diff --git "a/model_analysis.json" "b/model_analysis.json" new file mode 100644--- /dev/null +++ "b/model_analysis.json" @@ -0,0 +1,4683 @@ +{ + "layer_types": { + "transformer": 391 + }, + "parameter_counts": { + "transformer.time_embed.time_mlp.0.weight": 262144, + "transformer.time_embed.time_mlp.0.bias": 1024, + "transformer.time_embed.time_mlp.2.weight": 1048576, + "transformer.time_embed.time_mlp.2.bias": 1024, + "transformer.text_embed.text_embed.weight": 254600, + "transformer.input_embed.proj.weight": 307200, + "transformer.input_embed.proj.bias": 1024, + "transformer.input_embed.conv_pos_embed.conv1d.0.weight": 2031616, + "transformer.input_embed.conv_pos_embed.conv1d.0.bias": 1024, + "transformer.input_embed.conv_pos_embed.conv1d.2.weight": 2031616, + "transformer.input_embed.conv_pos_embed.conv1d.2.bias": 1024, + "transformer.layers.0.1.g": 1024, + "transformer.layers.0.2.to_q.weight": 1048576, + "transformer.layers.0.2.to_q.bias": 1024, + "transformer.layers.0.2.to_k.weight": 1048576, + "transformer.layers.0.2.to_k.bias": 1024, + "transformer.layers.0.2.to_v.weight": 1048576, + "transformer.layers.0.2.to_v.bias": 1024, + "transformer.layers.0.2.to_out.0.weight": 1048576, + "transformer.layers.0.2.to_out.0.bias": 1024, + "transformer.layers.0.3.g": 1024, + "transformer.layers.0.4.ff.0.0.weight": 4194304, + "transformer.layers.0.4.ff.0.0.bias": 4096, + "transformer.layers.0.4.ff.2.weight": 4194304, + "transformer.layers.0.4.ff.2.bias": 1024, + "transformer.layers.1.1.g": 1024, + "transformer.layers.1.2.to_q.weight": 1048576, + "transformer.layers.1.2.to_q.bias": 1024, + "transformer.layers.1.2.to_k.weight": 1048576, + "transformer.layers.1.2.to_k.bias": 1024, + "transformer.layers.1.2.to_v.weight": 1048576, + "transformer.layers.1.2.to_v.bias": 1024, + "transformer.layers.1.2.to_out.0.weight": 1048576, + "transformer.layers.1.2.to_out.0.bias": 1024, + "transformer.layers.1.3.g": 1024, + "transformer.layers.1.4.ff.0.0.weight": 4194304, + "transformer.layers.1.4.ff.0.0.bias": 4096, + "transformer.layers.1.4.ff.2.weight": 4194304, + "transformer.layers.1.4.ff.2.bias": 1024, + "transformer.layers.2.1.g": 1024, + "transformer.layers.2.2.to_q.weight": 1048576, + "transformer.layers.2.2.to_q.bias": 1024, + "transformer.layers.2.2.to_k.weight": 1048576, + "transformer.layers.2.2.to_k.bias": 1024, + "transformer.layers.2.2.to_v.weight": 1048576, + "transformer.layers.2.2.to_v.bias": 1024, + "transformer.layers.2.2.to_out.0.weight": 1048576, + "transformer.layers.2.2.to_out.0.bias": 1024, + "transformer.layers.2.3.g": 1024, + "transformer.layers.2.4.ff.0.0.weight": 4194304, + "transformer.layers.2.4.ff.0.0.bias": 4096, + "transformer.layers.2.4.ff.2.weight": 4194304, + "transformer.layers.2.4.ff.2.bias": 1024, + "transformer.layers.3.1.g": 1024, + "transformer.layers.3.2.to_q.weight": 1048576, + "transformer.layers.3.2.to_q.bias": 1024, + "transformer.layers.3.2.to_k.weight": 1048576, + "transformer.layers.3.2.to_k.bias": 1024, + "transformer.layers.3.2.to_v.weight": 1048576, + "transformer.layers.3.2.to_v.bias": 1024, + "transformer.layers.3.2.to_out.0.weight": 1048576, + "transformer.layers.3.2.to_out.0.bias": 1024, + "transformer.layers.3.3.g": 1024, + "transformer.layers.3.4.ff.0.0.weight": 4194304, + "transformer.layers.3.4.ff.0.0.bias": 4096, + "transformer.layers.3.4.ff.2.weight": 4194304, + "transformer.layers.3.4.ff.2.bias": 1024, + "transformer.layers.4.1.g": 1024, + "transformer.layers.4.2.to_q.weight": 1048576, + "transformer.layers.4.2.to_q.bias": 1024, + "transformer.layers.4.2.to_k.weight": 1048576, + "transformer.layers.4.2.to_k.bias": 1024, + "transformer.layers.4.2.to_v.weight": 1048576, + "transformer.layers.4.2.to_v.bias": 1024, + "transformer.layers.4.2.to_out.0.weight": 1048576, + "transformer.layers.4.2.to_out.0.bias": 1024, + "transformer.layers.4.3.g": 1024, + "transformer.layers.4.4.ff.0.0.weight": 4194304, + "transformer.layers.4.4.ff.0.0.bias": 4096, + "transformer.layers.4.4.ff.2.weight": 4194304, + "transformer.layers.4.4.ff.2.bias": 1024, + "transformer.layers.5.1.g": 1024, + "transformer.layers.5.2.to_q.weight": 1048576, + "transformer.layers.5.2.to_q.bias": 1024, + "transformer.layers.5.2.to_k.weight": 1048576, + "transformer.layers.5.2.to_k.bias": 1024, + "transformer.layers.5.2.to_v.weight": 1048576, + "transformer.layers.5.2.to_v.bias": 1024, + "transformer.layers.5.2.to_out.0.weight": 1048576, + "transformer.layers.5.2.to_out.0.bias": 1024, + "transformer.layers.5.3.g": 1024, + "transformer.layers.5.4.ff.0.0.weight": 4194304, + "transformer.layers.5.4.ff.0.0.bias": 4096, + "transformer.layers.5.4.ff.2.weight": 4194304, + "transformer.layers.5.4.ff.2.bias": 1024, + "transformer.layers.6.1.g": 1024, + "transformer.layers.6.2.to_q.weight": 1048576, + "transformer.layers.6.2.to_q.bias": 1024, + "transformer.layers.6.2.to_k.weight": 1048576, + "transformer.layers.6.2.to_k.bias": 1024, + "transformer.layers.6.2.to_v.weight": 1048576, + "transformer.layers.6.2.to_v.bias": 1024, + "transformer.layers.6.2.to_out.0.weight": 1048576, + "transformer.layers.6.2.to_out.0.bias": 1024, + "transformer.layers.6.3.g": 1024, + "transformer.layers.6.4.ff.0.0.weight": 4194304, + "transformer.layers.6.4.ff.0.0.bias": 4096, + "transformer.layers.6.4.ff.2.weight": 4194304, + "transformer.layers.6.4.ff.2.bias": 1024, + "transformer.layers.7.1.g": 1024, + "transformer.layers.7.2.to_q.weight": 1048576, + "transformer.layers.7.2.to_q.bias": 1024, + "transformer.layers.7.2.to_k.weight": 1048576, + "transformer.layers.7.2.to_k.bias": 1024, + "transformer.layers.7.2.to_v.weight": 1048576, + "transformer.layers.7.2.to_v.bias": 1024, + "transformer.layers.7.2.to_out.0.weight": 1048576, + "transformer.layers.7.2.to_out.0.bias": 1024, + "transformer.layers.7.3.g": 1024, + "transformer.layers.7.4.ff.0.0.weight": 4194304, + "transformer.layers.7.4.ff.0.0.bias": 4096, + "transformer.layers.7.4.ff.2.weight": 4194304, + "transformer.layers.7.4.ff.2.bias": 1024, + "transformer.layers.8.1.g": 1024, + "transformer.layers.8.2.to_q.weight": 1048576, + "transformer.layers.8.2.to_q.bias": 1024, + "transformer.layers.8.2.to_k.weight": 1048576, + "transformer.layers.8.2.to_k.bias": 1024, + "transformer.layers.8.2.to_v.weight": 1048576, + "transformer.layers.8.2.to_v.bias": 1024, + "transformer.layers.8.2.to_out.0.weight": 1048576, + "transformer.layers.8.2.to_out.0.bias": 1024, + "transformer.layers.8.3.g": 1024, + "transformer.layers.8.4.ff.0.0.weight": 4194304, + "transformer.layers.8.4.ff.0.0.bias": 4096, + "transformer.layers.8.4.ff.2.weight": 4194304, + "transformer.layers.8.4.ff.2.bias": 1024, + "transformer.layers.9.1.g": 1024, + "transformer.layers.9.2.to_q.weight": 1048576, + "transformer.layers.9.2.to_q.bias": 1024, + "transformer.layers.9.2.to_k.weight": 1048576, + "transformer.layers.9.2.to_k.bias": 1024, + "transformer.layers.9.2.to_v.weight": 1048576, + "transformer.layers.9.2.to_v.bias": 1024, + "transformer.layers.9.2.to_out.0.weight": 1048576, + "transformer.layers.9.2.to_out.0.bias": 1024, + "transformer.layers.9.3.g": 1024, + "transformer.layers.9.4.ff.0.0.weight": 4194304, + "transformer.layers.9.4.ff.0.0.bias": 4096, + "transformer.layers.9.4.ff.2.weight": 4194304, + "transformer.layers.9.4.ff.2.bias": 1024, + "transformer.layers.10.1.g": 1024, + "transformer.layers.10.2.to_q.weight": 1048576, + "transformer.layers.10.2.to_q.bias": 1024, + "transformer.layers.10.2.to_k.weight": 1048576, + "transformer.layers.10.2.to_k.bias": 1024, + "transformer.layers.10.2.to_v.weight": 1048576, + "transformer.layers.10.2.to_v.bias": 1024, + "transformer.layers.10.2.to_out.0.weight": 1048576, + "transformer.layers.10.2.to_out.0.bias": 1024, + "transformer.layers.10.3.g": 1024, + "transformer.layers.10.4.ff.0.0.weight": 4194304, + "transformer.layers.10.4.ff.0.0.bias": 4096, + "transformer.layers.10.4.ff.2.weight": 4194304, + "transformer.layers.10.4.ff.2.bias": 1024, + "transformer.layers.11.1.g": 1024, + "transformer.layers.11.2.to_q.weight": 1048576, + "transformer.layers.11.2.to_q.bias": 1024, + "transformer.layers.11.2.to_k.weight": 1048576, + "transformer.layers.11.2.to_k.bias": 1024, + "transformer.layers.11.2.to_v.weight": 1048576, + "transformer.layers.11.2.to_v.bias": 1024, + "transformer.layers.11.2.to_out.0.weight": 1048576, + "transformer.layers.11.2.to_out.0.bias": 1024, + "transformer.layers.11.3.g": 1024, + "transformer.layers.11.4.ff.0.0.weight": 4194304, + "transformer.layers.11.4.ff.0.0.bias": 4096, + "transformer.layers.11.4.ff.2.weight": 4194304, + "transformer.layers.11.4.ff.2.bias": 1024, + "transformer.layers.12.1.g": 1024, + "transformer.layers.12.2.to_q.weight": 