---
license: apache-2.0
---
# MiniMax-H3-NF4

This model is the **NF4 quantized version** of the video generation model [MiniMax-H3](https://modelscope.cn/models/MiniMax/MiniMax-H3), using the `bitsandbytes` 4-bit quantization method. When used with [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio), it enables model inference on devices with limited GPU memory and system RAM.

## Environment Setup

```shell
git clone https://github.com/modelscope/DiffSynth-Studio.git
cd DiffSynth-Studio
pip install -e ".[all]"
```

## Inference Code

### Enable Memory Management

Run the following code to perform inference using DiffSynth-Studio. Memory management will be automatically enabled. The actual GPU memory usage depends on available GPU memory, and the model can run with as little as 8GB of GPU memory.

FL2VA:

```python
import torch
from diffsynth.pipelines.minimax_h3_audio_video import MiniMaxH3Pipeline, ModelConfig
from diffsynth.utils.data.audio_video import write_video_audio
from modelscope import dataset_snapshot_download
from PIL import Image

vram_config = {
    "offload_dtype": "disk",
    "offload_device": "disk",
    "onload_dtype": torch.bfloat16,
    "onload_device": "cpu",
    "preparing_dtype": torch.bfloat16,
    "preparing_device": "cuda",
    "computation_dtype": torch.bfloat16,
    "computation_device": "cuda",
}

pipe = MiniMaxH3Pipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-fl2va-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-text-encoder-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="video_vae_nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="audio_vae_nf4.safetensors", **vram_config),
    ],
    processor_config=ModelConfig(model_id="MiniMax/MiniMax-H3", origin_file_pattern="FL2VA/processor/"),
    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 4,
)

prompt = "A girl is very happy, she is speaking in english: “I enjoy working with Diffsynth-Studio, it's a perfect framework.”"
video, audio = pipe(
    prompt=prompt,
    height=480, width=832, num_frames=124, num_inference_steps=50, seed=0,
)
write_video_audio(
    video=video, audio=audio,
    output_path="t2va.mp4", fps=24, audio_sample_rate=32000,
)
```

Ref2VA:

