From 6e7efd65604aff68ae6c96ad8aecf39d05dae95d Mon Sep 17 00:00:00 2001 From: opparco Date: Sat, 25 Feb 2023 02:43:26 +0900 Subject: [PATCH] verbose for debug --- composable_lora.py | 30 ++++++++++++++++-------------- 1 file changed, 16 insertions(+), 14 deletions(-) diff --git a/composable_lora.py b/composable_lora.py index bea85c3..34aac8b 100644 --- a/composable_lora.py +++ b/composable_lora.py @@ -32,12 +32,13 @@ def reset_counters(): def debug(*values: object, lora_layer_name: str): - if lora_layer_name.startswith("transformer_"): - if lora_layer_name.endswith("_11_mlp_fc2"): - print(*values) - elif lora_layer_name.startswith("diffusion_model_"): - if lora_layer_name.endswith("_11_1_proj_out"): - print(*values) + if verbose: + if lora_layer_name.startswith("transformer_"): + if lora_layer_name.endswith("_11_mlp_fc2"): + print(*values) + elif lora_layer_name.startswith("diffusion_model_"): + if lora_layer_name.endswith("_11_1_proj_out"): + print(*values) def lora_forward(compvis_module, input, res): @@ -80,12 +81,12 @@ def lora_forward(compvis_module, input, res): loras = prompt_loras[text_model_encoder_counter] multiplier = loras.get(lora.name, 0.0) if multiplier != 0.0: - # debug(f"c #{text_model_encoder_counter} lora.name={lora.name} mul={multiplier}", lora_layer_name=lora_layer_name) + debug(f"c #{text_model_encoder_counter} lora.name={lora.name} mul={multiplier}", lora_layer_name=lora_layer_name) res += multiplier * alpha * patch else: # uc if opt_uc_text_model_encoder and lora.multiplier != 0.0: - # debug(f"uc #{text_model_encoder_counter} lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) + debug(f"uc #{text_model_encoder_counter} lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) res += lora.multiplier * alpha * patch if get_learned_conditioning_prompt_schedules_counter != 1 and lora_layer_name.endswith("_11_mlp_fc2"): # last lora_layer_name of text_model_encoder @@ -103,13 +104,13 @@ def lora_forward(compvis_module, input, res): for p, loras in enumerate(prompt_loras): multiplier = loras.get(lora.name, 0.0) if multiplier != 0.0: - # debug(f"tensor #{b}.{p} lora.name={lora.name} mul={multiplier}", lora_layer_name=lora_layer_name) + debug(f"tensor #{b}.{p} lora.name={lora.name} mul={multiplier}", lora_layer_name=lora_layer_name) res[tensor_off] += multiplier * alpha * patch[tensor_off] tensor_off += 1 # uc if opt_uc_diffusion_model and lora.multiplier != 0.0: - # debug(f"uncond lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) + debug(f"uncond lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) res[uncond_off] += lora.multiplier * alpha * patch[uncond_off] uncond_off += 1 else: # "diffusion_model_" @@ -119,12 +120,12 @@ def lora_forward(compvis_module, input, res): loras = prompt_loras[diffusion_model_counter] multiplier = loras.get(lora.name, 0.0) if multiplier != 0.0: - # debug(f"c #{diffusion_model_counter} lora.name={lora.name} mul={multiplier}", lora_layer_name=lora_layer_name) + debug(f"c #{diffusion_model_counter} lora.name={lora.name} mul={multiplier}", lora_layer_name=lora_layer_name) res += multiplier * alpha * patch else: # uc if opt_uc_diffusion_model and lora.multiplier != 0.0: - # debug(f"uc {lora_layer_name} lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) + debug(f"uc {lora_layer_name} lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) res += lora.multiplier * alpha * patch if lora_layer_name.endswith("_11_1_proj_out"): # last lora_layer_name of diffusion_model @@ -135,7 +136,7 @@ def lora_forward(compvis_module, input, res): else: # default if lora.multiplier != 0.0: - # debug(f"DEFAULT {lora_layer_name} lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) + debug(f"DEFAULT {lora_layer_name} lora.name={lora.name} lora.mul={lora.multiplier}", lora_layer_name=lora_layer_name) res += lora.multiplier * alpha * patch return res @@ -154,7 +155,7 @@ def lora_get_learned_conditioning_prompt_schedules(prompts, steps): # # order: uc c # - # print(f"### get_learned_conditioning_prompt_schedules") + prompt_schedules = prompt_parser.get_learned_conditioning_prompt_schedules_before_lora(prompts, steps) get_learned_conditioning_prompt_schedules_counter += 1 @@ -165,6 +166,7 @@ def lora_get_learned_conditioning_prompt_schedules(prompts, steps): enabled = False opt_uc_text_model_encoder = False opt_uc_diffusion_model = False +verbose = False num_batches: int = 0 prompt_loras: List[Dict[str, float]] = []