FLUX.1-schnell-OrbitQuant-W3A3 / orbitquant_manifest.json
WaveCut's picture
Expand native comparison to ten complex prompts
181cf30 verified
Raw History Blame Contribute Delete
151 kB
{
"artifact_format": "orbitquant-v1",
"source_model_id": "black-forest-labs/FLUX.1-schnell",
"source_revision": "741f7c3ce8b383c54771c7003378a50191e9efe9",
"source_license": "apache-2.0",
"quant_method": "orbitquant",
"paper": "https://arxiv.org/abs/2607.02461",
"weight_bits": 3,
"activation_bits": 3,
"rotation": "rpbh",
"rotation_seed": 0,
"block_size": "paper",
"block_size_policy": "largest_power_of_two_dividing_dim",
"codebook": "lloyd_max",
"codebook_version": 2,
"row_norm_dtype": "bfloat16",
"runtime_mode": "auto_fused",
"activation_kernel_backend": "auto",
"activation_eps": 1e-10,
"adaln_group_size": 64,
"quantization_device": "cuda",
"weight_quantization_backend": "triton_cuda",
"quantization_staging_mode": "streaming",
"target_policy": "flux",
"adaln_policy": "int4_rtn_group64_bf16_activation",
"quantized_modules": [
"transformer_blocks.0.attn.to_q",
"transformer_blocks.0.attn.to_k",
"transformer_blocks.0.attn.to_v",
"transformer_blocks.0.attn.to_out.0",
"transformer_blocks.0.attn.add_q_proj",
"transformer_blocks.0.attn.add_k_proj",
"transformer_blocks.0.attn.add_v_proj",
"transformer_blocks.0.attn.to_add_out",
"transformer_blocks.0.ff.net.0.proj",
"transformer_blocks.0.ff.net.2",
"transformer_blocks.0.ff_context.net.0.proj",
"transformer_blocks.0.ff_context.net.2",
"transformer_blocks.1.attn.to_q",
"transformer_blocks.1.attn.to_k",
"transformer_blocks.1.attn.to_v",
"transformer_blocks.1.attn.to_out.0",
"transformer_blocks.1.attn.add_q_proj",
"transformer_blocks.1.attn.add_k_proj",
"transformer_blocks.1.attn.add_v_proj",
"transformer_blocks.1.attn.to_add_out",
"transformer_blocks.1.ff.net.0.proj",
"transformer_blocks.1.ff.net.2",
"transformer_blocks.1.ff_context.net.0.proj",
"transformer_blocks.1.ff_context.net.2",
"transformer_blocks.2.attn.to_q",
"transformer_blocks.2.attn.to_k",
"transformer_blocks.2.attn.to_v",
"transformer_blocks.2.attn.to_out.0",
"transformer_blocks.2.attn.add_q_proj",
"transformer_blocks.2.attn.add_k_proj",
"transformer_blocks.2.attn.add_v_proj",
"transformer_blocks.2.attn.to_add_out",
"transformer_blocks.2.ff.net.0.proj",
"transformer_blocks.2.ff.net.2",
"transformer_blocks.2.ff_context.net.0.proj",
"transformer_blocks.2.ff_context.net.2",
"transformer_blocks.3.attn.to_q",
"transformer_blocks.3.attn.to_k",
"transformer_blocks.3.attn.to_v",
"transformer_blocks.3.attn.to_out.0",
"transformer_blocks.3.attn.add_q_proj",
"transformer_blocks.3.attn.add_k_proj",
"transformer_blocks.3.attn.add_v_proj",
"transformer_blocks.3.attn.to_add_out",
"transformer_blocks.3.ff.net.0.proj",
"transformer_blocks.3.ff.net.2",
"transformer_blocks.3.ff_context.net.0.proj",
"transformer_blocks.3.ff_context.net.2",
"transformer_blocks.4.attn.to_q",
"transformer_blocks.4.attn.to_k",
"transformer_blocks.4.attn.to_v",
"transformer_blocks.4.attn.to_out.0",
"transformer_blocks.4.attn.add_q_proj",
"transformer_blocks.4.attn.add_k_proj",
"transformer_blocks.4.attn.add_v_proj",
"transformer_blocks.4.attn.to_add_out",
"transformer_blocks.4.ff.net.0.proj",
"transformer_blocks.4.ff.net.2",
"transformer_blocks.4.ff_context.net.0.proj",
"transformer_blocks.4.ff_context.net.2",
"transformer_blocks.5.attn.to_q",
"transformer_blocks.5.attn.to_k",
"transformer_blocks.5.attn.to_v",
"transformer_blocks.5.attn.to_out.0",
"transformer_blocks.5.attn.add_q_proj",
"transformer_blocks.5.attn.add_k_proj",
"transformer_blocks.5.attn.add_v_proj",
"transformer_blocks.5.attn.to_add_out",
"transformer_blocks.5.ff.net.0.proj",
"transformer_blocks.5.ff.net.2",
"transformer_blocks.5.ff_context.net.0.proj",
"transformer_blocks.5.ff_context.net.2",
"transformer_blocks.6.attn.to_q",
"transformer_blocks.6.attn.to_k",
"transformer_blocks.6.attn.to_v",
"transformer_blocks.6.attn.to_out.0",
"transformer_blocks.6.attn.add_q_proj",
"transformer_blocks.6.attn.add_k_proj",
"transformer_blocks.6.attn.add_v_proj",
"transformer_blocks.6.attn.to_add_out",
"transformer_blocks.6.ff.net.0.proj",
"transformer_blocks.6.ff.net.2",
"transformer_blocks.6.ff_context.net.0.proj",
"transformer_blocks.6.ff_context.net.2",
"transformer_blocks.7.attn.to_q",
"transformer_blocks.7.attn.to_k",
"transformer_blocks.7.attn.to_v",
"transformer_blocks.7.attn.to_out.0",
"transformer_blocks.7.attn.add_q_proj",
"transformer_blocks.7.attn.add_k_proj",
"transformer_blocks.7.attn.add_v_proj",
"transformer_blocks.7.attn.to_add_out",
"transformer_blocks.7.ff.net.0.proj",
"transformer_blocks.7.ff.net.2",
"transformer_blocks.7.ff_context.net.0.proj",
"transformer_blocks.7.ff_context.net.2",
"transformer_blocks.8.attn.to_q",
"transformer_blocks.8.attn.to_k",
"transformer_blocks.8.attn.to_v",
"transformer_blocks.8.attn.to_out.0",
"transformer_blocks.8.attn.add_q_proj",
"transformer_blocks.8.attn.add_k_proj",
"transformer_blocks.8.attn.add_v_proj",
"transformer_blocks.8.attn.to_add_out",
"transformer_blocks.8.ff.net.0.proj",
"transformer_blocks.8.ff.net.2",
"transformer_blocks.8.ff_context.net.0.proj",
"transformer_blocks.8.ff_context.net.2",
"transformer_blocks.9.attn.to_q",
"transformer_blocks.9.attn.to_k",
"transformer_blocks.9.attn.to_v",
"transformer_blocks.9.attn.to_out.0",
"transformer_blocks.9.attn.add_q_proj",
"transformer_blocks.9.attn.add_k_proj",
"transformer_blocks.9.attn.add_v_proj",
"transformer_blocks.9.attn.to_add_out",
"transformer_blocks.9.ff.net.0.proj",
"transformer_blocks.9.ff.net.2",
"transformer_blocks.9.ff_context.net.0.proj",
"transformer_blocks.9.ff_context.net.2",
"transformer_blocks.10.attn.to_q",
"transformer_blocks.10.attn.to_k",
"transformer_blocks.10.attn.to_v",
"transformer_blocks.10.attn.to_out.0",
"transformer_blocks.10.attn.add_q_proj",
"transformer_blocks.10.attn.add_k_proj",
"transformer_blocks.10.attn.add_v_proj",
"transformer_blocks.10.attn.to_add_out",
"transformer_blocks.10.ff.net.0.proj",
"transformer_blocks.10.ff.net.2",
"transformer_blocks.10.ff_context.net.0.proj",
"transformer_blocks.10.ff_context.net.2",
"transformer_blocks.11.attn.to_q",
"transformer_blocks.11.attn.to_k",
"transformer_blocks.11.attn.to_v",
"transformer_blocks.11.attn.to_out.0",
"transformer_blocks.11.attn.add_q_proj",
"transformer_blocks.11.attn.add_k_proj",
"transformer_blocks.11.attn.add_v_proj",
"transformer_blocks.11.attn.to_add_out",
"transformer_blocks.11.ff.net.0.proj",
"transformer_blocks.11.ff.net.2",
"transformer_blocks.11.ff_context.net.0.proj",
"transformer_blocks.11.ff_context.net.2",
"transformer_blocks.12.attn.to_q",
"transformer_blocks.12.attn.to_k",
"transformer_blocks.12.attn.to_v",
"transformer_blocks.12.attn.to_out.0",
"transformer_blocks.12.attn.add_q_proj",
"transformer_blocks.12.attn.add_k_proj",
"transformer_blocks.12.attn.add_v_proj",
"transformer_blocks.12.attn.to_add_out",
"transformer_blocks.12.ff.net.0.proj",
"transformer_blocks.12.ff.net.2",
"transformer_blocks.12.ff_context.net.0.proj",
"transformer_blocks.12.ff_context.net.2",
"transformer_blocks.13.attn.to_q",
"transformer_blocks.13.attn.to_k",
"transformer_blocks.13.attn.to_v",
"transformer_blocks.13.attn.to_out.0",
"transformer_blocks.13.attn.add_q_proj",
"transformer_blocks.13.attn.add_k_proj",
"transformer_blocks.13.attn.add_v_proj",
"transformer_blocks.13.attn.to_add_out",
"transformer_blocks.13.ff.net.0.proj",
"transformer_blocks.13.ff.net.2",
"transformer_blocks.13.ff_context.net.0.proj",
"transformer_blocks.13.ff_context.net.2",
"transformer_blocks.14.attn.to_q",
"transformer_blocks.14.attn.to_k",
"transformer_blocks.14.attn.to_v",
"transformer_blocks.14.attn.to_out.0",
"transformer_blocks.14.attn.add_q_proj",
"transformer_blocks.14.attn.add_k_proj",
"transformer_blocks.14.attn.add_v_proj",
"transformer_blocks.14.attn.to_add_out",
"transformer_blocks.14.ff.net.0.proj",
"transformer_blocks.14.ff.net.2",
"transformer_blocks.14.ff_context.net.0.proj",
"transformer_blocks.14.ff_context.net.2",
"transformer_blocks.15.attn.to_q",
"transformer_blocks.15.attn.to_k",
"transformer_blocks.15.attn.to_v",
"transformer_blocks.15.attn.to_out.0",
"transformer_blocks.15.attn.add_q_proj",
"transformer_blocks.15.attn.add_k_proj",
"transformer_blocks.15.attn.add_v_proj",
"transformer_blocks.15.attn.to_add_out",
"transformer_blocks.15.ff.net.0.proj",
"transformer_blocks.15.ff.net.2",
"transformer_blocks.15.ff_context.net.0.proj",
"transformer_blocks.15.ff_context.net.2",
"transformer_blocks.16.attn.to_q",
"transformer_blocks.16.attn.to_k",
"transformer_blocks.16.attn.to_v",
"transformer_blocks.16.attn.to_out.0",
"transformer_blocks.16.attn.add_q_proj",
"transformer_blocks.16.attn.add_k_proj",
"transformer_blocks.16.attn.add_v_proj",
"transformer_blocks.16.attn.to_add_out",
"transformer_blocks.16.ff.net.0.proj",
"transformer_blocks.16.ff.net.2",
"transformer_blocks.16.ff_context.net.0.proj",
"transformer_blocks.16.ff_context.net.2",
"transformer_blocks.17.attn.to_q",
"transformer_blocks.17.attn.to_k",
"transformer_blocks.17.attn.to_v",
"transformer_blocks.17.attn.to_out.0",
"transformer_blocks.17.attn.add_q_proj",
"transformer_blocks.17.attn.add_k_proj",
"transformer_blocks.17.attn.add_v_proj",
"transformer_blocks.17.attn.to_add_out",
"transformer_blocks.17.ff.net.0.proj",
"transformer_blocks.17.ff.net.2",
"transformer_blocks.17.ff_context.net.0.proj",
"transformer_blocks.17.ff_context.net.2",
"transformer_blocks.18.attn.to_q",
"transformer_blocks.18.attn.to_k",
"transformer_blocks.18.attn.to_v",
"transformer_blocks.18.attn.to_out.0",
"transformer_blocks.18.attn.add_q_proj",
"transformer_blocks.18.attn.add_k_proj",
"transformer_blocks.18.attn.add_v_proj",
"transformer_blocks.18.attn.to_add_out",
"transformer_blocks.18.ff.net.0.proj",
"transformer_blocks.18.ff.net.2",
"transformer_blocks.18.ff_context.net.0.proj",
"transformer_blocks.18.ff_context.net.2",
"single_transformer_blocks.0.proj_mlp",
"single_transformer_blocks.0.proj_out",
"single_transformer_blocks.0.attn.to_q",
"single_transformer_blocks.0.attn.to_k",
"single_transformer_blocks.0.attn.to_v",
"single_transformer_blocks.1.proj_mlp",
"single_transformer_blocks.1.proj_out",
"single_transformer_blocks.1.attn.to_q",
"single_transformer_blocks.1.attn.to_k",
"single_transformer_blocks.1.attn.to_v",
"single_transformer_blocks.2.proj_mlp",
"single_transformer_blocks.2.proj_out",
"single_transformer_blocks.2.attn.to_q",
"single_transformer_blocks.2.attn.to_k",
"single_transformer_blocks.2.attn.to_v",
"single_transformer_blocks.3.proj_mlp",
"single_transformer_blocks.3.proj_out",
"single_transformer_blocks.3.attn.to_q",
"single_transformer_blocks.3.attn.to_k",
"single_transformer_blocks.3.attn.to_v",
"single_transformer_blocks.4.proj_mlp",
"single_transformer_blocks.4.proj_out",
"single_transformer_blocks.4.attn.to_q",
"single_transformer_blocks.4.attn.to_k",
"single_transformer_blocks.4.attn.to_v",
"single_transformer_blocks.5.proj_mlp",
"single_transformer_blocks.5.proj_out",
"single_transformer_blocks.5.attn.to_q",
"single_transformer_blocks.5.attn.to_k",
"single_transformer_blocks.5.attn.to_v",
"single_transformer_blocks.6.proj_mlp",
"single_transformer_blocks.6.proj_out",
"single_transformer_blocks.6.attn.to_q",
"single_transformer_blocks.6.attn.to_k",
"single_transformer_blocks.6.attn.to_v",
"single_transformer_blocks.7.proj_mlp",
"single_transformer_blocks.7.proj_out",
"single_transformer_blocks.7.attn.to_q",
"single_transformer_blocks.7.attn.to_k",
"single_transformer_blocks.7.attn.to_v",
"single_transformer_blocks.8.proj_mlp",
"single_transformer_blocks.8.proj_out",
"single_transformer_blocks.8.attn.to_q",
"single_transformer_blocks.8.attn.to_k",
"single_transformer_blocks.8.attn.to_v",
"single_transformer_blocks.9.proj_mlp",
"single_transformer_blocks.9.proj_out",
"single_transformer_blocks.9.attn.to_q",
"single_transformer_blocks.9.attn.to_k",
"single_transformer_blocks.9.attn.to_v",
"single_transformer_blocks.10.proj_mlp",
"single_transformer_blocks.10.proj_out",
"single_transformer_blocks.10.attn.to_q",
"single_transformer_blocks.10.attn.to_k",
"single_transformer_blocks.10.attn.to_v",
"single_transformer_blocks.11.proj_mlp",
"single_transformer_blocks.11.proj_out",
"single_transformer_blocks.11.attn.to_q",
"single_transformer_blocks.11.attn.to_k",
"single_transformer_blocks.11.attn.to_v",
"single_transformer_blocks.12.proj_mlp",
"single_transformer_blocks.12.proj_out",
"single_transformer_blocks.12.attn.to_q",
"single_transformer_blocks.12.attn.to_k",
"single_transformer_blocks.12.attn.to_v",
"single_transformer_blocks.13.proj_mlp",
"single_transformer_blocks.13.proj_out",
"single_transformer_blocks.13.attn.to_q",
"single_transformer_blocks.13.attn.to_k",
"single_transformer_blocks.13.attn.to_v",
"single_transformer_blocks.14.proj_mlp",
"single_transformer_blocks.14.proj_out",
"single_transformer_blocks.14.attn.to_q",
"single_transformer_blocks.14.attn.to_k",
"single_transformer_blocks.14.attn.to_v",
"single_transformer_blocks.15.proj_mlp",
"single_transformer_blocks.15.proj_out",
"single_transformer_blocks.15.attn.to_q",
"single_transformer_blocks.15.attn.to_k",
"single_transformer_blocks.15.attn.to_v",
"single_transformer_blocks.16.proj_mlp",
"single_transformer_blocks.16.proj_out",
"single_transformer_blocks.16.attn.to_q",
"single_transformer_blocks.16.attn.to_k",
"single_transformer_blocks.16.attn.to_v",
"single_transformer_blocks.17.proj_mlp",
"single_transformer_blocks.17.proj_out",
"single_transformer_blocks.17.attn.to_q",
"single_transformer_blocks.17.attn.to_k",
"single_transformer_blocks.17.attn.to_v",
"single_transformer_blocks.18.proj_mlp",
"single_transformer_blocks.18.proj_out",
"single_transformer_blocks.18.attn.to_q",
"single_transformer_blocks.18.attn.to_k",
"single_transformer_blocks.18.attn.to_v",
"single_transformer_blocks.19.proj_mlp",
"single_transformer_blocks.19.proj_out",
"single_transformer_blocks.19.attn.to_q",
