Upload model

config.json

@@ -2,7 +2,6 @@
   "architectures": [
     "Sam2VideoModel"
   ],
-  "binarize_mask_from_pts_for_mem_enc": true,
   "enable_occlusion_spatial_embedding": true,
   "enable_temporal_pos_encoding_for_object_pointers": true,
   "fill_hole_area": 8,
@@ -30,9 +29,6 @@
   "mask_downsampler_stride": 2,
   "mask_downsampler_total_stride": 16,
   "max_object_pointers_in_encoder": 16,
-  "memory_attention_apply_pe_at_cross_attn_keys": true,
-  "memory_attention_apply_pe_at_cross_attn_queries": false,
-  "memory_attention_apply_pe_at_self_attn": false,
   "memory_attention_downsample_rate": 1,
   "memory_attention_dropout": 0.1,
   "memory_attention_feed_forward_hidden_act": "relu",
@@ -55,7 +51,6 @@
   "memory_fuser_layer_scale_init_value": 1e-06,
   "memory_fuser_num_layers": 2,
   "memory_fuser_padding": 3,
-  "memory_fuser_use_depthwise_conv": true,
   "model_type": "sam2",
   "multimask_max_pt_num": 1,
   "multimask_min_pt_num": 0,
@@ -64,7 +59,6 @@
   "non_overlap_masks": false,
   "non_overlap_masks_for_mem_enc": false,
   "num_maskmem": 7,
-  "preserve_temporal_direction_in_object_pointers": true,
   "project_temporal_pos_encoding_in_object_pointers": true,
   "prompt_encoder_config": {
     "hidden_act": "gelu",
@@ -80,7 +74,7 @@
   "sigmoid_bias_for_mem_enc": -10.0,
   "sigmoid_scale_for_mem_enc": 20.0,
   "torch_dtype": "float32",
-  "transformers_version": "4.55.0.dev0",
+  "transformers_version": "4.56.0.dev0",
   "vision_config": {
     "backbone_channel_list": [
       768,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c53ac6147ba7add4d91c213ee5b607331fd1cac5c4982f5cc28d3fe5caa57e1
-size 155908440
+oid sha256:8f64d3461b06b668002929bc378e57f1058fdca616c8e3d300f7e738f3456e42
+size 155908128