Push model using huggingface_hub.

config.json

@@ -3,13 +3,14 @@
   "allow_tf32": true,
   "async_reward_computation": false,
   "exp_name": "ddpo",
-  "log_with": "tensorboard",
+  "log_with": "wandb",
   "logdir": "ll",
   "max_workers": 2,
   "mixed_precision": "fp16",
   "negative_prompts": "",
   "num_checkpoint_limit": 5,
-  "num_epochs": 10,
+  "num_epochs": 4,
+  "offline_data_folder": "../data_offline/",
   "per_prompt_stat_tracking": true,
   "per_prompt_stat_tracking_buffer_size": 32,
   "per_prompt_stat_tracking_min_count": 16,
@@ -25,7 +26,7 @@
   "sample_eta": 1.0,
   "sample_guidance_scale": 5.0,
   "sample_num_batches_per_epoch": 4,
-  "sample_num_steps": 50,
+  "sample_num_steps": 5,
   "save_freq": 200,
   "seed": 0,
   "tracker_kwargs": {},

feature_extractor/preprocessor_config.json

@@ -1,21 +1,4 @@
 {
-  "_valid_processor_keys": [
-    "images",
-    "do_resize",
-    "size",
-    "resample",
-    "do_center_crop",
-    "crop_size",
-    "do_rescale",
-    "rescale_factor",
-    "do_normalize",
-    "image_mean",
-    "image_std",
-    "do_convert_rgb",
-    "return_tensors",
-    "data_format",
-    "input_data_format"
-  ],
   "crop_size": {
     "height": 224,
     "width": 224

model_index.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "StableDiffusionPipeline",
-  "_diffusers_version": "0.29.2",
-  "_name_or_path": "runwayml/stable-diffusion-v1-5",
+  "_diffusers_version": "0.30.2",
+  "_name_or_path": "stabilityai/stable-diffusion-2-1",
   "feature_extractor": [
     "transformers",
     "CLIPImageProcessor"
@@ -10,7 +10,7 @@
     null,
     null
   ],
-  "requires_safety_checker": true,
+  "requires_safety_checker": false,
   "safety_checker": [
     null,
     null

pytorch_lora_weights.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a7bb0d5ea4b9612cb59d0c0827ccd35b73d4e2f139cf022ee66e17a9987e5d6e
-size 3226184
+oid sha256:fdc85df99eb2e74b0cd18ce902d4d7573da787ac49f2ebaaf43e3fa3437606a0
+size 3357296

scheduler/scheduler_config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "DDIMScheduler",
-  "_diffusers_version": "0.29.2",
+  "_diffusers_version": "0.30.2",
   "beta_end": 0.012,
   "beta_schedule": "scaled_linear",
   "beta_start": 0.00085,
@@ -8,7 +8,7 @@
   "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,

text_encoder/config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "/root/.cache/huggingface/hub/models--runwayml--stable-diffusion-v1-5/snapshots/1d0c4ebf6ff58a5caecab40fa1406526bca4b5b9/text_encoder",
+  "_name_or_path": "/root/.cache/huggingface/hub/models--stabilityai--stable-diffusion-2-1/snapshots/5cae40e6a2745ae2b01ad92ae5043f95f23644d6/text_encoder",
   "architectures": [
     "CLIPTextModel"
   ],
@@ -7,19 +7,19 @@
   "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": 768,
+  "projection_dim": 512,
   "torch_dtype": "float16",
-  "transformers_version": "4.41.2",
+  "transformers_version": "4.44.2",
   "vocab_size": 49408
 }

text_encoder/model.safetensors

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

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": "0.29.2",
-  "_name_or_path": "/root/.cache/huggingface/hub/models--runwayml--stable-diffusion-v1-5/snapshots/1d0c4ebf6ff58a5caecab40fa1406526bca4b5b9/unet",
+  "_diffusers_version": "0.30.2",
+  "_name_or_path": "/root/.cache/huggingface/hub/models--stabilityai--stable-diffusion-2-1/snapshots/5cae40e6a2745ae2b01ad92ae5043f95f23644d6/unet",
   "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,
@@ -63,6 +68,6 @@
     "CrossAttnUpBlock2D",
     "CrossAttnUpBlock2D"
   ],
-  "upcast_attention": false,
-  "use_linear_projection": false
+  "upcast_attention": true,
+  "use_linear_projection": true
 }

unet/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd9c1226b798747c5df91a1cddf63a05b975869d85004c9d5ab666707e03bdc1
-size 1722353544
+oid sha256:450ff1a97176ededa9a2f1b66db627713fcf956cbc054e3a3ce5da9c27bae98d
+size 1735264088

vae/config.json

@@ -1,7 +1,7 @@
 {
   "_class_name": "AutoencoderKL",
-  "_diffusers_version": "0.29.2",
-  "_name_or_path": "/root/.cache/huggingface/hub/models--runwayml--stable-diffusion-v1-5/snapshots/1d0c4ebf6ff58a5caecab40fa1406526bca4b5b9/vae",
+  "_diffusers_version": "0.30.2",
+  "_name_or_path": "/root/.cache/huggingface/hub/models--stabilityai--stable-diffusion-2-1/snapshots/5cae40e6a2745ae2b01ad92ae5043f95f23644d6/vae",
   "act_fn": "silu",
   "block_out_channels": [
     128,
@@ -21,9 +21,10 @@
   "latents_mean": null,
   "latents_std": null,
   "layers_per_block": 2,
+  "mid_block_add_attention": true,
   "norm_num_groups": 32,
   "out_channels": 3,
-  "sample_size": 512,
+  "sample_size": 768,
   "scaling_factor": 0.18215,
   "shift_factor": null,
   "up_block_types": [

vae/diffusion_pytorch_model.safetensors

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