1048576, + "transformer.layers.12.2.to_q.bias": 1024, + "transformer.layers.12.2.to_k.weight": 1048576, + "transformer.layers.12.2.to_k.bias": 1024, + "transformer.layers.12.2.to_v.weight": 1048576, + "transformer.layers.12.2.to_v.bias": 1024, + "transformer.layers.12.2.to_out.0.weight": 1048576, + "transformer.layers.12.2.to_out.0.bias": 1024, + "transformer.layers.12.3.g": 1024, + "transformer.layers.12.4.ff.0.0.weight": 4194304, + "transformer.layers.12.4.ff.0.0.bias": 4096, + "transformer.layers.12.4.ff.2.weight": 4194304, + "transformer.layers.12.4.ff.2.bias": 1024, + "transformer.layers.13.0.weight": 2097152, + "transformer.layers.13.1.g": 1024, + "transformer.layers.13.2.to_q.weight": 1048576, + "transformer.layers.13.2.to_q.bias": 1024, + "transformer.layers.13.2.to_k.weight": 1048576, + "transformer.layers.13.2.to_k.bias": 1024, + "transformer.layers.13.2.to_v.weight": 1048576, + "transformer.layers.13.2.to_v.bias": 1024, + "transformer.layers.13.2.to_out.0.weight": 1048576, + "transformer.layers.13.2.to_out.0.bias": 1024, + "transformer.layers.13.3.g": 1024, + "transformer.layers.13.4.ff.0.0.weight": 4194304, + "transformer.layers.13.4.ff.0.0.bias": 4096, + "transformer.layers.13.4.ff.2.weight": 4194304, + "transformer.layers.13.4.ff.2.bias": 1024, + "transformer.layers.14.0.weight": 2097152, + "transformer.layers.14.1.g": 1024, + "transformer.layers.14.2.to_q.weight": 1048576, + "transformer.layers.14.2.to_q.bias": 1024, + "transformer.layers.14.2.to_k.weight": 1048576, + "transformer.layers.14.2.to_k.bias": 1024, + "transformer.layers.14.2.to_v.weight": 1048576, + "transformer.layers.14.2.to_v.bias": 1024, + "transformer.layers.14.2.to_out.0.weight": 1048576, + "transformer.layers.14.2.to_out.0.bias": 1024, + "transformer.layers.14.3.g": 1024, + "transformer.layers.14.4.ff.0.0.weight": 4194304, + "transformer.layers.14.4.ff.0.0.bias": 4096, + "transformer.layers.14.4.ff.2.weight": 4194304, + "transformer.layers.14.4.ff.2.bias": 1024, + "transformer.layers.15.0.weight": 2097152, + "transformer.layers.15.1.g": 1024, + "transformer.layers.15.2.to_q.weight": 1048576, + "transformer.layers.15.2.to_q.bias": 1024, + "transformer.layers.15.2.to_k.weight": 1048576, + "transformer.layers.15.2.to_k.bias": 1024, + "transformer.layers.15.2.to_v.weight": 1048576, + "transformer.layers.15.2.to_v.bias": 1024, + "transformer.layers.15.2.to_out.0.weight": 1048576, + "transformer.layers.15.2.to_out.0.bias": 1024, + "transformer.layers.15.3.g": 1024, + "transformer.layers.15.4.ff.0.0.weight": 4194304, + "transformer.layers.15.4.ff.0.0.bias": 4096, + "transformer.layers.15.4.ff.2.weight": 4194304, + "transformer.layers.15.4.ff.2.bias": 1024, + "transformer.layers.16.0.weight": 2097152, + "transformer.layers.16.1.g": 1024, + "transformer.layers.16.2.to_q.weight": 1048576, + "transformer.layers.16.2.to_q.bias": 1024, + "transformer.layers.16.2.to_k.weight": 1048576, + "transformer.layers.16.2.to_k.bias": 1024, + "transformer.layers.16.2.to_v.weight": 1048576, + "transformer.layers.16.2.to_v.bias": 1024, + "transformer.layers.16.2.to_out.0.weight": 1048576, + "transformer.layers.16.2.to_out.0.bias": 1024, + "transformer.layers.16.3.g": 1024, + "transformer.layers.16.4.ff.0.0.weight": 4194304, + "transformer.layers.16.4.ff.0.0.bias": 4096, + "transformer.layers.16.4.ff.2.weight": 4194304, + "transformer.layers.16.4.ff.2.bias": 1024, + "transformer.layers.17.0.weight": 2097152, + "transformer.layers.17.1.g": 1024, + "transformer.layers.17.2.to_q.weight": 1048576, + "transformer.layers.17.2.to_q.bias": 1024, + "transformer.layers.17.2.to_k.weight": 1048576, + "transformer.layers.17.2.to_k.bias": 1024, + "transformer.layers.17.2.to_v.weight": 1048576, + "transformer.layers.17.2.to_v.bias": 1024, + "transformer.layers.17.2.to_out.0.weight": 1048576, + "transformer.layers.17.2.to_out.0.bias": 1024, + "transformer.layers.17.3.g": 1024, + "transformer.layers.17.4.ff.0.0.weight": 4194304, + "transformer.layers.17.4.ff.0.0.bias": 4096, + "transformer.layers.17.4.ff.2.weight": 4194304, + "transformer.layers.17.4.ff.2.bias": 1024, + "transformer.layers.18.0.weight": 2097152, + "transformer.layers.18.1.g": 1024, + "transformer.layers.18.2.to_q.weight": 1048576, + "transformer.layers.18.2.to_q.bias": 1024, + "transformer.layers.18.2.to_k.weight": 1048576, + "transformer.layers.18.2.to_k.bias": 1024, + "transformer.layers.18.2.to_v.weight": 1048576, + "transformer.layers.18.2.to_v.bias": 1024, + "transformer.layers.18.2.to_out.0.weight": 1048576, + "transformer.layers.18.2.to_out.0.bias": 1024, + "transformer.layers.18.3.g": 1024, + "transformer.layers.18.4.ff.0.0.weight": 4194304, + "transformer.layers.18.4.ff.0.0.bias": 4096, + "transformer.layers.18.4.ff.2.weight": 4194304, + "transformer.layers.18.4.ff.2.bias": 1024, + "transformer.layers.19.0.weight": 2097152, + "transformer.layers.19.1.g": 1024, + "transformer.layers.19.2.to_q.weight": 1048576, + "transformer.layers.19.2.to_q.bias": 1024, + "transformer.layers.19.2.to_k.weight": 1048576, + "transformer.layers.19.2.to_k.bias": 1024, + "transformer.layers.19.2.to_v.weight": 1048576, + "transformer.layers.19.2.to_v.bias": 1024, + "transformer.layers.19.2.to_out.0.weight": 1048576, + "transformer.layers.19.2.to_out.0.bias": 1024, + "transformer.layers.19.3.g": 1024, + "transformer.layers.19.4.ff.0.0.weight": 4194304, + "transformer.layers.19.4.ff.0.0.bias": 4096, + "transformer.layers.19.4.ff.2.weight": 4194304, + "transformer.layers.19.4.ff.2.bias": 1024, + "transformer.layers.20.0.weight": 2097152, + "transformer.layers.20.1.g": 1024, + "transformer.layers.20.2.to_q.weight": 1048576, + "transformer.layers.20.2.to_q.bias": 1024, + "transformer.layers.20.2.to_k.weight": 1048576, + "transformer.layers.20.2.to_k.bias": 1024, + "transformer.layers.20.2.to_v.weight": 1048576, + "transformer.layers.20.2.to_v.bias": 1024, + "transformer.layers.20.2.to_out.0.weight": 1048576, + "transformer.layers.20.2.to_out.0.bias": 1024, + "transformer.layers.20.3.g": 1024, + "transformer.layers.20.4.ff.0.0.weight": 4194304, + "transformer.layers.20.4.ff.0.0.bias": 4096, + "transformer.layers.20.4.ff.2.weight": 4194304, + "transformer.layers.20.4.ff.2.bias": 1024, + "transformer.layers.21.0.weight": 2097152, + "transformer.layers.21.1.g": 1024, + "transformer.layers.21.2.to_q.weight": 1048576, + "transformer.layers.21.2.to_q.bias": 1024, + "transformer.layers.21.2.to_k.weight": 1048576, + "transformer.layers.21.2.to_k.bias": 1024, + "transformer.layers.21.2.to_v.weight": 1048576, + "transformer.layers.21.2.to_v.bias": 1024, + "transformer.layers.21.2.to_out.0.weight": 1048576, + "transformer.layers.21.2.to_out.0.bias": 1024, + "transformer.layers.21.3.g": 1024, + "transformer.layers.21.4.ff.0.0.weight": 4194304, + "transformer.layers.21.4.ff.0.0.bias": 4096, + "transformer.layers.21.4.ff.2.weight": 4194304, + "transformer.layers.21.4.ff.2.bias": 1024, + "transformer.layers.22.0.weight": 2097152, + "transformer.layers.22.1.g": 1024, + "transformer.layers.22.2.to_q.weight": 1048576, + "transformer.layers.22.2.to_q.bias": 1024, + "transformer.layers.22.2.to_k.weight": 1048576, + "transformer.layers.22.2.to_k.bias": 1024, + "transformer.layers.22.2.to_v.weight": 1048576, + "transformer.layers.22.2.to_v.bias": 1024, + "transformer.layers.22.2.to_out.0.weight": 1048576, + "transformer.layers.22.2.to_out.0.bias": 1024, + "transformer.layers.22.3.g": 1024, + "transformer.layers.22.4.ff.0.0.weight": 4194304, + "transformer.layers.22.4.ff.0.0.bias": 4096, + "transformer.layers.22.4.ff.2.weight": 4194304, + "transformer.layers.22.4.ff.2.bias": 1024, + "transformer.layers.23.0.weight": 2097152, + "transformer.layers.23.1.g": 1024, + "transformer.layers.23.2.to_q.weight": 1048576, + "transformer.layers.23.2.to_q.bias": 1024, + "transformer.layers.23.2.to_k.weight": 1048576, + "transformer.layers.23.2.to_k.bias": 1024, + "transformer.layers.23.2.to_v.weight": 1048576, + "transformer.layers.23.2.to_v.bias": 1024, + "transformer.layers.23.2.to_out.0.weight": 1048576, + "transformer.layers.23.2.to_out.0.bias": 1024, + "transformer.layers.23.3.g": 1024, + "transformer.layers.23.4.ff.0.0.weight": 4194304, + "transformer.layers.23.4.ff.0.0.bias": 4096, + "transformer.layers.23.4.ff.2.weight": 4194304, + "transformer.layers.23.4.ff.2.bias": 1024, + "transformer.layers.24.0.weight": 2097152, + "transformer.layers.24.1.g": 1024, + "transformer.layers.24.2.to_q.weight": 1048576, + "transformer.layers.24.2.to_q.bias": 1024, + "transformer.layers.24.2.to_k.weight": 1048576, + "transformer.layers.24.2.to_k.bias": 1024, + "transformer.layers.24.2.to_v.weight": 