"single_transformer_blocks.19.attn.to_k",
"single_transformer_blocks.19.attn.to_v",
"single_transformer_blocks.20.proj_mlp",
"single_transformer_blocks.20.proj_out",
"single_transformer_blocks.20.attn.to_q",
"single_transformer_blocks.20.attn.to_k",
"single_transformer_blocks.20.attn.to_v",
"single_transformer_blocks.21.proj_mlp",
"single_transformer_blocks.21.proj_out",
"single_transformer_blocks.21.attn.to_q",
"single_transformer_blocks.21.attn.to_k",
"single_transformer_blocks.21.attn.to_v",
"single_transformer_blocks.22.proj_mlp",
"single_transformer_blocks.22.proj_out",
"single_transformer_blocks.22.attn.to_q",
"single_transformer_blocks.22.attn.to_k",
"single_transformer_blocks.22.attn.to_v",
"single_transformer_blocks.23.proj_mlp",
"single_transformer_blocks.23.proj_out",
"single_transformer_blocks.23.attn.to_q",
"single_transformer_blocks.23.attn.to_k",
"single_transformer_blocks.23.attn.to_v",
"single_transformer_blocks.24.proj_mlp",
"single_transformer_blocks.24.proj_out",
"single_transformer_blocks.24.attn.to_q",
"single_transformer_blocks.24.attn.to_k",
"single_transformer_blocks.24.attn.to_v",
"single_transformer_blocks.25.proj_mlp",
"single_transformer_blocks.25.proj_out",
"single_transformer_blocks.25.attn.to_q",
"single_transformer_blocks.25.attn.to_k",
"single_transformer_blocks.25.attn.to_v",
"single_transformer_blocks.26.proj_mlp",
"single_transformer_blocks.26.proj_out",
"single_transformer_blocks.26.attn.to_q",
"single_transformer_blocks.26.attn.to_k",
"single_transformer_blocks.26.attn.to_v",
"single_transformer_blocks.27.proj_mlp",
"single_transformer_blocks.27.proj_out",
"single_transformer_blocks.27.attn.to_q",
"single_transformer_blocks.27.attn.to_k",
"single_transformer_blocks.27.attn.to_v",
"single_transformer_blocks.28.proj_mlp",
"single_transformer_blocks.28.proj_out",
"single_transformer_blocks.28.attn.to_q",
"single_transformer_blocks.28.attn.to_k",
"single_transformer_blocks.28.attn.to_v",
"single_transformer_blocks.29.proj_mlp",
"single_transformer_blocks.29.proj_out",
"single_transformer_blocks.29.attn.to_q",
"single_transformer_blocks.29.attn.to_k",
"single_transformer_blocks.29.attn.to_v",
"single_transformer_blocks.30.proj_mlp",
"single_transformer_blocks.30.proj_out",
"single_transformer_blocks.30.attn.to_q",
"single_transformer_blocks.30.attn.to_k",
"single_transformer_blocks.30.attn.to_v",
"single_transformer_blocks.31.proj_mlp",
"single_transformer_blocks.31.proj_out",
"single_transformer_blocks.31.attn.to_q",
"single_transformer_blocks.31.attn.to_k",
"single_transformer_blocks.31.attn.to_v",
"single_transformer_blocks.32.proj_mlp",
"single_transformer_blocks.32.proj_out",
"single_transformer_blocks.32.attn.to_q",
"single_transformer_blocks.32.attn.to_k",
"single_transformer_blocks.32.attn.to_v",
"single_transformer_blocks.33.proj_mlp",
"single_transformer_blocks.33.proj_out",
"single_transformer_blocks.33.attn.to_q",
"single_transformer_blocks.33.attn.to_k",
"single_transformer_blocks.33.attn.to_v",
"single_transformer_blocks.34.proj_mlp",
"single_transformer_blocks.34.proj_out",
"single_transformer_blocks.34.attn.to_q",
"single_transformer_blocks.34.attn.to_k",
"single_transformer_blocks.34.attn.to_v",
"single_transformer_blocks.35.proj_mlp",
"single_transformer_blocks.35.proj_out",
"single_transformer_blocks.35.attn.to_q",
"single_transformer_blocks.35.attn.to_k",
"single_transformer_blocks.35.attn.to_v",
"single_transformer_blocks.36.proj_mlp",
"single_transformer_blocks.36.proj_out",
"single_transformer_blocks.36.attn.to_q",
"single_transformer_blocks.36.attn.to_k",
"single_transformer_blocks.36.attn.to_v",
"single_transformer_blocks.37.proj_mlp",
"single_transformer_blocks.37.proj_out",
"single_transformer_blocks.37.attn.to_q",
"single_transformer_blocks.37.attn.to_k",
"single_transformer_blocks.37.attn.to_v"
],
"adaln_modules": [
"transformer_blocks.0.norm1.linear",
"transformer_blocks.0.norm1_context.linear",
"transformer_blocks.1.norm1.linear",
"transformer_blocks.1.norm1_context.linear",
"transformer_blocks.2.norm1.linear",
"transformer_blocks.2.norm1_context.linear",
"transformer_blocks.3.norm1.linear",
"transformer_blocks.3.norm1_context.linear",
"transformer_blocks.4.norm1.linear",
"transformer_blocks.4.norm1_context.linear",
"transformer_blocks.5.norm1.linear",
"transformer_blocks.5.norm1_context.linear",
"transformer_blocks.6.norm1.linear",
"transformer_blocks.6.norm1_context.linear",
"transformer_blocks.7.norm1.linear",
"transformer_blocks.7.norm1_context.linear",
"transformer_blocks.8.norm1.linear",
"transformer_blocks.8.norm1_context.linear",
"transformer_blocks.9.norm1.linear",
"transformer_blocks.9.norm1_context.linear",
"transformer_blocks.10.norm1.linear",
"transformer_blocks.10.norm1_context.linear",
"transformer_blocks.11.norm1.linear",
"transformer_blocks.11.norm1_context.linear",
"transformer_blocks.12.norm1.linear",
"transformer_blocks.12.norm1_context.linear",
"transformer_blocks.13.norm1.linear",
"transformer_blocks.13.norm1_context.linear",
"transformer_blocks.14.norm1.linear",
"transformer_blocks.14.norm1_context.linear",
"transformer_blocks.15.norm1.linear",
"transformer_blocks.15.norm1_context.linear",
"transformer_blocks.16.norm1.linear",
"transformer_blocks.16.norm1_context.linear",
"transformer_blocks.17.norm1.linear",
"transformer_blocks.17.norm1_context.linear",
"transformer_blocks.18.norm1.linear",
"transformer_blocks.18.norm1_context.linear",
"single_transformer_blocks.0.norm.linear",
"single_transformer_blocks.1.norm.linear",
"single_transformer_blocks.2.norm.linear",
"single_transformer_blocks.3.norm.linear",
"single_transformer_blocks.4.norm.linear",
"single_transformer_blocks.5.norm.linear",
"single_transformer_blocks.6.norm.linear",
"single_transformer_blocks.7.norm.linear",
"single_transformer_blocks.8.norm.linear",
"single_transformer_blocks.9.norm.linear",
"single_transformer_blocks.10.norm.linear",
"single_transformer_blocks.11.norm.linear",
"single_transformer_blocks.12.norm.linear",
"single_transformer_blocks.13.norm.linear",
"single_transformer_blocks.14.norm.linear",
"single_transformer_blocks.15.norm.linear",
"single_transformer_blocks.16.norm.linear",
"single_transformer_blocks.17.norm.linear",
"single_transformer_blocks.18.norm.linear",
"single_transformer_blocks.19.norm.linear",
"single_transformer_blocks.20.norm.linear",
"single_transformer_blocks.21.norm.linear",
"single_transformer_blocks.22.norm.linear",
"single_transformer_blocks.23.norm.linear",
"single_transformer_blocks.24.norm.linear",
"single_transformer_blocks.25.norm.linear",
"single_transformer_blocks.26.norm.linear",
"single_transformer_blocks.27.norm.linear",
"single_transformer_blocks.28.norm.linear",
"single_transformer_blocks.29.norm.linear",
"single_transformer_blocks.30.norm.linear",
"single_transformer_blocks.31.norm.linear",
"single_transformer_blocks.32.norm.linear",
"single_transformer_blocks.33.norm.linear",
"single_transformer_blocks.34.norm.linear",
"single_transformer_blocks.35.norm.linear",
"single_transformer_blocks.36.norm.linear",
"single_transformer_blocks.37.norm.linear"
],
"skipped_modules": [
"time_text_embed.timestep_embedder.linear_1",
"time_text_embed.timestep_embedder.linear_2",
"time_text_embed.text_embedder.linear_1",
"time_text_embed.text_embedder.linear_2",
"context_embedder",
"x_embedder",
"norm_out.linear",
"proj_out"
],
"module_shapes": {
"time_text_embed.timestep_embedder.linear_1.weight": [
3072,
256
],
"time_text_embed.timestep_embedder.linear_1.bias": [
3072
],
"time_text_embed.timestep_embedder.linear_2.weight": [
3072,
3072
],
"time_text_embed.timestep_embedder.linear_2.bias": [
3072
],
"time_text_embed.text_embedder.linear_1.weight": [
3072,
768
],
"time_text_embed.text_embedder.linear_1.bias": [
3072
],
"time_text_embed.text_embedder.linear_2.weight": [
3072,
3072
],
"time_text_embed.text_embedder.linear_2.bias": [
3072
],
"context_embedder.weight": [
3072,
4096
],
"context_embedder.bias": [
3072
],
"x_embedder.weight": [
3072,
64
],
"x_embedder.bias": [
3072
],
"transformer_blocks.0.norm1.linear.bias": [
18432
],
"transformer_blocks.0.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.0.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.0.norm1_context.linear.bias": [
18432
],
"transformer_blocks.0.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.0.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.0.attn.norm_q.weight": [
128
],
"transformer_blocks.0.attn.norm_k.weight": [
128
],
"transformer_blocks.0.attn.to_q.bias": [
3072
],
"transformer_blocks.0.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.to_q.row_norms": [
3072
],
"transformer_blocks.0.attn.to_k.bias": [
3072
],
"transformer_blocks.0.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.to_k.row_norms": [
3072
],
"transformer_blocks.0.attn.to_v.bias": [
3072
],
"transformer_blocks.0.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.to_v.row_norms": [
3072
],
"transformer_blocks.0.attn.to_out.0.bias": [
3072
],
"transformer_blocks.0.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.0.attn.norm_added_q.weight": [
128
],
"transformer_blocks.0.attn.norm_added_k.weight": [
128
],
"transformer_blocks.0.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.0.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.0.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.0.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.0.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.0.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.0.attn.to_add_out.bias": [
3072
],
"transformer_blocks.0.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.0.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.0.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.0.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.0.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.0.ff.net.2.bias": [
3072
],
"transformer_blocks.0.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.0.ff.net.2.row_norms": [
3072
],
"transformer_blocks.0.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.0.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.0.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.0.ff_context.net.2.bias": [
3072
],
"transformer_blocks.0.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.0.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.1.norm1.linear.bias": [
18432
],
"transformer_blocks.1.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.1.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.1.norm1_context.linear.bias": [
18432
],
"transformer_blocks.1.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.1.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.1.attn.norm_q.weight": [
128
],
"transformer_blocks.1.attn.norm_k.weight": [
128
],
"transformer_blocks.1.attn.to_q.bias": [
3072
],
"transformer_blocks.1.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.to_q.row_norms": [
3072
],
"transformer_blocks.1.attn.to_k.bias": [
3072
],
"transformer_blocks.1.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.to_k.row_norms": [
3072
],
"transformer_blocks.1.attn.to_v.bias": [
3072
],
"transformer_blocks.1.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.to_v.row_norms": [
3072
],
"transformer_blocks.1.attn.to_out.0.bias": [
3072
],
"transformer_blocks.1.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.1.attn.norm_added_q.weight": [
128
],
"transformer_blocks.1.attn.norm_added_k.weight": [
128
],
"transformer_blocks.1.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.1.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.1.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.1.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.1.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.1.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.1.attn.to_add_out.bias": [
3072
],
"transformer_blocks.1.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.1.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.1.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.1.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.1.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.1.ff.net.2.bias": [
3072
],
"transformer_blocks.1.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.1.ff.net.2.row_norms": [
3072
],
"transformer_blocks.1.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.1.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.1.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.1.ff_context.net.2.bias": [
3072
],
"transformer_blocks.1.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.1.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.2.norm1.linear.bias": [
18432
],
"transformer_blocks.2.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.2.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.2.norm1_context.linear.bias": [
18432
],
"transformer_blocks.2.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.2.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.2.attn.norm_q.weight": [
128
],
"transformer_blocks.2.attn.norm_k.weight": [
128
],
"transformer_blocks.2.attn.to_q.bias": [
3072
],
"transformer_blocks.2.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.to_q.row_norms": [
3072
],
"transformer_blocks.2.attn.to_k.bias": [
3072
],
"transformer_blocks.2.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.to_k.row_norms": [
3072
],
"transformer_blocks.2.attn.to_v.bias": [
3072
],
"transformer_blocks.2.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.to_v.row_norms": [
3072
],
"transformer_blocks.2.attn.to_out.0.bias": [
3072
],
"transformer_blocks.2.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.2.attn.norm_added_q.weight": [
128
],
"transformer_blocks.2.attn.norm_added_k.weight": [
128
],