<details>

<summary>Expand code</summary>

```python
import torch
from PIL import Image
from diffsynth.pipelines.minimax_h3_audio_video import MiniMaxH3Pipeline, ModelConfig
from diffsynth.utils.data.audio_video import write_video_audio
from diffsynth.utils.data.audio import read_audio
from diffsynth.utils.data import VideoData
from modelscope import dataset_snapshot_download

def align_frame_count(frame_count):
    current = max(int(frame_count), 1)
    while current % 17 != 5:
        current += 1
    return current

def read_video_with_fps(path, num_out_frames, height, width, fps=24):
    video = VideoData(path, height=height, width=width)
    frames = video.raw_data()
    src_fps = float(video.data.reader.get_meta_data()["fps"])
    out = []
    for k in range(num_out_frames):
        idx = int(round(k * src_fps / fps))
        if idx >= len(frames):
            break
        out.append(frames[idx])
    return out

vram_config = {
    "offload_dtype": "disk",
    "offload_device": "disk",
    "onload_dtype": torch.bfloat16,
    "onload_device": "cpu",
    "preparing_dtype": torch.bfloat16,
    "preparing_device": "cuda",
    "computation_dtype": torch.bfloat16,
    "computation_device": "cuda",
}
pipe = MiniMaxH3Pipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-ref2va-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-text-encoder-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="video_vae_nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="audio_vae_nf4.safetensors", **vram_config),
    ],
    processor_config=ModelConfig(model_id="MiniMax/MiniMax-H3", origin_file_pattern="Ref2VA/processor/"),
    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 5,
)

# Text + Reference Image -> Video + Audio
dataset_snapshot_download(dataset_id="DiffSynth-Studio/diffsynth_example_dataset", local_dir="data/diffsynth_example_dataset", allow_file_pattern="minimax_h3/MiniMax-H3-Ref2VA/*")
ref_image = Image.open("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/0.png").convert("RGB")
prompt = "A webpage, website UI design, web animation. The video demonstrates a smooth scrolling-down effect. A highly dynamic and energetic product homepage-style landing page UI/UX demo video, with the main focus being product image 1. The page features bold, slanted, oversized sans-serif typography for striking layout. The background contains fast-moving light effects combined with dynamically shifting dark carbon fiber or sport breathable mesh textures. The video shows a tightly paced, powerful scrolling-down animation, along with strong UI interactions such as visual zoom-in and color inversion on mouse hover."
video, audio = pipe(
    prompt=prompt,
    height=480, width=832, num_frames=124, num_inference_steps=50, seed=42,
    references=[{"type": "image", "image": Image.open("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/0.png").convert("RGB")}]
)
write_video_audio(
    video=video, audio=audio,
    output_path="ti2va.mp4", fps=24, audio_sample_rate=32000,
)

# Text + Reference Audio + Reference Video -> Video + Audio
ref_video = read_video_with_fps("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/video.mp4", 124, 480, 832)
ref_audio, sample_rate = read_audio("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/voice.mp3", duration=len(ref_video) / 24, resample=True, resample_rate=pipe.audio_vae.sample_rate)
prompt = "subject_definitions:\n<Subject 1> is the young man with short wavy blonde hair, wearing a bright pink suit jacket, matching pink trousers, an unbuttoned white shirt, and silver rings, holding a small black lamb in his arms in <Video 1>.\n<Video 1> is the source video for the editing task.\n<Audio 1> is the synchronized audio track of <Video 1>, providing the background music.\n<Audio 2> is the voice timbre reference for <Subject 1>'s voice, containing a spoken male voiceover.\n\nsummary:\n[video editing + audio reference + audio reuse] The target video is an edited version of <Video 1>. <Subject 1>, wearing a bright pink suit and holding a black lamb, stands in a grassy field with other white lambs in the background. The edit animates <Subject 1>'s face to speak the user-provided dialogue. <Audio 1> is partially reused as the continuous background music, while the target references the calm male voice timbre of <Audio 2> for <Subject 1>'s spoken lines.\n\nretention_analysis:\n<Subject 1> (appears in [Shot 1]): fully_preserved - the man retains his identity, wavy blonde hair, pink suit, white shirt, accessories, and the black lamb he holds, with his mouth newly animated to speak.\n<Video 1> (source video editing): fully_preserved - the original camera framing, warm golden hour lighting, grassy hill setting, and background white lambs are maintained while the central character is edited.\n<Audio 1>: partially_copy - the atmospheric background music from <Audio 1> is reused in the target video, mixed beneath the newly added spoken dialogue.\n<Audio 2>: reference - the target audio references the male voice timbre from <Audio 2> to generate <Subject 1>'s spoken dialogue.\n\ndetailed_description:\nThe target video is in realistic photographic style.\n[Shot 1] The shot begins from the source <Video 1>, showing <Subject 1>, a young man with short wavy blonde hair, wearing a bright pink suit jacket, matching pink trousers, and a casually unbuttoned white shirt. He stands confidently in a sunlit green pasture, gently holding a small black lamb securely in his arms. The warm, golden hour lighting casts soft shadows across his face and the bright pink fabric of his suit. Behind him, several white lambs stand and graze on the rolling grassy hill against a clear, pale blue sky. The atmospheric background music from <Audio 1> plays continuously throughout the scene. <Subject 1> physically speaks, his mouth movements naturally syncing to the new dialogue, with his voice timbre referencing the calm male delivery from <Audio 2>. Looking thoughtfully forward, <Subject 1> (S1) speaks softly, <d>[English] Follow the wind, live free.</d> As he delivers the line, he subtly shifts his weight, cradling the resting black lamb while the camera slowly pushes in. <Subject 1> (S1) continues his thought, <d>[English] Leave worries behind, enjoy the moment.</d> Exactly as his voice stops, his lips meet in a relaxed, peaceful smile, and his jaw ceases speaking motion. He then turns his gaze slightly away toward the horizon, gently stroking the black lamb's fleece with his fingers as the camera holds on this tranquil, sunlit state through the end of the video.\n\noverall_soundscape:\nThe soundscape consists of the continuous, atmospheric background music from <Audio 1>, overlaid with the clear, calm male dialogue spoken by the main character, referencing the voice timbre of <Audio 2>.\n\nnon_diegetic_music:\nThe atmospheric, sustained background music from <Audio 1> is reused as the continuous score, playing quietly beneath the spoken dialogue."
video, audio = pipe(
    prompt=prompt,
    height=480, width=832, num_frames=124, num_inference_steps=50, seed=42,
    references=[
        {"type": "video", "video": ref_video},
        {"type": "audio", "audio": ref_audio, "sample_rate": sample_rate},
    ],
)
write_video_audio(
    video=video, audio=audio,
    output_path="tav2va.mp4", fps=24, audio_sample_rate=32000,
)
```

