Files
Kohya-ss-sd-scripts/library/fp8_optimization_utils.py
Kohya S. 34e7138b6a Add/modify some implementation for anima (#2261)
* fix: update extend-exclude list in _typos.toml to include configs

* fix: exclude anima tests from pytest

* feat: add entry for 'temperal' in extend-words section of _typos.toml for Qwen-Image VAE

* fix: update default value for --discrete_flow_shift in anima training guide

* feat: add Qwen-Image VAE

* feat: simplify encode_tokens

* feat: use unified attention module, add wrapper for state dict compatibility

* feat: loading with dynamic fp8 optimization and LoRA support

* feat: add anima minimal inference script (WIP)

* format: format

* feat: simplify target module selection by regular expression patterns

* feat: kept caption dropout rate in cache and handle in training script

* feat: update train_llm_adapter and verbose default values to string type

* fix: use strategy instead of using tokenizers directly

* feat: add dtype property and all-zero mask handling in cross-attention in LLMAdapterTransformerBlock

* feat: support 5d tensor in get_noisy_model_input_and_timesteps

* feat: update loss calculation to support 5d tensor

* fix: update argument names in anima_train_utils to align with other archtectures

* feat: simplify Anima training script and update empty caption handling

* feat: support LoRA format without `net.` prefix

* fix: update to work fp8_scaled option

* feat: add regex-based learning rates and dimensions handling in create_network

* fix: improve regex matching for module selection and learning rates in LoRANetwork

* fix: update logging message for regex match in LoRANetwork

* fix: keep latents 4D except DiT call

* feat: enhance block swap functionality for inference and training in Anima model

* feat: refactor Anima training script

* feat: optimize VAE processing by adjusting tensor dimensions and data types

* fix: wait all block trasfer before siwtching offloader mode

* feat: update Anima training guide with new argument specifications and regex-based module selection. Thank you Claude!

* feat: support LORA for Qwen3

* feat: update Anima SAI model spec metadata handling

* fix: remove unused code

* feat: split CFG processing in do_sample function to reduce memory usage

* feat: add VAE chunking and caching options to reduce memory usage

* feat: optimize RMSNorm forward method and remove unused torch_attention_op

* Update library/strategy_anima.py

Use torch.all instead of all.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update library/safetensors_utils.py

Fix duplicated new_key for concat_hook.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update anima_minimal_inference.py

Remove unused code.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update anima_train.py

Remove unused import.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update library/anima_train_utils.py

Remove unused import.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* fix: review with Copilot

* feat: add script to convert LoRA format to ComfyUI compatible format (WIP, not tested yet)

* feat: add process_escape function to handle escape sequences in prompts

* feat: enhance LoRA weight handling in model loading and add text encoder loading function

* feat: improve ComfyUI conversion script with prefix constants and module name adjustments

* feat: update caption dropout documentation to clarify cache regeneration requirement

