Upload model

config.json

@@ -4,7 +4,6 @@
   ],
   "enable_occlusion_spatial_embedding": true,
   "enable_temporal_pos_encoding_for_object_pointers": true,
-  "fill_hole_area": 8,
   "image_size": 1024,
   "initializer_range": 0.02,
   "mask_decoder_config": {
@@ -51,15 +50,12 @@
   "memory_fuser_layer_scale_init_value": 1e-06,
   "memory_fuser_num_layers": 2,
   "memory_fuser_padding": 3,
-  "model_type": "sam2",
+  "model_type": "sam2_video",
   "multimask_max_pt_num": 1,
   "multimask_min_pt_num": 0,
   "multimask_output_for_tracking": true,
   "multimask_output_in_sam": true,
-  "non_overlap_masks": false,
-  "non_overlap_masks_for_mem_enc": false,
   "num_maskmem": 7,
-  "project_temporal_pos_encoding_in_object_pointers": true,
   "prompt_encoder_config": {
     "hidden_act": "gelu",
     "hidden_size": 256,
@@ -88,14 +84,25 @@
       "architectures": null,
       "bad_words_ids": null,
       "begin_suppress_tokens": null,
+      "blocks_per_stage": [
+        2,
+        6,
+        36,
+        4
+      ],
       "bos_token_id": null,
       "chunk_size_feed_forward": 0,
       "cross_attention_hidden_size": null,
       "decoder_start_token_id": null,
-      "dim_mul": 2.0,
       "diversity_penalty": 0.0,
       "do_sample": false,
       "early_stopping": false,
+      "embed_dim_per_stage": [
+        144,
+        288,
+        576,
+        1152
+      ],
       "encoder_no_repeat_ngram_size": 0,
       "eos_token_id": null,
       "exponential_decay_length_penalty": null,
@@ -107,7 +114,6 @@
         33,
         43
       ],
-      "head_mul": 2.0,
       "hidden_act": "gelu",
       "hidden_size": 144,
       "id2label": {
@@ -129,9 +135,16 @@
       "length_penalty": 1.0,
       "max_length": 20,
       "min_length": 0,
+      "mlp_ratio": 4.0,
       "model_type": "sam2_hiera_det_model",
       "no_repeat_ngram_size": 0,
-      "num_attention_heads": 2,
+      "num_attention_heads": 1,
+      "num_attention_heads_per_stage": [
+        2,
+        4,
+        8,
+        16
+      ],
       "num_beam_groups": 1,
       "num_beams": 1,
       "num_channels": 3,
@@ -165,12 +178,6 @@
       "return_dict": true,
       "return_dict_in_generate": false,
       "sep_token_id": null,
-      "stages": [
-        2,
-        6,
-        36,
-        4
-      ],
       "suppress_tokens": null,
       "task_specific_params": null,
       "temperature": 1.0,
@@ -188,7 +195,7 @@
         7,
         7
       ],
-      "window_spec": [
+      "window_size_per_stage": [
         8,
         4,
         16,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:506e3c92baf38f4fbb6c1da8b6da76f151d6a6724f54f10757fd3ed918b74f1f
-size 897897472
+oid sha256:dc407dce21301fd94abb395c5099b4f2c455fdc8a8f261ac3d0ea6d4cd197230
+size 897897416