| { |
| "_class_name": "HunyuanVideoTransformer3DModel", |
| "_diffusers_version": "0.37.0", |
| "_name_or_path": "/root/.cache/huggingface/hub/models--hunyuanvideo-community--HunyuanVideo/snapshots/e8c2aaa66fe3742a32c11a6766aecbf07c56e773/transformer", |
| "attention_head_dim": 128, |
| "guidance_embeds": true, |
| "image_condition_type": null, |
| "in_channels": 16, |
| "mlp_ratio": 4.0, |
| "num_attention_heads": 24, |
| "num_layers": 20, |
| "num_refiner_layers": 2, |
| "num_single_layers": 40, |
| "out_channels": 16, |
| "patch_size": 2, |
| "patch_size_t": 1, |
| "pooled_projection_dim": 768, |
| "qk_norm": "rms_norm", |
| "quantization_config": { |
| "ignore": [ |
| "double_blocks.0.img_attn_k_norm", |
| "double_blocks.0.img_attn_q_norm", |
| "double_blocks.0.img_mod.linear", |
| "double_blocks.0.txt_attn_k_norm", |
| "double_blocks.0.txt_attn_q_norm", |
| "double_blocks.0.txt_mod.linear", |
| "double_blocks.1.img_attn_k_norm", |
| "double_blocks.1.img_attn_q_norm", |
| "double_blocks.1.img_mod.linear", |
| "double_blocks.1.txt_attn_k_norm", |
| "double_blocks.1.txt_attn_q_norm", |
| "double_blocks.1.txt_mod.linear", |
| "double_blocks.10.img_attn_k_norm", |
| "double_blocks.10.img_attn_q_norm", |
| "double_blocks.10.img_mod.linear", |
| "double_blocks.10.txt_attn_k_norm", |
| "double_blocks.10.txt_attn_q_norm", |
| "double_blocks.10.txt_mod.linear", |
| "double_blocks.11.img_attn_k_norm", |
| "double_blocks.11.img_attn_q_norm", |
| "double_blocks.11.img_mod.linear", |
| "double_blocks.11.txt_attn_k_norm", |
| "double_blocks.11.txt_attn_q_norm", |
| "double_blocks.11.txt_mod.linear", |
| "double_blocks.12.img_attn_k_norm", |
| "double_blocks.12.img_attn_q_norm", |
| "double_blocks.12.img_mod.linear", |
| "double_blocks.12.txt_attn_k_norm", |
| "double_blocks.12.txt_attn_q_norm", |
| "double_blocks.12.txt_mod.linear", |
| "double_blocks.13.img_attn_k_norm", |
| "double_blocks.13.img_attn_q_norm", |
| "double_blocks.13.img_mod.linear", |
| "double_blocks.13.txt_attn_k_norm", |
| "double_blocks.13.txt_attn_q_norm", |
| "double_blocks.13.txt_mod.linear", |
| "double_blocks.14.img_attn_k_norm", |
| "double_blocks.14.img_attn_q_norm", |
| "double_blocks.14.img_mod.linear", |
| "double_blocks.14.txt_attn_k_norm", |
| "double_blocks.14.txt_attn_q_norm", |
| "double_blocks.14.txt_mod.linear", |
| "double_blocks.15.img_attn_k_norm", |
| "double_blocks.15.img_attn_q_norm", |
| "double_blocks.15.img_mod.linear", |
| "double_blocks.15.txt_attn_k_norm", |
| "double_blocks.15.txt_attn_q_norm", |
| "double_blocks.15.txt_mod.linear", |
| "double_blocks.16.img_attn_k_norm", |
| "double_blocks.16.img_attn_q_norm", |
| "double_blocks.16.img_mod.linear", |
| "double_blocks.16.txt_attn_k_norm", |
| "double_blocks.16.txt_attn_q_norm", |
| "double_blocks.16.txt_mod.linear", |
| "double_blocks.17.img_attn_k_norm", |
| "double_blocks.17.img_attn_q_norm", |
| "double_blocks.17.img_mod.linear", |
| "double_blocks.17.txt_attn_k_norm", |
| "double_blocks.17.txt_attn_q_norm", |
| "double_blocks.17.txt_mod.linear", |
| "double_blocks.18.img_attn_k_norm", |
| "double_blocks.18.img_attn_q_norm", |
| "double_blocks.18.img_mod.linear", |
| "double_blocks.18.txt_attn_k_norm", |
| "double_blocks.18.txt_attn_q_norm", |
| "double_blocks.18.txt_mod.linear", |
| "double_blocks.19.img_attn_k_norm", |
| "double_blocks.19.img_attn_q_norm", |
| "double_blocks.19.img_mod.linear", |
| "double_blocks.19.txt_attn_k_norm", |
| "double_blocks.19.txt_attn_q_norm", |
| "double_blocks.19.txt_mod.linear", |
| "double_blocks.2.img_attn_k_norm", |
| "double_blocks.2.img_attn_q_norm", |