* feat: add clarification on learning rate adjustments

* feat: add note on PyTorch version requirement to prevent NaN loss

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-02-13 08:15:06 +09:00

483 lines
20 KiB
Python

import os
from typing import List, Optional, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
import logging
from tqdm import tqdm
from library.device_utils import clean_memory_on_device
from library.safetensors_utils import MemoryEfficientSafeOpen, TensorWeightAdapter, WeightTransformHooks
from library.utils import setup_logging
setup_logging()
import logging
logger = logging.getLogger(__name__)
def calculate_fp8_maxval(exp_bits=4, mantissa_bits=3, sign_bits=1):
"""
Calculate the maximum representable value in FP8 format.
Default is E4M3 format (4-bit exponent, 3-bit mantissa, 1-bit sign). Only supports E4M3 and E5M2 with sign bit.
Args:
exp_bits (int): Number of exponent bits
mantissa_bits (int): Number of mantissa bits
sign_bits (int): Number of sign bits (0 or 1)
Returns:
float: Maximum value representable in FP8 format
"""
assert exp_bits + mantissa_bits + sign_bits == 8, "Total bits must be 8"
if exp_bits == 4 and mantissa_bits == 3 and sign_bits == 1:
return torch.finfo(torch.float8_e4m3fn).max
elif exp_bits == 5 and mantissa_bits == 2 and sign_bits == 1:
return torch.finfo(torch.float8_e5m2).max
else:
raise ValueError(f"Unsupported FP8 format: E{exp_bits}M{mantissa_bits} with sign_bits={sign_bits}")
# The following is a manual calculation method (wrong implementation for E5M2), kept for reference.
"""
# Calculate exponent bias
bias = 2 ** (exp_bits - 1) - 1
# Calculate maximum mantissa value
mantissa_max = 1.0
for i in range(mantissa_bits - 1):
mantissa_max += 2 ** -(i + 1)
# Calculate maximum value
max_value = mantissa_max * (2 ** (2**exp_bits - 1 - bias))
return max_value
"""
def quantize_fp8(tensor, scale, fp8_dtype, max_value, min_value):
"""
Quantize a tensor to FP8 format using PyTorch's native FP8 dtype support.
Args:
tensor (torch.Tensor): Tensor to quantize
scale (float or torch.Tensor): Scale factor
fp8_dtype (torch.dtype): Target FP8 dtype (torch.float8_e4m3fn or torch.float8_e5m2)
max_value (float): Maximum representable value in FP8
min_value (float): Minimum representable value in FP8
Returns:
torch.Tensor: Quantized tensor in FP8 format
"""
tensor = tensor.to(torch.float32) # ensure tensor is in float32 for division
# Create scaled tensor
tensor = torch.div(tensor, scale).nan_to_num_(0.0) # handle NaN values, equivalent to nonzero_mask in previous function
# Clamp tensor to range
tensor = tensor.clamp_(min=min_value, max=max_value)
# Convert to FP8 dtype
tensor = tensor.to(fp8_dtype)
return tensor
def optimize_state_dict_with_fp8(
state_dict: dict,
calc_device: Union[str, torch.device],
target_layer_keys: Optional[list[str]] = None,
exclude_layer_keys: Optional[list[str]] = None,
exp_bits: int = 4,
mantissa_bits: int = 3,
move_to_device: bool = False,
quantization_mode: str = "block",
block_size: Optional[int] = 64,
):
"""
Optimize Linear layer weights in a model's state dict to FP8 format. The state dict is modified in-place.
This function is a static version of load_safetensors_with_fp8_optimization without loading from files.
Args:
state_dict (dict): State dict to optimize, replaced in-place
calc_device (str): Device to quantize tensors on
target_layer_keys (list, optional): Layer key patterns to target (None for all Linear layers)
exclude_layer_keys (list, optional): Layer key patterns to exclude
exp_bits (int): Number of exponent bits
mantissa_bits (int): Number of mantissa bits
move_to_device (bool): Move optimized tensors to the calculating device
Returns:
dict: FP8 optimized state dict
"""
if exp_bits == 4 and mantissa_bits == 3:
fp8_dtype = torch.float8_e4m3fn
elif exp_bits == 5 and mantissa_bits == 2:
fp8_dtype = torch.float8_e5m2
else:
raise ValueError(f"Unsupported FP8 format: E{exp_bits}M{mantissa_bits}")
# Calculate FP8 max value
max_value = calculate_fp8_maxval(exp_bits, mantissa_bits)
min_value = -max_value # this function supports only signed FP8
# Create optimized state dict
optimized_count = 0
# Enumerate tarket keys
target_state_dict_keys = []
for key in state_dict.keys():
# Check if it's a weight key and matches target patterns
is_target = (target_layer_keys is None or any(pattern in key for pattern in target_layer_keys)) and key.endswith(".weight")
