472k

ema_model/config.json

@@ -0,0 +1,62 @@
+{
+  "_class_name": "MaskGiTUViT",
+  "_name_or_path": "/fsx/william/research-run-with-downsample-checkpoint-84000-unwrapped-model",
+  "_version": "0.0.1",
+  "add_cond_embeds": true,
+  "add_cross_attention": true,
+  "add_micro_cond_embeds": true,
+  "attention_dropout": 0.0,
+  "block_has_attention": [
+    true
+  ],
+  "block_num_heads": [
+    12
+  ],
+  "block_out_channels": [
+    768
+  ],
+  "codebook_size": 8192,
+  "cond_embed_dim": 768,
+  "decay": 0.9999,
+  "encoder_hidden_size": 768,
+  "ffn_type": "glu",
+  "force_down_up_sample": true,
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "in_channels": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 2816,
+  "inv_gamma": 1.0,
+  "layer_norm_before_mlm": false,
+  "layer_norm_embedddings": false,
+  "layer_norm_eps": 1e-06,
+  "learn_uncond_embeds": false,
+  "ln_elementwise_affine": true,
+  "mask_token_id": 8255,
+  "max_position_embeddings": 256,
+  "micro_cond_embed_dim": 1280,
+  "micro_cond_encode_dim": 256,
+  "min_decay": 0.0,
+  "norm_type": "rmsnorm",
+  "num_attention_heads": 16,
+  "num_classes": null,
+  "num_hidden_layers": 22,
+  "num_res_blocks": 3,
+  "num_vq_tokens": 1024,
+  "optimization_step": 266000,
+  "patch_size": 1,
+  "power": 0.6666666666666666,
+  "project_encoder_hidden_states": true,
+  "res_ffn_factor": 4,
+  "update_after_step": 0,
+  "use_bias": false,
+  "use_codebook_size_for_output": true,
+  "use_ema_warmup": false,
+  "use_empty_embeds_for_uncond": true,
+  "use_encoder_layernorm": false,
+  "use_normformer": false,
+  "use_position_embeddings": false,
+  "use_vannilla_resblock": false,
+  "vocab_size": 8256,
+  "xavier_init_embed": true
+}

ema_model/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89ef76de32a88da344645ea9a346ddd0d383939305926534b3d53828a74f00cf
+size 2433247453

unwrapped_model/config.json

@@ -0,0 +1,55 @@
+{
+  "_class_name": "MaskGiTUViT",
+  "_name_or_path": "/fsx/william/research-run-with-downsample-checkpoint-84000-unwrapped-model",
+  "_version": "0.0.1",
+  "add_cond_embeds": true,
+  "add_cross_attention": true,
+  "add_micro_cond_embeds": true,
+  "attention_dropout": 0.0,
+  "block_has_attention": [
+    true
+  ],
+  "block_num_heads": [
+    12
+  ],
+  "block_out_channels": [
+    768
+  ],
+  "codebook_size": 8192,
+  "cond_embed_dim": 768,
+  "encoder_hidden_size": 768,
+  "ffn_type": "glu",
+  "force_down_up_sample": true,
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "in_channels": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 2816,
+  "layer_norm_before_mlm": false,
+  "layer_norm_embedddings": false,
+  "layer_norm_eps": 1e-06,
+  "learn_uncond_embeds": false,
+  "ln_elementwise_affine": true,
+  "mask_token_id": 8255,
+  "max_position_embeddings": 256,
+  "micro_cond_embed_dim": 1280,
+  "micro_cond_encode_dim": 256,
+  "norm_type": "rmsnorm",
+  "num_attention_heads": 16,
+  "num_classes": null,
+  "num_hidden_layers": 22,
+  "num_res_blocks": 3,
+  "num_vq_tokens": 1024,
+  "patch_size": 1,
+  "project_encoder_hidden_states": true,
+  "res_ffn_factor": 4,
+  "use_bias": false,
+  "use_codebook_size_for_output": true,
+  "use_empty_embeds_for_uncond": true,
+  "use_encoder_layernorm": false,
+  "use_normformer": false,
+  "use_position_embeddings": false,
+  "use_vannilla_resblock": false,
+  "vocab_size": 8256,
+  "xavier_init_embed": true
+}

unwrapped_model/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47a02818dec5a3719741be44fa37b309a5a267889553dd70e8b4a96f608e5f36
+size 2433254429