| { |
| "add_skip_keys": false, |
| "dequantize_fp32": true, |
| "dynamic_loss_threshold": null, |
| "group_size": 0, |
| "hadamard_group_size": 128, |
| "is_integer": true, |
| "is_training": false, |
| "modules_dtype_dict": {}, |
| "modules_quant_config": {}, |
| "modules_to_not_convert": [ |
| ".txt_out", |
| ".proj_out", |
| ".x_embedder", |
| ".emb_in", |
| ".vid_in", |
| "patch_embedding", |
| ".t_embedder", |
| ".time_embed", |
| "multi_modal_projector", |
| "embed_tokens", |
| "lm_head", |
| "time_text_embed", |
| "norm", |
| ".context_embedder", |
| ".txt_in", |
| ".norm_out", |
| ".img_out", |
| ".final_layer", |
| "time_embedder", |
| ".vid_out", |
| "wte", |
| ".y_embedder", |
| ".emb_out", |
| "patch_embed", |
| "patch_emb", |
| ".condition_embedder", |
| ".img_in", |
| "layers.0.self_attn.norm_q.weight", |
| "layers.0.self_attn.norm_k.weight", |
| "layers.0.self_attn.norm_added_q.weight", |
| "layers.0.self_attn.norm_added_k.weight", |
| "layers.0.input_layernorm.weight", |
| "layers.0.input_layernorm_moe_gen.weight", |
| "layers.0.post_attention_layernorm.weight", |
| "layers.0.post_attention_layernorm_moe_gen.weight", |
| "layers.1.self_attn.norm_q.weight", |
| "layers.1.self_attn.norm_k.weight", |
| "layers.1.self_attn.norm_added_q.weight", |
| "layers.1.self_attn.norm_added_k.weight", |
| "layers.1.input_layernorm.weight", |
| "layers.1.input_layernorm_moe_gen.weight", |
| "layers.1.post_attention_layernorm.weight", |
| "layers.1.post_attention_layernorm_moe_gen.weight", |
| "layers.2.self_attn.norm_q.weight", |
| "layers.2.self_attn.norm_k.weight", |
| "layers.2.self_attn.norm_added_q.weight", |
| "layers.2.self_attn.norm_added_k.weight", |
| "layers.2.input_layernorm.weight", |
| "layers.2.input_layernorm_moe_gen.weight", |
| "layers.2.post_attention_layernorm.weight", |
| "layers.2.post_attention_layernorm_moe_gen.weight", |
| "layers.3.self_attn.norm_q.weight", |
| "layers.3.self_attn.norm_k.weight", |
| "layers.3.self_attn.norm_added_q.weight", |
| "layers.3.self_attn.norm_added_k.weight", |
| "layers.3.input_layernorm.weight", |
| "layers.3.input_layernorm_moe_gen.weight", |
| "layers.3.post_attention_layernorm.weight", |
| "layers.3.post_attention_layernorm_moe_gen.weight", |
| "layers.4.self_attn.norm_q.weight", |
| "layers.4.self_attn.norm_k.weight", |
| "layers.4.self_attn.norm_added_q.weight", |
| "layers.4.self_attn.norm_added_k.weight", |
| "layers.4.input_layernorm.weight", |
| "layers.4.input_layernorm_moe_gen.weight", |
| "layers.4.post_attention_layernorm.weight", |
| "layers.4.post_attention_layernorm_moe_gen.weight", |
| "layers.5.self_attn.norm_q.weight", |
| "layers.5.self_attn.norm_k.weight", |
| "layers.5.self_attn.norm_added_q.weight", |
| "layers.5.self_attn.norm_added_k.weight", |
| "layers.5.input_layernorm.weight", |
| "layers.5.input_layernorm_moe_gen.weight", |
| "layers.5.post_attention_layernorm.weight", |
| "layers.5.post_attention_layernorm_moe_gen.weight", |
| "layers.6.self_attn.norm_q.weight", |
| "layers.6.self_attn.norm_k.weight", |
| "layers.6.self_attn.norm_added_q.weight", |
| "layers.6.self_attn.norm_added_k.weight", |