1048576, + "transformer.layers.24.2.to_v.bias": 1024, + "transformer.layers.24.2.to_out.0.weight": 1048576, + "transformer.layers.24.2.to_out.0.bias": 1024, + "transformer.layers.24.3.g": 1024, + "transformer.layers.24.4.ff.0.0.weight": 4194304, + "transformer.layers.24.4.ff.0.0.bias": 4096, + "transformer.layers.24.4.ff.2.weight": 4194304, + "transformer.layers.24.4.ff.2.bias": 1024, + "transformer.layers.25.0.weight": 2097152, + "transformer.layers.25.1.g": 1024, + "transformer.layers.25.2.to_q.weight": 1048576, + "transformer.layers.25.2.to_q.bias": 1024, + "transformer.layers.25.2.to_k.weight": 1048576, + "transformer.layers.25.2.to_k.bias": 1024, + "transformer.layers.25.2.to_v.weight": 1048576, + "transformer.layers.25.2.to_v.bias": 1024, + "transformer.layers.25.2.to_out.0.weight": 1048576, + "transformer.layers.25.2.to_out.0.bias": 1024, + "transformer.layers.25.3.g": 1024, + "transformer.layers.25.4.ff.0.0.weight": 4194304, + "transformer.layers.25.4.ff.0.0.bias": 4096, + "transformer.layers.25.4.ff.2.weight": 4194304, + "transformer.layers.25.4.ff.2.bias": 1024, + "transformer.norm_out.g": 1024, + "transformer.proj_out.weight": 102400, + "transformer.proj_out.bias": 100 + }, + "important_layers": [ + "transformer.time_embed.time_mlp.0.weight", + "transformer.time_embed.time_mlp.2.weight", + "transformer.text_embed.text_embed.weight", + "transformer.input_embed.proj.weight", + "transformer.input_embed.conv_pos_embed.conv1d.0.weight", + "transformer.input_embed.conv_pos_embed.conv1d.2.weight", + "transformer.layers.0.2.to_q.weight", + "transformer.layers.0.2.to_k.weight", + "transformer.layers.0.2.to_v.weight", + "transformer.layers.0.2.to_out.0.weight", + "transformer.layers.0.4.ff.0.0.weight", + "transformer.layers.0.4.ff.2.weight", + "transformer.layers.1.2.to_q.weight", + "transformer.layers.1.2.to_k.weight", + "transformer.layers.1.2.to_v.weight", + "transformer.layers.1.2.to_out.0.weight", + "transformer.layers.1.4.ff.0.0.weight", + "transformer.layers.1.4.ff.2.weight", + "transformer.layers.2.2.to_q.weight", + "transformer.layers.2.2.to_k.weight", + "transformer.layers.2.2.to_v.weight", + "transformer.layers.2.2.to_out.0.weight", + "transformer.layers.2.4.ff.0.0.weight", + "transformer.layers.2.4.ff.2.weight", + "transformer.layers.3.2.to_q.weight", + "transformer.layers.3.2.to_k.weight", + "transformer.layers.3.2.to_v.weight", + "transformer.layers.3.2.to_out.0.weight", + "transformer.layers.3.4.ff.0.0.weight", + "transformer.layers.3.4.ff.2.weight", + "transformer.layers.4.2.to_q.weight", + "transformer.layers.4.2.to_k.weight", + "transformer.layers.4.2.to_v.weight", + "transformer.layers.4.2.to_out.0.weight", + "transformer.layers.4.4.ff.0.0.weight", + "transformer.layers.4.4.ff.2.weight", + "transformer.layers.5.2.to_q.weight", + "transformer.layers.5.2.to_k.weight", + "transformer.layers.5.2.to_v.weight", + "transformer.layers.5.2.to_out.0.weight", + "transformer.layers.5.4.ff.0.0.weight", + "transformer.layers.5.4.ff.2.weight", + "transformer.layers.6.2.to_q.weight", + "transformer.layers.6.2.to_k.weight", + "transformer.layers.6.2.to_v.weight", + "transformer.layers.6.2.to_out.0.weight", + "transformer.layers.6.4.ff.0.0.weight", + "transformer.layers.6.4.ff.2.weight", + "transformer.layers.7.2.to_q.weight", + "transformer.layers.7.2.to_k.weight", + "transformer.layers.7.2.to_v.weight", + "transformer.layers.7.2.to_out.0.weight", + "transformer.layers.7.4.ff.0.0.weight", + "transformer.layers.7.4.ff.2.weight", + "transformer.layers.8.4.ff.0.0.weight", + "transformer.layers.8.4.ff.2.weight", + "transformer.layers.9.4.ff.0.0.weight", + "transformer.layers.9.4.ff.2.weight", + "transformer.layers.10.4.ff.0.0.weight", + "transformer.layers.10.4.ff.2.weight", + "transformer.layers.11.4.ff.0.0.weight", + "transformer.layers.11.4.ff.2.weight", + "transformer.layers.12.4.ff.0.0.weight", + "transformer.layers.12.4.ff.2.weight", + "transformer.layers.13.0.weight", + "transformer.layers.13.4.ff.0.0.weight", + "transformer.layers.13.4.ff.2.weight", + "transformer.layers.14.0.weight", + "transformer.layers.14.4.ff.0.0.weight", + "transformer.layers.14.4.ff.2.weight", + "transformer.layers.15.0.weight", + "transformer.layers.15.4.ff.0.0.weight", + "transformer.layers.15.4.ff.2.weight", + "transformer.layers.16.4.ff.0.0.weight", + "transformer.layers.16.4.ff.2.weight", + "transformer.layers.17.4.ff.0.0.weight", + "transformer.layers.17.4.ff.2.weight", + "transformer.layers.18.4.ff.0.0.weight", + "transformer.layers.18.4.ff.2.weight", + "transformer.layers.19.4.ff.0.0.weight", + "transformer.layers.19.4.ff.2.weight", + "transformer.layers.20.4.ff.0.0.weight", + "transformer.layers.20.4.ff.2.weight", + "transformer.layers.21.4.ff.0.0.weight", + "transformer.layers.21.4.ff.2.weight", + "transformer.layers.22.4.ff.0.0.weight", + "transformer.layers.22.4.ff.2.weight", + "transformer.layers.23.4.ff.0.0.weight", + "transformer.layers.23.4.ff.2.weight", + "transformer.layers.24.4.ff.0.0.weight", + "transformer.layers.24.4.ff.2.weight", + "transformer.layers.25.4.ff.0.0.weight", + "transformer.layers.25.4.ff.2.weight" + ], + "bottleneck_layers": [], + "recommendations": { + "focus_layers": [ + "transformer.time_embed.time_mlp.0.weight", + "transformer.time_embed.time_mlp.2.weight", + "transformer.text_embed.text_embed.weight", + "transformer.input_embed.proj.weight", + "transformer.input_embed.conv_pos_embed.conv1d.0.weight", + "transformer.input_embed.conv_pos_embed.conv1d.2.weight", + "transformer.layers.0.2.to_q.weight", + "transformer.layers.0.2.to_k.weight", + "transformer.layers.0.2.to_v.weight", + "transformer.layers.0.2.to_out.0.weight", + "transformer.layers.0.4.ff.0.0.weight", + "transformer.layers.0.4.ff.2.weight", + "transformer.layers.1.2.to_q.weight", + "transformer.layers.1.2.to_k.weight", + "transformer.layers.1.2.to_v.weight", + "transformer.layers.1.2.to_out.0.weight", + "transformer.layers.1.4.ff.0.0.weight", + "transformer.layers.1.4.ff.2.weight", + "transformer.layers.2.2.to_q.weight", + "transformer.layers.2.2.to_k.weight", + "transformer.layers.2.2.to_v.weight", + "transformer.layers.2.2.to_out.0.weight", + "transformer.layers.2.4.ff.0.0.weight", + "transformer.layers.2.4.ff.2.weight", + "transformer.layers.3.2.to_q.weight", + "transformer.layers.3.2.to_k.weight", + "transformer.layers.3.2.to_v.weight", + "transformer.layers.3.2.to_out.0.weight", + "transformer.layers.3.4.ff.0.0.weight", + "transformer.layers.3.4.ff.2.weight", + "transformer.layers.4.2.to_q.weight", + "transformer.layers.4.2.to_k.weight", + "transformer.layers.4.2.to_v.weight", + "transformer.layers.4.2.to_out.0.weight", + "transformer.layers.4.4.ff.0.0.weight", + "transformer.layers.4.4.ff.2.weight", + "transformer.layers.5.2.to_q.weight", + "transformer.layers.5.2.to_k.weight", + "transformer.layers.5.2.to_v.weight", + "transformer.layers.5.2.to_out.0.weight", + "transformer.layers.5.4.ff.0.0.weight", + "transformer.layers.5.4.ff.2.weight", + "transformer.layers.6.2.to_q.weight", + "transformer.layers.6.2.to_k.weight", + "transformer.layers.6.2.to_v.weight", + "transformer.layers.6.2.to_out.0.weight", + "transformer.layers.6.4.ff.0.0.weight", + "transformer.layers.6.4.ff.2.weight", + "transformer.layers.7.2.to_q.weight", + "transformer.layers.7.2.to_k.weight", + "transformer.layers.7.2.to_v.weight", + "transformer.layers.7.2.to_out.0.weight", + "transformer.layers.7.4.ff.0.0.weight", + "transformer.layers.7.4.ff.2.weight", + "transformer.layers.8.4.ff.0.0.weight", + "transformer.layers.8.4.ff.2.weight", + "transformer.layers.9.4.ff.0.0.weight", + "transformer.layers.9.4.ff.2.weight", + "transformer.layers.10.4.ff.0.0.weight", + "transformer.layers.10.4.ff.2.weight", + "transformer.layers.11.4.ff.0.0.weight", + "transformer.layers.11.4.ff.2.weight", + "transformer.layers.12.4.ff.0.0.weight", + "transformer.layers.12.4.ff.2.weight", + "transformer.layers.13.0.weight", + "transformer.layers.13.4.ff.0.0.weight", + "transformer.layers.13.4.ff.2.weight", + "transformer.layers.14.0.weight", + "transformer.layers.14.4.ff.0.0.weight", + "transformer.layers.14.4.ff.2.weight", + "transformer.layers.15.0.weight", + "transformer.layers.15.4.ff.0.0.weight", + "transformer.layers.15.4.ff.2.weight", + "transformer.layers.16.4.ff.0.0.weight", + "transformer.layers.16.4.ff.2.weight", + "transformer.layers.17.4.ff.0.0.weight", + "transformer.layers.17.4.ff.2.weight", + "transformer.layers.18.4.ff.0.0.weight", + "transformer.layers.18.4.ff.2.weight", + "transformer.layers.19.4.ff.0.0.weight", + "transformer.layers.19.4.ff.2.weight", + "transformer.layers.20.4.ff.0.0.weight", + "transformer.layers.20.4.ff.2.weight", + "transformer.layers.21.4.ff.0.0.weight", + "transformer.layers.21.4.ff.2.weight", + "transformer.layers.22.4.ff.0.0.weight", + "transformer.layers.22.4.ff.2.weight", + "transformer.layers.23.4.ff.0.0.weight", + "transformer.layers.23.4.ff.2.weight", + "transformer.layers.24.4.ff.0.0.weight", + "transformer.layers.24.4.ff.2.weight", + "transformer.layers.25.4.ff.0.0.weight", + "transformer.layers.25.4.ff.2.weight" + ] + }, + "total_parameters": 