</details>

[Comfy-Org/MiniMax-H3](https://www.modelscope.cn/models/Comfy-Org/MiniMax-H3) provides a pruned variant of MiniMax-H3, which replaces the DiT's timestep embedding MLP with a lookup table, reducing the input dimension of `adaln_proj.linear` from 2688 to 8 and bringing the total number of model parameters down to approximately 20B. We also provide NF4 quantized weights for this version, with each DiT file being around 9.8 GB.

Usage is identical to the above; simply replace the DiT's `origin_file_pattern` with the pruned version. The text encoder and both VAEs still reuse the same set of weights.

FL2VA:

<details>

<summary>Expand code</summary>

```python
import torch
from diffsynth.pipelines.minimax_h3_audio_video import MiniMaxH3Pipeline, ModelConfig
from diffsynth.utils.data.audio_video import write_video_audio
from modelscope import dataset_snapshot_download
from PIL import Image

vram_config = {
    "offload_dtype": "disk",
    "offload_device": "disk",
    "onload_dtype": torch.bfloat16,
    "onload_device": "cpu",
    "preparing_dtype": torch.bfloat16,
    "preparing_device": "cuda",
    "computation_dtype": torch.bfloat16,
    "computation_device": "cuda",
}
pipe = MiniMaxH3Pipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-fl2va-pruned-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-text-encoder-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="video_vae_nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="audio_vae_nf4.safetensors", **vram_config),
    ],
    processor_config=ModelConfig(model_id="MiniMax/MiniMax-H3", origin_file_pattern="FL2VA/processor/"),
    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 4,
)
```

prompt = "A girl is very happy, she is speaking in english: “I enjoy working with Diffsynth-Studio, it's a perfect framework.”"
video, audio = pipe(
    prompt=prompt,
    height=480, width=832, num_frames=124, num_inference_steps=50, seed=0,
)
write_video_audio(
    video=video, audio=audio,
    output_path="t2va.mp4", fps=24, audio_sample_rate=32000,
)
```

</details>

Ref2VA:

<details>

<summary>Expand code</summary>

```python
import torch
from PIL import Image
from diffsynth.pipelines.minimax_h3_audio_video import MiniMaxH3Pipeline, ModelConfig
from diffsynth.utils.data.audio_video import write_video_audio
from diffsynth.utils.data.audio import read_audio
from diffsynth.utils.data import VideoData
from modelscope import dataset_snapshot_download

def align_frame_count(frame_count):
    current = max(int(frame_count), 1)
    while current % 17 != 5:
        current += 1
    return current

def read_video_with_fps(path, num_out_frames, height, width, fps=24):
    video = VideoData(path, height=height, width=width)
    frames = video.raw_data()
    src_fps = float(video.data.reader.get_meta_data()["fps"])
    out = []
    for k in range(num_out_frames):
        idx = int(round(k * src_fps / fps))
        if idx >= len(frames):
            break
        out.append(frames[idx])
    return out

