Steps 10k

logs/text2image-fine-tune/1714081180.478312/events.out.tfevents.1714081180.r-nroggendorff-jupyter-hmr6yyzh-23682-zxs8o.419.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:becfa0d0394d69bf7977cdcfa01c9a23d152fcc0cd5ff2db55d124dadc0d7fd7
+size 2249

logs/text2image-fine-tune/1714081180.4926512/hparams.yml

@@ -0,0 +1,51 @@
+adam_beta1: 0.9
+adam_beta2: 0.999
+adam_epsilon: 1.0e-08
+adam_weight_decay: 0.01
+allow_tf32: false
+cache_dir: null
+caption_column: text
+center_crop: true
+checkpointing_steps: 100000
+checkpoints_total_limit: null
+dataloader_num_workers: 0
+dataset_config_name: null
+dataset_name: nroggendorff/zelda
+enable_xformers_memory_efficient_attention: false
+gradient_accumulation_steps: 1
+gradient_checkpointing: true
+hub_model_id: null
+hub_token: null
+image_column: image
+input_perturbation: 0
+learning_rate: 0.01
+local_rank: -1
+logging_dir: logs
+lr_scheduler: constant
+lr_warmup_steps: 0
+max_grad_norm: 1.0
+max_train_samples: null
+max_train_steps: 10000
+mixed_precision: fp16
+noise_offset: 0
+non_ema_revision: null
+num_train_epochs: 93
+output_dir: sd-zelda-model
+prediction_type: null
+pretrained_model_name_or_path: stabilityai/stable-diffusion-2-1
+push_to_hub: false
+random_flip: true
+report_to: tensorboard
+resolution: 512
+resume_from_checkpoint: null
+revision: null
+scale_lr: false
+seed: null
+snr_gamma: null
+tracker_project_name: text2image-fine-tune
+train_batch_size: 1
+train_data_dir: null
+use_8bit_adam: false
+use_ema: true
+validation_epochs: 5
+variant: null

logs/text2image-fine-tune/1714081217.8511393/events.out.tfevents.1714081217.r-nroggendorff-jupyter-hmr6yyzh-23682-zxs8o.488.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c4a5303c99efd9ea0c528dfb723ac69a95ed014c8b4c82f2438f934f2aa2dd0
+size 2249

logs/text2image-fine-tune/1714081217.8651016/hparams.yml

@@ -0,0 +1,51 @@
+adam_beta1: 0.9
+adam_beta2: 0.999
+adam_epsilon: 1.0e-08
+adam_weight_decay: 0.01
+allow_tf32: false
+cache_dir: null
+caption_column: text
+center_crop: true
+checkpointing_steps: 100000
+checkpoints_total_limit: null
+dataloader_num_workers: 0
+dataset_config_name: null
+dataset_name: nroggendorff/zelda
+enable_xformers_memory_efficient_attention: false
+gradient_accumulation_steps: 1
+gradient_checkpointing: true
+hub_model_id: null
+hub_token: null
+image_column: image
+input_perturbation: 0
+learning_rate: 1.0e-05
+local_rank: -1
+logging_dir: logs
+lr_scheduler: constant
+lr_warmup_steps: 0
+max_grad_norm: 1.0
+max_train_samples: null
+max_train_steps: 10000
+mixed_precision: fp16
+noise_offset: 0
+non_ema_revision: null
+num_train_epochs: 93
+output_dir: sd-zelda-model
+prediction_type: null
+pretrained_model_name_or_path: stabilityai/stable-diffusion-2-1
+push_to_hub: false
+random_flip: true
+report_to: tensorboard
+resolution: 512
+resume_from_checkpoint: null
+revision: null
+scale_lr: false
+seed: null
+snr_gamma: null
+tracker_project_name: text2image-fine-tune
+train_batch_size: 1
+train_data_dir: null
+use_8bit_adam: false
+use_ema: true
+validation_epochs: 5
+variant: null

logs/text2image-fine-tune/1714156890.6629379/events.out.tfevents.1714156890.r-nroggendorff-jupyter-9e6ft6d4-e8cf3-drep1.284.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8aa55e56b716f84ae4ef19984f515a62e364217f104e2309bc26d0ae92578ff4
+size 2247