is_excluded = exclude_layer_keys is not None and any(pattern in key for pattern in exclude_layer_keys)
is_target = is_target and not is_excluded
if is_target and isinstance(state_dict[key], torch.Tensor):
target_state_dict_keys.append(key)
# Process each key
for key in tqdm(target_state_dict_keys):
value = state_dict[key]
# Save original device and dtype
original_device = value.device
original_dtype = value.dtype
# Move to calculation device
if calc_device is not None:
value = value.to(calc_device)
quantized_weight, scale_tensor = quantize_weight(key, value, fp8_dtype, max_value, min_value, quantization_mode, block_size)
# Add to state dict using original key for weight and new key for scale
fp8_key = key # Maintain original key
scale_key = key.replace(".weight", ".scale_weight")
if not move_to_device:
quantized_weight = quantized_weight.to(original_device)
# keep scale shape: [1] or [out,1] or [out, num_blocks, 1]. We can determine the quantization mode from the shape of scale_weight in the patched model.
scale_tensor = scale_tensor.to(dtype=original_dtype, device=quantized_weight.device)
state_dict[fp8_key] = quantized_weight
state_dict[scale_key] = scale_tensor
optimized_count += 1
if calc_device is not None: # optimized_count % 10 == 0 and
# free memory on calculation device
clean_memory_on_device(calc_device)
logger.info(f"Number of optimized Linear layers: {optimized_count}")
return state_dict
def quantize_weight(
key: str,
tensor: torch.Tensor,
fp8_dtype: torch.dtype,
max_value: float,
min_value: float,
quantization_mode: str = "block",
block_size: int = 64,
):
original_shape = tensor.shape
# Determine quantization mode
if quantization_mode == "block":
if tensor.ndim != 2:
quantization_mode = "tensor" # fallback to per-tensor
else:
out_features, in_features = tensor.shape
if in_features % block_size != 0:
quantization_mode = "channel" # fallback to per-channel
logger.warning(
f"Layer {key} with shape {tensor.shape} is not divisible by block_size {block_size}, fallback to per-channel quantization."
)
else:
num_blocks = in_features // block_size
tensor = tensor.contiguous().view(out_features, num_blocks, block_size) # [out, num_blocks, block_size]
elif quantization_mode == "channel":
if tensor.ndim != 2:
quantization_mode = "tensor" # fallback to per-tensor
# Calculate scale factor (per-tensor or per-output-channel with percentile or max)
# value shape is expected to be [out_features, in_features] for Linear weights
if quantization_mode == "channel" or quantization_mode == "block":
# row-wise percentile to avoid being dominated by outliers
# result shape: [out_features, 1] or [out_features, num_blocks, 1]
scale_dim = 1 if quantization_mode == "channel" else 2
abs_w = torch.abs(tensor)
# shape: [out_features, 1] or [out_features, num_blocks, 1]
row_max = torch.max(abs_w, dim=scale_dim, keepdim=True).values
scale = row_max / max_value
else:
# per-tensor
tensor_max = torch.max(torch.abs(tensor).view(-1))
scale = tensor_max / max_value
# print(f"Optimizing {key} with scale: {scale}")
# numerical safety
scale = torch.clamp(scale, min=1e-8)
scale = scale.to(torch.float32) # ensure scale is in float32 for division
# Quantize weight to FP8 (scale can be scalar or [out,1], broadcasting works)
quantized_weight = quantize_fp8(tensor, scale, fp8_dtype, max_value, min_value)
# If block-wise, restore original shape
if quantization_mode == "block":
quantized_weight = quantized_weight.view(original_shape) # restore to original shape [out, in]
return quantized_weight, scale
def load_safetensors_with_fp8_optimization(
model_files: List[str],
calc_device: Union[str, torch.device],
target_layer_keys=None,
exclude_layer_keys=None,
exp_bits=4,
mantissa_bits=3,
move_to_device=False,
weight_hook=None,
quantization_mode: str = "block",
block_size: Optional[int] = 64,
disable_numpy_memmap: bool = False,
weight_transform_hooks: Optional[WeightTransformHooks] = None,
) -> dict:
"""
Load weight tensors from safetensors files and merge LoRA weights into the state dict with explicit FP8 optimization.
Args:
model_files (list[str]): List of model files to load
calc_device (str or torch.device): Device to quantize tensors on