vram_config = {
    "offload_dtype": "disk",
    "offload_device": "disk",
    "onload_dtype": torch.bfloat16,
    "onload_device": "cpu",
    "preparing_dtype": torch.bfloat16,
    "preparing_device": "cuda",
    "computation_dtype": torch.bfloat16,
    "computation_device": "cuda",
}
pipe = MiniMaxH3Pipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-ref2va-pruned-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="minimax-h3-text-encoder-nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="video_vae_nf4.safetensors", **vram_config),
        ModelConfig(model_id="DiffSynth-Studio/MiniMax-H3-NF4", origin_file_pattern="audio_vae_nf4.safetensors", **vram_config),
    ],
    processor_config=ModelConfig(model_id="MiniMax/MiniMax-H3", origin_file_pattern="Ref2VA/processor/"),
    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 5,
)

# Text + Reference Image -> Video + Audio
dataset_snapshot_download(dataset_id="DiffSynth-Studio/diffsynth_example_dataset", local_dir="data/diffsynth_example_dataset", allow_file_pattern="minimax_h3/MiniMax-H3-Ref2VA/*")
ref_image = Image.open("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/0.png").convert("RGB")
prompt = "A website page, web UI design, website motion effects. The video demonstrates a smooth scrolling-down effect on a webpage. An intense, dynamic product landing page UI/UX demonstration video with strong visual impact, centered around Product Image 1. The layout features bold, slanted, oversized sans-serif typography. The background includes dynamic lighting effects conveying speed, along with evolving dark carbon fiber or athletic mesh textures. The video showcases a fast-paced, powerful scrolling animation, as well as strong UI interactions such as dramatic visual zooming and color inversion on mouse hover."
video, audio = pipe(
    prompt=prompt,
    height=480, width=832, num_frames=124, num_inference_steps=50, seed=42,
    references=[{"type": "image", "image": Image.open("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/0.png").convert("RGB")}]
)
write_video_audio(
    video=video, audio=audio,
    output_path="ti2va.mp4", fps=24, audio_sample_rate=32000,
)

# Text + Reference Audio + Reference Video -> Video + Audio
ref_video = read_video_with_fps("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/video.mp4", 124, 480, 832)
ref_audio, sample_rate = read_audio("data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/voice.mp3", duration=len(ref_video) / 24, resample=True, resample_rate=pipe.audio_vae.sample_rate)
prompt = "subject_definitions:\n<Subject 1> is the young man with short wavy blonde hair, wearing a bright pink suit jacket, matching pink trousers, an unbuttoned white shirt, and silver rings, holding a small black lamb in his arms in <Video 1>.\n<Video 1> is the source video for the editing task.\n<Audio 1> is the synchronized audio track of <Video 1>, providing the background music.\n<Audio 2> is the voice timbre reference for <Subject 1>'s voice, containing a spoken male voiceover.\n\nsummary:\n[video editing + audio reference + audio reuse] The target video is an edited version of <Video 1>. <Subject 1>, wearing a bright pink suit and holding a black lamb, stands in a grassy field with other white lambs in the background. The edit animates <Subject 1>'s face to speak the user-provided dialogue. <Audio 1> is partially reused as the continuous background music, while the target references the calm male voice timbre of <Audio 2> for <Subject 1>'s spoken lines.\n\nretention_analysis:\n<Subject 1> (appears in [Shot 1]): fully_preserved - the man retains his identity, wavy blonde hair, pink suit, white shirt, accessories, and the black lamb he holds, with his mouth newly animated to speak.\n<Video 1> (source video editing): fully_preserved - the original camera framing, warm golden hour lighting, grassy hill setting, and background white lambs are maintained while the central character is edited.\n<Audio 1>: partially_copy - the atmospheric background music from <Audio 1> is reused in the target video, mixed beneath the newly added spoken dialogue.\n<Audio 2>: reference - the target audio references the male voice timbre from <Audio 2> to generate <Subject 1>'s spoken dialogue.\n\ndetailed_description:\nThe target video is in realistic photographic style.\n[Shot 1] The shot begins from the source <Video 1>, showing <Subject 1>, a young man with short wavy blonde hair, wearing a bright pink suit jacket, matching pink trousers, and a casually unbuttoned white shirt. He stands confidently in a sunlit green pasture, gently holding a small black lamb securely in his arms. The warm, golden hour lighting casts soft shadows across his face and the bright pink fabric of his suit. Behind him, several white lambs stand and graze on the rolling grassy hill against a clear, pale blue sky. The atmospheric background music from <Audio 1> plays continuously throughout the scene. <Subject 1> physically speaks, his mouth movements naturally syncing to the new dialogue, with his voice timbre referencing the calm male delivery from <Audio 2>. Looking thoughtfully forward, <Subject 1> (S1) speaks softly, <d>[English] Follow the wind, live free.</d> As he delivers the line, he subtly shifts his weight, cradling the resting black lamb while the camera slowly pushes in. <Subject 1> (S1) continues his thought, <d>[English] Leave worries behind, enjoy the moment.</d> Exactly as his voice stops, his lips meet in a relaxed, peaceful smile, and his jaw ceases speaking motion. He then turns his gaze slightly away toward the horizon, gently stroking the black lamb's fleece with his fingers as the camera holds on this tranquil, sunlit state through the end of the video.\n\noverall_soundscape:\nThe soundscape consists of the continuous, atmospheric background music from <Audio 1>, overlaid with the clear, calm male dialogue spoken by the main character, referencing the voice timbre of <Audio 2>.\n\nnon_diegetic_music:\nThe atmospheric, sustained background music from <Audio 1> is reused as the continuous score, playing quietly beneath the spoken dialogue."
video, audio = pipe(
    prompt=prompt,
    height=480, width=832, num_frames=124, num_inference_steps=50, seed=42,
    references=[
        {"type": "video", "video": ref_video},
        {"type": "audio", "audio": ref_audio, "sample_rate": sample_rate},
    ],
)
write_video_audio(
    video=video, audio=audio,
    output_path="tav2va.mp4", fps=24, audio_sample_rate=32000,
)