391, + "total_elements": 360755948, + "param_ranges": { + "transformer.time_embed.time_mlp.0.weight": { + "min": -0.43053385615348816, + "max": 0.2987181544303894, + "mean": -0.0025508857797831297, + "std": 0.04255492985248566, + "sparsity": 0.0, + "shape": [ + 1024, + 256 + ] + }, + "transformer.time_embed.time_mlp.0.bias": { + "min": -0.06311193853616714, + "max": 0.10768741369247437, + "mean": 0.0006200151983648539, + "std": 0.03410356491804123, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.time_embed.time_mlp.2.weight": { + "min": -0.41268399357795715, + "max": 0.8365581035614014, + "mean": -0.00020668540673796088, + "std": 0.02410803735256195, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.time_embed.time_mlp.2.bias": { + "min": -0.11543754488229752, + "max": 0.3218643069267273, + "mean": -0.0009378742543049157, + "std": 0.019571715965867043, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.text_embed.text_embed.weight": { + "min": -2.7987403869628906, + "max": 2.8775689601898193, + "mean": -0.0003620539791882038, + "std": 0.6153795123100281, + "sparsity": 0.0, + "shape": [ + 2546, + 100 + ] + }, + "transformer.input_embed.proj.weight": { + "min": -0.2798178493976593, + "max": 0.38195931911468506, + "mean": 0.0004235386732034385, + "std": 0.04274815320968628, + "sparsity": 0.0, + "shape": [ + 1024, + 300 + ] + }, + "transformer.input_embed.proj.bias": { + "min": -0.22243480384349823, + "max": 0.20970797538757324, + "mean": -0.004494894295930862, + "std": 0.04093479365110397, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.input_embed.conv_pos_embed.conv1d.0.weight": { + "min": -0.42797601222991943, + "max": 0.47545987367630005, + "mean": 3.68623682334146e-06, + "std": 0.024507373571395874, + "sparsity": 0.0, + "shape": [ + 1024, + 64, + 31 + ] + }, + "transformer.input_embed.conv_pos_embed.conv1d.0.bias": { + "min": -0.32538774609565735, + "max": 0.15757951140403748, + "mean": -0.046732865273952484, + "std": 0.05161404609680176, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.input_embed.conv_pos_embed.conv1d.2.weight": { + "min": -0.4106280207633972, + "max": 0.35474810004234314, + "mean": -0.000128601081087254, + "std": 0.02359883114695549, + "sparsity": 0.0, + "shape": [ + 1024, + 64, + 31 + ] + }, + "transformer.input_embed.conv_pos_embed.conv1d.2.bias": { + "min": -0.22982768714427948, + "max": 0.2626851797103882, + "mean": -0.029157839715480804, + "std": 0.04937523230910301, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.0.1.g": { + "min": 0.2546607255935669, + "max": 0.8210369348526001, + "mean": 0.5255380868911743, + "std": 0.08102277666330338, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.0.2.to_q.weight": { + "min": -0.29706501960754395, + "max": 0.26598596572875977, + "mean": -0.0004244564042892307, + "std": 0.03210071846842766, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.0.2.to_q.bias": { + "min": -0.09280094504356384, + "max": 0.12531320750713348, + "mean": 0.0006500966264866292, + "std": 0.025744492188096046, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.0.2.to_k.weight": { + "min": -0.2907920181751251, + "max": 0.2819848656654358, + "mean": -7.519756036344916e-05, + "std": 0.030932072550058365, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.0.2.to_k.bias": { + "min": -5.9063615798950195, + "max": 5.821039199829102, + "mean": -0.009349350817501545, + "std": 1.2963582277297974, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.0.2.to_v.weight": { + "min": -0.4252597391605377, + "max": 0.3442302644252777, + "mean": 9.807890455704182e-05, + "std": 0.029951980337500572, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.0.2.to_v.bias": { + "min": -0.02886926755309105, + "max": 0.027612265199422836, + "mean": -0.0003159886400680989, + "std": 0.012566552497446537, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.0.2.to_out.0.weight": { + "min": -0.4545641541481018, + "max": 0.4486750364303589, + "mean": 2.288275572936982e-05, + "std": 0.023853391408920288, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.0.2.to_out.0.bias": { + "min": -0.08885892480611801, + "max": 0.09123405814170837, + "mean": 0.002273206366226077, + "std": 0.019519906491041183, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.0.3.g": { + "min": 0.26680853962898254, + "max": 1.0574053525924683, + "mean": 0.5312761068344116, + "std": 0.10467371344566345, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.0.4.ff.0.0.weight": { + "min": -0.575035810470581, + "max": 0.6089199781417847, + "mean": -0.00043114880099892616, + "std": 0.03859530761837959, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.0.4.ff.0.0.bias": { + "min": -0.18226587772369385, + "max": 0.04570382833480835, + "mean": -0.029475372284650803, + "std": 0.04265210032463074, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.0.4.ff.2.weight": { + "min": -1.1675195693969727, + "max": 1.6349984407424927, + "mean": 0.00032014260068535805, + "std": 0.02769290842115879, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.0.4.ff.2.bias": { + "min": -0.16255705058574677, + "max": 0.20596350729465485, + "mean": -0.021122729405760765, + "std": 0.0279533751308918, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.1.1.g": { + "min": 0.2242354154586792, + "max": 0.8446622490882874, + "mean": 0.4876382350921631, + "std": 0.07536358386278152, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.1.2.to_q.weight": { + "min": -0.2560102641582489, + "max": 0.3063015341758728, + "mean": -8.342660294147208e-06, + "std": 0.03346908837556839, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.1.2.to_q.bias": { + "min": -0.09542153775691986, + "max": 0.11059843748807907, + "mean": 6.575271254405379e-05, + "std": 0.026967303827404976, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.1.2.to_k.weight": { + "min": -0.29774004220962524, + "max": 0.296736478805542, + "mean": 5.098901965538971e-05, + "std": 0.03253892436623573, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.1.2.to_k.bias": { + "min": -5.170334339141846, + "max": 5.090466022491455, + "mean": -0.014626836404204369, + "std": 1.1584166288375854, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.1.2.to_v.weight": { + "min": -0.34492507576942444, + "max": 0.3436436355113983, + "mean": 7.888609980000183e-05, + "std": 0.030058223754167557, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.1.2.to_v.bias": { + "min": -0.03619777783751488, + "max": 0.033210255205631256, + "mean": -0.00014313205610960722, + "std": 0.013021216727793217, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.1.2.to_out.0.weight": { + "min": -0.31573694944381714, + "max": 0.37568625807762146, + "mean": -2.092823342536576e-05, + "std": 0.024055050686001778, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.1.2.to_out.0.bias": { + "min": -0.10554195195436478, + "max": 0.12217912822961807, + "mean": -0.001965724630281329, + "std": 0.02885899320244789, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.1.3.g": { + "min": 0.31185638904571533, + "max": 1.1226844787597656, + "mean": 0.6664173007011414, + "std": 0.09809636324644089, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.1.4.ff.0.0.weight": { + "min": -0.8725683689117432, + "max": 0.6277270317077637, + "mean": 0.001675453968346119, + "std": 0.04743659123778343, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.1.4.ff.0.0.bias": { + "min": -0.27133694291114807, + "max": 0.034276124089956284, + "mean": -0.04661266878247261, + "std": 0.04062533751130104, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.1.4.ff.2.weight": { + "min": -0.9219098687171936, + "max": 0.9648231863975525, + "mean": 0.0010219970718026161, + "std": 0.04070163145661354, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.1.4.ff.2.bias": { + "min": -0.14473342895507812, + "max": 0.07504827529191971, + "mean": -0.009093794040381908, + "std": 0.025712795555591583, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.2.1.g": { + "min": 0.23969869315624237, + "max": 0.7134895920753479, + "mean": 0.4472740888595581, + "std": 0.05947508662939072, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.2.2.to_q.weight": { + "min": -0.273428350687027, + "max": 0.2982955574989319, + "mean": 8.738919859752059e-06, + "std": 0.03547058627009392, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.2.2.to_q.bias": { + "min": -0.11919642984867096, + "max": 0.11864279955625534, + "mean": 0.0007499873172491789, + "std": 0.027633123099803925, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.2.2.to_k.weight": { + "min": -0.2816983759403229, + "max": 0.2803042232990265, + "mean": -7.669557089684531e-05, + "std": 0.03509991616010666, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.2.2.to_k.bias": { + "min": -2.514967679977417, + "max": 2.5269885063171387, + "mean": 0.026808204129338264, + "std": 0.587342381477356, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.2.2.to_v.weight": { + "min": -0.22164961695671082, + "max": 0.27225953340530396, + "mean": 2.8316171665210277e-06, + "std": 0.03073127381503582, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.2.2.to_v.bias": { + "min": -0.033373475074768066, + "max": 0.031244782730937004, + "mean": 0.00011742905917344615, + "std": 0.012399322353303432, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.2.2.to_out.0.weight": { + "min": -0.23557811975479126, + "max": 0.23209546506404877, + "mean": 5.68200193811208e-05, + "std": 0.025696886703372, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.2.2.to_out.0.bias": { + "min": -0.13604120910167694, + "max": 0.1282019019126892, + "mean": -0.005500663537532091, + "std": 0.0399833545088768, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.2.3.g": { + "min": 0.3545033931732178, + "max": 1.174311876296997, + "mean": 0.7105965614318848, + "std": 0.10393685102462769, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.2.4.ff.0.0.weight": { + "min": -0.6180550456047058, + "max": 0.555590033531189, + "mean": 0.0011606597108766437, + "std": 0.046113595366477966, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.2.4.ff.0.0.bias": { + "min": -0.1892954707145691, + "max": 0.024854592978954315, + "mean": -0.034856364130973816, + "std": 0.028640495613217354, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.2.4.ff.2.weight": { + "min": -1.1327383518218994, + "max": 0.972023606300354, + "mean": 0.00035934254992753267, + "std": 0.04234174266457558, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.2.4.ff.2.bias": { + "min": -0.5991520285606384, + "max": 0.06305119395256042, + "mean": -0.004881403874605894, + "std": 0.02864677459001541, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.3.1.g": { + "min": 0.3754131495952606, + "max": 0.9439838528633118, + "mean": 0.5927106142044067, + "std": 0.0675281211733818, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.3.2.to_q.weight": { + "min": -0.3920239806175232, + "max": 0.36984747648239136, + "mean": 7.029663538560271e-05, + "std": 0.03718528896570206, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.3.2.to_q.bias": { + "min": -0.11922823637723923, + "max": 0.13680268824100494, + "mean": 0.0009289362351410091, + "std": 0.029231850057840347, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.3.2.to_k.weight": { + "min": -0.6204254031181335, + "max": 0.5099692940711975, + "mean": 1.5338478988269344e-05, + "std": 0.03643814101815224, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.3.2.to_k.bias": { + "min": -8.20434856414795, + "max": 8.80774211883545, + "mean": -0.10939832031726837, + "std": 1.7015736103057861, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.3.2.to_v.weight": { + "min": -0.2770227789878845, + "max": 0.2399866282939911, + "mean": 5.2325925935292616e-05, + "std": 0.032612308859825134, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.3.2.to_v.bias": { + "min": -0.05187734216451645, + "max": 0.039576977491378784, + "mean": 9.007145126815885e-05, + "std": 0.01296569500118494, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.3.2.to_out.0.weight": { + "min": -0.23111766576766968, + "max": 0.23510430753231049, + "mean": -2.2175441699801013e-05, + "std": 0.029389016330242157, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.3.2.to_out.0.bias": { + "min": -0.20455272495746613, + "max": 0.10541031509637833, + "mean": -0.0040219868533313274, + "std": 0.03264109417796135, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.3.3.g": { + "min": 0.3396739959716797, + "max": 1.0156350135803223, + "mean": 0.7007465362548828, + "std": 0.09685582667589188, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.3.4.ff.0.0.weight": { + "min": -0.5659961700439453, + "max": 0.8350182771682739, + "mean": 0.0004152171895839274, + "std": 0.042294152081012726, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.3.4.ff.0.0.bias": { + "min": -0.2122603803873062, + "max": 0.03037133999168873, + "mean": -0.03219597041606903, + "std": 0.026528161019086838, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.3.4.ff.2.weight": { + "min": -0.7565364837646484, + "max": 0.7206384539604187, + "mean": -1.6425212379544973e-05, + "std": 0.03683505579829216, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.3.4.ff.2.bias": { + "min": -0.2637326717376709, + "max": 0.10635162889957428, + "mean": -0.003013473004102707, + "std": 0.028875315561890602, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.4.1.g": { + "min": 0.2841384708881378, + "max": 0.6960581541061401, + "mean": 0.4994935393333435, + "std": 0.046687543392181396, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.4.2.to_q.weight": { + "min": -0.27911630272865295, + "max": 0.23450130224227905, + "mean": -0.00011085053847637028, + "std": 0.038756489753723145, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.4.2.to_q.bias": { + "min": -0.15422014892101288, + "max": 0.1267157793045044, + "mean": -0.0022325206082314253, + "std": 0.03337828069925308, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.4.2.to_k.weight": { + "min": -0.4152379035949707, + "max": 0.6604457497596741, + "mean": -1.880790659924969e-05, + "std": 0.0390951968729496, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.4.2.to_k.bias": { + "min": -4.242863178253174, + "max": 4.727988243103027, + "mean": -0.020436234772205353, + "std": 1.0083643198013306, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.4.2.to_v.weight": { + "min": -0.24566201865673065, + "max": 0.2078404426574707, + "mean": 4.393987182993442e-05, + "std": 0.03396216034889221, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.4.2.to_v.bias": { + "min": -0.03461671993136406, + "max": 0.04490647837519646, + "mean": -1.8480626749806106e-05, + "std": 0.012636142782866955, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.4.2.to_out.0.weight": { + "min": -0.201488196849823, + "max": 0.20679476857185364, + "mean": -2.9119719329173677e-05, + "std": 0.03102005459368229, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.4.2.to_out.0.bias": { + "min": -0.2001321166753769, + "max": 0.11347545683383942, + "mean": -0.0028973689768463373, + "std": 0.03452814370393753, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.4.3.g": { + "min": 0.3668859004974365, + "max": 1.0606576204299927, + "mean": 0.6705638766288757, + "std": 0.06651072949171066, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.4.4.ff.0.0.weight": { + "min": -0.3992147445678711, + "max": 0.5030191540718079, + "mean": -3.829112756648101e-05, + "std": 0.04113021492958069, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.4.4.ff.0.0.bias": { + "min": -0.12906894087791443, + "max": 0.02686660923063755, + "mean": -0.030545957386493683, + "std": 0.02190822921693325, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.4.4.ff.2.weight": { + "min": -0.45001378655433655, + "max": 0.43416494131088257, + "mean": 7.559473306173459e-05, + "std": 0.03489038348197937, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.4.4.ff.2.bias": { + "min": -0.26795703172683716, + "max": 0.07305809110403061, + "mean": -0.0010922406800091267, + "std": 0.023138197138905525, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.5.1.g": { + "min": 0.28738605976104736, + "max": 0.6873639225959778, + "mean": 0.5246094465255737, + "std": 0.04773576930165291, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.5.2.to_q.weight": { + "min": -0.2230275720357895, + "max": 0.22428689897060394, + "mean": 1.5606414308422245e-05, + "std": 0.03894846886396408, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.5.2.to_q.bias": { + "min": -0.1365431845188141, + "max": 0.1094546914100647, + "mean": 0.0002404236583970487, + "std": 0.02924003079533577, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.5.2.to_k.weight": { + "min": -0.375844806432724, + "max": 0.4382041096687317, + "mean": -9.796498488867655e-06, + "std": 