| "layers.6.input_layernorm.weight", |
| "layers.6.input_layernorm_moe_gen.weight", |
| "layers.6.post_attention_layernorm.weight", |
| "layers.6.post_attention_layernorm_moe_gen.weight", |
| "layers.7.self_attn.norm_q.weight", |
| "layers.7.self_attn.norm_k.weight", |
| "layers.7.self_attn.norm_added_q.weight", |
| "layers.7.self_attn.norm_added_k.weight", |
| "layers.7.input_layernorm.weight", |
| "layers.7.input_layernorm_moe_gen.weight", |
| "layers.7.post_attention_layernorm.weight", |
| "layers.7.post_attention_layernorm_moe_gen.weight", |
| "layers.8.self_attn.norm_q.weight", |
| "layers.8.self_attn.norm_k.weight", |
| "layers.8.self_attn.norm_added_q.weight", |
| "layers.8.self_attn.norm_added_k.weight", |
| "layers.8.input_layernorm.weight", |
| "layers.8.input_layernorm_moe_gen.weight", |
| "layers.8.post_attention_layernorm.weight", |
| "layers.8.post_attention_layernorm_moe_gen.weight", |
| "layers.9.self_attn.norm_q.weight", |
| "layers.9.self_attn.norm_k.weight", |
| "layers.9.self_attn.norm_added_q.weight", |
| "layers.9.self_attn.norm_added_k.weight", |
| "layers.9.input_layernorm.weight", |
| "layers.9.input_layernorm_moe_gen.weight", |
| "layers.9.post_attention_layernorm.weight", |
| "layers.9.post_attention_layernorm_moe_gen.weight", |
| "layers.10.self_attn.norm_q.weight", |
| "layers.10.self_attn.norm_k.weight", |
| "layers.10.self_attn.norm_added_q.weight", |
| "layers.10.self_attn.norm_added_k.weight", |
| "layers.10.input_layernorm.weight", |
| "layers.10.input_layernorm_moe_gen.weight", |
| "layers.10.post_attention_layernorm.weight", |
| "layers.10.post_attention_layernorm_moe_gen.weight", |
| "layers.11.self_attn.norm_q.weight", |
| "layers.11.self_attn.norm_k.weight", |
| "layers.11.self_attn.norm_added_q.weight", |
| "layers.11.self_attn.norm_added_k.weight", |
| "layers.11.input_layernorm.weight", |
| "layers.11.input_layernorm_moe_gen.weight", |
| "layers.11.post_attention_layernorm.weight", |
| "layers.11.post_attention_layernorm_moe_gen.weight", |
| "layers.12.self_attn.norm_q.weight", |
| "layers.12.self_attn.norm_k.weight", |
| "layers.12.self_attn.norm_added_q.weight", |
| "layers.12.self_attn.norm_added_k.weight", |
| "layers.12.input_layernorm.weight", |
| "layers.12.input_layernorm_moe_gen.weight", |
| "layers.12.post_attention_layernorm.weight", |
| "layers.12.post_attention_layernorm_moe_gen.weight", |
| "layers.13.self_attn.norm_q.weight", |
| "layers.13.self_attn.norm_k.weight", |
| "layers.13.self_attn.norm_added_q.weight", |
| "layers.13.self_attn.norm_added_k.weight", |
| "layers.13.input_layernorm.weight", |
| "layers.13.input_layernorm_moe_gen.weight", |
| "layers.13.post_attention_layernorm.weight", |
| "layers.13.post_attention_layernorm_moe_gen.weight", |
| "layers.14.self_attn.norm_q.weight", |
| "layers.14.self_attn.norm_k.weight", |
| "layers.14.self_attn.norm_added_q.weight", |
| "layers.14.self_attn.norm_added_k.weight", |
| "layers.14.input_layernorm.weight", |
| "layers.14.input_layernorm_moe_gen.weight", |
| "layers.14.post_attention_layernorm.weight", |
| "layers.14.post_attention_layernorm_moe_gen.weight", |
