r""" Generates video frames from input image and text prompt using diffusion process. Args: input_prompt (`str`):

r""" Generates video frames from text prompt using diffusion process. Args: input_prompt (`str`): Text pr

Chunk generation method tailored for sequential training Args: noise: noise tensor for a single chunk [batch_siz

Generate image/videos from noise and compute the DMD loss. The noisy input to the generator is backward simulated. This remov

Get the calibrated kwargs for the given model and scale rules. If this model has not yet been calibrated with these scale rules, an e

Get the kwargs for tasks that should be used to calibrate the given model for this PTQ method before running evaluation.

Get the size of a single element, in bytes, for a parameter.

Get the size of a single element, in bytes, for a parameter.

Get the size of a single element, in bytes, for a parameter.

Get the size of a single element, in bytes, for a parameter.

Get the packing factor for a parameter.

Prepare this layer for post-training quantization by quantizing the weight, storing the quantized weight, and deleting the original w

Get the quantized parameters for the layer.

Get the quantized parameters for the layer.

Prepare this layer for post-training quantization by quantizing the weight, storing the quantized weight, and deleting the original w

Return weight conversions for the gpt oss model.

The name of the group experiments are being run in. If this is not None and an experiment with this group name and matching parameter

Whether ANY owned timestep bucket at ``block_pos`` has samples (2D only).

Initialize the default distributed group when it is not yet ready.

Return True if the PyTorch backend is available on the current machine.

Return True if the CUTLASS backend is available on the current machine.

Return True if the Transformer Engine backend is available on the current machine.

Return True if the Triton backend is available on the current machine.

Return True if the PyTorch backend is available on the current machine.

Return True if the CUDA backend is available on the current machine.

Return True if the given module type is a quantized module.

Return True if the CUTLASS backend supports the given inputs and output data type.

Return True if the Transformer Engine backend supports the given input and quantization configuration.

Return True if the Triton backend supports the given input and quantization configuration.

Return True if the PyTorch backend supports the given input and quantization configuration.

Return True if the CUDA backend supports the given input and quantization configuration.

Load LoRA weights from state dict Args: lora_model: The LoRA-wrapped model lora_state_dict: LoRA state dict to load m

Restore buckets from a serialized state. Args: state: dict produced by ``state_dict``. strict_offset: when True (defa

Log GPU memory usage at a given training stage.

Optimize a model with SpinQuant on Modal.

Return high precision parameters to be quantized and deleted.

Return high precision parameters to be quantized and deleted.

Return high precision parameters to be quantized and deleted.

Return high precision parameters to be quantized and deleted.

Return high precision parameters to be quantized and deleted.

Create the per-block prompt list and latent noise for one text prompt.

Quantize a model using SmoothQuant.

Export a quantized model with SpinQuant.

Return a model without any quantization.

Quantize a model using round-to-nearest quantization.

Quantize a model with GPTQ.

Quantize a model using AWQ.