0.039285723119974136, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.5.2.to_k.bias": { + "min": -3.8503658771514893, + "max": 5.0051727294921875, + "mean": 0.009742870926856995, + "std": 0.8458123803138733, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.5.2.to_v.weight": { + "min": -0.22342294454574585, + "max": 0.22070662677288055, + "mean": -2.869974196073599e-07, + "std": 0.03440912440419197, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.5.2.to_v.bias": { + "min": -0.04364994913339615, + "max": 0.03587768226861954, + "mean": -0.00025836972054094076, + "std": 0.012079192325472832, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.5.2.to_out.0.weight": { + "min": -0.21351207792758942, + "max": 0.18924757838249207, + "mean": -1.7089078028220683e-05, + "std": 0.03153553605079651, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.5.2.to_out.0.bias": { + "min": -0.18106123805046082, + "max": 0.12093079835176468, + "mean": -0.0023932361509650946, + "std": 0.04127350077033043, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.5.3.g": { + "min": 0.4227260649204254, + "max": 0.9448354244232178, + "mean": 0.662743330001831, + "std": 0.05696980655193329, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.5.4.ff.0.0.weight": { + "min": -0.37132784724235535, + "max": 0.4766311049461365, + "mean": -8.210691157728434e-05, + "std": 0.0408891923725605, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.5.4.ff.0.0.bias": { + "min": -0.20899881422519684, + "max": 0.027237456291913986, + "mean": -0.03024902194738388, + "std": 0.02138604037463665, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.5.4.ff.2.weight": { + "min": -0.3416992723941803, + "max": 0.735672652721405, + "mean": 8.195172995328903e-05, + "std": 0.03476583957672119, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.5.4.ff.2.bias": { + "min": -0.2404329776763916, + "max": 0.05046902596950531, + "mean": -0.001188310096040368, + "std": 0.020469345152378082, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.6.1.g": { + "min": 0.3061925768852234, + "max": 0.654449999332428, + "mean": 0.5251765251159668, + "std": 0.04624079912900925, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.6.2.to_q.weight": { + "min": -0.3049762547016144, + "max": 0.21794484555721283, + "mean": 7.015730807324871e-05, + "std": 0.03949474170804024, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.6.2.to_q.bias": { + "min": -0.14950910210609436, + "max": 0.13137659430503845, + "mean": 0.000338978337822482, + "std": 0.030483614653348923, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.6.2.to_k.weight": { + "min": -0.2578408122062683, + "max": 0.20263631641864777, + "mean": 3.113361162832007e-05, + "std": 0.03948460891842842, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.6.2.to_k.bias": { + "min": -2.339005708694458, + "max": 2.378676176071167, + "mean": -0.026260126382112503, + "std": 0.45006638765335083, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.6.2.to_v.weight": { + "min": -0.18907499313354492, + "max": 0.21106310188770294, + "mean": 3.715493221534416e-05, + "std": 0.03479326516389847, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.6.2.to_v.bias": { + "min": -0.031842123717069626, + "max": 0.03563522920012474, + "mean": -0.00019889514078386128, + "std": 0.012288383208215237, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.6.2.to_out.0.weight": { + "min": -0.18906620144844055, + "max": 0.17065204679965973, + "mean": -6.830548954894766e-05, + "std": 0.032169949263334274, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.6.2.to_out.0.bias": { + "min": -0.13967929780483246, + "max": 0.13765227794647217, + "mean": -0.0025106696411967278, + "std": 0.051296915858983994, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.6.3.g": { + "min": 0.4670410752296448, + "max": 0.9571460485458374, + "mean": 0.668942928314209, + "std": 0.052938032895326614, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.6.4.ff.0.0.weight": { + "min": -0.32453829050064087, + "max": 0.3093876242637634, + "mean": -9.305285857408307e-07, + "std": 0.04094512388110161, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.6.4.ff.0.0.bias": { + "min": -0.12521511316299438, + "max": 0.025563344359397888, + "mean": -0.030704183503985405, + "std": 0.01984286867082119, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.6.4.ff.2.weight": { + "min": -0.4407917857170105, + "max": 0.4464106857776642, + "mean": 9.500519081484526e-05, + "std": 0.03511863574385643, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.6.4.ff.2.bias": { + "min": -0.22505341470241547, + "max": 0.051904987543821335, + "mean": -0.0011818333296105266, + "std": 0.018484966829419136, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.7.1.g": { + "min": 0.33911338448524475, + "max": 0.7404670715332031, + "mean": 0.5587128400802612, + "std": 0.04148301109671593, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.7.2.to_q.weight": { + "min": -0.27348393201828003, + "max": 0.2790311872959137, + "mean": 2.0330318875494413e-05, + "std": 0.041056688874959946, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.7.2.to_q.bias": { + "min": -0.13706564903259277, + "max": 0.14011380076408386, + "mean": 0.0004902533255517483, + "std": 0.026642272248864174, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.7.2.to_k.weight": { + "min": -0.4915930926799774, + "max": 0.35670116543769836, + "mean": 8.893256745068356e-05, + "std": 0.04069444537162781, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.7.2.to_k.bias": { + "min": -2.300570249557495, + "max": 1.7478224039077759, + "mean": -0.021113090217113495, + "std": 0.5004414319992065, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.7.2.to_v.weight": { + "min": -0.21829766035079956, + "max": 0.19811730086803436, + "mean": -4.052483200212009e-05, + "std": 0.034232448786497116, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.7.2.to_v.bias": { + "min": -0.041401587426662445, + "max": 0.038982585072517395, + "mean": -0.00013965339167043567, + "std": 0.012888636440038681, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.7.2.to_out.0.weight": { + "min": -0.17814268171787262, + "max": 0.1835789680480957, + "mean": 4.7900641220621765e-05, + "std": 0.031555790454149246, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.7.2.to_out.0.bias": { + "min": -0.1802123337984085, + "max": 0.1839253157377243, + "mean": -0.0022146895062178373, + "std": 0.05485367402434349, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.7.3.g": { + "min": 0.47431865334510803, + "max": 1.0268715620040894, + "mean": 0.6453023552894592, + "std": 0.05052410438656807, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.7.4.ff.0.0.weight": { + "min": -0.27204346656799316, + "max": 0.30987846851348877, + "mean": 0.00011226898641325533, + "std": 0.04068146273493767, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.7.4.ff.0.0.bias": { + "min": -0.10556552559137344, + "max": 0.02664870023727417, + "mean": -0.02952779270708561, + "std": 0.017950357869267464, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.7.4.ff.2.weight": { + "min": -0.3395896553993225, + "max": 0.3302164077758789, + "mean": 5.2438736020121723e-05, + "std": 0.03441261127591133, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.7.4.ff.2.bias": { + "min": -0.18205386400222778, + "max": 0.04234303906559944, + "mean": -0.0010605738498270512, + "std": 0.01722128316760063, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.8.1.g": { + "min": 0.3254714906215668, + "max": 0.6875306367874146, + "mean": 0.5112907886505127, + "std": 0.03710601106286049, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.8.2.to_q.weight": { + "min": -0.23404580354690552, + "max": 0.22564062476158142, + "mean": -3.628679769462906e-05, + "std": 0.03917597234249115, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.8.2.to_q.bias": { + "min": -0.1157650500535965, + "max": 0.13217955827713013, + "mean": 0.00015458319103345275, + "std": 0.02921123616397381, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.8.2.to_k.weight": { + "min": -0.3531610369682312, + "max": 0.28566646575927734, + "mean": 7.01215958542889e-06, + "std": 0.03924458101391792, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.8.2.to_k.bias": { + "min": -4.1371684074401855, + "max": 3.5479142665863037, + "mean": -0.011608399450778961, + "std": 0.6831862926483154, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.8.2.to_v.weight": { + "min": -0.2113451212644577, + "max": 0.20978450775146484, + "mean": 3.466297494014725e-05, + "std": 0.03448467701673508, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.8.2.to_v.bias": { + "min": -0.0358961820602417, + "max": 