| "layers.15.self_attn.norm_q.weight", |
| "layers.15.self_attn.norm_k.weight", |
| "layers.15.self_attn.norm_added_q.weight", |
| "layers.15.self_attn.norm_added_k.weight", |
| "layers.15.input_layernorm.weight", |
| "layers.15.input_layernorm_moe_gen.weight", |
| "layers.15.post_attention_layernorm.weight", |
| "layers.15.post_attention_layernorm_moe_gen.weight", |
| "layers.16.self_attn.norm_q.weight", |
| "layers.16.self_attn.norm_k.weight", |
| "layers.16.self_attn.norm_added_q.weight", |
| "layers.16.self_attn.norm_added_k.weight", |
| "layers.16.input_layernorm.weight", |
| "layers.16.input_layernorm_moe_gen.weight", |
| "layers.16.post_attention_layernorm.weight", |
| "layers.16.post_attention_layernorm_moe_gen.weight", |
| "layers.17.self_attn.norm_q.weight", |
| "layers.17.self_attn.norm_k.weight", |
| "layers.17.self_attn.norm_added_q.weight", |
| "layers.17.self_attn.norm_added_k.weight", |
| "layers.17.input_layernorm.weight", |
| "layers.17.input_layernorm_moe_gen.weight", |
| "layers.17.post_attention_layernorm.weight", |
| "layers.17.post_attention_layernorm_moe_gen.weight", |
| "layers.18.self_attn.norm_q.weight", |
| "layers.18.self_attn.norm_k.weight", |
| "layers.18.self_attn.norm_added_q.weight", |
| "layers.18.self_attn.norm_added_k.weight", |
| "layers.18.input_layernorm.weight", |
| "layers.18.input_layernorm_moe_gen.weight", |
| "layers.18.post_attention_layernorm.weight", |
| "layers.18.post_attention_layernorm_moe_gen.weight", |
| "layers.19.self_attn.norm_q.weight", |
| "layers.19.self_attn.norm_k.weight", |
| "layers.19.self_attn.norm_added_q.weight", |
| "layers.19.self_attn.norm_added_k.weight", |
| "layers.19.input_layernorm.weight", |
| "layers.19.input_layernorm_moe_gen.weight", |
| "layers.19.post_attention_layernorm.weight", |
| "layers.19.post_attention_layernorm_moe_gen.weight", |
| "layers.20.self_attn.norm_q.weight", |
| "layers.20.self_attn.norm_k.weight", |
| "layers.20.self_attn.norm_added_q.weight", |
| "layers.20.self_attn.norm_added_k.weight", |
| "layers.20.input_layernorm.weight", |
| "layers.20.input_layernorm_moe_gen.weight", |
| "layers.20.post_attention_layernorm.weight", |
| "layers.20.post_attention_layernorm_moe_gen.weight", |
| "layers.21.self_attn.norm_q.weight", |
| "layers.21.self_attn.norm_k.weight", |
| "layers.21.self_attn.norm_added_q.weight", |
| "layers.21.self_attn.norm_added_k.weight", |
| "layers.21.input_layernorm.weight", |
| "layers.21.input_layernorm_moe_gen.weight", |
| "layers.21.post_attention_layernorm.weight", |
| "layers.21.post_attention_layernorm_moe_gen.weight", |
| "layers.22.self_attn.norm_q.weight", |
| "layers.22.self_attn.norm_k.weight", |
| "layers.22.self_attn.norm_added_q.weight", |
| "layers.22.self_attn.norm_added_k.weight", |
| "layers.22.input_layernorm.weight", |
| "layers.22.input_layernorm_moe_gen.weight", |
| "layers.22.post_attention_layernorm.weight", |
| "layers.22.post_attention_layernorm_moe_gen.weight", |
| "layers.23.self_attn.norm_q.weight", |
| "layers.23.self_attn.norm_k.weight", |
| "layers.23.self_attn.norm_added_q.weight", |
| "layers.23.self_attn.norm_added_k.weight", |
