Hugging Face's logo

laion
/
DALLE2-PyTorch

Model card Files Community
6
DALLE2-PyTorch
File size: 4,747 Bytes 
8a996ed
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
669093a
af5f7d2
8a996ed
 
 
 
 
 
 
 
af5f7d2
8a996ed
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
 
{
    "seed": 1,
    
    "decoder": {
        "unets": [
            {
                "dim": 32,
                "cond_dim": 512,
                "image_embed_dim": 768,
                "text_embed_dim": 768,
                "cond_on_text_encodings": true,
                "channels": 3,
                "dim_mults": [1, 2, 3, 4],
                "num_resnet_blocks": 4,
                "attn_heads": 8,
                "attn_dim_head": 64,
                "sparse_attn": true,
                "memory_efficient": true,
		        "self_attn": [false, true, true, true]
            },
            {
                "dim": 32,
                "cond_dim": 512,
                "image_embed_dim": 768,
                "text_embed_dim": 768,
                "cond_on_text_encodings": true,
                "channels": 3,
                "dim_mults": [1, 2, 3, 4],
                "num_resnet_blocks": 4,
                "attn_heads": 8,
                "attn_dim_head": 64,
                "sparse_attn": true,
                "memory_efficient": true,
		        "self_attn": [false, true, true, true]
            },
            {
                "dim": 192,
                "cond_dim": 512,
                "image_embed_dim": 768,
                "text_embed_dim": 768,
                "cond_on_text_encodings": true,
                "init_cross_embed": false,
                "channels": 3,
                "dim_mults": [1, 2, 3, 4],
                "num_resnet_blocks": 3,
                "attn_heads": 8,
                "attn_dim_head": 64,
                "sparse_attn": false,
                "memory_efficient": true,
                "self_attn": [false, false, false, false]
            }
        ],
        "clip": {
            "make": "openai",
            "model": "ViT-L/14"
        },
        "image_sizes": [64, 256, 1024],
        "random_crop_sizes": [null, null, 256],
        "channels": 3,
        "timesteps": 1000,
        "loss_type": "l2",
        "beta_schedule": ["cosine", "cosine", "cosine"],
        "learned_variance": true,
        "text_cond_drop_prob": 0.0,
        "image_cond_drop_prob": 0.0
    },
    "data": {
        "webdataset_base_url": "pipe:aws s3 cp --quiet s3://s-datasets/laion-high-resolution/{}.tar -",
        "num_workers": 6,
        "batch_size": 8,
        "start_shard": 0,
        "end_shard": 17535,
        "shard_width": 5,
        "index_width": 4,
        "splits": {
            "train": 0.75,
            "val": 0.15,
            "test": 0.1
        },
        "shuffle_train": false,
        "resample_train": true,
        "preprocessing": {
            "RandomResizedCrop": {
                "size": [1024, 1024],
                "scale": [0.75, 1.0],
                "ratio": [1.0, 1.0]
            },
            "ToTensor": true
        }
    },
    "train": {
        "epochs": 1000,
        "lr": 1.2e-4,
        "wd": 0.0,
        "max_grad_norm": 0.5,
        "save_every_n_samples": 2000000,
        "n_sample_images": 2,
        "device": "cuda:0",
        "epoch_samples": 10000000,
        "validation_samples": 100000,
        "use_ema": true,
        "ema_beta": 0.9999,
        "unet_training_mask": [false, false, true]
    },
    "evaluate": {
        "n_evaluation_samples": 2,
        "FID": {
            "feature": 64
        },
        "LPIPS": {
            "net_type": "vgg",
            "reduction": "mean"
        }
    },
    "tracker": {
        "data_path": "/fsx/aidan/new/multinode/experiments/decoder_1024/.tracker-data",
        "overwrite_data_path": true,

        "log": {
            "log_type": "wandb",

            "wandb_entity": "Veldrovive",
            "wandb_project": "upsamplers_1024",
            "wandb_resume": false,

            "auto_resume": true,
            "verbose": true
        },

        "load": {
            "load_from": null,
            "only_auto_resume": true,
	        "file_path": "/fsx/aidan/new/multinode/experiments/decoder_1024/models/checkpoints/latest.pth"
        },

        "save": [
        {
            "save_to": "huggingface",
            "huggingface_repo": "laion/DALLE2-PyTorch",
            
            "save_meta_to": "upsampler/1024/v1.0.3/",
            "save_latest_to": "upsampler/1024/v1.0.3/latest.pth",

            "save_type": "model"
        },{
            "save_to": "huggingface",
            "huggingface_repo": "laion/DALLE2-PyTorch",
            
            "save_latest_to": "upsampler/1024/v1.0.2/checkpoints/latest.pth",

            "save_type": "checkpoint"
        },{
            "save_to": "local",
            "save_latest_to": "/fsx/aidan/new/multinode/experiments/decoder_1024/models/checkpoints/latest.pth",

            "save_type": "checkpoint"
        }]
    }
}