"transformer_blocks.2.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.2.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.2.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.2.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.2.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.2.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.2.attn.to_add_out.bias": [
3072
],
"transformer_blocks.2.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.2.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.2.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.2.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.2.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.2.ff.net.2.bias": [
3072
],
"transformer_blocks.2.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.2.ff.net.2.row_norms": [
3072
],
"transformer_blocks.2.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.2.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.2.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.2.ff_context.net.2.bias": [
3072
],
"transformer_blocks.2.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.2.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.3.norm1.linear.bias": [
18432
],
"transformer_blocks.3.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.3.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.3.norm1_context.linear.bias": [
18432
],
"transformer_blocks.3.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.3.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.3.attn.norm_q.weight": [
128
],
"transformer_blocks.3.attn.norm_k.weight": [
128
],
"transformer_blocks.3.attn.to_q.bias": [
3072
],
"transformer_blocks.3.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.to_q.row_norms": [
3072
],
"transformer_blocks.3.attn.to_k.bias": [
3072
],
"transformer_blocks.3.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.to_k.row_norms": [
3072
],
"transformer_blocks.3.attn.to_v.bias": [
3072
],
"transformer_blocks.3.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.to_v.row_norms": [
3072
],
"transformer_blocks.3.attn.to_out.0.bias": [
3072
],
"transformer_blocks.3.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.3.attn.norm_added_q.weight": [
128
],
"transformer_blocks.3.attn.norm_added_k.weight": [
128
],
"transformer_blocks.3.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.3.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.3.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.3.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.3.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.3.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.3.attn.to_add_out.bias": [
3072
],
"transformer_blocks.3.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.3.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.3.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.3.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.3.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.3.ff.net.2.bias": [
3072
],
"transformer_blocks.3.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.3.ff.net.2.row_norms": [
3072
],
"transformer_blocks.3.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.3.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.3.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.3.ff_context.net.2.bias": [
3072
],
"transformer_blocks.3.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.3.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.4.norm1.linear.bias": [
18432
],
"transformer_blocks.4.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.4.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.4.norm1_context.linear.bias": [
18432
],
"transformer_blocks.4.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.4.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.4.attn.norm_q.weight": [
128
],
"transformer_blocks.4.attn.norm_k.weight": [
128
],
"transformer_blocks.4.attn.to_q.bias": [
3072
],
"transformer_blocks.4.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.to_q.row_norms": [
3072
],
"transformer_blocks.4.attn.to_k.bias": [
3072
],
"transformer_blocks.4.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.to_k.row_norms": [
3072
],
"transformer_blocks.4.attn.to_v.bias": [
3072
],
"transformer_blocks.4.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.to_v.row_norms": [
3072
],
"transformer_blocks.4.attn.to_out.0.bias": [
3072
],
"transformer_blocks.4.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.4.attn.norm_added_q.weight": [
128
],
"transformer_blocks.4.attn.norm_added_k.weight": [
128
],
"transformer_blocks.4.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.4.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.4.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.4.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.4.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.4.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.4.attn.to_add_out.bias": [
3072
],
"transformer_blocks.4.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.4.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.4.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.4.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.4.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.4.ff.net.2.bias": [
3072
],
"transformer_blocks.4.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.4.ff.net.2.row_norms": [
3072
],
"transformer_blocks.4.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.4.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.4.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.4.ff_context.net.2.bias": [
3072
],
"transformer_blocks.4.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.4.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.5.norm1.linear.bias": [
18432
],
"transformer_blocks.5.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.5.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.5.norm1_context.linear.bias": [
18432
],
"transformer_blocks.5.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.5.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.5.attn.norm_q.weight": [
128
],
"transformer_blocks.5.attn.norm_k.weight": [
128
],
"transformer_blocks.5.attn.to_q.bias": [
3072
],
"transformer_blocks.5.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.to_q.row_norms": [
3072
],
"transformer_blocks.5.attn.to_k.bias": [
3072
],
"transformer_blocks.5.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.to_k.row_norms": [
3072
],
"transformer_blocks.5.attn.to_v.bias": [
3072
],
"transformer_blocks.5.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.to_v.row_norms": [
3072
],
"transformer_blocks.5.attn.to_out.0.bias": [
3072
],
"transformer_blocks.5.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.5.attn.norm_added_q.weight": [
128
],
"transformer_blocks.5.attn.norm_added_k.weight": [
128
],
"transformer_blocks.5.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.5.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.5.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.5.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.5.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.5.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.5.attn.to_add_out.bias": [
3072
],
"transformer_blocks.5.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.5.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.5.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.5.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.5.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.5.ff.net.2.bias": [
3072
],
"transformer_blocks.5.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.5.ff.net.2.row_norms": [
3072
],
"transformer_blocks.5.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.5.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.5.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.5.ff_context.net.2.bias": [
3072
],
"transformer_blocks.5.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.5.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.6.norm1.linear.bias": [
18432
],
"transformer_blocks.6.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.6.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.6.norm1_context.linear.bias": [
18432
],
"transformer_blocks.6.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.6.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.6.attn.norm_q.weight": [
128
],
"transformer_blocks.6.attn.norm_k.weight": [
128
],
"transformer_blocks.6.attn.to_q.bias": [
3072
],
"transformer_blocks.6.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.to_q.row_norms": [
3072
],
"transformer_blocks.6.attn.to_k.bias": [
3072
],
"transformer_blocks.6.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.to_k.row_norms": [
3072
],
"transformer_blocks.6.attn.to_v.bias": [
3072
],
"transformer_blocks.6.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.to_v.row_norms": [
3072
],
"transformer_blocks.6.attn.to_out.0.bias": [
3072
],
"transformer_blocks.6.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.6.attn.norm_added_q.weight": [
128
],
"transformer_blocks.6.attn.norm_added_k.weight": [
128
],
"transformer_blocks.6.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.6.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.6.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.6.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.6.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.6.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.6.attn.to_add_out.bias": [
3072
],
"transformer_blocks.6.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.6.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.6.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.6.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.6.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.6.ff.net.2.bias": [
3072
],
"transformer_blocks.6.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.6.ff.net.2.row_norms": [
3072
],
"transformer_blocks.6.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.6.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.6.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.6.ff_context.net.2.bias": [
3072
],
"transformer_blocks.6.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.6.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.7.norm1.linear.bias": [
18432
],
"transformer_blocks.7.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.7.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.7.norm1_context.linear.bias": [
18432
],
"transformer_blocks.7.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.7.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.7.attn.norm_q.weight": [
128
],
"transformer_blocks.7.attn.norm_k.weight": [
128
],
"transformer_blocks.7.attn.to_q.bias": [
3072
],
"transformer_blocks.7.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.to_q.row_norms": [
3072
],
"transformer_blocks.7.attn.to_k.bias": [
3072
],
"transformer_blocks.7.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.to_k.row_norms": [
3072
],
"transformer_blocks.7.attn.to_v.bias": [
3072
],
"transformer_blocks.7.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.to_v.row_norms": [
3072
],
"transformer_blocks.7.attn.to_out.0.bias": [
3072
],
"transformer_blocks.7.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.7.attn.norm_added_q.weight": [
128
],
"transformer_blocks.7.attn.norm_added_k.weight": [
128
],
"transformer_blocks.7.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.7.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.7.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.7.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.7.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.7.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.7.attn.to_add_out.bias": [
3072
],
"transformer_blocks.7.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.7.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.7.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.7.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.7.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.7.ff.net.2.bias": [
3072
],
"transformer_blocks.7.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.7.ff.net.2.row_norms": [
3072
],
"transformer_blocks.7.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.7.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.7.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.7.ff_context.net.2.bias": [
3072
],
"transformer_blocks.7.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.7.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.8.norm1.linear.bias": [
18432
],
"transformer_blocks.8.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.8.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.8.norm1_context.linear.bias": [
18432
],
"transformer_blocks.8.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.8.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.8.attn.norm_q.weight": [
128
],
"transformer_blocks.8.attn.norm_k.weight": [
128
],
"transformer_blocks.8.attn.to_q.bias": [
3072
],
"transformer_blocks.8.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.to_q.row_norms": [
3072
],
"transformer_blocks.8.attn.to_k.bias": [
3072
],
"transformer_blocks.8.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.to_k.row_norms": [
3072
],
"transformer_blocks.8.attn.to_v.bias": [
3072
],
"transformer_blocks.8.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.to_v.row_norms": [
3072
],
"transformer_blocks.8.attn.to_out.0.bias": [
3072
],
"transformer_blocks.8.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.8.attn.norm_added_q.weight": [
128
],
"transformer_blocks.8.attn.norm_added_k.weight": [
128
],
"transformer_blocks.8.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.8.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.8.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.8.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.8.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.8.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.8.attn.to_add_out.bias": [
