Upload decoder_config.json with huggingface_hub

decoder_config.json

@@ -0,0 +1,116 @@
+{
+    "seed": 1,
+    
+    "decoder": {
+        "unets": [
+            {
+                "dim": 416,
+                "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
+            }
+        ],
+        "clip": {
+            "make": "openai",
+            "model": "ViT-L/14"
+        },
+        "image_sizes": [64],
+        "channels": 3,
+        "timesteps": 1000,
+        "loss_type": "l2",
+        "beta_schedule": ["cosine"],
+        "learned_variance": true
+    },
+    "data": {
+        "webdataset_base_url": "pipe:aws s3 cp --quiet s3://s-datasets/laion-aesthetic/data/laion2B-en-aesthetic/{}.tar -", 
+        "embeddings_url": "s3://s-datasets/laion-aesthetic/ordered_embeddings/",
+        "num_workers": 12,
+        "batch_size": 22,
+        "start_shard": 0,
+        "end_shard": 5247,
+        "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": [64, 64],
+                "scale": [0.75, 1.0],
+                "ratio": [1.0, 1.0]
+            },
+            "ToTensor": true
+        }
+    },
+    "train": {
+        "epochs": 1000,
+        "lr":1.2e-4,
+        "wd": 0.00,
+        "max_grad_norm": 0.5,
+        "save_every_n_samples": 200000,
+        "n_sample_images": 10,
+        "device": "cuda:0",
+        "epoch_samples": 2000000,
+        "validation_samples": 100000,
+        "use_ema": true,
+        "ema_beta": 0.9999,
+        "ema_inv_gamma":0.75,
+        "save_all": false,
+        "save_latest": true,
+        "save_best": true,
+        "unet_training_mask": [true]
+    },
+    "evaluate": {
+        "n_evaluation_samples": 30,
+        "FID": {
+            "feature": 64
+        },
+        "LPIPS": {
+            "net_type": "vgg",
+            "reduction": "mean"
+        }
+    },
+    "tracker": {
+        "data_path": ".tracker-data",
+        "overwrite_data_path": true,
+
+        "log": {
+            "log_type": "wandb",
+
+            "wandb_entity": "nousr_laion",
+            "wandb_project": "dalle2_train_decoder",
+
+            "verbose": true
+        },
+
+        "load": {
+            "load_from": null
+        },
+
+        "save": [{
+            "save_to": "wandb"
+        },
+        {
+            "save_to": "huggingface",
+            "huggingface_repo": "laion/DALLE2-PyTorch",
+
+            "save_all": false,
+            "save_latest": true,
+            "save_best": true,
+
+            "save_type": "model"
+        }]
+    }
+}