| "layers.23.input_layernorm.weight", |
| "layers.23.input_layernorm_moe_gen.weight", |
| "layers.23.post_attention_layernorm.weight", |
| "layers.23.post_attention_layernorm_moe_gen.weight", |
| "layers.24.self_attn.norm_q.weight", |
| "layers.24.self_attn.norm_k.weight", |
| "layers.24.self_attn.norm_added_q.weight", |
| "layers.24.self_attn.norm_added_k.weight", |
| "layers.24.input_layernorm.weight", |
| "layers.24.input_layernorm_moe_gen.weight", |
| "layers.24.post_attention_layernorm.weight", |
| "layers.24.post_attention_layernorm_moe_gen.weight", |
| "layers.25.self_attn.norm_q.weight", |
| "layers.25.self_attn.norm_k.weight", |
| "layers.25.self_attn.norm_added_q.weight", |
| "layers.25.self_attn.norm_added_k.weight", |
| "layers.25.input_layernorm.weight", |
| "layers.25.input_layernorm_moe_gen.weight", |
| "layers.25.post_attention_layernorm.weight", |
| "layers.25.post_attention_layernorm_moe_gen.weight", |
| "layers.26.self_attn.norm_q.weight", |
| "layers.26.self_attn.norm_k.weight", |
| "layers.26.self_attn.norm_added_q.weight", |
| "layers.26.self_attn.norm_added_k.weight", |
| "layers.26.input_layernorm.weight", |
| "layers.26.input_layernorm_moe_gen.weight", |
| "layers.26.post_attention_layernorm.weight", |
| "layers.26.post_attention_layernorm_moe_gen.weight", |
| "layers.27.self_attn.norm_q.weight", |
| "layers.27.self_attn.norm_k.weight", |
| "layers.27.self_attn.norm_added_q.weight", |
| "layers.27.self_attn.norm_added_k.weight", |
| "layers.27.input_layernorm.weight", |
| "layers.27.input_layernorm_moe_gen.weight", |
| "layers.27.post_attention_layernorm.weight", |
| "layers.27.post_attention_layernorm_moe_gen.weight", |
| "layers.28.self_attn.norm_q.weight", |
| "layers.28.self_attn.norm_k.weight", |
| "layers.28.self_attn.norm_added_q.weight", |
| "layers.28.self_attn.norm_added_k.weight", |
| "layers.28.input_layernorm.weight", |
| "layers.28.input_layernorm_moe_gen.weight", |
| "layers.28.post_attention_layernorm.weight", |
| "layers.28.post_attention_layernorm_moe_gen.weight", |
| "layers.29.self_attn.norm_q.weight", |
| "layers.29.self_attn.norm_k.weight", |
| "layers.29.self_attn.norm_added_q.weight", |
| "layers.29.self_attn.norm_added_k.weight", |
| "layers.29.input_layernorm.weight", |
| "layers.29.input_layernorm_moe_gen.weight", |
| "layers.29.post_attention_layernorm.weight", |
| "layers.29.post_attention_layernorm_moe_gen.weight", |
| "layers.30.self_attn.norm_q.weight", |
| "layers.30.self_attn.norm_k.weight", |
| "layers.30.self_attn.norm_added_q.weight", |
| "layers.30.self_attn.norm_added_k.weight", |
| "layers.30.input_layernorm.weight", |
| "layers.30.input_layernorm_moe_gen.weight", |
| "layers.30.post_attention_layernorm.weight", |
| "layers.30.post_attention_layernorm_moe_gen.weight", |
| "layers.31.self_attn.norm_q.weight", |
| "layers.31.self_attn.norm_k.weight", |
| "layers.31.self_attn.norm_added_q.weight", |
| "layers.31.self_attn.norm_added_k.weight", |
| "layers.31.input_layernorm.weight", |
| "layers.31.input_layernorm_moe_gen.weight", |
| "layers.31.post_attention_layernorm.weight", |
