Instructions to use deathlegionteam/LEGION-Video-Gen with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use deathlegionteam/LEGION-Video-Gen with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("deathlegionteam/LEGION-Video-Gen", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
| { | |
| "_class_name": "HunyuanVideo15Pipeline", | |
| "_diffusers_version": "0.36.0.dev0", | |
| "_name_or_path": "deathlegionteam/LEGION-Video-Gen", | |
| "architectures": [ | |
| "HunyuanVideo15ForConditionalGeneration" | |
| ], | |
| "model_type": "hunyuan_video", | |
| "license": "apache-2.0", | |
| "library_name": "diffusers", | |
| "pipeline_tag": "text-to-video", | |
| "base_model": "deathlegionteam/LEGION-Video-Gen", | |
| "guider": [ | |
| "diffusers", | |
| "ClassifierFreeGuidance" | |
| ], | |
| "scheduler": [ | |
| "diffusers", | |
| "FlowMatchEulerDiscreteScheduler" | |
| ], | |
| "text_encoder": [ | |
| "transformers", | |
| "Qwen2_5_VLTextModel" | |
| ], | |
| "text_encoder_2": [ | |
| "transformers", | |
| "T5EncoderModel" | |
| ], | |
| "tokenizer": [ | |
| "transformers", | |
| "Qwen2TokenizerFast" | |
| ], | |
| "tokenizer_2": [ | |
| "transformers", | |
| "ByT5Tokenizer" | |
| ], | |
| "transformer": [ | |
| "diffusers", | |
| "HunyuanVideo15Transformer3DModel" | |
| ], | |
| "vae": [ | |
| "diffusers", | |
| "AutoencoderKLHunyuanVideo15" | |
| ], | |
| "transformer_config": { | |
| "_class_name": "HunyuanVideo15Transformer3DModel", | |
| "attention_head_dim": 128, | |
| "image_embed_dim": 1152, | |
| "in_channels": 65, | |
| "mlp_ratio": 4.0, | |
| "num_attention_heads": 16, | |
| "num_layers": 54, | |
| "num_refiner_layers": 2, | |
| "out_channels": 32, | |
| "patch_size": 1, | |
| "patch_size_t": 1, | |
| "qk_norm": "rms_norm", | |
| "rope_axes_dim": [16, 56, 56], | |
| "rope_theta": 256.0, | |
| "target_size": 640, | |
| "task_type": "t2v", | |
| "text_embed_2_dim": 1472, | |
| "text_embed_dim": 3584, | |
| "use_meanflow": false | |
| }, | |
| "vae_config": { | |
| "_class_name": "AutoencoderKLHunyuanVideo15", | |
| "in_channels": 3, | |
| "out_channels": 3, | |
| "latent_channels": 32, | |
| "block_out_channels": [128, 256, 512, 512], | |
| "layers_per_block": 2, | |
| "norm_num_groups": 32, | |
| "scaling_factor": 0.476986, | |
| "force_upcast": true, | |
| "use_quant_conv": false | |
| }, | |
| "scheduler_config": { | |
| "_class_name": "FlowMatchEulerDiscreteScheduler", | |
| "num_train_timesteps": 1000, | |
| "shift": 7.0, | |
| "use_dynamic_shifting": false, | |
| "base_shift": 0.5, | |
| "max_shift": 1.15, | |
| "base_image_seq_len": 256, | |
| "max_image_seq_len": 4096 | |
| } | |
| } |