| "double_blocks.2.img_mod.linear", |
| "double_blocks.2.txt_attn_k_norm", |
| "double_blocks.2.txt_attn_q_norm", |
| "double_blocks.2.txt_mod.linear", |
| "double_blocks.3.img_attn_k_norm", |
| "double_blocks.3.img_attn_q_norm", |
| "double_blocks.3.img_mod.linear", |
| "double_blocks.3.txt_attn_k_norm", |
| "double_blocks.3.txt_attn_q_norm", |
| "double_blocks.3.txt_mod.linear", |
| "double_blocks.4.img_attn_k_norm", |
| "double_blocks.4.img_attn_q_norm", |
| "double_blocks.4.img_mod.linear", |
| "double_blocks.4.txt_attn_k_norm", |
| "double_blocks.4.txt_attn_q_norm", |
| "double_blocks.4.txt_mod.linear", |
| "double_blocks.5.img_attn_k_norm", |
| "double_blocks.5.img_attn_q_norm", |
| "double_blocks.5.img_mod.linear", |
| "double_blocks.5.txt_attn_k_norm", |
| "double_blocks.5.txt_attn_q_norm", |
| "double_blocks.5.txt_mod.linear", |
| "double_blocks.6.img_attn_k_norm", |
| "double_blocks.6.img_attn_q_norm", |
| "double_blocks.6.img_mod.linear", |
| "double_blocks.6.txt_attn_k_norm", |
| "double_blocks.6.txt_attn_q_norm", |
| "double_blocks.6.txt_mod.linear", |
| "double_blocks.7.img_attn_k_norm", |
| "double_blocks.7.img_attn_q_norm", |
| "double_blocks.7.img_mod.linear", |
| "double_blocks.7.txt_attn_k_norm", |
| "double_blocks.7.txt_attn_q_norm", |
| "double_blocks.7.txt_mod.linear", |
| "double_blocks.8.img_attn_k_norm", |
| "double_blocks.8.img_attn_q_norm", |
| "double_blocks.8.img_mod.linear", |
| "double_blocks.8.txt_attn_k_norm", |
| "double_blocks.8.txt_attn_q_norm", |
| "double_blocks.8.txt_mod.linear", |
| "double_blocks.9.img_attn_k_norm", |
| "double_blocks.9.img_attn_q_norm", |
| "double_blocks.9.img_mod.linear", |
| "double_blocks.9.txt_attn_k_norm", |
| "double_blocks.9.txt_attn_q_norm", |
| "double_blocks.9.txt_mod.linear", |
| "final_layer.adaLN_modulation.linear", |
| "final_layer.linear", |
| "guidance_in.mlp.fc_in", |
| "guidance_in.mlp.fc_out", |
| "img_in.proj", |
| "single_blocks.0.k_norm", |
| "single_blocks.0.modulation.linear", |
| "single_blocks.0.q_norm", |
| "single_blocks.1.k_norm", |
| "single_blocks.1.modulation.linear", |
| "single_blocks.1.q_norm", |
| "single_blocks.10.k_norm", |
| "single_blocks.10.modulation.linear", |
| "single_blocks.10.q_norm", |
| "single_blocks.11.k_norm", |
| "single_blocks.11.modulation.linear", |
| "single_blocks.11.q_norm", |
| "single_blocks.12.k_norm", |
| "single_blocks.12.modulation.linear", |
| "single_blocks.12.q_norm", |
| "single_blocks.13.k_norm", |
| "single_blocks.13.modulation.linear", |
| "single_blocks.13.q_norm", |
| "single_blocks.14.k_norm", |
| "single_blocks.14.modulation.linear", |
| "single_blocks.14.q_norm", |
| "single_blocks.15.k_norm", |
| "single_blocks.15.modulation.linear", |
| "single_blocks.15.q_norm", |
| "single_blocks.16.k_norm", |
| "single_blocks.16.modulation.linear", |
| "single_blocks.16.q_norm", |
| "single_blocks.17.k_norm", |
| "single_blocks.17.modulation.linear", |
| "single_blocks.17.q_norm", |
| "single_blocks.18.k_norm", |
| "single_blocks.18.modulation.linear", |
| "single_blocks.18.q_norm", |
| "single_blocks.19.k_norm", |
| "single_blocks.19.modulation.linear", |
| "single_blocks.19.q_norm", |
| "single_blocks.2.k_norm", |
| "single_blocks.2.modulation.linear", |
| "single_blocks.2.q_norm", |
| "single_blocks.20.k_norm", |
| "single_blocks.20.modulation.linear", |
| "single_blocks.20.q_norm", |
| "single_blocks.21.k_norm", |
| "single_blocks.21.modulation.linear", |
| "single_blocks.21.q_norm", |
| "single_blocks.22.k_norm", |
| "single_blocks.22.modulation.linear", |
| "single_blocks.22.q_norm", |
| "single_blocks.23.k_norm", |