</details>

### Extreme Hardware Optimization

If your computing hardware has extremely limited performance, we support enabling direct disk-to-GPU-memory access. With this configuration, tensors in the model will be loaded sequentially from disk into GPU memory as needed for computation, allowing the model to run with only 8GB of RAM:

```diff
vram_config = {
+   "offload_dtype": "disk",
+   "offload_device": "disk",
+   "onload_dtype": "disk",
+   "onload_device": "disk",
+   "preparing_dtype": "disk",
+   "preparing_device": "disk",
+   "computation_dtype": torch.bfloat16,
+   "computation_device": "cuda",
}
pipe = MiniMaxH3Pipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=...,
    processor_config=...,
+   vram_limit=0,
)
```

If you wish to run model inference on Mac M-series chips, we also provide support—although this is not recommended:

```diff
vram_config = {
+   "offload_dtype": "disk",
+   "offload_device": "disk",
+   "onload_dtype": "disk",
+   "onload_device": "disk",
+   "preparing_dtype": "disk",
+   "preparing_device": "disk",
+   "computation_dtype": torch.bfloat16,
+   "computation_device": "mps",
}
pipe = MiniMaxH3Pipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
+   device="mps",
    model_configs=...,
    processor_config=...,
+   vram_limit=0,
)
```

## Training Code

This quantized model supports LoRA training. Follow the steps below to launch the training job.

Download the example dataset:

```shell
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "minimax_h3/MiniMax-H3-FL2VA/*" --local_dir ./data/diffsynth_example_dataset
```

