Initial commit: Add FaceLift model weights

README.md

@@ -3,3 +3,29 @@ license: other
 license_name: adobe-research-license
 license_link: LICENSE
 ---
+
+# FaceLift: Single Image 3D Face Reconstruction
+
+**ICCV 2025** | [Paper](https://arxiv.org/abs/2412.17812) | [Code](https://github.com/weijielyu/FaceLift) | [Project Page](https://www.wlyu.me/FaceLift/)
+
+FaceLift transforms a single facial image into a high-fidelity 3D Gaussian head representation. This repository contains the pretrained model checkpoints.
+
+For usage instructions, see the [official code repository](https://github.com/weijielyu/FaceLift).
+
+## Citation
+
+```bibtex
+@misc{lyu2025facelift,
+    title={FaceLift: Learning Generalizable Single Image 3D Face Reconstruction from Synthetic Heads}, 
+    author={Weijie Lyu and Yi Zhou and Ming-Hsuan Yang and Zhixin Shu},
+    year={2025},
+    eprint={2412.17812},
+    archivePrefix={arXiv},
+    primaryClass={cs.CV},
+    url={https://arxiv.org/abs/2412.17812}
+}
+```
+
+## License
+
+Model weights are licensed under the Adobe Research License. See [LICENSE](LICENSE) for details.

gslrm/ckpt_0000000000021125.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:705e9aa37c7ea964bb06f25db060c12676a93d9ce6a365ac8b50128a3dc12662
+size 3723107450

mvdiffusion/pipeckpts/feature_extractor/preprocessor_config.json

@@ -0,0 +1,27 @@
+{
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
+  "do_center_crop": true,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "CLIPImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 224
+  }
+}

mvdiffusion/pipeckpts/image_encoder/config.json

@@ -0,0 +1,23 @@
+{
+  "_name_or_path": "stabilityai/stable-diffusion-2-1-unclip",
+  "architectures": [
+    "CLIPVisionModelWithProjection"
+  ],
+  "attention_dropout": 0.0,
+  "dropout": 0.0,
+  "hidden_act": "gelu",
+  "hidden_size": 1280,
+  "image_size": 224,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "intermediate_size": 5120,
+  "layer_norm_eps": 1e-05,
+  "model_type": "clip_vision_model",
+  "num_attention_heads": 16,
+  "num_channels": 3,
+  "num_hidden_layers": 32,
+  "patch_size": 14,
+  "projection_dim": 1024,
+  "torch_dtype": "float16",
+  "transformers_version": "4.44.2"
+}

mvdiffusion/pipeckpts/image_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae616c24393dd1854372b0639e5541666f7521cbe219669255e865cb7f89466a
+size 1264217240

mvdiffusion/pipeckpts/image_noising_scheduler/scheduler_config.json

@@ -0,0 +1,19 @@
+{
+  "_class_name": "DDPMScheduler",
+  "_diffusers_version": "0.30.3",
+  "beta_end": 0.02,
+  "beta_schedule": "squaredcos_cap_v2",
+  "beta_start": 0.0001,
+  "clip_sample": true,
+  "clip_sample_range": 1.0,
+  "dynamic_thresholding_ratio": 0.995,
+  "num_train_timesteps": 1000,
+  "prediction_type": "epsilon",
+  "rescale_betas_zero_snr": false,
+  "sample_max_value": 1.0,
+  "steps_offset": 0,
+  "thresholding": false,
+  "timestep_spacing": "leading",
+  "trained_betas": null,
+  "variance_type": "fixed_small"
+}

mvdiffusion/pipeckpts/image_normalizer/config.json

@@ -0,0 +1,6 @@
+{
+  "_class_name": "StableUnCLIPImageNormalizer",
+  "_diffusers_version": "0.30.3",
+  "_name_or_path": "stabilityai/stable-diffusion-2-1-unclip",
+  "embedding_dim": 1024
+}

mvdiffusion/pipeckpts/image_normalizer/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a726e7159498c008a99fad02cb52280065ee93432f2af192a8511b7ddcf997a3
+size 8368

mvdiffusion/pipeckpts/model_index.json

@@ -0,0 +1,40 @@
+{
+  "_class_name": "StableUnCLIPImg2ImgPipeline",
+  "_diffusers_version": "0.30.3",
+  "feature_extractor": [
+    "transformers",
+    "CLIPImageProcessor"
+  ],
+  "image_encoder": [
+    "transformers",
+    "CLIPVisionModelWithProjection"
+  ],
+  "image_noising_scheduler": [
+    "diffusers",
+    "DDPMScheduler"
+  ],
+  "image_normalizer": [
+    "stable_diffusion",
+    "StableUnCLIPImageNormalizer"
+  ],
+  "scheduler": [
+    "diffusers",
+    "DDIMScheduler"
+  ],
+  "text_encoder": [
+    "transformers",
+    "CLIPTextModel"
+  ],
+  "tokenizer": [
+    "transformers",
+    "CLIPTokenizer"
+  ],
+  "unet": [
+    "mvdiffusion.models.unet_mv2d_condition",
+    "UNetMV2DConditionModel"
+  ],
+  "vae": [
+    "diffusers",
+    "AutoencoderKL"
+  ]
+}

mvdiffusion/pipeckpts/scheduler/scheduler_config.json

@@ -0,0 +1,20 @@
+{
+  "_class_name": "DDIMScheduler",
+  "_diffusers_version": "0.30.3",
+  "beta_end": 0.012,
+  "beta_schedule": "scaled_linear",
+  "beta_start": 0.00085,
+  "clip_sample": false,
+  "clip_sample_range": 1.0,
+  "dynamic_thresholding_ratio": 0.995,
+  "num_train_timesteps": 1000,
+  "prediction_type": "v_prediction",
+  "rescale_betas_zero_snr": false,
+  "sample_max_value": 1.0,
+  "set_alpha_to_one": false,
+  "skip_prk_steps": true,
+  "steps_offset": 1,
+  "thresholding": false,
+  "timestep_spacing": "leading",
+  "trained_betas": null
+}