logs/text2image-fine-tune/1714156890.6823308/hparams.yml

@@ -0,0 +1,51 @@
+adam_beta1: 0.9
+adam_beta2: 0.999
+adam_epsilon: 1.0e-08
+adam_weight_decay: 0.01
+allow_tf32: false
+cache_dir: null
+caption_column: text
+center_crop: true
+checkpointing_steps: 100000
+checkpoints_total_limit: null
+dataloader_num_workers: 0
+dataset_config_name: null
+dataset_name: nroggendorff/zelda
+enable_xformers_memory_efficient_attention: false
+gradient_accumulation_steps: 1
+gradient_checkpointing: true
+hub_model_id: null
+hub_token: null
+image_column: image
+input_perturbation: 0
+learning_rate: 1.0e-05
+local_rank: -1
+logging_dir: logs
+lr_scheduler: constant
+lr_warmup_steps: 0
+max_grad_norm: 1.0
+max_train_samples: null
+max_train_steps: 10000
+mixed_precision: fp16
+noise_offset: 0
+non_ema_revision: null
+num_train_epochs: 93
+output_dir: sd-zelda-model
+prediction_type: null
+pretrained_model_name_or_path: runwayml/stable-diffusion-v1-5
+push_to_hub: false
+random_flip: true
+report_to: tensorboard
+resolution: 512
+resume_from_checkpoint: null
+revision: null
+scale_lr: false
+seed: null
+snr_gamma: null
+tracker_project_name: text2image-fine-tune
+train_batch_size: 1
+train_data_dir: null
+use_8bit_adam: false
+use_ema: true
+validation_epochs: 5
+variant: null

logs/text2image-fine-tune/events.out.tfevents.1714081180.r-nroggendorff-jupyter-hmr6yyzh-23682-zxs8o.419.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de15feb7e10ea1c86846d275bd6291a15f368b92637aa2bbc25ef90be5e55864
+size 88

logs/text2image-fine-tune/events.out.tfevents.1714081217.r-nroggendorff-jupyter-hmr6yyzh-23682-zxs8o.488.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2199798a65632f19251fb5df47395764d5c4bc402a8e3fd59487775b6dfb8ff9
+size 88

logs/text2image-fine-tune/events.out.tfevents.1714156890.r-nroggendorff-jupyter-9e6ft6d4-e8cf3-drep1.284.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e4b3c82a65c56c29d7f16c68c2e1bd3e1fc5512290eb7e6781ea6e92da9fd2b
+size 489961

model_index.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "StableDiffusionPipeline",
   "_diffusers_version": "0.28.0.dev0",
-  "_name_or_path": "stabilityai/stable-diffusion-2-1",
+  "_name_or_path": "runwayml/stable-diffusion-v1-5",
   "feature_extractor": [
     "transformers",
     "CLIPImageProcessor"
@@ -10,14 +10,14 @@
     null,
     null
   ],
-  "requires_safety_checker": false,
+  "requires_safety_checker": true,
   "safety_checker": [
-    null,
-    null
+    "stable_diffusion",
+    "StableDiffusionSafetyChecker"
   ],
   "scheduler": [
     "diffusers",
-    "DDIMScheduler"
+    "PNDMScheduler"
   ],
   "text_encoder": [
     "transformers",

safety_checker/config.json

@@ -15,7 +15,7 @@
     "num_attention_heads": 12
   },
   "torch_dtype": "float32",
-  "transformers_version": "4.40.0",
+  "transformers_version": "4.40.1",
   "vision_config": {
     "dropout": 0.0,
     "hidden_size": 1024,

scheduler/scheduler_config.json

@@ -1,20 +1,15 @@
 {
-  "_class_name": "DDIMScheduler",
+  "_class_name": "PNDMScheduler",
   "_diffusers_version": "0.28.0.dev0",
   "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,
+  "prediction_type": "epsilon",
   "set_alpha_to_one": false,
   "skip_prk_steps": true,
   "steps_offset": 1,
-  "thresholding": false,
   "timestep_spacing": "leading",
   "trained_betas": null
 }

