From 5a22015e6c28774385136191fa346813d983dd8f Mon Sep 17 00:00:00 2001 From: "l.y@live.cn" Date: Sun, 15 Jun 2025 15:14:12 +0800 Subject: [PATCH] fix FLOAT node error --- aiia_float_nodes.py | 247 +++++++++++++++++++++++--------------------- 1 file changed, 130 insertions(+), 117 deletions(-) diff --git a/aiia_float_nodes.py b/aiia_float_nodes.py index ad1091d..a993508 100644 --- a/aiia_float_nodes.py +++ b/aiia_float_nodes.py @@ -1,4 +1,4 @@ -# --- START OF FILE aiia_float_nodes.py --- +# --- START OF FILE aiia_float_nodes.py (FIXED for MP3 Codec & Tensor Dim) --- import torch import os @@ -20,16 +20,16 @@ from tqdm import tqdm # 导入 tqdm def _patched_decode_for_in_memory_stack( self_float_model, # FLOATModel 实例 (float_pipe.G) - s_r: torch.Tensor, - s_r_feats: list, + s_r: torch.Tensor, + s_r_feats: list, r_d: torch.Tensor ) -> dict: # 返回 {'d_hat': cpu_stacked_tensor_tchw} - + T_prime = r_d.shape[1] B = r_d.shape[0] # 应该总是 1 - + comfy_pbar = ProgressBar(T_prime) # ComfyUI 进度条 - + console_pbar_desc = "[FLOAT In-Memory] Processing Frames" # 尝试获取更具体的描述，如果 self_float_model 是可预期的类型 if hasattr(self_float_model, '__class__') and hasattr(self_float_model.__class__, '__name__'): @@ -40,14 +40,14 @@ def _patched_decode_for_in_memory_stack( with tqdm(total=T_prime, desc=console_pbar_desc, unit="frame") as console_pbar: processed_frames_cpu_list = [] - opt = self_float_model.opt - FRAMES_PER_GPU_CHUNK = getattr(opt, 'decode_gpu_chunk_size', 32) - + opt = self_float_model.opt + FRAMES_PER_GPU_CHUNK = getattr(opt, 'decode_gpu_chunk_size', 32) + gpu_frame_buffer = [] for t_idx in range(T_prime): - current_motion_vector = r_d[:, t_idx] - s_r_plus_motion = s_r + current_motion_vector + current_motion_vector = r_d[:, t_idx] + s_r_plus_motion = s_r + current_motion_vector img_t_gpu_raw, _ = self_float_model.motion_autoencoder.dec(s_r_plus_motion, alpha=None, feats=s_r_feats) img_t_gpu_clamped = torch.clamp(img_t_gpu_raw, -1, 1) # 值域 [-1, 1] @@ -58,11 +58,11 @@ def _patched_decode_for_in_memory_stack( for frame_gpu in gpu_frame_buffer: processed_frames_cpu_list.append(frame_gpu.cpu()) gpu_frame_buffer = [] # 清空 buffer - + comfy_pbar.update(1) console_pbar.update(1) # 更新 tqdm 进度条 - del r_d, s_r, s_r_feats + del r_d, s_r, s_r_feats if not processed_frames_cpu_list: print(f"警告: [PatchedDecodeInMemory] 未生成任何帧。") @@ -70,39 +70,39 @@ def _patched_decode_for_in_memory_stack( return {'d_hat': torch.empty((0, img_c, img_h, img_w), device='cpu')} try: d_hat_stacked_cpu = torch.stack(processed_frames_cpu_list, dim=0) # (T, C, H, W) - return {'d_hat': d_hat_stacked_cpu} - except RuntimeError as e_cpu_stack: + return {'d_hat': d_hat_stacked_cpu} + except RuntimeError as e_cpu_stack: print(f"错误: [PatchedDecodeInMemory] CPU堆叠错误: {e_cpu_stack}") raise def _patched_decode_and_save_to_disk( - self_float_model, - s_r: torch.Tensor, - s_r_feats: list, + self_float_model, + s_r: torch.Tensor, + s_r_feats: list, r_d: torch.Tensor, - output_frames_dir: str, + output_frames_dir: str, node_name_log_prefix: str # 这个可以用作 tqdm 的 desc ) -> dict: - + T_prime = r_d.shape[1] B = r_d.shape[0] # 应该总是 1 (由原始代码的 squeeze(0) 暗示) - + comfy_pbar = ProgressBar(T_prime) # ComfyUI 进度条 - + # 使用传入的 node_name_log_prefix 作为基础描述，并添加操作说明 console_pbar_desc = f"{node_name_log_prefix} Saving Frames" with tqdm(total=T_prime, desc=console_pbar_desc, unit="frame") as console_pbar: opt = self_float_model.opt FRAMES_PER_GPU_CHUNK_FOR_PROCESSING = getattr(opt, 'frames_per_gpu_chunk_for_processing', 16) - + gpu_frame_buffer = []; saved_frame_count = 0 self_float_model._last_run_saved_frames = 0 for t_idx in range(T_prime): - current_motion_vector = r_d[:, t_idx] - s_r_plus_motion = s_r + current_motion_vector + current_motion_vector = r_d[:, t_idx] + s_r_plus_motion = s_r + current_motion_vector img_t_gpu_raw, _ = self_float_model.motion_autoencoder.dec(s_r_plus_motion, alpha=None, feats=s_r_feats) img_t_gpu_clamped = torch.clamp(img_t_gpu_raw, -1, 1) gpu_frame_buffer.append(img_t_gpu_clamped.squeeze(0) if B == 1 else img_t_gpu_clamped[0]) @@ -111,39 +111,34 @@ def _patched_decode_and_save_to_disk( (t_idx == T_prime - 1 and len(gpu_frame_buffer) > 0): if gpu_frame_buffer: # 确保 buffer 不为空 current_gpu_chunk_to_process = torch.stack(gpu_frame_buffer, dim=0) if len(gpu_frame_buffer) > 1 else gpu_frame_buffer[0].unsqueeze(0) - gpu_frame_buffer = [] - - chunk_cpu_chw = current_gpu_chunk_to_process.cpu(); del current_gpu_chunk_to_process + gpu_frame_buffer = [] + + chunk_cpu_chw = current_gpu_chunk_to_process.cpu(); del current_gpu_chunk_to_process chunk_cpu_hwc_float_0_1 = ((chunk_cpu_chw.permute(0, 2, 3, 1).clamp(-1,1) + 1.0) / 2.0) - + for frame_idx_in_chunk in range(chunk_cpu_hwc_float_0_1.shape[0]): frame_to_save_np = (chunk_cpu_hwc_float_0_1[frame_idx_in_chunk].numpy() * 255).astype(np.uint8) filename = f"frame_{saved_frame_count:06d}.png" filepath = os.path.join(output_frames_dir, filename) - try: + try: Image.fromarray(frame_to_save_np).save(filepath) saved_frame_count += 1 - except Exception as e_save: - # 对于保存单个帧的错误，可以选择不打断 tqdm，只打印警告 - # console_pbar.write(f"警告: 保存帧 {filepath} 失败: {e_save}") - # 或者如果希望错误更明显，可以打印到标准错误流 - print(f"警告: [{node_name_log_prefix}] 保存帧 {filepath} 失败: {e_save}", file=sys.stderr if 'sys' in globals() else None) # pylint: disable=undefined-variable - del chunk_cpu_chw, chunk_cpu_hwc_float_0_1 - + except Exception as e_save: + import sys + print(f"警告: [{node_name_log_prefix}] 保存帧 {filepath} 失败: {e_save}", file=sys.stderr) + del chunk_cpu_chw, chunk_cpu_hwc_float_0_1 + comfy_pbar.update(1) console_pbar.update(1) # 更新 tqdm 进度条 - + del r_d, s_r, s_r_feats - # tqdm 会在 with 块结束时自动清理/关闭 - # 如果有任何最终信息要在进度条下方打印，确保在 with 块之后 - # （tqdm 默认会这样做，除非 leave=False 且没有其他输出） - + print(f"信息: [{node_name_log_prefix}] 已处理并尝试保存 {saved_frame_count} 帧到 {output_frames_dir}") self_float_model._last_run_saved_frames = saved_frame_count - + img_c = getattr(opt, 'output_nc', 3); img_h = getattr(opt, 'input_size', 64); img_w = getattr(opt, 'input_size', 64) batch_size_for_placeholder = B if B > 0 else 1 # 确保批次大小至少为1 - d_hat_placeholder_btchw = torch.empty((batch_size_for_placeholder, 0, img_c, img_h, img_w), device='cpu') + d_hat_placeholder_btchw = torch.empty((batch_size_for_placeholder, 0, img_c, img_h, img_w), device='cpu') return {'d_hat': d_hat_placeholder_btchw} @@ -163,43 +158,58 @@ class AIIA_FloatProcess_InMemory: print(f"错误: [{self.