3072
],
"transformer_blocks.8.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.8.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.8.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.8.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.8.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.8.ff.net.2.bias": [
3072
],
"transformer_blocks.8.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.8.ff.net.2.row_norms": [
3072
],
"transformer_blocks.8.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.8.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.8.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.8.ff_context.net.2.bias": [
3072
],
"transformer_blocks.8.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.8.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.9.norm1.linear.bias": [
18432
],
"transformer_blocks.9.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.9.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.9.norm1_context.linear.bias": [
18432
],
"transformer_blocks.9.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.9.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.9.attn.norm_q.weight": [
128
],
"transformer_blocks.9.attn.norm_k.weight": [
128
],
"transformer_blocks.9.attn.to_q.bias": [
3072
],
"transformer_blocks.9.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.to_q.row_norms": [
3072
],
"transformer_blocks.9.attn.to_k.bias": [
3072
],
"transformer_blocks.9.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.to_k.row_norms": [
3072
],
"transformer_blocks.9.attn.to_v.bias": [
3072
],
"transformer_blocks.9.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.to_v.row_norms": [
3072
],
"transformer_blocks.9.attn.to_out.0.bias": [
3072
],
"transformer_blocks.9.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.9.attn.norm_added_q.weight": [
128
],
"transformer_blocks.9.attn.norm_added_k.weight": [
128
],
"transformer_blocks.9.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.9.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.9.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.9.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.9.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.9.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.9.attn.to_add_out.bias": [
3072
],
"transformer_blocks.9.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.9.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.9.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.9.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.9.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.9.ff.net.2.bias": [
3072
],
"transformer_blocks.9.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.9.ff.net.2.row_norms": [
3072
],
"transformer_blocks.9.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.9.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.9.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.9.ff_context.net.2.bias": [
3072
],
"transformer_blocks.9.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.9.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.10.norm1.linear.bias": [
18432
],
"transformer_blocks.10.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.10.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.10.norm1_context.linear.bias": [
18432
],
"transformer_blocks.10.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.10.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.10.attn.norm_q.weight": [
128
],
"transformer_blocks.10.attn.norm_k.weight": [
128
],
"transformer_blocks.10.attn.to_q.bias": [
3072
],
"transformer_blocks.10.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.to_q.row_norms": [
3072
],
"transformer_blocks.10.attn.to_k.bias": [
3072
],
"transformer_blocks.10.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.to_k.row_norms": [
3072
],
"transformer_blocks.10.attn.to_v.bias": [
3072
],
"transformer_blocks.10.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.to_v.row_norms": [
3072
],
"transformer_blocks.10.attn.to_out.0.bias": [
3072
],
"transformer_blocks.10.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.10.attn.norm_added_q.weight": [
128
],
"transformer_blocks.10.attn.norm_added_k.weight": [
128
],
"transformer_blocks.10.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.10.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.10.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.10.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.10.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.10.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.10.attn.to_add_out.bias": [
3072
],
"transformer_blocks.10.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.10.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.10.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.10.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.10.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.10.ff.net.2.bias": [
3072
],
"transformer_blocks.10.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.10.ff.net.2.row_norms": [
3072
],
"transformer_blocks.10.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.10.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.10.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.10.ff_context.net.2.bias": [
3072
],
"transformer_blocks.10.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.10.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.11.norm1.linear.bias": [
18432
],
"transformer_blocks.11.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.11.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.11.norm1_context.linear.bias": [
18432
],
"transformer_blocks.11.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.11.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.11.attn.norm_q.weight": [
128
],
"transformer_blocks.11.attn.norm_k.weight": [
128
],
"transformer_blocks.11.attn.to_q.bias": [
3072
],
"transformer_blocks.11.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.to_q.row_norms": [
3072
],
"transformer_blocks.11.attn.to_k.bias": [
3072
],
"transformer_blocks.11.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.to_k.row_norms": [
3072
],
"transformer_blocks.11.attn.to_v.bias": [
3072
],
"transformer_blocks.11.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.to_v.row_norms": [
3072
],
"transformer_blocks.11.attn.to_out.0.bias": [
3072
],
"transformer_blocks.11.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.11.attn.norm_added_q.weight": [
128
],
"transformer_blocks.11.attn.norm_added_k.weight": [
128
],
"transformer_blocks.11.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.11.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.11.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.11.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.11.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.11.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.11.attn.to_add_out.bias": [
3072
],
"transformer_blocks.11.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.11.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.11.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.11.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.11.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.11.ff.net.2.bias": [
3072
],
"transformer_blocks.11.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.11.ff.net.2.row_norms": [
3072
],
"transformer_blocks.11.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.11.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.11.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.11.ff_context.net.2.bias": [
3072
],
"transformer_blocks.11.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.11.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.12.norm1.linear.bias": [
18432
],
"transformer_blocks.12.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.12.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.12.norm1_context.linear.bias": [
18432
],
"transformer_blocks.12.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.12.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.12.attn.norm_q.weight": [
128
],
"transformer_blocks.12.attn.norm_k.weight": [
128
],
"transformer_blocks.12.attn.to_q.bias": [
3072
],
"transformer_blocks.12.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.to_q.row_norms": [
3072
],
"transformer_blocks.12.attn.to_k.bias": [
3072
],
"transformer_blocks.12.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.to_k.row_norms": [
3072
],
"transformer_blocks.12.attn.to_v.bias": [
3072
],
"transformer_blocks.12.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.to_v.row_norms": [
3072
],
"transformer_blocks.12.attn.to_out.0.bias": [
3072
],
"transformer_blocks.12.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.12.attn.norm_added_q.weight": [
128
],
"transformer_blocks.12.attn.norm_added_k.weight": [
128
],
"transformer_blocks.12.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.12.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.12.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.12.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.12.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.12.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.12.attn.to_add_out.bias": [
3072
],
"transformer_blocks.12.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.12.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.12.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.12.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.12.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.12.ff.net.2.bias": [
3072
],
"transformer_blocks.12.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.12.ff.net.2.row_norms": [
3072
],
"transformer_blocks.12.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.12.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.12.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.12.ff_context.net.2.bias": [
3072
],
"transformer_blocks.12.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.12.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.13.norm1.linear.bias": [
18432
],
"transformer_blocks.13.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.13.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.13.norm1_context.linear.bias": [
18432
],
"transformer_blocks.13.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.13.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.13.attn.norm_q.weight": [
128
],
"transformer_blocks.13.attn.norm_k.weight": [
128
],
"transformer_blocks.13.attn.to_q.bias": [
3072
],
"transformer_blocks.13.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.to_q.row_norms": [
3072
],
"transformer_blocks.13.attn.to_k.bias": [
3072
],
"transformer_blocks.13.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.to_k.row_norms": [
3072
],
"transformer_blocks.13.attn.to_v.bias": [
3072
],
"transformer_blocks.13.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.to_v.row_norms": [
3072
],
"transformer_blocks.13.attn.to_out.0.bias": [
3072
],
"transformer_blocks.13.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.13.attn.norm_added_q.weight": [
128
],
"transformer_blocks.13.attn.norm_added_k.weight": [
128
],
"transformer_blocks.13.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.13.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.13.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.13.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.13.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.13.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.13.attn.to_add_out.bias": [
3072
],
"transformer_blocks.13.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.13.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.13.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.13.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.13.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.13.ff.net.2.bias": [
3072
],
"transformer_blocks.13.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.13.ff.net.2.row_norms": [
3072
],
"transformer_blocks.13.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.13.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.13.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.13.ff_context.net.2.bias": [
3072
],
"transformer_blocks.13.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.13.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.14.norm1.linear.bias": [
18432
],
"transformer_blocks.14.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.14.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.14.norm1_context.linear.bias": [
18432
],
"transformer_blocks.14.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.14.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.14.attn.norm_q.weight": [
128
],
"transformer_blocks.14.attn.norm_k.weight": [
128
],
"transformer_blocks.14.attn.to_q.bias": [
3072
],
"transformer_blocks.14.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.to_q.row_norms": [
3072
],
"transformer_blocks.14.attn.to_k.bias": [
3072
],
"transformer_blocks.14.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.to_k.row_norms": [
3072
],
"transformer_blocks.14.attn.to_v.bias": [
3072
],
"transformer_blocks.14.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.to_v.row_norms": [
3072
],
"transformer_blocks.14.attn.to_out.0.bias": [
3072
],
"transformer_blocks.14.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.14.attn.norm_added_q.weight": [
128
],
"transformer_blocks.14.attn.norm_added_k.weight": [
128
],