| "layers.31.post_attention_layernorm_moe_gen.weight", |
| "layers.32.self_attn.norm_q.weight", |
| "layers.32.self_attn.norm_k.weight", |
| "layers.32.self_attn.norm_added_q.weight", |
| "layers.32.self_attn.norm_added_k.weight", |
| "layers.32.input_layernorm.weight", |
| "layers.32.input_layernorm_moe_gen.weight", |
| "layers.32.post_attention_layernorm.weight", |
| "layers.32.post_attention_layernorm_moe_gen.weight", |
| "layers.33.self_attn.norm_q.weight", |
| "layers.33.self_attn.norm_k.weight", |
| "layers.33.self_attn.norm_added_q.weight", |
| "layers.33.self_attn.norm_added_k.weight", |
| "layers.33.input_layernorm.weight", |
| "layers.33.input_layernorm_moe_gen.weight", |
| "layers.33.post_attention_layernorm.weight", |
| "layers.33.post_attention_layernorm_moe_gen.weight", |
| "layers.34.self_attn.norm_q.weight", |
| "layers.34.self_attn.norm_k.weight", |
| "layers.34.self_attn.norm_added_q.weight", |
| "layers.34.self_attn.norm_added_k.weight", |
| "layers.34.input_layernorm.weight", |
| "layers.34.input_layernorm_moe_gen.weight", |
| "layers.34.post_attention_layernorm.weight", |
| "layers.34.post_attention_layernorm_moe_gen.weight", |
| "layers.35.self_attn.norm_q.weight", |
| "layers.35.self_attn.norm_k.weight", |
| "layers.35.self_attn.norm_added_q.weight", |
| "layers.35.self_attn.norm_added_k.weight", |
| "layers.35.input_layernorm.weight", |
| "layers.35.input_layernorm_moe_gen.weight", |
| "layers.35.post_attention_layernorm.weight", |
| "layers.35.post_attention_layernorm_moe_gen.weight", |
| "layers.36.self_attn.norm_q.weight", |
| "layers.36.self_attn.norm_k.weight", |
| "layers.36.self_attn.norm_added_q.weight", |
| "layers.36.self_attn.norm_added_k.weight", |
| "layers.36.input_layernorm.weight", |
| "layers.36.input_layernorm_moe_gen.weight", |
| "layers.36.post_attention_layernorm.weight", |
| "layers.36.post_attention_layernorm_moe_gen.weight", |
| "layers.37.self_attn.norm_q.weight", |
| "layers.37.self_attn.norm_k.weight", |
| "layers.37.self_attn.norm_added_q.weight", |
| "layers.37.self_attn.norm_added_k.weight", |
| "layers.37.input_layernorm.weight", |
| "layers.37.input_layernorm_moe_gen.weight", |
| "layers.37.post_attention_layernorm.weight", |
| "layers.37.post_attention_layernorm_moe_gen.weight", |
| "layers.38.self_attn.norm_q.weight", |
| "layers.38.self_attn.norm_k.weight", |
| "layers.38.self_attn.norm_added_q.weight", |
| "layers.38.self_attn.norm_added_k.weight", |
| "layers.38.input_layernorm.weight", |
| "layers.38.input_layernorm_moe_gen.weight", |
| "layers.38.post_attention_layernorm.weight", |
| "layers.38.post_attention_layernorm_moe_gen.weight", |
| "layers.39.self_attn.norm_q.weight", |
| "layers.39.self_attn.norm_k.weight", |
| "layers.39.self_attn.norm_added_q.weight", |
| "layers.39.self_attn.norm_added_k.weight", |
| "layers.39.input_layernorm.weight", |
| "layers.39.input_layernorm_moe_gen.weight", |
| "layers.39.post_attention_layernorm.weight", |
| "layers.39.post_attention_layernorm_moe_gen.weight", |
| "layers.40.self_attn.norm_q.weight", |
| "layers.40.self_attn.norm_k.weight", |
| "layers.40.self_attn.norm_added_q.weight", |
