mirror of
https://github.com/comfyanonymous/ComfyUI.git
synced 2026-01-16 01:00:49 +08:00
f75765721d
Implement a model patcher and caster for aimdo. A new ModelPatcher implementation which backs onto comfy-aimdo to implement varying model load levels that can be adjusted during model use. The patcher defers all load processes to lazily load the model during use (e.g. the first step of a ksampler) and automatically negotiates a load level during the inference to maximize VRAM usage without OOMing. If inference requires more VRAM than is available weights are offloaded to make space before the OOM happens. As for loading the weight onto the GPU, that happens via comfy_cast_weights which is now used in all cases. cast_bias_weight checks whether the VBAR assigned to the model has space for the weight (based on the same load priority semantics as the original ModelPatcher). If it does, the VRAM as returned by the Aimdo allocator is used as the parameter GPU side. The caster is responsible for populating the weight data. This is done using the usual offload_stream (which mean we now have asynchronous load overlapping first use compute). Pinning works a little differently. When a weight is detected during load as unable to fit, a pin is allocated at the time of casting and the weight as used by the layer is DMAd back to the the pin using the GPU DMA TX engine, also using the asynchronous offload streams. This means you get to pin the Lora modified and requantized weights which can be a major speedup for offload+quantize+lora use cases, This works around the JIT Lora + FP8 exclusion and brings FP8MM to heavy offloading users (who probably really need it with more modest GPUs). There is a performance risk in that a CPU+RAM patch has been replace with a GPU+RAM patch but my initial performance results look good. Most users as likely to have a GPU that outruns their CPU in these woods. Some common code is written to consolidate a layers tensors for aimdo mapping, pinning, and DMA transfers. interpret_gathered_like() allows unpacking a raw buffer as a set of tensors. This is used consistently to bundle and pack weights, quantization metadata (QuantizedTensor bits) and biases into one payload for DMA in the load process reducing Cuda overhead a little. Some Quantization metadata was missing async offload is some cases which is now added. This also pins quantization metadata and consolidates the number of cuda_host_register calls (which can be expensive). |
||
|---|---|---|
| .. | ||
| audio_encoders | ||
| cldm | ||
| comfy_types | ||
| extra_samplers | ||
| image_encoders | ||
| k_diffusion | ||
| ldm | ||
| sd1_tokenizer | ||
| t2i_adapter | ||
| taesd | ||
| text_encoders | ||
| weight_adapter | ||
| checkpoint_pickle.py | ||
| cli_args.py | ||
| clip_config_bigg.json | ||
| clip_model.py | ||
| clip_vision_config_g.json | ||
| clip_vision_config_h.json | ||
| clip_vision_config_vitl_336_llava.json | ||
| clip_vision_config_vitl_336.json | ||
| clip_vision_config_vitl.json | ||
| clip_vision_siglip2_base_naflex.json | ||
| clip_vision_siglip_384.json | ||
| clip_vision_siglip_512.json | ||
| clip_vision.py | ||
| conds.py | ||
| context_windows.py | ||
| controlnet.py | ||
| diffusers_convert.py | ||
| diffusers_load.py | ||
| float.py | ||
| gligen.py | ||
| hooks.py | ||
| latent_formats.py | ||
| lora_convert.py | ||
| lora.py | ||
| memory_management.py | ||
| model_base.py | ||
| model_detection.py | ||
| model_management.py | ||
| model_patcher.py | ||
| model_sampling.py | ||
| nested_tensor.py | ||
| ops.py | ||
| options.py | ||
| patcher_extension.py | ||
| pinned_memory.py | ||
| pixel_space_convert.py | ||
| quant_ops.py | ||
| rmsnorm.py | ||
| sample.py | ||
| sampler_helpers.py | ||
| samplers.py | ||
| sd1_clip_config.json | ||
| sd1_clip.py | ||
| sd.py | ||
| sdxl_clip.py | ||
| supported_models_base.py | ||
| supported_models.py | ||
| utils.py | ||