"transformer_blocks.14.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.14.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.14.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.14.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.14.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.14.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.14.attn.to_add_out.bias": [
3072
],
"transformer_blocks.14.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.14.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.14.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.14.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.14.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.14.ff.net.2.bias": [
3072
],
"transformer_blocks.14.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.14.ff.net.2.row_norms": [
3072
],
"transformer_blocks.14.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.14.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.14.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.14.ff_context.net.2.bias": [
3072
],
"transformer_blocks.14.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.14.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.15.norm1.linear.bias": [
18432
],
"transformer_blocks.15.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.15.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.15.norm1_context.linear.bias": [
18432
],
"transformer_blocks.15.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.15.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.15.attn.norm_q.weight": [
128
],
"transformer_blocks.15.attn.norm_k.weight": [
128
],
"transformer_blocks.15.attn.to_q.bias": [
3072
],
"transformer_blocks.15.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.to_q.row_norms": [
3072
],
"transformer_blocks.15.attn.to_k.bias": [
3072
],
"transformer_blocks.15.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.to_k.row_norms": [
3072
],
"transformer_blocks.15.attn.to_v.bias": [
3072
],
"transformer_blocks.15.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.to_v.row_norms": [
3072
],
"transformer_blocks.15.attn.to_out.0.bias": [
3072
],
"transformer_blocks.15.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.15.attn.norm_added_q.weight": [
128
],
"transformer_blocks.15.attn.norm_added_k.weight": [
128
],
"transformer_blocks.15.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.15.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.15.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.15.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.15.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.15.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.15.attn.to_add_out.bias": [
3072
],
"transformer_blocks.15.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.15.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.15.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.15.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.15.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.15.ff.net.2.bias": [
3072
],
"transformer_blocks.15.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.15.ff.net.2.row_norms": [
3072
],
"transformer_blocks.15.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.15.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.15.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.15.ff_context.net.2.bias": [
3072
],
"transformer_blocks.15.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.15.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.16.norm1.linear.bias": [
18432
],
"transformer_blocks.16.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.16.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.16.norm1_context.linear.bias": [
18432
],
"transformer_blocks.16.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.16.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.16.attn.norm_q.weight": [
128
],
"transformer_blocks.16.attn.norm_k.weight": [
128
],
"transformer_blocks.16.attn.to_q.bias": [
3072
],
"transformer_blocks.16.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.to_q.row_norms": [
3072
],
"transformer_blocks.16.attn.to_k.bias": [
3072
],
"transformer_blocks.16.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.to_k.row_norms": [
3072
],
"transformer_blocks.16.attn.to_v.bias": [
3072
],
"transformer_blocks.16.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.to_v.row_norms": [
3072
],
"transformer_blocks.16.attn.to_out.0.bias": [
3072
],
"transformer_blocks.16.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.16.attn.norm_added_q.weight": [
128
],
"transformer_blocks.16.attn.norm_added_k.weight": [
128
],
"transformer_blocks.16.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.16.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.16.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.16.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.16.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.16.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.16.attn.to_add_out.bias": [
3072
],
"transformer_blocks.16.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.16.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.16.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.16.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.16.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.16.ff.net.2.bias": [
3072
],
"transformer_blocks.16.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.16.ff.net.2.row_norms": [
3072
],
"transformer_blocks.16.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.16.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.16.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.16.ff_context.net.2.bias": [
3072
],
"transformer_blocks.16.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.16.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.17.norm1.linear.bias": [
18432
],
"transformer_blocks.17.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.17.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.17.norm1_context.linear.bias": [
18432
],
"transformer_blocks.17.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.17.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.17.attn.norm_q.weight": [
128
],
"transformer_blocks.17.attn.norm_k.weight": [
128
],
"transformer_blocks.17.attn.to_q.bias": [
3072
],
"transformer_blocks.17.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.to_q.row_norms": [
3072
],
"transformer_blocks.17.attn.to_k.bias": [
3072
],
"transformer_blocks.17.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.to_k.row_norms": [
3072
],
"transformer_blocks.17.attn.to_v.bias": [
3072
],
"transformer_blocks.17.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.to_v.row_norms": [
3072
],
"transformer_blocks.17.attn.to_out.0.bias": [
3072
],
"transformer_blocks.17.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.17.attn.norm_added_q.weight": [
128
],
"transformer_blocks.17.attn.norm_added_k.weight": [
128
],
"transformer_blocks.17.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.17.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.17.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.17.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.17.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.17.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.17.attn.to_add_out.bias": [
3072
],
"transformer_blocks.17.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.17.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.17.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.17.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.17.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.17.ff.net.2.bias": [
3072
],
"transformer_blocks.17.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.17.ff.net.2.row_norms": [
3072
],
"transformer_blocks.17.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.17.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.17.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.17.ff_context.net.2.bias": [
3072
],
"transformer_blocks.17.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.17.ff_context.net.2.row_norms": [
3072
],
"transformer_blocks.18.norm1.linear.bias": [
18432
],
"transformer_blocks.18.norm1.linear.packed_weight": [
28311552
],
"transformer_blocks.18.norm1.linear.scales": [
18432,
48
],
"transformer_blocks.18.norm1_context.linear.bias": [
18432
],
"transformer_blocks.18.norm1_context.linear.packed_weight": [
28311552
],
"transformer_blocks.18.norm1_context.linear.scales": [
18432,
48
],
"transformer_blocks.18.attn.norm_q.weight": [
128
],
"transformer_blocks.18.attn.norm_k.weight": [
128
],
"transformer_blocks.18.attn.to_q.bias": [
3072
],
"transformer_blocks.18.attn.to_q.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.to_q.row_norms": [
3072
],
"transformer_blocks.18.attn.to_k.bias": [
3072
],
"transformer_blocks.18.attn.to_k.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.to_k.row_norms": [
3072
],
"transformer_blocks.18.attn.to_v.bias": [
3072
],
"transformer_blocks.18.attn.to_v.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.to_v.row_norms": [
3072
],
"transformer_blocks.18.attn.to_out.0.bias": [
3072
],
"transformer_blocks.18.attn.to_out.0.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.to_out.0.row_norms": [
3072
],
"transformer_blocks.18.attn.norm_added_q.weight": [
128
],
"transformer_blocks.18.attn.norm_added_k.weight": [
128
],
"transformer_blocks.18.attn.add_q_proj.bias": [
3072
],
"transformer_blocks.18.attn.add_q_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.add_q_proj.row_norms": [
3072
],
"transformer_blocks.18.attn.add_k_proj.bias": [
3072
],
"transformer_blocks.18.attn.add_k_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.add_k_proj.row_norms": [
3072
],
"transformer_blocks.18.attn.add_v_proj.bias": [
3072
],
"transformer_blocks.18.attn.add_v_proj.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.add_v_proj.row_norms": [
3072
],
"transformer_blocks.18.attn.to_add_out.bias": [
3072
],
"transformer_blocks.18.attn.to_add_out.packed_weight_indices": [
3538944
],
"transformer_blocks.18.attn.to_add_out.row_norms": [
3072
],
"transformer_blocks.18.ff.net.0.proj.bias": [
12288
],
"transformer_blocks.18.ff.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.18.ff.net.0.proj.row_norms": [
12288
],
"transformer_blocks.18.ff.net.2.bias": [
3072
],
"transformer_blocks.18.ff.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.18.ff.net.2.row_norms": [
3072
],
"transformer_blocks.18.ff_context.net.0.proj.bias": [
12288
],
"transformer_blocks.18.ff_context.net.0.proj.packed_weight_indices": [
14155776
],
"transformer_blocks.18.ff_context.net.0.proj.row_norms": [
12288
],
"transformer_blocks.18.ff_context.net.2.bias": [
3072
],
"transformer_blocks.18.ff_context.net.2.packed_weight_indices": [
14155776
],
"transformer_blocks.18.ff_context.net.2.row_norms": [
3072
],
"single_transformer_blocks.0.norm.linear.bias": [
9216
],
"single_transformer_blocks.0.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.0.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.0.proj_mlp.bias": [
12288
],
"single_transformer_blocks.0.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.0.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.0.proj_out.bias": [
3072
],
"single_transformer_blocks.0.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.0.proj_out.row_norms": [
3072
],
"single_transformer_blocks.0.attn.norm_q.weight": [
128
],
"single_transformer_blocks.0.attn.norm_k.weight": [
128
],
"single_transformer_blocks.0.attn.to_q.bias": [
3072
],
"single_transformer_blocks.0.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.0.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.0.attn.to_k.bias": [
3072
],
"single_transformer_blocks.0.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.0.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.0.attn.to_v.bias": [
3072
],
"single_transformer_blocks.0.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.0.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.1.norm.linear.bias": [
9216
],
"single_transformer_blocks.1.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.1.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.1.proj_mlp.bias": [
12288
],
"single_transformer_blocks.1.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.1.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.1.proj_out.bias": [
3072
],
"single_transformer_blocks.1.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.1.proj_out.row_norms": [
3072
],
"single_transformer_blocks.1.attn.norm_q.weight": [
128
],
"single_transformer_blocks.1.attn.norm_k.weight": [
128
],
"single_transformer_blocks.1.attn.to_q.bias": [
3072
],
"single_transformer_blocks.1.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.1.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.1.attn.to_k.bias": [
3072
],
"single_transformer_blocks.1.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.1.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.1.attn.to_v.bias": [
3072
],
"single_transformer_blocks.1.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.1.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.2.norm.linear.bias": [
9216
],
"single_transformer_blocks.2.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.2.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.2.proj_mlp.bias": [
12288
],
"single_transformer_blocks.2.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.2.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.2.proj_out.bias": [
3072
],
"single_transformer_blocks.2.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.2.proj_out.row_norms": [
3072
],
"single_transformer_blocks.2.attn.norm_q.weight": [
128
],
"single_transformer_blocks.2.attn.norm_k.weight": [
128
],
"single_transformer_blocks.2.attn.to_q.bias": [
3072
],
"single_transformer_blocks.2.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.2.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.2.attn.to_k.bias": [
