Upload 4 files

README.md

@@ -0,0 +1,16 @@
+---
+license: apache-2.0
+tags:
+- trl
+- ddpo
+- diffusers
+- reinforcement-learning
+- text-to-image
+- stable-diffusion
+---
+
+# TRL DDPO Model
+
+This is a diffusion model that has been fine-tuned with reinforcement learning to
+ guide the model outputs according to a value, function, or human feedback. The model can be used for image generation conditioned with text.
+

config.json

@@ -0,0 +1,58 @@
+{
+  "accelerator_kwargs": {},
+  "allow_tf32": true,
+  "artistic_log_on": true,
+  "async_reward_computation": false,
+  "exp_name": "o2o",
+  "global_step": 0,
+  "high_reward": 100.0,
+  "huggingface_note": "RePic250_2on",
+  "log_with": "wandb",
+  "logdir": "ll",
+  "low_reward": 50.0,
+  "max_workers": 2,
+  "mixed_precision": "fp16",
+  "negative_prompts": "",
+  "num_checkpoint_limit": 10,
+  "num_epochs": 250,
+  "offpolicy_sample_batch_size": 1,
+  "pass_images": 0,
+  "per_prompt_stat_tracking": true,
+  "per_prompt_stat_tracking_buffer_size": 32,
+  "per_prompt_stat_tracking_min_count": 16,
+  "project_kwargs": {
+    "automatic_checkpoint_naming": true,
+    "logging_dir": "./outputs/logs",
+    "project_dir": "./outputs/",
+    "total_limit": 5
+  },
+  "resolution": 768,
+  "resume_from": "",
+  "reward_function_usage": false,
+  "run_name": "",
+  "sample_batch_size": 3,
+  "sample_eta": 1.0,
+  "sample_guidance_scale": 5.0,
+  "sample_num_batches_per_epoch": 1,
+  "sample_num_steps": 50,
+  "save_freq": 1000,
+  "seed": 1,
+  "tracker_kwargs": {},
+  "tracker_project_name": "Dev",
+  "train_adam_beta1": 0.9,
+  "train_adam_beta2": 0.999,
+  "train_adam_epsilon": 1e-08,
+  "train_adam_weight_decay": 0.0001,
+  "train_adv_clip_max": 5,
+  "train_batch_size": 3,
+  "train_cfg": true,
+  "train_clip_range": 0.0001,
+  "train_gradient_accumulation_steps": 1,
+  "train_learning_rate": 0.0003,
+  "train_max_grad_norm": 1.0,
+  "train_num_inner_epochs": 1,
+  "train_timestep_fraction": 1.0,
+  "train_use_8bit_adam": false,
+  "valid_batch_size": 1,
+  "valid_size": 10
+}

model_index.json

@@ -0,0 +1,38 @@
+{
+  "_class_name": "StableDiffusionPipeline",
+  "_diffusers_version": "0.32.2",
+  "_name_or_path": "stabilityai/stable-diffusion-2-1",
+  "feature_extractor": [
+    "transformers",
+    "CLIPImageProcessor"
+  ],
+  "image_encoder": [
+    null,
+    null
+  ],
+  "requires_safety_checker": false,
+  "safety_checker": [
+    null,
+    null
+  ],
+  "scheduler": [
+    "diffusers",
+    "DDIMScheduler"
+  ],
+  "text_encoder": [
+    "transformers",
+    "CLIPTextModel"
+  ],
+  "tokenizer": [
+    "transformers",
+    "CLIPTokenizer"
+  ],
+  "unet": [
+    "diffusers",
+    "UNet2DConditionModel"
+  ],
+  "vae": [
+    "diffusers",
+    "AutoencoderKL"
+  ]
+}

pytorch_lora_weights.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f25a035e4921c5a044a34caffe4f48276476fda6079c7c235ff71f7262b2e6b
+size 3357296