End of training

README.md

@@ -1,7 +1,7 @@
 
 ---
 license: creativeml-openrail-m
-base_model: CompVis/stable-diffusion-v1-4
+base_model: stabilityai/stable-diffusion-2
 instance_prompt: a photo of oumka dog
 tags:
 - stable-diffusion
@@ -14,7 +14,7 @@ inference: true
     
 # DreamBooth - moteloumka/model
 
-This is a dreambooth model derived from CompVis/stable-diffusion-v1-4. The weights were trained on a photo of oumka dog using [DreamBooth](https://dreambooth.github.io/).
+This is a dreambooth model derived from stabilityai/stable-diffusion-2. The weights were trained on a photo of oumka dog using [DreamBooth](https://dreambooth.github.io/).
 You can find some example images in the following. 
 
 

checkpoint-500/optimizer.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fa9218b1e89c651fe2a84b268f783fe19c455fee23c1aa420c8151d5ed6bee59
-size 1725110394
+oid sha256:68fd90e79dd314f5668f05682b8c109485d033a861e86b6e0ee86a1be005419c
+size 1737914938

checkpoint-500/random_states_0.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d724bf61ddfdc729f8b27bad9265688fa880113f0e79ea937d53cd8306617845
+oid sha256:8340ab1fca5cb015a1e989e6cd364510ffda7065a62330b6b14e9e9740bcd0ea
 size 14344

checkpoint-500/unet/config.json

@@ -1,12 +1,17 @@
 {
   "_class_name": "UNet2DConditionModel",
   "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "CompVis/stable-diffusion-v1-4",
+  "_name_or_path": "stabilityai/stable-diffusion-2",
   "act_fn": "silu",
   "addition_embed_type": null,
   "addition_embed_type_num_heads": 64,
   "addition_time_embed_dim": null,
-  "attention_head_dim": 8,
+  "attention_head_dim": [
+    5,
+    10,
+    20,
+    20
+  ],
   "attention_type": "default",
   "block_out_channels": [
     320,
@@ -19,7 +24,7 @@
   "class_embeddings_concat": false,
   "conv_in_kernel": 3,
   "conv_out_kernel": 3,
-  "cross_attention_dim": 768,
+  "cross_attention_dim": 1024,
   "cross_attention_norm": null,
   "down_block_types": [
     "CrossAttnDownBlock2D",
@@ -50,7 +55,7 @@
   "resnet_skip_time_act": false,
   "resnet_time_scale_shift": "default",
   "reverse_transformer_layers_per_block": null,
-  "sample_size": 64,
+  "sample_size": 96,
   "time_cond_proj_dim": null,
   "time_embedding_act_fn": null,
   "time_embedding_dim": null,
@@ -64,5 +69,5 @@
     "CrossAttnUpBlock2D"
   ],
   "upcast_attention": false,
-  "use_linear_projection": false
+  "use_linear_projection": true
 }

checkpoint-500/unet/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f685aac566c4157aed5baee015981785816ba98ce0a465cf45d750141552f34c
-size 3438167536
+oid sha256:09ef6482e5ed82e2d444337aa0299ff948852ac84457e07ee56f4a73bbb825c1
+size 3463726504

logs/dreambooth/1704626737.7465522/events.out.tfevents.1704626737.n0gsyprqec.991.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a69b87980ddd7eea78e8777912b1954ebe549989a6fd56d60a7b1e60fbf33f30
+size 2925

logs/dreambooth/1704626737.748845/hparams.yml

@@ -0,0 +1,60 @@
+adam_beta1: 0.9
+adam_beta2: 0.999
+adam_epsilon: 1.0e-08
+adam_weight_decay: 0.01
+allow_tf32: false
+center_crop: false
+checkpointing_steps: 500
+checkpoints_total_limit: null
+class_data_dir: /notebooks/diffusers/examples/dreambooth/gen_oumka
+class_labels_conditioning: null
+class_prompt: large brown hound
+dataloader_num_workers: 0
+enable_xformers_memory_efficient_attention: false
+gradient_accumulation_steps: 2
+gradient_checkpointing: true
+hub_model_id: null
+hub_token: null
+instance_data_dir: /notebooks/diffusers/examples/dreambooth/oumka
+instance_prompt: a photo of oumka dog
+learning_rate: 5.0e-06
+local_rank: -1
+logging_dir: logs
+lr_num_cycles: 1
+lr_power: 1.0
+lr_scheduler: constant
+lr_warmup_steps: 0
+max_grad_norm: 1.0
+max_train_steps: 800
+mixed_precision: null
+num_class_images: 400
+num_train_epochs: 4
+num_validation_images: 4
+offset_noise: false
+output_dir: /notebooks/diffusers/examples/dreambooth/model
+pre_compute_text_embeddings: false
+pretrained_model_name_or_path: stabilityai/stable-diffusion-2
+prior_generation_precision: null
+prior_loss_weight: 1.0
+push_to_hub: true
+report_to: tensorboard
+resolution: 512
+resume_from_checkpoint: null
+revision: null
+sample_batch_size: 4
+scale_lr: false
+seed: null
+set_grads_to_none: false
+skip_save_text_encoder: false
+snr_gamma: null
+text_encoder_use_attention_mask: false
+tokenizer_max_length: null
+tokenizer_name: null
+train_batch_size: 1
+train_text_encoder: false
+use_8bit_adam: true
+validation_prompt: null
+validation_scheduler: DPMSolverMultistepScheduler
+validation_steps: 100
+variant: null
+with_prior_preservation: true