0.04827914386987686, + "mean": 0.000792390201240778, + "std": 0.012867480516433716, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.8.2.to_out.0.weight": { + "min": -0.21074581146240234, + "max": 0.19335627555847168, + "mean": -1.3081223642075201e-06, + "std": 0.031695783138275146, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.8.2.to_out.0.bias": { + "min": -0.18677425384521484, + "max": 0.17732204496860504, + "mean": -0.002835639752447605, + "std": 0.05864328145980835, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.8.3.g": { + "min": 0.47452113032341003, + "max": 1.0454236268997192, + "mean": 0.651544451713562, + "std": 0.05015648156404495, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.8.4.ff.0.0.weight": { + "min": -0.2484469711780548, + "max": 0.3293426036834717, + "mean": 0.00018075718253385276, + "std": 0.04056986793875694, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.8.4.ff.0.0.bias": { + "min": -0.12511543929576874, + "max": 0.024807237088680267, + "mean": -0.03050871379673481, + "std": 0.017624877393245697, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.8.4.ff.2.weight": { + "min": -0.42189329862594604, + "max": 0.4829460680484772, + "mean": -1.433467332390137e-06, + "std": 0.0353967621922493, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.8.4.ff.2.bias": { + "min": -0.1517147570848465, + "max": 0.043470486998558044, + "mean": 4.9440553993918e-05, + "std": 0.014891887083649635, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.9.1.g": { + "min": 0.31546592712402344, + "max": 0.6829473972320557, + "mean": 0.552940845489502, + "std": 0.0407881923019886, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.9.2.to_q.weight": { + "min": -0.20687490701675415, + "max": 0.22027458250522614, + "mean": 3.187588299624622e-05, + "std": 0.03829912096261978, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.9.2.to_q.bias": { + "min": -0.1380588412284851, + "max": 0.11287239193916321, + "mean": 2.8096917958464473e-05, + "std": 0.025843404233455658, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.9.2.to_k.weight": { + "min": -0.40360599756240845, + "max": 0.37176549434661865, + "mean": 2.5846293283393607e-05, + "std": 0.03817979246377945, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.9.2.to_k.bias": { + "min": -3.7753050327301025, + "max": 2.8720550537109375, + "mean": 0.001174271572381258, + "std": 0.5172262787818909, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.9.2.to_v.weight": { + "min": -0.203634575009346, + "max": 0.19783173501491547, + "mean": 2.9641731089213863e-05, + "std": 0.034296903759241104, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.9.2.to_v.bias": { + "min": -0.050782062113285065, + "max": 0.039943333715200424, + "mean": -0.00042034429498016834, + "std": 0.01341927982866764, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.9.2.to_out.0.weight": { + "min": -0.1968999058008194, + "max": 0.20258377492427826, + "mean": -1.2486772902775556e-05, + "std": 0.03180483356118202, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.9.2.to_out.0.bias": { + "min": -0.19323131442070007, + "max": 0.19526611268520355, + "mean": -0.002963971346616745, + "std": 0.06255338340997696, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.9.3.g": { + "min": 0.34893983602523804, + "max": 1.0871814489364624, + "mean": 0.6672742962837219, + "std": 0.05565904080867767, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.9.4.ff.0.0.weight": { + "min": -0.22620373964309692, + "max": 0.251870721578598, + "mean": 0.00035865549580194056, + "std": 0.040759552270174026, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.9.4.ff.0.0.bias": { + "min": -0.09141312539577484, + "max": 0.043738093227148056, + "mean": -0.03009146638214588, + "std": 0.017630403861403465, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.9.4.ff.2.weight": { + "min": -0.3538018465042114, + "max": 0.30474764108657837, + "mean": -4.393163908389397e-05, + "std": 0.03712210804224014, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.9.4.ff.2.bias": { + "min": -0.16189776360988617, + "max": 0.06336814165115356, + "mean": -8.093340147752315e-05, + "std": 0.019419532269239426, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.10.1.g": { + "min": 0.34863826632499695, + "max": 0.7244340777397156, + "mean": 0.5424437522888184, + "std": 0.039265505969524384, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.10.2.to_q.weight": { + "min": -0.21948550641536713, + "max": 0.22342580556869507, + "mean": -1.1189426913915668e-05, + "std": 0.039230361580848694, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.10.2.to_q.bias": { + "min": -0.11868271231651306, + "max": 0.17081572115421295, + "mean": 0.00028613023459911346, + "std": 0.025137728080153465, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.10.2.to_k.weight": { + "min": -0.24677123129367828, + "max": 0.30096495151519775, + "mean": -3.686630952870473e-05, + "std": 0.03892983868718147, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.10.2.to_k.bias": { + "min": -3.5091044902801514, + "max": 3.718792676925659, + "mean": 0.01584971882402897, + "std": 0.7831407189369202, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.10.2.to_v.weight": { + "min": -0.21897639334201813, + "max": 0.23756206035614014, + "mean": -1.3331029549590312e-05, + "std": 0.036302708089351654, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.10.2.to_v.bias": { + "min": -0.047262854874134064, + "max": 0.05141079053282738, + "mean": 0.00047719862777739763, + "std": 0.013516917824745178, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.10.2.to_out.0.weight": { + "min": -0.2142534703016281, + "max": 0.21756578981876373, + "mean": 5.647125362884253e-05, + "std": 0.03361497074365616, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.10.2.to_out.0.bias": { + "min": -0.21157103776931763, + "max": 0.23160234093666077, + "mean": -0.005100839305669069, + "std": 0.06188952922821045, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.10.3.g": { + "min": 0.3621518015861511, + "max": 1.1046018600463867, + "mean": 0.6994094252586365, + "std": 0.0540287047624588, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.10.4.ff.0.0.weight": { + "min": -0.23531383275985718, + "max": 0.24546286463737488, + "mean": 0.0004634457582142204, + "std": 0.0412684828042984, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.10.4.ff.0.0.bias": { + "min": -0.0981631875038147, + "max": 0.06831478327512741, + "mean": -0.031439878046512604, + "std": 0.01814098283648491, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.10.4.ff.2.weight": { + "min": -0.30264386534690857, + "max": 0.3523462414741516, + "mean": -8.214355329982936e-05, + "std": 0.0402742475271225, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.10.4.ff.2.bias": { + "min": -0.15258914232254028, + "max": 0.14998860657215118, + "mean": 0.0002567686606198549, + "std": 0.023048948496580124, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.11.1.g": { + "min": 0.9985297322273254, + "max": 1.007304310798645, + "mean": 0.9998952150344849, + "std": 0.0011818050406873226, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.11.2.to_q.weight": { + "min": -0.031265806406736374, + "max": 0.0312703475356102, + "mean": -1.928816709551029e-05, + "std": 0.018041111528873444, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.11.2.to_q.bias": { + "min": -0.031225843355059624, + "max": 0.030984507873654366, + "mean": -0.001084179850295186, + "std": 0.01795078068971634, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.11.2.to_k.weight": { + "min": -0.031264401972293854, + "max": 0.03126936033368111, + "mean": 3.5438486065686448e-06, + "std": 0.018041551113128662, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.11.2.to_k.bias": { + "min": -0.031160537153482437, + "max": 0.031171930953860283, + "mean": 0.00033398409141227603, + "std": 0.01806296594440937, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.11.2.to_v.weight": { + "min": 0.0, + "max": 0.0, + "mean": 0.0, + "std": 0.0, + "sparsity": 1.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.11.2.to_v.bias": { + "min": 0.0, + "max": 0.0, + "mean": 0.0, + "std": 0.0, + "sparsity": 1.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.11.2.to_out.0.weight": { + "min": 0.0, + "max": 0.0, + "mean": 0.0, + "std": 0.0, + "sparsity": 1.