text_encoder/config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "stabilityai/stable-diffusion-2-1",
+  "_name_or_path": "runwayml/stable-diffusion-v1-5",
   "architectures": [
     "CLIPTextModel"
   ],
@@ -7,18 +7,18 @@
   "bos_token_id": 0,
   "dropout": 0.0,
   "eos_token_id": 2,
-  "hidden_act": "gelu",
-  "hidden_size": 1024,
+  "hidden_act": "quick_gelu",
+  "hidden_size": 768,
   "initializer_factor": 1.0,
   "initializer_range": 0.02,
-  "intermediate_size": 4096,
+  "intermediate_size": 3072,
   "layer_norm_eps": 1e-05,
   "max_position_embeddings": 77,
   "model_type": "clip_text_model",
-  "num_attention_heads": 16,
-  "num_hidden_layers": 23,
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
   "pad_token_id": 1,
-  "projection_dim": 512,
+  "projection_dim": 768,
   "torch_dtype": "float16",
   "transformers_version": "4.40.1",
   "vocab_size": 49408

text_encoder/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc1827c465450322616f06dea41596eac7d493f4e95904dcb51f0fc745c4e13f
-size 680820392
+oid sha256:660c6f5b1abae9dc498ac2d21e1347d2abdb0cf6c0c0c8576cd796491d9a6cdd
+size 246144152

tokenizer/special_tokens_map.json

@@ -13,7 +13,7 @@
     "rstrip": false,
     "single_word": false
   },
-  "pad_token": "!",
+  "pad_token": "<|endoftext|>",
   "unk_token": {
     "content": "<|endoftext|>",
     "lstrip": false,

tokenizer/tokenizer_config.json

@@ -1,14 +1,6 @@
 {
   "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,
@@ -32,7 +24,7 @@
   "eos_token": "<|endoftext|>",
   "errors": "replace",
   "model_max_length": 77,
-  "pad_token": "!",
+  "pad_token": "<|endoftext|>",
   "tokenizer_class": "CLIPTokenizer",
   "unk_token": "<|endoftext|>"
 }

unet/config.json

@@ -1,17 +1,12 @@
 {
   "_class_name": "UNet2DConditionModel",
   "_diffusers_version": "0.28.0.dev0",
-  "_name_or_path": "stabilityai/stable-diffusion-2-1",
+  "_name_or_path": "runwayml/stable-diffusion-v1-5",
   "act_fn": "silu",
   "addition_embed_type": null,
   "addition_embed_type_num_heads": 64,
   "addition_time_embed_dim": null,
-  "attention_head_dim": [
-    5,
-    10,
-    20,
-    20
-  ],
+  "attention_head_dim": 8,
   "attention_type": "default",
   "block_out_channels": [
     320,
@@ -24,7 +19,7 @@
   "class_embeddings_concat": false,
   "conv_in_kernel": 3,
   "conv_out_kernel": 3,
-  "cross_attention_dim": 1024,
+  "cross_attention_dim": 768,
   "cross_attention_norm": null,
   "down_block_types": [
     "CrossAttnDownBlock2D",
@@ -55,7 +50,7 @@
   "resnet_skip_time_act": false,
   "resnet_time_scale_shift": "default",
   "reverse_transformer_layers_per_block": null,
-  "sample_size": 96,
+  "sample_size": 64,
   "time_cond_proj_dim": null,
   "time_embedding_act_fn": null,
   "time_embedding_dim": null,
@@ -68,6 +63,6 @@
     "CrossAttnUpBlock2D",
     "CrossAttnUpBlock2D"
   ],
-  "upcast_attention": true,
-  "use_linear_projection": true
+  "upcast_attention": false,
+  "use_linear_projection": false
 }

unet/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:243b7ad476b2821c0d93a7150810e86acd3d317a428338e2cda7a0ff556a7866
-size 3463726504
+oid sha256:389c0327e81edd5b0ab7a8d633340654e99c7ff92c533428cccfb33762f5c04e
+size 3438167536

vae/config.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "AutoencoderKL",
   "_diffusers_version": "0.28.0.dev0",
-  "_name_or_path": "stabilityai/stable-diffusion-2-1",
+  "_name_or_path": "runwayml/stable-diffusion-v1-5",
   "act_fn": "silu",
   "block_out_channels": [
     128,
@@ -23,7 +23,7 @@
   "layers_per_block": 2,
   "norm_num_groups": 32,
   "out_channels": 3,
-  "sample_size": 768,
+  "sample_size": 512,
   "scaling_factor": 0.18215,
   "up_block_types": [
     "UpDecoderBlock2D",

vae/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e4c08995484ee61270175e9e7a072b66a6e4eeb5f0c266667fe1f45b90daf9a
+oid sha256:4fbcf0ebe55a0984f5a5e00d8c4521d52359af7229bb4d81890039d2aa16dd7c
 size 167335342