| "layers.40.self_attn.norm_added_k.weight", |
| "layers.40.input_layernorm.weight", |
| "layers.40.input_layernorm_moe_gen.weight", |
| "layers.40.post_attention_layernorm.weight", |
| "layers.40.post_attention_layernorm_moe_gen.weight", |
| "layers.41.self_attn.norm_q.weight", |
| "layers.41.self_attn.norm_k.weight", |
| "layers.41.self_attn.norm_added_q.weight", |
| "layers.41.self_attn.norm_added_k.weight", |
| "layers.41.input_layernorm.weight", |
| "layers.41.input_layernorm_moe_gen.weight", |
| "layers.41.post_attention_layernorm.weight", |
| "layers.41.post_attention_layernorm_moe_gen.weight", |
| "layers.42.self_attn.norm_q.weight", |
| "layers.42.self_attn.norm_k.weight", |
| "layers.42.self_attn.norm_added_q.weight", |
| "layers.42.self_attn.norm_added_k.weight", |
| "layers.42.input_layernorm.weight", |
| "layers.42.input_layernorm_moe_gen.weight", |
| "layers.42.post_attention_layernorm.weight", |
| "layers.42.post_attention_layernorm_moe_gen.weight", |
| "layers.43.self_attn.norm_q.weight", |
| "layers.43.self_attn.norm_k.weight", |
| "layers.43.self_attn.norm_added_q.weight", |
| "layers.43.self_attn.norm_added_k.weight", |
| "layers.43.input_layernorm.weight", |
| "layers.43.input_layernorm_moe_gen.weight", |
| "layers.43.post_attention_layernorm.weight", |
| "layers.43.post_attention_layernorm_moe_gen.weight", |
| "layers.44.self_attn.norm_q.weight", |
| "layers.44.self_attn.norm_k.weight", |
| "layers.44.self_attn.norm_added_q.weight", |
| "layers.44.self_attn.norm_added_k.weight", |
| "layers.44.input_layernorm.weight", |
| "layers.44.input_layernorm_moe_gen.weight", |
| "layers.44.post_attention_layernorm.weight", |
| "layers.44.post_attention_layernorm_moe_gen.weight", |
| "layers.45.self_attn.norm_q.weight", |
| "layers.45.self_attn.norm_k.weight", |
| "layers.45.self_attn.norm_added_q.weight", |
| "layers.45.self_attn.norm_added_k.weight", |
| "layers.45.input_layernorm.weight", |
| "layers.45.input_layernorm_moe_gen.weight", |
| "layers.45.post_attention_layernorm.weight", |
| "layers.45.post_attention_layernorm_moe_gen.weight", |
| "layers.46.self_attn.norm_q.weight", |
| "layers.46.self_attn.norm_k.weight", |
| "layers.46.self_attn.norm_added_q.weight", |
| "layers.46.self_attn.norm_added_k.weight", |
| "layers.46.input_layernorm.weight", |
| "layers.46.input_layernorm_moe_gen.weight", |
| "layers.46.post_attention_layernorm.weight", |
| "layers.46.post_attention_layernorm_moe_gen.weight", |
| "layers.47.self_attn.norm_q.weight", |
| "layers.47.self_attn.norm_k.weight", |
| "layers.47.self_attn.norm_added_q.weight", |
| "layers.47.self_attn.norm_added_k.weight", |
| "layers.47.input_layernorm.weight", |
| "layers.47.input_layernorm_moe_gen.weight", |
| "layers.47.post_attention_layernorm.weight", |
| "layers.47.post_attention_layernorm_moe_gen.weight", |
| "layers.48.self_attn.norm_q.weight", |
| "layers.48.self_attn.norm_k.weight", |
| "layers.48.self_attn.norm_added_q.weight", |
| "layers.48.self_attn.norm_added_k.weight", |
| "layers.48.input_layernorm.weight", |
| "layers.48.input_layernorm_moe_gen.weight", |
| "layers.48.post_attention_layernorm.weight", |
