Upload folder using huggingface_hub

config.json

@@ -0,0 +1,59 @@
+{
+  "architectures": [
+    "VASTDiT3"
+  ],
+  "audio_hidden_size": 1152,
+  "audio_in_channels": 8,
+  "audio_input_size": [
+    null,
+    null
+  ],
+  "audio_patch_size": [
+    4,
+    1
+  ],
+  "caption_channels": 4096,
+  "class_dropout_prob": 0.1,
+  "depth": 28,
+  "drop_path": 0.0,
+  "enable_flash_attn": true,
+  "enable_layernorm_kernel": false,
+  "enable_sequence_parallelism": false,
+  "freeze_audio_branch": true,
+  "freeze_video_branch": true,
+  "freeze_y_embedder": true,
+  "hidden_size": 1152,
+  "in_channels": 4,
+  "input_size": [
+    null,
+    null,
+    null
+  ],
+  "input_sq_size": 512,
+  "mlp_ratio": 4.0,
+  "model_max_length": 300,
+  "model_type": "VASTDiT3",
+  "num_heads": 16,
+  "only_infer_audio": false,
+  "only_train_audio": false,
+  "only_train_temporal": false,
+  "patch_size": [
+    1,
+    2,
+    2
+  ],
+  "pred_sigma": true,
+  "qk_norm": true,
+  "skip_y_embedder": false,
+  "spatial_prior_len": 32,
+  "st_prior_channel": 128,
+  "temporal_prior_len": 32,
+  "torch_dtype": "bfloat16",
+  "train_st_prior_attn": true,
+  "train_va_cross_attn": true,
+  "transformers_version": "4.39.3",
+  "weight_init_from": [
+    "./checkpoints/JavisDiT-audio-v0",
+    "./checkpoints/OpenSora-STDiT-v3/model.safetensors"
+  ]
+}

pytorch_model-00001.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b7d7f3e974f97aa3f0baada67acf3dea7e0f07170b25c3a5bd711fbcc97957c
+size 1070739616

pytorch_model-00002.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44aa3a44fea82c261141ce239bf5eb543df43d49f60786894c35ed1c4f062877
+size 1070893412

pytorch_model-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:687337c9c2e91acc40cfb05ac5e320f1319227ad73fd6485bb59b36f7feb9237
+size 1072626064

pytorch_model-00004.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2992414f3a9f0350c7dd16da8147fe0135f24b3f632ef27a509c38862c04f149
+size 1073558938

pytorch_model-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f3adf60c440ffad9c09c8b04df0fbb3e35ddce5f448520e311d9f02b0ef9001
+size 1073642040

pytorch_model-00006.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff17a37ad45e5c7f3b8d1f4e2875e68b3804515e6e45c01e3745ebd0796ae58a
+size 1073723768

pytorch_model-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fe6d177f16105770036ab4c2011cb99af48691ba27cc450d925ddd44a2c158f
+size 1010299692