__class__.NODE_NAME}] {error_message_text}") return (torch.zeros((1, 64, 64, 3), dtype=torch.float32),) - def floatprocess_in_memory(self, float_pipe, ref_image, ref_audio, - a_cfg_scale, r_cfg_scale, e_cfg_scale, - fps, emotion, crop_input_image, seed, nfe, - device_override: str = "default", + def floatprocess_in_memory(self, float_pipe, ref_image, ref_audio, + a_cfg_scale, r_cfg_scale, e_cfg_scale, + fps, emotion, crop_input_image, seed, nfe, + device_override: str = "default", decode_gpu_chunk_size: int = 32): node_name_log = f"[{self.__class__.NODE_NAME}]" print(f"{node_name_log} 流程开始 (内存输出模式)。") start_time_process = time.time() _default_error_tuple = self._create_error_image("未知错误 (初始化或预处理失败)", log_message=False) - return_value = _default_error_tuple + return_value = _default_error_tuple if float_pipe is None or not hasattr(float_pipe, 'opt') or not hasattr(float_pipe.G, 'decode_latent_into_image'): return self._create_error_image("float_pipe 无效或不完整", log_message=True) - + processing_device = torch.device(device_override) if device_override != "default" else torch.device("cuda" if torch.cuda.is_available() else "cpu") print(f"{node_name_log} 本次运行将在设备上: {processing_device}") - + original_decode_method = None original_opt_rank_backup = getattr(float_pipe.opt, 'rank', None) - original_opt_fps_backup = getattr(float_pipe.opt, 'fps', None) + original_opt_fps_backup = getattr(float_pipe.opt, 'fps', None) original_opt_decode_chunk_backup = getattr(float_pipe.opt, 'decode_gpu_chunk_size', None) with tempfile.TemporaryDirectory(prefix="aiia_fp_inmem_") as temp_run_dir: try: - audio_waveform_to_save = ref_audio['waveform'].squeeze(0); - if audio_waveform_to_save.ndim > 1 and audio_waveform_to_save.shape[0] > 1: audio_waveform_to_save = audio_waveform_to_save[0] - audio_save_path = os.path.join(temp_run_dir, "temp_audio.wav"); torchaudio.save(audio_save_path, audio_waveform_to_save.cpu(), ref_audio["sample_rate"]) + # --- START OF AUDIO FIX for In-Memory Node --- + waveform_2d = ref_audio['waveform'].squeeze(0) + if waveform_2d.shape[0] > 1: + audio_waveform_to_save = waveform_2d[0:1, :] + else: + audio_waveform_to_save = waveform_2d + + audio_save_path = os.path.join(temp_run_dir, "temp_audio.wav") + torchaudio.save( + audio_save_path, + audio_waveform_to_save.cpu(), + ref_audio["sample_rate"], + encoding="PCM_S", + bits_per_sample=16 + ) + # --- END OF AUDIO FIX for In-Memory Node --- + ref_image_chw = ref_image[0].permute(2, 0, 1).cpu(); image_save_path = os.path.join(temp_run_dir, "temp_ref_image.png") vutils.save_image(ref_image_chw, image_save_path, normalize=False) + if hasattr(float_pipe.opt, 'rank'): float_pipe.opt.rank = processing_device.index if processing_device.type == 'cuda' and processing_device.index is not None else (0 if processing_device.type == 'cuda' else -1) if hasattr(float_pipe.opt, 