3072
],
"single_transformer_blocks.2.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.2.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.2.attn.to_v.bias": [
3072
],
"single_transformer_blocks.2.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.2.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.3.norm.linear.bias": [
9216
],
"single_transformer_blocks.3.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.3.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.3.proj_mlp.bias": [
12288
],
"single_transformer_blocks.3.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.3.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.3.proj_out.bias": [
3072
],
"single_transformer_blocks.3.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.3.proj_out.row_norms": [
3072
],
"single_transformer_blocks.3.attn.norm_q.weight": [
128
],
"single_transformer_blocks.3.attn.norm_k.weight": [
128
],
"single_transformer_blocks.3.attn.to_q.bias": [
3072
],
"single_transformer_blocks.3.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.3.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.3.attn.to_k.bias": [
3072
],
"single_transformer_blocks.3.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.3.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.3.attn.to_v.bias": [
3072
],
"single_transformer_blocks.3.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.3.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.4.norm.linear.bias": [
9216
],
"single_transformer_blocks.4.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.4.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.4.proj_mlp.bias": [
12288
],
"single_transformer_blocks.4.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.4.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.4.proj_out.bias": [
3072
],
"single_transformer_blocks.4.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.4.proj_out.row_norms": [
3072
],
"single_transformer_blocks.4.attn.norm_q.weight": [
128
],
"single_transformer_blocks.4.attn.norm_k.weight": [
128
],
"single_transformer_blocks.4.attn.to_q.bias": [
3072
],
"single_transformer_blocks.4.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.4.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.4.attn.to_k.bias": [
3072
],
"single_transformer_blocks.4.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.4.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.4.attn.to_v.bias": [
3072
],
"single_transformer_blocks.4.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.4.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.5.norm.linear.bias": [
9216
],
"single_transformer_blocks.5.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.5.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.5.proj_mlp.bias": [
12288
],
"single_transformer_blocks.5.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.5.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.5.proj_out.bias": [
3072
],
"single_transformer_blocks.5.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.5.proj_out.row_norms": [
3072
],
"single_transformer_blocks.5.attn.norm_q.weight": [
128
],
"single_transformer_blocks.5.attn.norm_k.weight": [
128
],
"single_transformer_blocks.5.attn.to_q.bias": [
3072
],
"single_transformer_blocks.5.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.5.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.5.attn.to_k.bias": [
3072
],
"single_transformer_blocks.5.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.5.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.5.attn.to_v.bias": [
3072
],
"single_transformer_blocks.5.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.5.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.6.norm.linear.bias": [
9216
],
"single_transformer_blocks.6.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.6.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.6.proj_mlp.bias": [
12288
],
"single_transformer_blocks.6.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.6.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.6.proj_out.bias": [
3072
],
"single_transformer_blocks.6.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.6.proj_out.row_norms": [
3072
],
"single_transformer_blocks.6.attn.norm_q.weight": [
128
],
"single_transformer_blocks.6.attn.norm_k.weight": [
128
],
"single_transformer_blocks.6.attn.to_q.bias": [
3072
],
"single_transformer_blocks.6.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.6.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.6.attn.to_k.bias": [
3072
],
"single_transformer_blocks.6.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.6.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.6.attn.to_v.bias": [
3072
],
"single_transformer_blocks.6.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.6.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.7.norm.linear.bias": [
9216
],
"single_transformer_blocks.7.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.7.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.7.proj_mlp.bias": [
12288
],
"single_transformer_blocks.7.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.7.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.7.proj_out.bias": [
3072
],
"single_transformer_blocks.7.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.7.proj_out.row_norms": [
3072
],
"single_transformer_blocks.7.attn.norm_q.weight": [
128
],
"single_transformer_blocks.7.attn.norm_k.weight": [
128
],
"single_transformer_blocks.7.attn.to_q.bias": [
3072
],
"single_transformer_blocks.7.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.7.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.7.attn.to_k.bias": [
3072
],
"single_transformer_blocks.7.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.7.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.7.attn.to_v.bias": [
3072
],
"single_transformer_blocks.7.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.7.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.8.norm.linear.bias": [
9216
],
"single_transformer_blocks.8.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.8.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.8.proj_mlp.bias": [
12288
],
"single_transformer_blocks.8.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.8.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.8.proj_out.bias": [
3072
],
"single_transformer_blocks.8.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.8.proj_out.row_norms": [
3072
],
"single_transformer_blocks.8.attn.norm_q.weight": [
128
],
"single_transformer_blocks.8.attn.norm_k.weight": [
128
],
"single_transformer_blocks.8.attn.to_q.bias": [
3072
],
"single_transformer_blocks.8.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.8.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.8.attn.to_k.bias": [
3072
],
"single_transformer_blocks.8.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.8.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.8.attn.to_v.bias": [
3072
],
"single_transformer_blocks.8.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.8.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.9.norm.linear.bias": [
9216
],
"single_transformer_blocks.9.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.9.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.9.proj_mlp.bias": [
12288
],
"single_transformer_blocks.9.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.9.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.9.proj_out.bias": [
3072
],
"single_transformer_blocks.9.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.9.proj_out.row_norms": [
3072
],
"single_transformer_blocks.9.attn.norm_q.weight": [
128
],
"single_transformer_blocks.9.attn.norm_k.weight": [
128
],
"single_transformer_blocks.9.attn.to_q.bias": [
3072
],
"single_transformer_blocks.9.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.9.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.9.attn.to_k.bias": [
3072
],
"single_transformer_blocks.9.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.9.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.9.attn.to_v.bias": [
3072
],
"single_transformer_blocks.9.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.9.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.10.norm.linear.bias": [
9216
],
"single_transformer_blocks.10.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.10.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.10.proj_mlp.bias": [
12288
],
"single_transformer_blocks.10.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.10.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.10.proj_out.bias": [
3072
],
"single_transformer_blocks.10.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.10.proj_out.row_norms": [
3072
],
"single_transformer_blocks.10.attn.norm_q.weight": [
128
],
"single_transformer_blocks.10.attn.norm_k.weight": [
128
],
"single_transformer_blocks.10.attn.to_q.bias": [
3072
],
"single_transformer_blocks.10.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.10.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.10.attn.to_k.bias": [
3072
],
"single_transformer_blocks.10.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.10.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.10.attn.to_v.bias": [
3072
],
"single_transformer_blocks.10.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.10.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.11.norm.linear.bias": [
9216
],
"single_transformer_blocks.11.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.11.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.11.proj_mlp.bias": [
12288
],
"single_transformer_blocks.11.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.11.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.11.proj_out.bias": [
3072
],
"single_transformer_blocks.11.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.11.proj_out.row_norms": [
3072
],
"single_transformer_blocks.11.attn.norm_q.weight": [
128
],
"single_transformer_blocks.11.attn.norm_k.weight": [
128
],
"single_transformer_blocks.11.attn.to_q.bias": [
3072
],
"single_transformer_blocks.11.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.11.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.11.attn.to_k.bias": [
3072
],
"single_transformer_blocks.11.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.11.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.11.attn.to_v.bias": [
3072
],
"single_transformer_blocks.11.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.11.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.12.norm.linear.bias": [
9216
],
"single_transformer_blocks.12.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.12.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.12.proj_mlp.bias": [
12288
],
"single_transformer_blocks.12.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.12.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.12.proj_out.bias": [
3072
],
"single_transformer_blocks.12.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.12.proj_out.row_norms": [
3072
],
"single_transformer_blocks.12.attn.norm_q.weight": [
128
],
"single_transformer_blocks.12.attn.norm_k.weight": [
128
],
"single_transformer_blocks.12.attn.to_q.bias": [
3072
],
"single_transformer_blocks.12.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.12.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.12.attn.to_k.bias": [
3072
],
"single_transformer_blocks.12.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.12.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.12.attn.to_v.bias": [
3072
],
"single_transformer_blocks.12.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.12.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.13.norm.linear.bias": [
9216
],
"single_transformer_blocks.13.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.13.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.13.proj_mlp.bias": [
12288
],
"single_transformer_blocks.13.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.13.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.13.proj_out.bias": [
3072
],
"single_transformer_blocks.13.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.13.proj_out.row_norms": [
3072
],
"single_transformer_blocks.13.attn.norm_q.weight": [
128
],
"single_transformer_blocks.13.attn.norm_k.weight": [
128
],
"single_transformer_blocks.13.attn.to_q.bias": [
3072
],
"single_transformer_blocks.13.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.13.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.13.attn.to_k.bias": [
3072
],
"single_transformer_blocks.13.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.13.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.13.attn.to_v.bias": [
3072
],
"single_transformer_blocks.13.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.13.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.14.norm.linear.bias": [
9216
],
"single_transformer_blocks.14.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.14.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.14.proj_mlp.bias": [
12288
],
"single_transformer_blocks.14.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.14.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.14.proj_out.bias": [
3072
],
"single_transformer_blocks.14.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.14.proj_out.row_norms": [
3072
],
"single_transformer_blocks.14.attn.norm_q.weight": [
128
],
"single_transformer_blocks.14.attn.norm_k.weight": [
128
],
"single_transformer_blocks.14.attn.to_q.bias": [
3072
],
"single_transformer_blocks.14.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.14.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.14.attn.to_k.bias": [
3072
],
"single_transformer_blocks.14.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.14.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.14.attn.to_v.bias": [
