Skip to content

minimax_h3_checkpoint_fp8

Serialized block-FP8 execution for the MiniMax-H3 Qwen3-VL encoder.

Classes

fastvideo.models.encoders.minimax_h3_checkpoint_fp8.MiniMaxH3SerializedFP8Config 

MiniMaxH3SerializedFP8Config(weight_block_size: tuple[int, int])

Bases: QuantizationConfig

Serialized 128x128 block-FP8 contract for the H3 text encoder.

Source code in fastvideo/models/encoders/minimax_h3_checkpoint_fp8.py 
27
28
29
30
31
32
33
34
def __init__(self, weight_block_size: tuple[int, int]) -> None:
    super().__init__()
    if weight_block_size != (128, 128):
        raise ValueError("MiniMax-H3 serialized FP8 requires weight_block_size=[128, 128], "
                         f"got {list(weight_block_size)}")
    self.weight_block_size = weight_block_size
    self.is_checkpoint_fp8_serialized = True
    self.activation_scheme = "dynamic"

fastvideo.models.encoders.minimax_h3_checkpoint_fp8.MiniMaxH3SerializedFP8LinearMethod 

MiniMaxH3SerializedFP8LinearMethod(weight_block_size: tuple[int, int])

Bases: LinearMethodBase

Execute serialized 128x128 block-FP8 weights without re-quantizing them.

Source code in fastvideo/models/encoders/minimax_h3_checkpoint_fp8.py 
356
357
358
def __init__(self, weight_block_size: tuple[int, int]) -> None:
    super().__init__()
    self.weight_block_size = weight_block_size

Functions: