
| 43 |
+
"return_metadata": false,
|
| 44 |
+
"size": {
|
| 45 |
+
"shortest_edge": 292
|
| 46 |
+
},
|
| 47 |
+
"video_processor_type": "VJEPA2VideoProcessor"
|
| 48 |
+
}
|