3072
],
"single_transformer_blocks.14.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.14.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.15.norm.linear.bias": [
9216
],
"single_transformer_blocks.15.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.15.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.15.proj_mlp.bias": [
12288
],
"single_transformer_blocks.15.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.15.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.15.proj_out.bias": [
3072
],
"single_transformer_blocks.15.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.15.proj_out.row_norms": [
3072
],
"single_transformer_blocks.15.attn.norm_q.weight": [
128
],
"single_transformer_blocks.15.attn.norm_k.weight": [
128
],
"single_transformer_blocks.15.attn.to_q.bias": [
3072
],
"single_transformer_blocks.15.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.15.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.15.attn.to_k.bias": [
3072
],
"single_transformer_blocks.15.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.15.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.15.attn.to_v.bias": [
3072
],
"single_transformer_blocks.15.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.15.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.16.norm.linear.bias": [
9216
],
"single_transformer_blocks.16.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.16.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.16.proj_mlp.bias": [
12288
],
"single_transformer_blocks.16.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.16.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.16.proj_out.bias": [
3072
],
"single_transformer_blocks.16.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.16.proj_out.row_norms": [
3072
],
"single_transformer_blocks.16.attn.norm_q.weight": [
128
],
"single_transformer_blocks.16.attn.norm_k.weight": [
128
],
"single_transformer_blocks.16.attn.to_q.bias": [
3072
],
"single_transformer_blocks.16.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.16.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.16.attn.to_k.bias": [
3072
],
"single_transformer_blocks.16.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.16.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.16.attn.to_v.bias": [
3072
],
"single_transformer_blocks.16.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.16.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.17.norm.linear.bias": [
9216
],
"single_transformer_blocks.17.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.17.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.17.proj_mlp.bias": [
12288
],
"single_transformer_blocks.17.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.17.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.17.proj_out.bias": [
3072
],
"single_transformer_blocks.17.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.17.proj_out.row_norms": [
3072
],
"single_transformer_blocks.17.attn.norm_q.weight": [
128
],
"single_transformer_blocks.17.attn.norm_k.weight": [
128
],
"single_transformer_blocks.17.attn.to_q.bias": [
3072
],
"single_transformer_blocks.17.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.17.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.17.attn.to_k.bias": [
3072
],
"single_transformer_blocks.17.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.17.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.17.attn.to_v.bias": [
3072
],
"single_transformer_blocks.17.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.17.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.18.norm.linear.bias": [
9216
],
"single_transformer_blocks.18.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.18.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.18.proj_mlp.bias": [
12288
],
"single_transformer_blocks.18.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.18.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.18.proj_out.bias": [
3072
],
"single_transformer_blocks.18.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.18.proj_out.row_norms": [
3072
],
"single_transformer_blocks.18.attn.norm_q.weight": [
128
],
"single_transformer_blocks.18.attn.norm_k.weight": [
128
],
"single_transformer_blocks.18.attn.to_q.bias": [
3072
],
"single_transformer_blocks.18.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.18.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.18.attn.to_k.bias": [
3072
],
"single_transformer_blocks.18.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.18.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.18.attn.to_v.bias": [
3072
],
"single_transformer_blocks.18.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.18.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.19.norm.linear.bias": [
9216
],
"single_transformer_blocks.19.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.19.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.19.proj_mlp.bias": [
12288
],
"single_transformer_blocks.19.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.19.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.19.proj_out.bias": [
3072
],
"single_transformer_blocks.19.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.19.proj_out.row_norms": [
3072
],
"single_transformer_blocks.19.attn.norm_q.weight": [
128
],
"single_transformer_blocks.19.attn.norm_k.weight": [
128
],
"single_transformer_blocks.19.attn.to_q.bias": [
3072
],
"single_transformer_blocks.19.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.19.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.19.attn.to_k.bias": [
3072
],
"single_transformer_blocks.19.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.19.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.19.attn.to_v.bias": [
3072
],
"single_transformer_blocks.19.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.19.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.20.norm.linear.bias": [
9216
],
"single_transformer_blocks.20.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.20.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.20.proj_mlp.bias": [
12288
],
"single_transformer_blocks.20.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.20.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.20.proj_out.bias": [
3072
],
"single_transformer_blocks.20.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.20.proj_out.row_norms": [
3072
],
"single_transformer_blocks.20.attn.norm_q.weight": [
128
],
"single_transformer_blocks.20.attn.norm_k.weight": [
128
],
"single_transformer_blocks.20.attn.to_q.bias": [
3072
],
"single_transformer_blocks.20.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.20.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.20.attn.to_k.bias": [
3072
],
"single_transformer_blocks.20.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.20.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.20.attn.to_v.bias": [
3072
],
"single_transformer_blocks.20.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.20.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.21.norm.linear.bias": [
9216
],
"single_transformer_blocks.21.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.21.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.21.proj_mlp.bias": [
12288
],
"single_transformer_blocks.21.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.21.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.21.proj_out.bias": [
3072
],
"single_transformer_blocks.21.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.21.proj_out.row_norms": [
3072
],
"single_transformer_blocks.21.attn.norm_q.weight": [
128
],
"single_transformer_blocks.21.attn.norm_k.weight": [
128
],
"single_transformer_blocks.21.attn.to_q.bias": [
3072
],
"single_transformer_blocks.21.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.21.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.21.attn.to_k.bias": [
3072
],
"single_transformer_blocks.21.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.21.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.21.attn.to_v.bias": [
3072
],
"single_transformer_blocks.21.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.21.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.22.norm.linear.bias": [
9216
],
"single_transformer_blocks.22.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.22.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.22.proj_mlp.bias": [
12288
],
"single_transformer_blocks.22.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.22.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.22.proj_out.bias": [
3072
],
"single_transformer_blocks.22.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.22.proj_out.row_norms": [
3072
],
"single_transformer_blocks.22.attn.norm_q.weight": [
128
],
"single_transformer_blocks.22.attn.norm_k.weight": [
128
],
"single_transformer_blocks.22.attn.to_q.bias": [
3072
],
"single_transformer_blocks.22.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.22.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.22.attn.to_k.bias": [
3072
],
"single_transformer_blocks.22.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.22.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.22.attn.to_v.bias": [
3072
],
"single_transformer_blocks.22.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.22.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.23.norm.linear.bias": [
9216
],
"single_transformer_blocks.23.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.23.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.23.proj_mlp.bias": [
12288
],
"single_transformer_blocks.23.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.23.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.23.proj_out.bias": [
3072
],
"single_transformer_blocks.23.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.23.proj_out.row_norms": [
3072
],
"single_transformer_blocks.23.attn.norm_q.weight": [
128
],
"single_transformer_blocks.23.attn.norm_k.weight": [
128
],
"single_transformer_blocks.23.attn.to_q.bias": [
3072
],
"single_transformer_blocks.23.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.23.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.23.attn.to_k.bias": [
3072
],
"single_transformer_blocks.23.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.23.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.23.attn.to_v.bias": [
3072
],
"single_transformer_blocks.23.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.23.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.24.norm.linear.bias": [
9216
],
"single_transformer_blocks.24.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.24.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.24.proj_mlp.bias": [
12288
],
"single_transformer_blocks.24.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.24.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.24.proj_out.bias": [
3072
],
"single_transformer_blocks.24.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.24.proj_out.row_norms": [
3072
],
"single_transformer_blocks.24.attn.norm_q.weight": [
128
],
"single_transformer_blocks.24.attn.norm_k.weight": [
128
],
"single_transformer_blocks.24.attn.to_q.bias": [
3072
],
"single_transformer_blocks.24.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.24.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.24.attn.to_k.bias": [
3072
],
"single_transformer_blocks.24.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.24.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.24.attn.to_v.bias": [
3072
],
"single_transformer_blocks.24.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.24.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.25.norm.linear.bias": [
9216
],
"single_transformer_blocks.25.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.25.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.25.proj_mlp.bias": [
12288
],
"single_transformer_blocks.25.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.25.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.25.proj_out.bias": [
3072
],
"single_transformer_blocks.25.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.25.proj_out.row_norms": [
3072
],
"single_transformer_blocks.25.attn.norm_q.weight": [
128
],
"single_transformer_blocks.25.attn.norm_k.weight": [
128
],
"single_transformer_blocks.25.attn.to_q.bias": [
3072
],
"single_transformer_blocks.25.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.25.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.25.attn.to_k.bias": [
3072
],
"single_transformer_blocks.25.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.25.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.25.attn.to_v.bias": [
3072
],
"single_transformer_blocks.25.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.25.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.26.norm.linear.bias": [
9216
],
"single_transformer_blocks.26.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.26.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.26.proj_mlp.bias": [
12288
],
"single_transformer_blocks.26.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.26.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.26.proj_out.bias": [
3072
],
"single_transformer_blocks.26.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.26.proj_out.row_norms": [
3072
],
"single_transformer_blocks.26.attn.norm_q.weight": [
128
],
"single_transformer_blocks.26.attn.norm_k.weight": [
128
],
"single_transformer_blocks.26.attn.to_q.bias": [
3072
],
"single_transformer_blocks.26.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.26.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.26.attn.to_k.bias": [
3072
],
"single_transformer_blocks.26.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.26.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.26.attn.to_v.bias": [
3072
],
