panopstor
/

EveryDream

Text-to-Image Diffusers StableDiffusionPipeline Inference Endpoints

Model card Files Files and versions Community

EveryDream / v1-finetune_blacklight.yaml

panopstor's picture

Upload v1-finetune_blacklight.yaml

73f7537 over 1 year ago

raw history blame contribute delete

No virus

3.4 kB

	model:
	base_learning_rate: 1.0e-6
	target: ldm.models.diffusion.ddpm.LatentDiffusion
	params:
	linear_start: 0.00085
	linear_end: 0.0120
	num_timesteps_cond: 1
	log_every_t: 200
	timesteps: 1000
	first_stage_key: image
	cond_stage_key: caption
	image_size: 64
	channels: 4
	cond_stage_trainable: true # Note: different from the one we trained before
	conditioning_key: crossattn
	monitor: val/loss_simple_ema
	scale_factor: 0.18215
	use_ema: False
	unfreeze_model: True
	model_lr: 1.0e-6

	unet_config:
	target: ldm.modules.diffusionmodules.openaimodel.UNetModel
	params:
	image_size: 32 # unused
	in_channels: 4
	out_channels: 4
	model_channels: 320
	attention_resolutions: [ 4, 2, 1 ]
	num_res_blocks: 2
	channel_mult: [ 1, 2, 4, 4 ]
	num_heads: 8
	use_spatial_transformer: True
	transformer_depth: 1
	context_dim: 768
	use_checkpoint: True
	legacy: False

	first_stage_config:
	target: ldm.models.autoencoder.AutoencoderKL
	params:
	embed_dim: 4
	monitor: val/rec_loss
	ddconfig:
	double_z: true
	z_channels: 4
	resolution: 512
	in_channels: 3
	out_ch: 3
	ch: 128
	ch_mult:
	- 1
	- 2
	- 4
	- 4
	num_res_blocks: 2
	attn_resolutions: []
	dropout: 0.0
	lossconfig:
	target: torch.nn.Identity

	cond_stage_config:
	target: ldm.modules.encoders.modules.FrozenCLIPEmbedder

	data:
	target: main.DataModuleFromConfig
	params:
	batch_size: 2 # prefer highest possible without getting CUDA Out of Memory error
	num_workers: 8 # number of dataloader threads
	wrap: falsegit
	train:
	target: ldm.data.every_dream.EveryDreamBatch
	params:
	repeats: 70 # for micro models with 20-50 training images of one subject, try 30-60 repeats with 3-4 epochs (max_epochs below)
	flip_p: 0.5 # use 0.5 to randomly flip images each repeat, not recommended unless very low training data < 20
	debug_level: 1 # 1 to print if images are dropped due to multiple-aspect ratio images
	conditional_dropout: 0.15 # drop conditioning (caption) completely this often
	validation:
	target: ldm.data.ed_validate.EDValidateBatch
	params:
	repeats: 1
	test:
	target: ldm.data.ed_validate.EDValidateBatch
	params:
	repeats: 1

	lightning:
	modelcheckpoint:
	params:
	every_n_epochs: 1 # produce a ckpt every epoch, leave at 1!
	#every_n_train_steps: 1400 # can only use epoch or train step checkpoints
	save_top_k: 99 # save the best N ckpts according to loss, can reduce to save disk space but not recommended
	save_last: False
	filename: "{epoch:02d}-{step:05d}"
	callbacks:
	image_logger:
	target: main.ImageLogger
	params:
	batch_frequency: 150
	max_images: 16
	increase_log_steps: False

	trainer:
	benchmark: True
	max_epochs: 1 # suggest 3-4+ and adjust repeats above, this will give you a few ckpts to test
	max_steps: 99000 # better to end on epochs not steps, especially with >500 images to ensure even distribution, but you can set this if you really want...
	check_val_every_n_epoch: 1
	gpus: 0,