mvdiffusion/pipeckpts/text_encoder/config.json

@@ -0,0 +1,25 @@
+{
+  "_name_or_path": "stabilityai/stable-diffusion-2-1-unclip",
+  "architectures": [
+    "CLIPTextModel"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "dropout": 0.0,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_size": 1024,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 77,
+  "model_type": "clip_text_model",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 23,
+  "pad_token_id": 1,
+  "projection_dim": 512,
+  "torch_dtype": "float16",
+  "transformers_version": "4.44.2",
+  "vocab_size": 49408
+}

mvdiffusion/pipeckpts/text_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc1827c465450322616f06dea41596eac7d493f4e95904dcb51f0fc745c4e13f
+size 680820392

mvdiffusion/pipeckpts/tokenizer/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<|startoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "!",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

mvdiffusion/pipeckpts/tokenizer/tokenizer_config.json

@@ -0,0 +1,38 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "!",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "49406": {
+      "content": "<|startoftext|>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "49407": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|startoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "do_lower_case": true,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "model_max_length": 77,
+  "pad_token": "!",
+  "tokenizer_class": "CLIPTokenizer",
+  "unk_token": "<|endoftext|>"
+}

mvdiffusion/pipeckpts/unet/config.json

@@ -0,0 +1,82 @@
+{
+  "_class_name": "UNetMV2DConditionModel",
+  "_diffusers_version": "0.30.3",
+  "_name_or_path": "stabilityai/stable-diffusion-2-1-unclip",
+  "act_fn": "silu",
+  "addition_channels": [
+    1280,
+    1280,
+    1280
+  ],
+  "addition_downsample": false,
+  "addition_embed_type": null,
+  "addition_embed_type_num_heads": 64,
+  "addition_time_embed_dim": null,
+  "attention_head_dim": [
+    5,
+    10,
+    20,
+    20
+  ],
+  "block_out_channels": [
+    320,
+    640,
+    1280,
+    1280
+  ],
+  "cd_attention_last": false,
+  "cd_attention_mid": false,
+  "center_input_sample": false,
+  "class_embed_type": "projection",
+  "class_embeddings_concat": false,
+  "conv_in_kernel": 3,
+  "conv_out_kernel": 3,
+  "cross_attention_dim": 1024,
+  "cross_attention_norm": null,
+  "down_block_types": [
+    "CrossAttnDownBlockMV2D",
+    "CrossAttnDownBlockMV2D",
+    "CrossAttnDownBlockMV2D",
+    "DownBlock2D"
+  ],
+  "downsample_padding": 1,
+  "dual_cross_attention": false,
+  "encoder_hid_dim": null,
+  "encoder_hid_dim_type": null,
+  "flip_sin_to_cos": true,
+  "freq_shift": 0,
+  "in_channels": 8,
+  "layers_per_block": 2,
+  "mid_block_only_cross_attention": null,
+  "mid_block_scale_factor": 1,
+  "mid_block_type": "UNetMidBlockMV2DCrossAttn",
+  "multiview_attention": true,
+  "norm_eps": 1e-05,
+  "norm_num_groups": 32,
+  "num_attention_heads": null,
+  "num_class_embeds": null,
+  "num_views": 6,
+  "only_cross_attention": false,
+  "out_channels": 4,
+  "projection_class_embeddings_input_dim": 2048,
+  "resnet_out_scale_factor": 1.0,
+  "resnet_skip_time_act": false,
+  "resnet_time_scale_shift": "default",
+  "sample_size": 64,
+  "selfattn_block": "custom",
+  "sparse_mv_attention": true,
+  "time_cond_proj_dim": null,
+  "time_embedding_act_fn": null,
+  "time_embedding_dim": null,
+  "time_embedding_type": "positional",
+  "timestep_post_act": null,
+  "transformer_layers_per_block": 1,
+  "up_block_types": [
+    "UpBlock2D",
+    "CrossAttnUpBlockMV2D",
+    "CrossAttnUpBlockMV2D",
+    "CrossAttnUpBlockMV2D"
+  ],
+  "upcast_attention": true,
+  "use_linear_projection": true
+}

mvdiffusion/pipeckpts/unet/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2389a09216381a2020ac751c58735408feb4e76797c6f2393e66c7a0cedc1629
+size 3480822704

mvdiffusion/pipeckpts/vae/config.json

@@ -0,0 +1,38 @@
+{
+  "_class_name": "AutoencoderKL",
+  "_diffusers_version": "0.30.3",
+  "_name_or_path": "stabilityai/stable-diffusion-2-1-unclip",
+  "act_fn": "silu",
+  "block_out_channels": [
+    128,
+    256,
+    512,
+    512
+  ],
+  "down_block_types": [
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D"
+  ],
+  "force_upcast": true,
+  "in_channels": 3,
+  "latent_channels": 4,
+  "latents_mean": null,
+  "latents_std": null,
+  "layers_per_block": 2,
+  "mid_block_add_attention": true,
+  "norm_num_groups": 32,
+  "out_channels": 3,
+  "sample_size": 768,
+  "scaling_factor": 0.18215,
+  "shift_factor": null,
+  "up_block_types": [
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D"
+  ],
+  "use_post_quant_conv": true,
+  "use_quant_conv": true
+}

mvdiffusion/pipeckpts/vae/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e4c08995484ee61270175e9e7a072b66a6e4eeb5f0c266667fe1f45b90daf9a
+size 167335342