'fps'): float_pipe.opt.fps = float(fps) float_pipe.opt.decode_gpu_chunk_size = decode_gpu_chunk_size print(f"{node_name_log} opt 更新: rank={getattr(float_pipe.opt, 'rank', 'N/A')}, fps={getattr(float_pipe.opt, 'fps', 'N/A')}, decode_chunk={getattr(float_pipe.opt, 'decode_gpu_chunk_size', 'N/A')}") - + model_current_device_before_move = next(float_pipe.G.parameters()).device if model_current_device_before_move != processing_device: float_pipe.G.to(processing_device) print(f"{node_name_log} 模型移至: {processing_device}") @@ -207,17 +217,17 @@ class AIIA_FloatProcess_InMemory: original_decode_method = float_pipe.G.decode_latent_into_image float_pipe.G.decode_latent_into_image = types.MethodType(_patched_decode_for_in_memory_stack, float_pipe.G) print(f"信息: {node_name_log} 已替换 decode_latent_into_image 为内存堆叠版本。") - + print(f"{node_name_log} 开始运行推理...") images_thwc_cpu_float01 = float_pipe.run_inference( res_video_path=None, ref_path=image_save_path, audio_path=audio_save_path, a_cfg_scale=a_cfg_scale, r_cfg_scale=r_cfg_scale, e_cfg_scale=e_cfg_scale, emo=None if emotion == "none" else emotion, - no_crop=not crop_input_image, nfe=nfe, seed=seed, verbose=False + no_crop=not crop_input_image, nfe=nfe, seed=seed, verbose=False ) if not isinstance(images_thwc_cpu_float01, torch.Tensor): images_thwc_cpu_float01 = torch.from_numpy(images_thwc_cpu_float01.astype(np.float32)) - + print(f"信息: {node_name_log} 推理完成。输出图像序列形状: {images_thwc_cpu_float01.shape if images_thwc_cpu_float01 is not None else 'None'}") return_value = (images_thwc_cpu_float01,) except Exception as e_proc_inner: @@ -226,11 +236,9 @@ class AIIA_FloatProcess_InMemory: finally: if original_decode_method and hasattr(float_pipe.G, 'decode_latent_into_image'): current_decode_method = getattr(float_pipe.G, 'decode_latent_into_image', None) - # 检查当前绑定方法的 __func__ 是否是我们的 patch 函数 - if hasattr(current_decode_method, '__func__') and current_decode_method.__func__ is _patched_decode_for_in_memory_stack: + if hasattr(current_decode_method, '__func__') and current_decode_method.__func__ is _patched_decode_for_in_memory_stack: float_pipe.G.decode_latent_into_image = original_decode_method print(f"信息: {node_name_log} (finally) 已恢复 float_pipe.G.decode_latent_into_image 为原始方法。") - # 添加一个 elif 以防万一我们的 patch 方式改变了或者上面没匹配上，但方法确实不是原始的了 elif current_decode_method is not original_decode_method: print(f"警告: {node_name_log} (finally) decode_latent_into_image 不是预期中的 patch 方法，但也不是原始方法。仍尝试恢复为原始方法。") float_pipe.G.decode_latent_into_image = original_decode_method @@ -238,20 +246,20 @@ class AIIA_FloatProcess_InMemory: if original_opt_rank_backup is not None: float_pipe.opt.rank = original_opt_rank_backup if original_opt_fps_backup is not None: float_pipe.opt.fps = original_opt_fps_backup if original_opt_decode_chunk_backup is not None : float_pipe.opt.decode_gpu_chunk_size = original_opt_decode_chunk_backup - elif hasattr(float_pipe.opt, 'decode_gpu_chunk_size'): + elif hasattr(float_pipe.opt, 'decode_gpu_chunk_size'): try: del float_pipe.opt.decode_gpu_chunk_size - except AttributeError: # 以防万一 opt 