| "single_blocks.23.modulation.linear", |
| "single_blocks.23.q_norm", |
| "single_blocks.24.k_norm", |
| "single_blocks.24.modulation.linear", |
| "single_blocks.24.q_norm", |
| "single_blocks.25.k_norm", |
| "single_blocks.25.modulation.linear", |
| "single_blocks.25.q_norm", |
| "single_blocks.26.k_norm", |
| "single_blocks.26.modulation.linear", |
| "single_blocks.26.q_norm", |
| "single_blocks.27.k_norm", |
| "single_blocks.27.modulation.linear", |
| "single_blocks.27.q_norm", |
| "single_blocks.28.k_norm", |
| "single_blocks.28.modulation.linear", |
| "single_blocks.28.q_norm", |
| "single_blocks.29.k_norm", |
| "single_blocks.29.modulation.linear", |
| "single_blocks.29.q_norm", |
| "single_blocks.3.k_norm", |
| "single_blocks.3.modulation.linear", |
| "single_blocks.3.q_norm", |
| "single_blocks.30.k_norm", |
| "single_blocks.30.modulation.linear", |
| "single_blocks.30.q_norm", |
| "single_blocks.31.k_norm", |
| "single_blocks.31.modulation.linear", |
| "single_blocks.31.q_norm", |
| "single_blocks.32.k_norm", |
| "single_blocks.32.modulation.linear", |
| "single_blocks.32.q_norm", |
| "single_blocks.33.k_norm", |
| "single_blocks.33.modulation.linear", |
| "single_blocks.33.q_norm", |
| "single_blocks.34.k_norm", |
| "single_blocks.34.modulation.linear", |
| "single_blocks.34.q_norm", |
| "single_blocks.35.k_norm", |
| "single_blocks.35.modulation.linear", |
| "single_blocks.35.q_norm", |
| "single_blocks.36.k_norm", |
| "single_blocks.36.modulation.linear", |
| "single_blocks.36.q_norm", |
| "single_blocks.37.k_norm", |
| "single_blocks.37.modulation.linear", |
| "single_blocks.37.q_norm", |
| "single_blocks.38.k_norm", |
| "single_blocks.38.modulation.linear", |
| "single_blocks.38.q_norm", |
| "single_blocks.39.k_norm", |
| "single_blocks.39.modulation.linear", |
| "single_blocks.39.q_norm", |
| "single_blocks.4.k_norm", |
| "single_blocks.4.modulation.linear", |
| "single_blocks.4.q_norm", |
| "single_blocks.5.k_norm", |
| "single_blocks.5.modulation.linear", |
| "single_blocks.5.q_norm", |
| "single_blocks.6.k_norm", |
| "single_blocks.6.modulation.linear", |
| "single_blocks.6.q_norm", |
| "single_blocks.7.k_norm", |
| "single_blocks.7.modulation.linear", |
| "single_blocks.7.q_norm", |
| "single_blocks.8.k_norm", |
| "single_blocks.8.modulation.linear", |
| "single_blocks.8.q_norm", |
| "single_blocks.9.k_norm", |
| "single_blocks.9.modulation.linear", |
| "single_blocks.9.q_norm", |
| "time_in.mlp.fc_in", |
| "time_in.mlp.fc_out", |
| "txt_in.c_embedder.fc_in", |
| "txt_in.c_embedder.fc_out", |
| "txt_in.input_embedder", |
| "txt_in.refiner_blocks.0.adaLN_modulation.linear", |
| "txt_in.refiner_blocks.0.mlp.fc_in", |
| "txt_in.refiner_blocks.0.mlp.fc_out", |
| "txt_in.refiner_blocks.0.norm1", |
| "txt_in.refiner_blocks.0.norm2", |
| "txt_in.refiner_blocks.0.self_attn_proj", |
| "txt_in.refiner_blocks.0.self_attn_qkv", |
| "txt_in.refiner_blocks.1.adaLN_modulation.linear", |
| "txt_in.refiner_blocks.1.mlp.fc_in", |
| "txt_in.refiner_blocks.1.mlp.fc_out", |
| "txt_in.refiner_blocks.1.norm1", |
| "txt_in.refiner_blocks.1.norm2", |
| "txt_in.refiner_blocks.1.self_attn_proj", |
| "txt_in.refiner_blocks.1.self_attn_qkv", |
| "txt_in.t_embedder.mlp.fc_in", |
| "txt_in.t_embedder.mlp.fc_out", |
| "vector_in.fc_in", |
| "vector_in.fc_out" |
| ], |
| "producer": { |
| "name": "modelopt", |
| "version": "0.43.0rc2.dev66+gf7557221e.d20260407" |
| }, |
| "quant_algo": "FP8", |
| "quant_method": "modelopt", |
| "quant_type": "FP8" |
| }, |
| "rope_axes_dim": [ |
| 16, |
| 56, |
| 56 |
| ], |
| "rope_theta": 256.0, |
| "text_embed_dim": 4096 |
| } |