Training configuration suitable for data center GPUs (e.g., Nvidia H20): Run the following script to start LoRA training, which requires 48GB of GPU memory.

```shell
accelerate launch examples/minimax_h3/model_training/train.py \
  --dataset_base_path data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-FL2VA \
  --dataset_metadata_path data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-FL2VA/metadata.csv \
  --data_file_keys "video,input_audio" \
  --extra_inputs "input_audio" \
  --height 480 \
  --width 832 \
  --num_frames 124 \
  --dataset_repeat 100 \
  --model_id_with_origin_paths "DiffSynth-Studio/MiniMax-H3-NF4:minimax-h3-text-encoder-nf4.safetensors,DiffSynth-Studio/MiniMax-H3-NF4:minimax-h3-fl2va-nf4.safetensors,DiffSynth-Studio/MiniMax-H3-NF4:video_vae_nf4.safetensors,DiffSynth-Studio/MiniMax-H3-NF4:audio_vae_nf4.safetensors" \
  --learning_rate 1e-4 \
  --num_epochs 5 \
  --remove_prefix_in_ckpt "pipe.dit." \
  --output_path "./models/train/MiniMax-H3-T2VA-nf4" \
  --lora_base_model "dit" \
  --lora_target_modules "qkv_proj,out_proj" \
  --lora_rank 32 \
  --use_gradient_checkpointing \
  --find_unused_parameters
```

Training configuration suitable for consumer-grade GPUs (e.g., Nvidia RTX 4090): Run the above script to start a two-stage split training with gradient checkpointing and offload enabled, requiring 24GB of VRAM.

```shell
accelerate launch examples/minimax_h3/model_training/train.py \
  --dataset_base_path data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-FL2VA \
  --dataset_metadata_path data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-FL2VA/metadata.csv \
  --data_file_keys "video,input_audio" \
  --extra_inputs "input_audio" \
  --height 480 \
  --width 832 \
  --num_frames 124 \
  --dataset_repeat 1 \
  --model_id_with_origin_paths "DiffSynth-Studio/MiniMax-H3-NF4:minimax-h3-text-encoder-nf4.safetensors,DiffSynth-Studio/MiniMax-H3-NF4:video_vae_nf4.safetensors,DiffSynth-Studio/MiniMax-H3-NF4:audio_vae_nf4.safetensors" \
  --learning_rate 1e-4 \
  --num_epochs 1 \
  --remove_prefix_in_ckpt "pipe.dit." \
  --output_path "./models/train/MiniMax-H3-T2VA-nf4-split-cache" \
  --lora_base_model "dit" \
  --lora_target_modules "qkv_proj,out_proj" \
  --lora_rank 32 \
  --use_gradient_checkpointing \
  --use_gradient_checkpointing_offload \
  --task "sft:data_process"

accelerate launch examples/minimax_h3/model_training/train.py \
  --dataset_base_path "./models/train/MiniMax-H3-T2VA-nf4-split-cache" \
  --data_file_keys "video,input_audio" \
  --extra_inputs "input_audio" \
  --height 480 \
  --width 832 \
  --num_frames 124 \
  --dataset_repeat 100 \
  --model_id_with_origin_paths "DiffSynth-Studio/MiniMax-H3-NF4:minimax-h3-fl2va-nf4.safetensors" \
  --learning_rate 1e-4 \
  --num_epochs 5 \
  --remove_prefix_in_ckpt "pipe.dit." \
  --output_path "./models/train/MiniMax-H3-T2VA-nf4" \
  --lora_base_model "dit" \
  --lora_target_modules "qkv_proj,out_proj" \
  --lora_rank 32 \
  --use_gradient_checkpointing \
  --use_gradient_checkpointing_offload \
  --find_unused_parameters \
  --task "sft:train"
```

## References

* DiffSynth-Studio Documentation: [Minimax-H3](https://diffsynth-studio-doc.readthedocs.io/en/latest/Model_Details/MiniMax-H3.html)
* DiffSynth-Studio Documentation: [VRAM Management](https://diffsynth-studio-doc.readthedocs.io/en/latest/Pipeline_Usage/VRAM_management.html)
* DiffSynth-Studio Documentation: [Two-Stage Split Training](https://diffsynth-studio-doc.readthedocs.io/en/latest/Training/Split_Training.html)
* DiffSynth-Studio Documentation: [Low VRAM Training](https://diffsynth-studio-doc.readthedocs.io/en/latest/Pipeline_Usage/Model_Training.html#low-vram-training)