不是 OmegaConf 或普通 dict - pass - + except AttributeError: + pass + current_g_device_after_proc = next(float_pipe.G.parameters()).device if current_g_device_after_proc.type == 'cuda': - try: + try: float_pipe.G.to(torch.device("cpu")) torch.cuda.empty_cache() - except Exception as e_to_cpu: + except Exception as e_to_cpu: print(f"{node_name_log} (finally) 模型移至CPU或清空缓存时出错: {e_to_cpu}") - + end_time_process = time.time() print(f"{node_name_log} 方法总执行耗时: {end_time_process - start_time_process:.2f} 秒。") return return_value @@ -261,38 +269,36 @@ class AIIA_FloatProcess_ToDisk: NODE_NAME = "AIIA Float Process (To Disk)" CATEGORY = "AIIA/FLOAT" FUNCTION = "floatprocess_to_disk" - RETURN_TYPES = ("STRING", "INT") + RETURN_TYPES = ("STRING", "INT") RETURN_NAMES = ("frames_output_directory", "saved_frame_count") @classmethod def INPUT_TYPES(cls): - # ... (INPUT_TYPES 不变) ... base_inputs = AIIA_FloatProcess_InMemory.INPUT_TYPES() base_inputs["optional"]["output_subdir_name"] = ("STRING", {"default": "float_frames_AIIA", "tooltip": "在ComfyUI输出目录下创建的子目录名"}) if "decode_gpu_chunk_size" not in base_inputs["optional"]: base_inputs["optional"]["decode_gpu_chunk_size"] = ("INT", {"default": 32, "min":1, "max":128, "step":1, "tooltip":"GPU解码后一次处理并保存多少帧。影响显存和IO。"}) - else: - base_inputs["optional"]["decode_gpu_chunk_size"][1]["default"] = 32 + else: + base_inputs["optional"]["decode_gpu_chunk_size"][1]["default"] = 32 base_inputs["optional"]["decode_gpu_chunk_size"][1]["tooltip"] = "(To Disk) GPU解码后一次处理并保存多少帧。影响显存和IO。" return base_inputs - + def _create_error_string_count(self, error_message_text: str, log_message: bool = True) -> tuple: - # ... (不变) ... if log_message: print(f"错误: [{self.__class__.NODE_NAME}] {error_message_text}") return (f"错误: {error_message_text}", 0) - def floatprocess_to_disk(self, float_pipe, ref_image, ref_audio, - a_cfg_scale, r_cfg_scale, e_cfg_scale, - fps, emotion, crop_input_image, seed, nfe, + def floatprocess_to_disk(self, float_pipe, ref_image, ref_audio, + a_cfg_scale, r_cfg_scale, e_cfg_scale, + fps, emotion, crop_input_image, seed, nfe, device_override: str = "default", output_subdir_name: str = "float_frames_AIIA", decode_gpu_chunk_size: int = 16): - + node_name_log = f"[{self.__class__.NODE_NAME}]" print(f"{node_name_log} 流程开始 (输出到磁盘模式)。") start_time_process = time.time() - + _default_error_tuple = self._create_error_string_count("未知错误 (初始化或预处理失败)", log_message=False) return_value = _default_error_tuple @@ -306,7 +312,7 @@ class AIIA_FloatProcess_ToDisk: _intermediate_wrapper_for_patch = None # 声明以便 finally 块可以引用 original_opt_rank_backup = getattr(float_pipe.opt, 'rank', None) - original_opt_fps_backup = getattr(float_pipe.opt, 'fps', None) + original_opt_fps_backup = getattr(float_pipe.opt, 'fps', None) original_opt_frames_per_gpu_chunk_backup = getattr(float_pipe.opt, 'frames_per_gpu_chunk_for_processing', None) output_node_main_dir = folder_paths.get_output_directory() @@ -320,9 +326,23 @@ class AIIA_FloatProcess_ToDisk: with tempfile.TemporaryDirectory(prefix="aiia_fp_disk_input_") as input_temp_dir: try: - audio_waveform_to_save = ref_audio['waveform'].squeeze(0); - if audio_waveform_to_save.ndim > 1 and audio_waveform_to_save.shape[0] > 1: audio_waveform_to_save = audio_waveform_to_save[0] - audio_save_path = os.path.join(input_temp_dir, "temp_audio.wav"); torchaudio.save(audio_save_path, audio_waveform_to_save.cpu(), ref_audio["sample_rate"]) + # --- START OF AUDIO FIX for To-Disk Node --- + waveform_2d = ref_audio['waveform'].squeeze(0) + if waveform_2d.shape[0] > 1: + audio_waveform_to_save = waveform_2d[0:1, :] + else: + audio_waveform_to_save = waveform_2d + + audio_save_path = os.path.join(input_temp_dir, "temp_audio.wav") + torchaudio.save( + audio_save_path, + audio_waveform_to_save.cpu(), + ref_audio["sample_rate"], + encoding="PCM_S", + bits_per_sample=16 + ) + # --- END OF AUDIO FIX for To-Disk Node --- + ref_image_chw = ref_image[0].permute(2, 0, 1).cpu(); image_save_path = os.path.join(input_temp_dir, "temp_ref_image.png") vutils.save_image(ref_image_chw, image_save_path, normalize=False) @@ -331,44 +351,40 @@ class AIIA_FloatProcess_ToDisk: if hasattr(float_pipe.opt, 'fps'): float_pipe.opt.fps = float(fps) float_pipe.opt.frames_per_gpu_chunk_for_processing = decode_gpu_chunk_size print(f"{node_name_log} opt 更新: rank={getattr(float_pipe.opt, 'rank', 'N/A')}, fps={getattr(float_pipe.opt, 'fps', 'N/A')}, frames_chunk_for_processing={getattr(float_pipe.opt, 'frames_per_gpu_chunk_for_processing', 'N/A')}") - - model_current_device_before_move = next(float_pipe.G.parameters()).device + + model_current_device_before_move = next(float_pipe.G.parameters()).device if model_current_device_before_move != processing_device: float_pipe.G.to(processing_device) print(f"{node_name_log} 模型移至: {processing_device}") original_decode_method = float_pipe.G.decode_latent_into_image - - # _intermediate_wrapper_for_patch_local 需要定义在这里，以便捕获 frames_output_directory_final 和 node_name_log - # 修改签名以明确接受关键字参数 + def _intermediate_wrapper_for_patch_local(actual_self, *, s_r, s_r_feats, r_d): - # actual_self 在这里就是 float_pipe.G - # frames_output_directory_final 和 node_name_log 是从外部作用域捕获的 (闭包) return _patched_decode_and_save_to_disk( - actual_self, # 传递 self (模型实例) - s_r=s_r, # 以关键字参数传递 - s_r_feats=s_r_feats, # 以关键字参数传递 - r_d=r_d, # 以关键字参数传递 - output_frames_dir=frames_output_directory_final, # 我们自己的参数 - node_name_log_prefix=node_name_log # 我们自己的参数 + actual_self, + s_r=s_r, + s_r_feats=s_r_feats, + r_d=r_d, + output_frames_dir=frames_output_directory_final, + node_name_log_prefix=node_name_log ) - + _intermediate_wrapper_for_patch = _intermediate_wrapper_for_patch_local float_pipe.G.decode_latent_into_image = types.MethodType(_intermediate_wrapper_for_patch, float_pipe.G) print(f"信息: {node_name_log} 已替换 float_pipe.G.decode_latent_into_image 为磁盘保存版本。") - + print(f"{node_name_log} 开始运行推理 (帧将保存到磁盘)...") - _ = float_pipe.run_inference( # 这里会调用被 patch 的方法 + _ = float_pipe.run_inference( res_video_path=None, ref_path=image_save_path, audio_path=audio_save_path, a_cfg_scale=a_cfg_scale, r_cfg_scale=r_cfg_scale, e_cfg_scale=e_cfg_scale, emo=None if