lvfm

lvfm/FMvT-S2-16f8/config.json

@@ -0,0 +1,36 @@
+{
+  "_class_name": "DiffuserSTDiT",
+  "_diffusers_version": "0.30.3",
+  "_name_or_path": "experiments/lvfm_STDiT-S_16f8_all/checkpoint-1000000/denoiser_ema",
+  "caption_channels": 1,
+  "class_dropout_prob": 0.0,
+  "decay": 0.9999,
+  "depth": 12,
+  "drop_path": 0.0,
+  "enable_flashattn": false,
+  "hidden_size": 384,
+  "in_channels": 32,
+  "input_size": [
+    64,
+    14,
+    14
+  ],
+  "inv_gamma": 1.0,
+  "min_decay": 0.0,
+  "mlp_ratio": 4.0,
+  "model_max_length": 1,
+  "no_temporal_pos_emb": false,
+  "num_heads": 6,
+  "optimization_step": 1,
+  "out_channels": 16,
+  "patch_size": [
+    1,
+    2,
+    2
+  ],
+  "power": 0.6666666666666666,
+  "space_scale": 1.0,
+  "time_scale": 1.0,
+  "update_after_step": 0,
+  "use_ema_warmup": false
+}

lvfm/FMvT-S2-16f8/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:334747fe2ce2230479f477c3b0c9762474558d4bab31045f9b64086e2b5b6293
+size 147609372

lvfm/FMvT-S2-4f4/config.json

@@ -0,0 +1,36 @@
+{
+  "_class_name": "DiffuserSTDiT",
+  "_diffusers_version": "0.30.3",
+  "_name_or_path": "experiments/lvfm_STDiT-S2_4f4_all/checkpoint-980000/denoiser_ema",
+  "caption_channels": 1,
+  "class_dropout_prob": 0.0,
+  "decay": 0.9999,
+  "depth": 12,
+  "drop_path": 0.0,
+  "enable_flashattn": false,
+  "hidden_size": 384,
+  "in_channels": 8,
+  "input_size": [
+    64,
+    28,
+    28
+  ],
+  "inv_gamma": 1.0,
+  "min_decay": 0.0,
+  "mlp_ratio": 4.0,
+  "model_max_length": 1,
+  "no_temporal_pos_emb": false,
+  "num_heads": 6,
+  "optimization_step": 20001,
+  "out_channels": 4,
+  "patch_size": [
+    1,
+    2,
+    2
+  ],
+  "power": 0.6666666666666666,
+  "space_scale": 1.0,
+  "time_scale": 1.0,
+  "update_after_step": 0,
+  "use_ema_warmup": false
+}

lvfm/FMvT-S2-4f4/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d8179f081db6c3af1e9b66476b76b0f63775ed27d7338bad48f79b7063b0cad
+size 147613788

lvfm/FMvT-S4-4f4/config.json

@@ -0,0 +1,36 @@
+{
+  "_class_name": "DiffuserSTDiT",
+  "_diffusers_version": "0.30.3",
+  "_name_or_path": "experiments/lvfm_STDiT-S4_4f4_all/checkpoint-950000/denoiser_ema",
+  "caption_channels": 1,
+  "class_dropout_prob": 0.0,
+  "decay": 0.9999,
+  "depth": 12,
+  "drop_path": 0.0,
+  "enable_flashattn": false,
+  "hidden_size": 384,
+  "in_channels": 8,
+  "input_size": [
+    64,
+    28,
+    28
+  ],
+  "inv_gamma": 1.0,
+  "min_decay": 0.0,
+  "mlp_ratio": 4.0,
+  "model_max_length": 1,
+  "no_temporal_pos_emb": false,
+  "num_heads": 6,
+  "optimization_step": 50001,
+  "out_channels": 4,
+  "patch_size": [
+    1,
+    4,
+    4
+  ],
+  "power": 0.6666666666666666,
+  "space_scale": 1.0,
+  "time_scale": 1.0,
+  "update_after_step": 0,
+  "use_ema_warmup": false
+}

lvfm/FMvT-S4-4f4/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f005f889f06011c343ba866c6742b70553698792a1582d1b674923dcfe61ca0
+size 147609364

lvfm/STUNet-S-16f8/config.json

@@ -0,0 +1,45 @@
+{
+  "_class_name": "UNetSTIC",
+  "_diffusers_version": "0.30.3",
+  "_name_or_path": "experiments/lvfm_UNetSTIC-S_16f8_all/checkpoint-80000/denoiser_ema",
+  "addition_time_embed_dim": 1,
+  "block_out_channels": [
+    64,
+    128,
+    192,
+    256
+  ],
+  "cross_attention_dim": 1,
+  "decay": 0.9999,
+  "down_block_types": [
+    "CrossAttnDownBlockSpatioTemporal",
+    "CrossAttnDownBlockSpatioTemporal",
+    "CrossAttnDownBlockSpatioTemporal",
+    "DownBlockSpatioTemporal"
+  ],
+  "in_channels": 32,
+  "inv_gamma": 1.0,
+  "layers_per_block": 2,
+  "min_decay": 0.0,
+  "num_attention_heads": [
+    8,
+    16,
+    16,
+    32
+  ],
+  "num_frames": 64,
+  "optimization_step": 20001,
+  "out_channels": 16,
+  "power": 0.6666666666666666,
+  "projection_class_embeddings_input_dim": 1,
+  "sample_size": 14,
+  "transformer_layers_per_block": 1,
+  "up_block_types": [
+    "UpBlockSpatioTemporal",
+    "CrossAttnUpBlockSpatioTemporal",
+    "CrossAttnUpBlockSpatioTemporal",
+    "CrossAttnUpBlockSpatioTemporal"
+  ],
+  "update_after_step": 0,
+  "use_ema_warmup": false
+}

lvfm/STUNet-S-16f8/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0b9557a143942a9035d4864c3160bad58e7aea97cd6a3596ed5d845bee77121
+size 183327136

lvfm/STUNet-S-4f4/config.json

@@ -0,0 +1,44 @@
+{
+  "_class_name": "UNetSTIC",
+  "_diffusers_version": "0.30.3",
+  "addition_time_embed_dim": 1,
+  "block_out_channels": [
+    64,
+    128,
+    192,
+    256
+  ],
+  "cross_attention_dim": 1,
+  "decay": 0.9999,
+  "down_block_types": [
+    "CrossAttnDownBlockSpatioTemporal",
+    "CrossAttnDownBlockSpatioTemporal",
+    "CrossAttnDownBlockSpatioTemporal",
+    "DownBlockSpatioTemporal"
+  ],
+  "in_channels": 8,
+  "inv_gamma": 1.0,
+  "layers_per_block": 2,
+  "min_decay": 0.0,
+  "num_attention_heads": [
+    8,
+    16,
+    16,
+    32
+  ],
+  "num_frames": 64,
+  "optimization_step": 50001,
+  "out_channels": 4,
+  "power": 0.6666666666666666,
+  "projection_class_embeddings_input_dim": 1,
+  "sample_size": 28,
+  "transformer_layers_per_block": 1,
+  "up_block_types": [
+    "UpBlockSpatioTemporal",
+    "CrossAttnUpBlockSpatioTemporal",
+    "CrossAttnUpBlockSpatioTemporal",
+    "CrossAttnUpBlockSpatioTemporal"
+  ],
+  "update_after_step": 0,
+  "use_ema_warmup": false
+}

lvfm/STUNet-S-4f4/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:905c3805763567fb7cb113cda48d8cee3296ccecefe213dc9525ba7dbe909a58
+size 183244088