config.json

{
    "vae": {
        "_class_name": "AutoencoderKL",
        "_diffusers_version": "0.26.1",
        "_name_or_path": "damo-vilab/text-to-video-ms-1.7b",
        "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,
        "layers_per_block": 2,
        "norm_num_groups": 32,
        "out_channels": 3,
        "sample_size": 512,
        "scaling_factor": 0.18215,
        "up_block_types": [
            "UpDecoderBlock2D",
            "UpDecoderBlock2D",
            "UpDecoderBlock2D",
            "UpDecoderBlock2D"
        ]
    },
    "unet": {
        "_class_name": "UNet3DConditionModel",
        "_diffusers_version": "0.26.1",
        "_name_or_path": "damo-vilab/text-to-video-ms-1.7b",
        "act_fn": "silu",
        "attention_head_dim": 64,
        "block_out_channels": [
            320,
            640,
            1280,
            1280
        ],
        "concat": false,
        "cross_attention_dim": 1024,
        "down_block_types": [
            "CrossAttnDownBlock3D",
            "CrossAttnDownBlock3D",
            "CrossAttnDownBlock3D",
            "DownBlock3D"
        ],
        "downsample_padding": 1,
        "in_channels": 4,
        "layers_per_block": 2,
        "merging_mode": "attention_cross_attention",
        "mid_block_scale_factor": 1,
        "norm_eps": 0.00001,
        "norm_num_groups": 32,
        "out_channels": 4,
        "sample_size": 32,
        "up_block_types": [
            "UpBlock3D",
            "CrossAttnUpBlock3D",
            "CrossAttnUpBlock3D",
            "CrossAttnUpBlock3D"
        ],
        "use_channel_expansion": false,
        "use_fps_conditioning": false,
        "use_image_embedding": true,
        "use_image_tokens": true,
        "use_repeat_context_img": true
    },
    "resampler": {
        "_class_name": "ImageEmbeddingContextResampler",
        "_diffusers_version": "0.26.1",
        "cross_attention_dim": 1024,
        "expansion_factor": 16,
        "inner_dim": 1280
    },
    "controlnet": {
        "_class_name": "ControlNetModel",
        "_diffusers_version": "0.26.1",
        "act_fn": "silu",
        "attention_head_dim": 64,
        "block_out_channels": [
            320,
            640,
            1280,
            1280
        ],
        "class_embed_type": null,
        "conditioning_embedding_out_channels": [
            32,
            96,
            256,
            512
        ],
        "controlnet_conditioning_channel_order": "rgb",
        "cross_attention_dim": 1024,
        "down_block_types": [
            "CrossAttnDownBlock3D",
            "CrossAttnDownBlock3D",
            "CrossAttnDownBlock3D",
            "DownBlock3D"
        ],
        "downsample_controlnet_cond": true,
        "downsample_padding": 1,
        "flip_sin_to_cos": true,
        "frame_expansion": "none",
        "freq_shift": 0,
        "global_pool_conditions": false,
        "in_channels": 4,
        "layers_per_block": 2,
        "merging_mode": "addition",
        "mid_block_scale_factor": 1,
        "norm_eps": 0.00001,
        "norm_num_groups": 32,
        "num_class_embeds": null,
        "num_frames": 8,
        "num_frames_conditioning": 8,
        "num_tranformers": 1,
        "only_cross_attention": false,
        "projection_class_embeddings_input_dim": null,
        "resnet_time_scale_shift": "default",
        "upcast_attention": false,
        "use_controlnet_mask": false,
        "use_image_embedding": false,
        "use_image_encoder_normalization": false,
        "use_image_tokens": false,
        "use_linear_projection": false,
        "use_repeat_context_img": true,
        "zero_conv_mode": "Identity"
    },
    "text_encoder": {
        "_name_or_path": "damo-vilab/text-to-video-ms-1.7b",
        "architectures": [
            "CLIPTextModel"
        ],
        "attention_dropout": 0,
        "bos_token_id": 0,
        "dropout": 0,
        "eos_token_id": 2,
        "hidden_act": "gelu",
        "hidden_size": 1024,
        "initializer_factor": 1,
        "initializer_range": 0.02,
        "intermediate_size": 4096,
        "layer_norm_eps": 0.00001,
        "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": "float32",
        "transformers_version": "4.39.0",
        "vocab_size": 49408
    },
    "tokenizer": {
        "model": "ali-vilab/text-to-video-ms-1.7b",
        "subfolder": "tokenizer"
    },
    "scheduler": {
        "_class_name": "DDIMScheduler",
        "_diffusers_version": "0.26.1",
        "beta_end": 0.012,
        "beta_schedule": "scaled_linear",
        "beta_start": 0.00085,
        "clip_sample": false,
        "clip_sample_range": 1,
        "dynamic_thresholding_ratio": 0.995,
        "num_train_timesteps": 1000,
        "prediction_type": "epsilon",
        "rescale_betas_zero_snr": false,
        "sample_max_value": 1,
        "set_alpha_to_one": false,
        "skip_prk_steps": true,
        "steps_offset": 1,
        "thresholding": false,
        "timestep_spacing": "leading",
        "trained_betas": null
    },
    "num_frames": 16,
    "num_frames_conditioning": 8,
    "temp_attend_on_uncond_include_past": false,
    "temp_attend_on_neighborhood_of_condition_frames": false,
    "temporal_self_attention_mask_included_itself": false,
    "temporal_self_attention_only_on_conditioning": false,
    "spatial_attend_on_condition_frames": false,
    "image_encoder_version": "laion2b_s32b_b79k"
}