| "layers.48.post_attention_layernorm_moe_gen.weight", |
| "layers.49.self_attn.norm_q.weight", |
| "layers.49.self_attn.norm_k.weight", |
| "layers.49.self_attn.norm_added_q.weight", |
| "layers.49.self_attn.norm_added_k.weight", |
| "layers.49.input_layernorm.weight", |
| "layers.49.input_layernorm_moe_gen.weight", |
| "layers.49.post_attention_layernorm.weight", |
| "layers.49.post_attention_layernorm_moe_gen.weight", |
| "layers.50.self_attn.norm_q.weight", |
| "layers.50.self_attn.norm_k.weight", |
| "layers.50.self_attn.norm_added_q.weight", |
| "layers.50.self_attn.norm_added_k.weight", |
| "layers.50.input_layernorm.weight", |
| "layers.50.input_layernorm_moe_gen.weight", |
| "layers.50.post_attention_layernorm.weight", |
| "layers.50.post_attention_layernorm_moe_gen.weight", |
| "layers.51.self_attn.norm_q.weight", |
| "layers.51.self_attn.norm_k.weight", |
| "layers.51.self_attn.norm_added_q.weight", |
| "layers.51.self_attn.norm_added_k.weight", |
| "layers.51.input_layernorm.weight", |
| "layers.51.input_layernorm_moe_gen.weight", |
| "layers.51.post_attention_layernorm.weight", |
| "layers.51.post_attention_layernorm_moe_gen.weight", |
| "layers.52.self_attn.norm_q.weight", |
| "layers.52.self_attn.norm_k.weight", |
| "layers.52.self_attn.norm_added_q.weight", |
| "layers.52.self_attn.norm_added_k.weight", |
| "layers.52.input_layernorm.weight", |
| "layers.52.input_layernorm_moe_gen.weight", |
| "layers.52.post_attention_layernorm.weight", |
| "layers.52.post_attention_layernorm_moe_gen.weight", |
| "layers.53.self_attn.norm_q.weight", |
| "layers.53.self_attn.norm_k.weight", |
| "layers.53.self_attn.norm_added_q.weight", |
| "layers.53.self_attn.norm_added_k.weight", |
| "layers.53.input_layernorm.weight", |
| "layers.53.input_layernorm_moe_gen.weight", |
| "layers.53.post_attention_layernorm.weight", |
| "layers.53.post_attention_layernorm_moe_gen.weight", |
| "layers.54.self_attn.norm_q.weight", |
| "layers.54.self_attn.norm_k.weight", |
| "layers.54.self_attn.norm_added_q.weight", |
| "layers.54.self_attn.norm_added_k.weight", |
| "layers.54.input_layernorm.weight", |
| "layers.54.input_layernorm_moe_gen.weight", |
| "layers.54.post_attention_layernorm.weight", |
| "layers.54.post_attention_layernorm_moe_gen.weight", |
| "layers.55.self_attn.norm_q.weight", |
| "layers.55.self_attn.norm_k.weight", |
| "layers.55.self_attn.norm_added_q.weight", |
| "layers.55.self_attn.norm_added_k.weight", |
| "layers.55.input_layernorm.weight", |
| "layers.55.input_layernorm_moe_gen.weight", |
| "layers.55.post_attention_layernorm.weight", |
| "layers.55.post_attention_layernorm_moe_gen.weight", |
| "layers.56.self_attn.norm_q.weight", |
| "layers.56.self_attn.norm_k.weight", |
| "layers.56.self_attn.norm_added_q.weight", |
| "layers.56.self_attn.norm_added_k.weight", |
| "layers.56.input_layernorm.weight", |
| "layers.56.input_layernorm_moe_gen.weight", |
| "layers.56.post_attention_layernorm.weight", |
| "layers.56.post_attention_layernorm_moe_gen.weight", |
| "layers.57.self_attn.norm_q.weight", |
| "layers.57.self_attn.norm_k.weight", |
| "layers.57.self_attn.norm_added_q.weight", |