emotion == "none" else emotion, - no_crop=not crop_input_image, nfe=nfe, seed=seed, verbose=False + no_crop=not crop_input_image, nfe=nfe, seed=seed, verbose=False ) - + actual_saved_frames = getattr(float_pipe.G, '_last_run_saved_frames', 0) - if hasattr(float_pipe.G, '_last_run_saved_frames'): - delattr(float_pipe.G, '_last_run_saved_frames') - + if hasattr(float_pipe.G, '_last_run_saved_frames'): + delattr(float_pipe.G, '_last_run_saved_frames') + if actual_saved_frames > 0: print(f"信息: {node_name_log} 推理完成。{actual_saved_frames} 帧已保存到 {frames_output_directory_final}") return_value = (frames_output_directory_final, actual_saved_frames) @@ -381,34 +397,33 @@ class AIIA_FloatProcess_ToDisk: finally: if original_decode_method and hasattr(float_pipe.G, 'decode_latent_into_image'): current_decode_method = getattr(float_pipe.G, 'decode_latent_into_image', None) - # 检查 _intermediate_wrapper_for_patch 是否已被定义 (即 try 块中 patch 代码已执行) if _intermediate_wrapper_for_patch is not None and \ hasattr(current_decode_method, '__func__') and \ - current_decode_method.__func__ is _intermediate_wrapper_for_patch: # 比较函数对象本身 + current_decode_method.__func__ is _intermediate_wrapper_for_patch: float_pipe.G.decode_latent_into_image = original_decode_method print(f"信息: {node_name_log} (finally) 已恢复 float_pipe.G.decode_latent_into_image 为原始方法。") elif current_decode_method is not original_decode_method: print(f"警告: {node_name_log} (finally) decode_latent_into_image 不是预期中的 patch 方法，但也不是原始方法。仍尝试恢复为原始方法。") float_pipe.G.decode_latent_into_image = original_decode_method - + if original_opt_rank_backup is not None: float_pipe.opt.rank = original_opt_rank_backup if original_opt_fps_backup is not None: float_pipe.opt.fps = original_opt_fps_backup - if original_opt_frames_per_gpu_chunk_backup is not None : + if original_opt_frames_per_gpu_chunk_backup is not None : float_pipe.opt.frames_per_gpu_chunk_for_processing = original_opt_frames_per_gpu_chunk_backup - elif hasattr(float_pipe.opt, 'frames_per_gpu_chunk_for_processing'): + elif hasattr(float_pipe.opt, 'frames_per_gpu_chunk_for_processing'): try: del float_pipe.opt.frames_per_gpu_chunk_for_processing except AttributeError: pass - + current_g_device_after_proc = next(float_pipe.G.parameters()).device if current_g_device_after_proc.type == 'cuda': - try: + try: float_pipe.G.to(torch.device("cpu")) torch.cuda.empty_cache() - except Exception as e_to_cpu: + except Exception as e_to_cpu: print(f"{node_name_log} (finally) 模型移至CPU或清空缓存时出错: {e_to_cpu}") - + end_time_process = time.time() print(f"{node_name_log} 方法总执行耗时: {end_time_process - start_time_process:.2f} 秒。") return return_value @@ -422,6 +437,4 @@ NODE_DISPLAY_NAME_MAPPINGS = { "AIIA_FloatProcess_InMemory": "Float Process (AIIA In-Memory)", "AIIA_FloatProcess_ToDisk": "Float Process (AIIA To-Disk for Long Audio)", } -print(f"--- AIIA FLOAT Process Nodes (InMemory & ToDisk) Loaded ---") - -# --- END OF FILE aiia_float_nodes.py --- \ No newline at end of file +print(f"--- AIIA FLOAT Process Nodes (InMemory & ToDisk) Loaded ---") \ No newline at end of file