logs/dreambooth/events.out.tfevents.1704626737.n0gsyprqec.991.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64fe96eca41542aafa0c802ec4ff3cd857c4ff78c708b489f1ce23675f7489c4
+size 133926

model_index.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "StableDiffusionPipeline",
   "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "CompVis/stable-diffusion-v1-4",
+  "_name_or_path": "stabilityai/stable-diffusion-2",
   "feature_extractor": [
     "transformers",
     "CLIPImageProcessor"
@@ -10,14 +10,14 @@
     null,
     null
   ],
-  "requires_safety_checker": true,
+  "requires_safety_checker": false,
   "safety_checker": [
-    "stable_diffusion",
-    "StableDiffusionSafetyChecker"
+    null,
+    null
   ],
   "scheduler": [
     "diffusers",
-    "PNDMScheduler"
+    "DDIMScheduler"
   ],
   "text_encoder": [
     "transformers",

scheduler/scheduler_config.json

@@ -1,15 +1,20 @@
 {
-  "_class_name": "PNDMScheduler",
+  "_class_name": "DDIMScheduler",
   "_diffusers_version": "0.26.0.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": "epsilon",
+  "prediction_type": "v_prediction",
+  "rescale_betas_zero_snr": false,
+  "sample_max_value": 1.0,
   "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": "CompVis/stable-diffusion-v1-4",
+  "_name_or_path": "stabilityai/stable-diffusion-2",
   "architectures": [
     "CLIPTextModel"
   ],
@@ -7,16 +7,16 @@
   "bos_token_id": 0,
   "dropout": 0.0,
   "eos_token_id": 2,
-  "hidden_act": "quick_gelu",
-  "hidden_size": 768,
+  "hidden_act": "gelu",
+  "hidden_size": 1024,
   "initializer_factor": 1.0,
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
   "max_position_embeddings": 77,
   "model_type": "clip_text_model",
-  "num_attention_heads": 12,
-  "num_hidden_layers": 12,
+  "num_attention_heads": 16,
+  "num_hidden_layers": 23,
   "pad_token_id": 1,
   "projection_dim": 512,
   "torch_dtype": "float32",

text_encoder/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:778d02eb9e707c3fbaae0b67b79ea0d1399b52e624fb634f2f19375ae7c047c3
-size 492265168
+oid sha256:67e013543d4fac905c882e2993d86a2d454ee69dc9e8f37c0c23d33a48959d15
+size 1361596304

tokenizer/special_tokens_map.json

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

tokenizer/tokenizer_config.json

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

unet/config.json

@@ -1,12 +1,17 @@
 {
   "_class_name": "UNet2DConditionModel",
   "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "CompVis/stable-diffusion-v1-4",
+  "_name_or_path": "stabilityai/stable-diffusion-2",
   "act_fn": "silu",
   "addition_embed_type": null,
   "addition_embed_type_num_heads": 64,
   "addition_time_embed_dim": null,
-  "attention_head_dim": 8,
+  "attention_head_dim": [
+    5,
+    10,
+    20,
+    20
+  ],
   "attention_type": "default",
   "block_out_channels": [
     320,
@@ -19,7 +24,7 @@
   "class_embeddings_concat": false,
   "conv_in_kernel": 3,
   "conv_out_kernel": 3,
-  "cross_attention_dim": 768,
+  "cross_attention_dim": 1024,
   "cross_attention_norm": null,
   "down_block_types": [
     "CrossAttnDownBlock2D",
@@ -50,7 +55,7 @@
   "resnet_skip_time_act": false,
   "resnet_time_scale_shift": "default",
   "reverse_transformer_layers_per_block": null,
-  "sample_size": 64,
+  "sample_size": 96,
   "time_cond_proj_dim": null,
   "time_embedding_act_fn": null,
   "time_embedding_dim": null,
@@ -64,5 +69,5 @@
     "CrossAttnUpBlock2D"
   ],
   "upcast_attention": false,
-  "use_linear_projection": false
+  "use_linear_projection": true
 }

unet/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a7c49827f5bfd3c851f59e4a29ad3dc169a5efb9833838d967f8b13595db52c
-size 3438167536
+oid sha256:6e141f16d4b571cab267865099e6821c819be7f56ad70497d43761b75780b2dc
+size 3463726504

vae/config.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "AutoencoderKL",
   "_diffusers_version": "0.26.0.dev0",
-  "_name_or_path": "/root/.cache/huggingface/hub/models--CompVis--stable-diffusion-v1-4/snapshots/133a221b8aa7292a167afc5127cb63fb5005638b/vae",
+  "_name_or_path": "/root/.cache/huggingface/hub/models--stabilityai--stable-diffusion-2/snapshots/1e128c8891e52218b74cde8f26dbfc701cb99d79/vae",
   "act_fn": "silu",
   "block_out_channels": [
     128,
@@ -21,7 +21,7 @@
   "layers_per_block": 2,
   "norm_num_groups": 32,
   "out_channels": 3,
-  "sample_size": 512,
+  "sample_size": 768,
   "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:b4d2b5932bb4151e54e694fd31ccf51fca908223c9485bd56cd0e1d83ad94c49
+oid sha256:2aa1f43011b553a4cba7f37456465cdbd48aab7b54b9348b890e8058ea7683ec
 size 334643268