target_layer_keys (list, optional): Layer key patterns to target for optimization (None for all Linear layers)
exclude_layer_keys (list, optional): Layer key patterns to exclude from optimization
exp_bits (int): Number of exponent bits
mantissa_bits (int): Number of mantissa bits
move_to_device (bool): Move optimized tensors to the calculating device
weight_hook (callable, optional): Function to apply to each weight tensor before optimization
quantization_mode (str): Quantization mode, "tensor", "channel", or "block"
block_size (int, optional): Block size for block-wise quantization (used if quantization_mode is "block")
disable_numpy_memmap (bool): Disable numpy memmap when loading safetensors
weight_transform_hooks (WeightTransformHooks, optional): Hooks for weight transformation during loading
Returns:
dict: FP8 optimized state dict
"""
if exp_bits == 4 and mantissa_bits == 3:
fp8_dtype = torch.float8_e4m3fn
elif exp_bits == 5 and mantissa_bits == 2:
fp8_dtype = torch.float8_e5m2
else:
raise ValueError(f"Unsupported FP8 format: E{exp_bits}M{mantissa_bits}")
# Calculate FP8 max value
max_value = calculate_fp8_maxval(exp_bits, mantissa_bits)
min_value = -max_value # this function supports only signed FP8
# Define function to determine if a key is a target key. target means fp8 optimization, not for weight hook.
def is_target_key(key):
# Check if weight key matches target patterns and does not match exclude patterns
is_target = (target_layer_keys is None or any(pattern in key for pattern in target_layer_keys)) and key.endswith(".weight")
is_excluded = exclude_layer_keys is not None and any(pattern in key for pattern in exclude_layer_keys)
return is_target and not is_excluded
# Create optimized state dict
optimized_count = 0
# Process each file
state_dict = {}
for model_file in model_files:
with MemoryEfficientSafeOpen(model_file, disable_numpy_memmap=disable_numpy_memmap) as original_f:
f = TensorWeightAdapter(weight_transform_hooks, original_f) if weight_transform_hooks is not None else original_f
keys = f.keys()
for key in tqdm(keys, desc=f"Loading {os.path.basename(model_file)}", unit="key"):
value = f.get_tensor(key)
# Save original device
original_device = value.device # usually cpu
if weight_hook is not None:
# Apply weight hook if provided
value = weight_hook(key, value, keep_on_calc_device=(calc_device is not None))
if not is_target_key(key):
target_device = calc_device if (calc_device is not None and move_to_device) else original_device
value = value.to(target_device)
state_dict[key] = value
continue
# Move to calculation device
if calc_device is not None:
value = value.to(calc_device)
original_dtype = value.dtype
if original_dtype.itemsize == 1:
raise ValueError(
f"Layer {key} is already in {original_dtype} format. `--fp8_scaled` optimization should not be applied. Please use fp16/bf16/float32 model weights."
+ f" / レイヤー {key} は既に{original_dtype}形式です。`--fp8_scaled` 最適化は適用できません。FP16/BF16/Float32のモデル重みを使用してください。"
)
quantized_weight, scale_tensor = quantize_weight(
key, value, fp8_dtype, max_value, min_value, quantization_mode, block_size
)
# Add to state dict using original key for weight and new key for scale
fp8_key = key # Maintain original key
scale_key = key.replace(".weight", ".scale_weight")
assert fp8_key != scale_key, "FP8 key and scale key must be different"
if not move_to_device:
quantized_weight = quantized_weight.to(original_device)
# keep scale shape: [1] or [out,1] or [out, num_blocks, 1]. We can determine the quantization mode from the shape of scale_weight in the patched model.
scale_tensor = scale_tensor.to(dtype=original_dtype, device=quantized_weight.device)
state_dict[fp8_key] = quantized_weight
state_dict[scale_key] = scale_tensor
optimized_count += 1
if calc_device is not None and optimized_count % 10 == 0:
# free memory on calculation device
clean_memory_on_device(calc_device)
logger.info(f"Number of optimized Linear layers: {optimized_count}")
return state_dict
def fp8_linear_forward_patch(self: nn.Linear, x, use_scaled_mm=False, max_value=None):
"""
Patched forward method for Linear layers with FP8 weights.
Args:
self: Linear layer instance
x (torch.Tensor): Input tensor
use_scaled_mm (bool): Use scaled_mm for FP8 Linear layers, requires SM 8.9+ (RTX 40 series)
max_value (float): Maximum value for FP8 quantization. If None, no quantization is applied for input tensor.