"single_transformer_blocks.26.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.26.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.27.norm.linear.bias": [
9216
],
"single_transformer_blocks.27.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.27.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.27.proj_mlp.bias": [
12288
],
"single_transformer_blocks.27.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.27.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.27.proj_out.bias": [
3072
],
"single_transformer_blocks.27.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.27.proj_out.row_norms": [
3072
],
"single_transformer_blocks.27.attn.norm_q.weight": [
128
],
"single_transformer_blocks.27.attn.norm_k.weight": [
128
],
"single_transformer_blocks.27.attn.to_q.bias": [
3072
],
"single_transformer_blocks.27.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.27.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.27.attn.to_k.bias": [
3072
],
"single_transformer_blocks.27.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.27.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.27.attn.to_v.bias": [
3072
],
"single_transformer_blocks.27.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.27.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.28.norm.linear.bias": [
9216
],
"single_transformer_blocks.28.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.28.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.28.proj_mlp.bias": [
12288
],
"single_transformer_blocks.28.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.28.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.28.proj_out.bias": [
3072
],
"single_transformer_blocks.28.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.28.proj_out.row_norms": [
3072
],
"single_transformer_blocks.28.attn.norm_q.weight": [
128
],
"single_transformer_blocks.28.attn.norm_k.weight": [
128
],
"single_transformer_blocks.28.attn.to_q.bias": [
3072
],
"single_transformer_blocks.28.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.28.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.28.attn.to_k.bias": [
3072
],
"single_transformer_blocks.28.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.28.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.28.attn.to_v.bias": [
3072
],
"single_transformer_blocks.28.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.28.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.29.norm.linear.bias": [
9216
],
"single_transformer_blocks.29.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.29.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.29.proj_mlp.bias": [
12288
],
"single_transformer_blocks.29.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.29.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.29.proj_out.bias": [
3072
],
"single_transformer_blocks.29.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.29.proj_out.row_norms": [
3072
],
"single_transformer_blocks.29.attn.norm_q.weight": [
128
],
"single_transformer_blocks.29.attn.norm_k.weight": [
128
],
"single_transformer_blocks.29.attn.to_q.bias": [
3072
],
"single_transformer_blocks.29.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.29.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.29.attn.to_k.bias": [
3072
],
"single_transformer_blocks.29.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.29.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.29.attn.to_v.bias": [
3072
],
"single_transformer_blocks.29.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.29.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.30.norm.linear.bias": [
9216
],
"single_transformer_blocks.30.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.30.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.30.proj_mlp.bias": [
12288
],
"single_transformer_blocks.30.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.30.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.30.proj_out.bias": [
3072
],
"single_transformer_blocks.30.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.30.proj_out.row_norms": [
3072
],
"single_transformer_blocks.30.attn.norm_q.weight": [
128
],
"single_transformer_blocks.30.attn.norm_k.weight": [
128
],
"single_transformer_blocks.30.attn.to_q.bias": [
3072
],
"single_transformer_blocks.30.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.30.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.30.attn.to_k.bias": [
3072
],
"single_transformer_blocks.30.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.30.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.30.attn.to_v.bias": [
3072
],
"single_transformer_blocks.30.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.30.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.31.norm.linear.bias": [
9216
],
"single_transformer_blocks.31.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.31.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.31.proj_mlp.bias": [
12288
],
"single_transformer_blocks.31.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.31.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.31.proj_out.bias": [
3072
],
"single_transformer_blocks.31.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.31.proj_out.row_norms": [
3072
],
"single_transformer_blocks.31.attn.norm_q.weight": [
128
],
"single_transformer_blocks.31.attn.norm_k.weight": [
128
],
"single_transformer_blocks.31.attn.to_q.bias": [
3072
],
"single_transformer_blocks.31.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.31.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.31.attn.to_k.bias": [
3072
],
"single_transformer_blocks.31.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.31.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.31.attn.to_v.bias": [
3072
],
"single_transformer_blocks.31.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.31.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.32.norm.linear.bias": [
9216
],
"single_transformer_blocks.32.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.32.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.32.proj_mlp.bias": [
12288
],
"single_transformer_blocks.32.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.32.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.32.proj_out.bias": [
3072
],
"single_transformer_blocks.32.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.32.proj_out.row_norms": [
3072
],
"single_transformer_blocks.32.attn.norm_q.weight": [
128
],
"single_transformer_blocks.32.attn.norm_k.weight": [
128
],
"single_transformer_blocks.32.attn.to_q.bias": [
3072
],
"single_transformer_blocks.32.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.32.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.32.attn.to_k.bias": [
3072
],
"single_transformer_blocks.32.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.32.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.32.attn.to_v.bias": [
3072
],
"single_transformer_blocks.32.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.32.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.33.norm.linear.bias": [
9216
],
"single_transformer_blocks.33.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.33.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.33.proj_mlp.bias": [
12288
],
"single_transformer_blocks.33.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.33.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.33.proj_out.bias": [
3072
],
"single_transformer_blocks.33.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.33.proj_out.row_norms": [
3072
],
"single_transformer_blocks.33.attn.norm_q.weight": [
128
],
"single_transformer_blocks.33.attn.norm_k.weight": [
128
],
"single_transformer_blocks.33.attn.to_q.bias": [
3072
],
"single_transformer_blocks.33.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.33.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.33.attn.to_k.bias": [
3072
],
"single_transformer_blocks.33.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.33.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.33.attn.to_v.bias": [
3072
],
"single_transformer_blocks.33.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.33.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.34.norm.linear.bias": [
9216
],
"single_transformer_blocks.34.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.34.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.34.proj_mlp.bias": [
12288
],
"single_transformer_blocks.34.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.34.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.34.proj_out.bias": [
3072
],
"single_transformer_blocks.34.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.34.proj_out.row_norms": [
3072
],
"single_transformer_blocks.34.attn.norm_q.weight": [
128
],
"single_transformer_blocks.34.attn.norm_k.weight": [
128
],
"single_transformer_blocks.34.attn.to_q.bias": [
3072
],
"single_transformer_blocks.34.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.34.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.34.attn.to_k.bias": [
3072
],
"single_transformer_blocks.34.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.34.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.34.attn.to_v.bias": [
3072
],
"single_transformer_blocks.34.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.34.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.35.norm.linear.bias": [
9216
],
"single_transformer_blocks.35.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.35.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.35.proj_mlp.bias": [
12288
],
"single_transformer_blocks.35.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.35.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.35.proj_out.bias": [
3072
],
"single_transformer_blocks.35.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.35.proj_out.row_norms": [
3072
],
"single_transformer_blocks.35.attn.norm_q.weight": [
128
],
"single_transformer_blocks.35.attn.norm_k.weight": [
128
],
"single_transformer_blocks.35.attn.to_q.bias": [
3072
],
"single_transformer_blocks.35.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.35.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.35.attn.to_k.bias": [
3072
],
"single_transformer_blocks.35.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.35.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.35.attn.to_v.bias": [
3072
],
"single_transformer_blocks.35.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.35.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.36.norm.linear.bias": [
9216
],
"single_transformer_blocks.36.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.36.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.36.proj_mlp.bias": [
12288
],
"single_transformer_blocks.36.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.36.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.36.proj_out.bias": [
3072
],
"single_transformer_blocks.36.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.36.proj_out.row_norms": [
3072
],
"single_transformer_blocks.36.attn.norm_q.weight": [
128
],
"single_transformer_blocks.36.attn.norm_k.weight": [
128
],
"single_transformer_blocks.36.attn.to_q.bias": [
3072
],
"single_transformer_blocks.36.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.36.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.36.attn.to_k.bias": [
3072
],
"single_transformer_blocks.36.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.36.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.36.attn.to_v.bias": [
3072
],
"single_transformer_blocks.36.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.36.attn.to_v.row_norms": [
3072
],
"single_transformer_blocks.37.norm.linear.bias": [
9216
],
"single_transformer_blocks.37.norm.linear.packed_weight": [
14155776
],
"single_transformer_blocks.37.norm.linear.scales": [
9216,
48
],
"single_transformer_blocks.37.proj_mlp.bias": [
12288
],
"single_transformer_blocks.37.proj_mlp.packed_weight_indices": [
14155776
],
"single_transformer_blocks.37.proj_mlp.row_norms": [
12288
],
"single_transformer_blocks.37.proj_out.bias": [
3072
],
"single_transformer_blocks.37.proj_out.packed_weight_indices": [
17694720
],
"single_transformer_blocks.37.proj_out.row_norms": [
3072
],
"single_transformer_blocks.37.attn.norm_q.weight": [
128
],
"single_transformer_blocks.37.attn.norm_k.weight": [
128
],
"single_transformer_blocks.37.attn.to_q.bias": [
3072
],
"single_transformer_blocks.37.attn.to_q.packed_weight_indices": [
3538944
],
"single_transformer_blocks.37.attn.to_q.row_norms": [
3072
],
"single_transformer_blocks.37.attn.to_k.bias": [
3072
],
"single_transformer_blocks.37.attn.to_k.packed_weight_indices": [
3538944
],
"single_transformer_blocks.37.attn.to_k.row_norms": [
3072
],
"single_transformer_blocks.37.attn.to_v.bias": [
3072
],
"single_transformer_blocks.37.attn.to_v.packed_weight_indices": [
3538944
],
"single_transformer_blocks.37.attn.to_v.row_norms": [
3072
],
"norm_out.linear.weight": [
6144,
3072
],
"norm_out.linear.bias": [
6144
],
"proj_out.weight": [
64,
3072
],
"proj_out.bias": [
64
]
},
"checksums": {
"assets/image_generation_comparison_matrix.webp": "bbe9d01bdf1585c7adc80cc343713b4df251c917ccbaf7815cc2907034280e9b",
"benchmark/summary.json": "bdbd546bdf4dbe0f8019dceb13ac8fb5048bc0b72241fb4d80c2aab0b2c99892",
"model.safetensors": "89e634e5f808e2d62584a0f7ae56d0f2f9cf4499e5d1e72c838b652984663a79",
"model_index.json": "648e26f9c65c66a13209d4c3d66fa1e754e144532c0164cb8de07b0c17a902ce",
"orbitquant_codebooks.safetensors": "3971b0c8b7b93a038a5e0f97eebb9082280aee9f2647832d5cafc6bfb45b35a0",
"orbitquant_rotations.safetensors": "ac6045b8b9ea4f802bc0aa394caea3300e527af02bb7743517b8dcc7cbf5d31a",
"prompts.json": "881003847ef4d052e0820f2d7264a36041d576f0222c64fd551bed8d7078151b",
"quantization_config.json": "04923fb7f0988981756ed01a552fbdfcea1ade674cba1c66e11cf0043674fae8"
}
}