| "layers.57.self_attn.norm_added_k.weight", |
| "layers.57.input_layernorm.weight", |
| "layers.57.input_layernorm_moe_gen.weight", |
| "layers.57.post_attention_layernorm.weight", |
| "layers.57.post_attention_layernorm_moe_gen.weight", |
| "layers.58.self_attn.norm_q.weight", |
| "layers.58.self_attn.norm_k.weight", |
| "layers.58.self_attn.norm_added_q.weight", |
| "layers.58.self_attn.norm_added_k.weight", |
| "layers.58.input_layernorm.weight", |
| "layers.58.input_layernorm_moe_gen.weight", |
| "layers.58.post_attention_layernorm.weight", |
| "layers.58.post_attention_layernorm_moe_gen.weight", |
| "layers.59.self_attn.norm_q.weight", |
| "layers.59.self_attn.norm_k.weight", |
| "layers.59.self_attn.norm_added_q.weight", |
| "layers.59.self_attn.norm_added_k.weight", |
| "layers.59.input_layernorm.weight", |
| "layers.59.input_layernorm_moe_gen.weight", |
| "layers.59.post_attention_layernorm.weight", |
| "layers.59.post_attention_layernorm_moe_gen.weight", |
| "layers.60.self_attn.norm_q.weight", |
| "layers.60.self_attn.norm_k.weight", |
| "layers.60.self_attn.norm_added_q.weight", |
| "layers.60.self_attn.norm_added_k.weight", |
| "layers.60.input_layernorm.weight", |
| "layers.60.input_layernorm_moe_gen.weight", |
| "layers.60.post_attention_layernorm.weight", |
| "layers.60.post_attention_layernorm_moe_gen.weight", |
| "layers.61.self_attn.norm_q.weight", |
| "layers.61.self_attn.norm_k.weight", |
| "layers.61.self_attn.norm_added_q.weight", |
| "layers.61.self_attn.norm_added_k.weight", |
| "layers.61.input_layernorm.weight", |
| "layers.61.input_layernorm_moe_gen.weight", |
| "layers.61.post_attention_layernorm.weight", |
| "layers.61.post_attention_layernorm_moe_gen.weight", |
| "layers.62.self_attn.norm_q.weight", |
| "layers.62.self_attn.norm_k.weight", |
| "layers.62.self_attn.norm_added_q.weight", |
| "layers.62.self_attn.norm_added_k.weight", |
| "layers.62.input_layernorm.weight", |
| "layers.62.input_layernorm_moe_gen.weight", |
| "layers.62.post_attention_layernorm.weight", |
| "layers.62.post_attention_layernorm_moe_gen.weight", |
| "layers.63.self_attn.norm_q.weight", |
| "layers.63.self_attn.norm_k.weight", |
| "layers.63.self_attn.norm_added_q.weight", |
| "layers.63.self_attn.norm_added_k.weight", |
| "layers.63.input_layernorm.weight", |
| "layers.63.input_layernorm_moe_gen.weight", |
| "layers.63.post_attention_layernorm.weight", |
| "layers.63.post_attention_layernorm_moe_gen.weight", |
| "norm_moe_gen.weight" |
| ], |
| "modules_to_not_use_matmul": [], |
| "non_blocking": false, |
| "quant_conv": false, |
| "quant_embedding": false, |
| "quant_method": "sdnq", |
| "quantization_device": null, |
| "quantized_matmul_dtype": null, |
| "return_device": null, |
| "sdnq_version": "0.1.9", |
| "svd_rank": 32, |
| "svd_steps": 8, |
| "use_dynamic_quantization": false, |
| "use_grad_ckpt": true, |
| "use_hadamard": false, |
| "use_quantized_matmul": true, |
| "use_quantized_matmul_conv": false, |
| "use_static_quantization": true, |
| "use_stochastic_rounding": false, |
| "use_svd": false, |
| "weights_dtype": "int8" |
| } |
|
|