Returns:
torch.Tensor: Result of linear transformation
"""
if use_scaled_mm:
# **not tested**
# _scaled_mm only works for per-tensor scale for now (per-channel scale does not work in certain cases)
if self.scale_weight.ndim != 1:
raise ValueError("scaled_mm only supports per-tensor scale_weight for now.")
input_dtype = x.dtype
original_weight_dtype = self.scale_weight.dtype
target_dtype = self.weight.dtype
# assert x.ndim == 3, "Input tensor must be 3D (batch_size, seq_len, hidden_dim)"
if max_value is None:
# no input quantization
scale_x = torch.tensor(1.0, dtype=torch.float32, device=x.device)
else:
# calculate scale factor for input tensor
scale_x = (torch.max(torch.abs(x.flatten())) / max_value).to(torch.float32)
# quantize input tensor to FP8: this seems to consume a lot of memory
fp8_max_value = torch.finfo(target_dtype).max
fp8_min_value = torch.finfo(target_dtype).min
x = quantize_fp8(x, scale_x, target_dtype, fp8_max_value, fp8_min_value)
original_shape = x.shape
x = x.reshape(-1, x.shape[-1]).to(target_dtype)
weight = self.weight.t()
scale_weight = self.scale_weight.to(torch.float32)
if self.bias is not None:
# float32 is not supported with bias in scaled_mm
o = torch._scaled_mm(x, weight, out_dtype=original_weight_dtype, bias=self.bias, scale_a=scale_x, scale_b=scale_weight)
else:
o = torch._scaled_mm(x, weight, out_dtype=input_dtype, scale_a=scale_x, scale_b=scale_weight)
o = o.reshape(original_shape[0], original_shape[1], -1) if len(original_shape) == 3 else o.reshape(original_shape[0], -1)
return o.to(input_dtype)
else:
# Dequantize the weight
original_dtype = self.scale_weight.dtype
if self.scale_weight.ndim < 3:
# per-tensor or per-channel quantization, we can broadcast
dequantized_weight = self.weight.to(original_dtype) * self.scale_weight
else:
# block-wise quantization, need to reshape weight to match scale shape for broadcasting
out_features, num_blocks, _ = self.scale_weight.shape
dequantized_weight = self.weight.to(original_dtype).contiguous().view(out_features, num_blocks, -1)
dequantized_weight = dequantized_weight * self.scale_weight
dequantized_weight = dequantized_weight.view(self.weight.shape)
# Perform linear transformation
if self.bias is not None:
output = F.linear(x, dequantized_weight, self.bias)
else:
output = F.linear(x, dequantized_weight)
return output
def apply_fp8_monkey_patch(model, optimized_state_dict, use_scaled_mm=False):
"""
Apply monkey patching to a model using FP8 optimized state dict.
Args:
model (nn.Module): Model instance to patch
optimized_state_dict (dict): FP8 optimized state dict
use_scaled_mm (bool): Use scaled_mm for FP8 Linear layers, requires SM 8.9+ (RTX 40 series)
Returns:
nn.Module: The patched model (same instance, modified in-place)
"""
# # Calculate FP8 float8_e5m2 max value
# max_value = calculate_fp8_maxval(5, 2)
max_value = None # do not quantize input tensor
# Find all scale keys to identify FP8-optimized layers
scale_keys = [k for k in optimized_state_dict.keys() if k.endswith(".scale_weight")]
# Enumerate patched layers
patched_module_paths = set()
scale_shape_info = {}
for scale_key in scale_keys:
# Extract module path from scale key (remove .scale_weight)
module_path = scale_key.rsplit(".scale_weight", 1)[0]
patched_module_paths.add(module_path)
# Store scale shape information
scale_shape_info[module_path] = optimized_state_dict[scale_key].shape
patched_count = 0
# Apply monkey patch to each layer with FP8 weights
for name, module in model.named_modules():
# Check if this module has a corresponding scale_weight
has_scale = name in patched_module_paths
# Apply patch if it's a Linear layer with FP8 scale
if isinstance(module, nn.Linear) and has_scale:
# register the scale_weight as a buffer to load the state_dict
# module.register_buffer("scale_weight", torch.tensor(1.0, dtype=module.weight.dtype))
scale_shape = scale_shape_info[name]
module.register_buffer("scale_weight", torch.ones(scale_shape, dtype=module.weight.dtype))
# Create a new forward method with the patched version.
def new_forward(self, x):
return fp8_linear_forward_patch(self, x, use_scaled_mm, max_value)
# Bind method to module
module.forward = new_forward.__get__(module, type(module))
patched_count += 1
logger.info(f"Number of monkey-patched Linear layers: {patched_count}")
return model