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.11.2.to_out.0.bias": { + "min": -0.0006154034635983407, + "max": 0.00041452725417912006, + "mean": 1.3732544630329357e-06, + "std": 0.00013773542013950646, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.11.3.g": { + "min": 0.9981350898742676, + "max": 1.0061345100402832, + "mean": 1.0003111362457275, + "std": 0.0018558463780209422, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.11.4.ff.0.0.weight": { + "min": -0.032749444246292114, + "max": 0.03284144029021263, + "mean": -6.684205800411291e-06, + "std": 0.01804272271692753, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.11.4.ff.0.0.bias": { + "min": -0.03275851905345917, + "max": 0.03259003907442093, + "mean": -0.00013117710477672517, + "std": 0.017956379801034927, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.11.4.ff.2.weight": { + "min": -0.0011779898777604103, + "max": 0.001155506120994687, + "mean": 3.63817605375516e-07, + "std": 0.00021426456805784255, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.11.4.ff.2.bias": { + "min": -0.0005257476586848497, + "max": 0.0003992951533291489, + "mean": 2.2647066089120926e-06, + "std": 0.00012679416977334768, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.12.1.g": { + "min": 0.3831771910190582, + "max": 0.7203002572059631, + "mean": 0.5807632207870483, + "std": 0.039030127227306366, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.12.2.to_q.weight": { + "min": -0.238657608628273, + "max": 0.1965981125831604, + "mean": 2.6105446522706188e-05, + "std": 0.03746547922492027, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.12.2.to_q.bias": { + "min": -0.11904074251651764, + "max": 0.16665399074554443, + "mean": 0.0009819172555580735, + "std": 0.027577750384807587, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.12.2.to_k.weight": { + "min": -0.2464642971754074, + "max": 0.5006471276283264, + "mean": -5.0186910812044516e-05, + "std": 0.03762289881706238, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.12.2.to_k.bias": { + "min": -3.946474552154541, + "max": 3.7734150886535645, + "mean": -0.0035824859514832497, + "std": 0.681806743144989, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.12.2.to_v.weight": { + "min": -0.22754359245300293, + "max": 0.25217491388320923, + "mean": -1.1530558367667254e-05, + "std": 0.03743445873260498, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.12.2.to_v.bias": { + "min": -0.07182253897190094, + "max": 0.0808083638548851, + "mean": -0.000513089878950268, + "std": 0.015668950974941254, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.12.2.to_out.0.weight": { + "min": -0.22810040414333344, + "max": 0.2579977512359619, + "mean": -2.8758044209098443e-05, + "std": 0.03542134538292885, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.12.2.to_out.0.bias": { + "min": -0.20080700516700745, + "max": 0.2153109759092331, + "mean": -0.005534037947654724, + "std": 0.0683637484908104, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.12.3.g": { + "min": 0.4053238332271576, + "max": 1.1908336877822876, + "mean": 0.7380030155181885, + "std": 0.05547412484884262, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.12.4.ff.0.0.weight": { + "min": -0.2215573787689209, + "max": 0.24592049419879913, + "mean": 0.000521159905474633, + "std": 0.041335850954055786, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.12.4.ff.0.0.bias": { + "min": -0.10345371812582016, + "max": 0.024234607815742493, + "mean": -0.032675523310899734, + "std": 0.018910475075244904, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.12.4.ff.2.weight": { + "min": -0.4504980742931366, + "max": 0.42334607243537903, + "mean": -0.0004341673047747463, + "std": 0.04689619690179825, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.12.4.ff.2.bias": { + "min": -0.2517021596431732, + "max": 0.4706237316131592, + "mean": 0.0032027317211031914, + "std": 0.04455312713980675, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.13.0.weight": { + "min": -0.3171294331550598, + "max": 0.33335307240486145, + "mean": -2.5211516913259402e-05, + "std": 0.021287426352500916, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.13.1.g": { + "min": 0.3245099186897278, + "max": 0.6862163543701172, + "mean": 0.5710394978523254, + "std": 0.04481911659240723, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.13.2.to_q.weight": { + "min": -0.1647811233997345, + "max": 0.1747460514307022, + "mean": -4.884982990915887e-05, + "std": 0.03318081423640251, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.13.2.to_q.bias": { + "min": -0.18714644014835358, + "max": 0.1431918442249298, + "mean": 4.32572269346565e-05, + "std": 0.029710030183196068, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.13.2.to_k.weight": { + "min": -0.3816376030445099, + "max": 0.24683159589767456, + "mean": -9.986059922084678e-06, + "std": 0.0327618382871151, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.13.2.to_k.bias": { + "min": -3.6597650051116943, + "max": 3.293627977371216, + "mean": -0.014285150915384293, + "std": 0.9855467677116394, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.13.2.to_v.weight": { + "min": -0.23503181338310242, + "max": 0.24772128462791443, + "mean": -1.80145725607872e-05, + "std": 0.04169723764061928, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.13.2.to_v.bias": { + "min": -0.07281922549009323, + "max": 0.1548185795545578, + "mean": 0.0006660926737822592, + "std": 0.025179805234074593, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.13.2.to_out.0.weight": { + "min": -0.2665092945098877, + "max": 0.2483654022216797, + "mean": -1.536182753625326e-05, + "std": 0.04013803228735924, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.13.2.to_out.0.bias": { + "min": -0.18998344242572784, + "max": 0.1951427161693573, + "mean": -0.0012352201156318188, + "std": 0.06669348478317261, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.13.3.g": { + "min": 0.32910633087158203, + "max": 1.0014653205871582, + "mean": 0.7192941308021545, + "std": 0.05263138189911842, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.13.4.ff.0.0.weight": { + "min": -0.23228409886360168, + "max": 0.24597151577472687, + "mean": 0.00018284631369169801, + "std": 0.040899865329265594, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.13.4.ff.0.0.bias": { + "min": -0.11449356377124786, + "max": 0.019026821479201317, + "mean": -0.042487140744924545, + "std": 0.018874552100896835, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.13.4.ff.2.weight": { + "min": -0.39081379771232605, + "max": 0.4084374010562897, + "mean": -2.154261528630741e-05, + "std": 0.04853346198797226, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.13.4.ff.2.bias": { + "min": -0.6939337849617004, + "max": 0.4130322337150574, + "mean": 0.0008477974915876985, + "std": 0.06032131612300873, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.14.0.weight": { + "min": -0.0010364858899265528, + "max": 1.000504493713379, + "mean": 0.00048820534721016884, + "std": 0.022089021280407906, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.14.1.g": { + "min": 0.997757077217102, + "max": 1.0054128170013428, + "mean": 0.9996482133865356, + "std": 0.0006391748902387917, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.14.2.to_q.weight": { + "min": -0.031263865530490875, + "max": 0.03126693516969681, + "mean": -2.1029807612649165e-05, + "std": 0.018032483756542206, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.14.2.to_q.bias": { + "min": -0.031225642189383507, + "max": 0.031231923028826714, + "mean": -0.000677043863106519, + "std": 0.017827108502388, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.14.2.to_k.weight": { + "min": -0.031264521181583405, + "max": 0.03126373142004013, + "mean": -8.835060725687072e-06, + "std": 0.018031509593129158, + "sparsity": 9.5367431640625e-07, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.14.2.to_k.bias": { + "min": -0.031228171661496162, + "max": 0.031247133389115334, + "mean": -0.0007299243006855249, + "std": 0.017942015081644058, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.14.2.to_v.weight": { + "min": 0.0, + "max": 0.0, + "mean": 0.0, + "std": 0.0, + "sparsity": 1.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.14.2.to_v.bias": { + "min": 0.0, + "max": 0.0, + "mean": 0.0, + "std": 0.0, + "sparsity": 1.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.14.2.to_out.0.weight": { + "min": 0.0, + "max": 0.0, + "mean": 0.0, + "std": 0.0, + "sparsity": 1.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.14.2.to_out.0.bias": { + "min": -0.00041853971197269857, + "max": 0.0003325868456158787, + "mean": -3.1447550554730697e-06, + "std": 0.0001163617562269792, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.14.3.g": { + "min": 0.9978469014167786, + "max": 1.0084865093231201, + "mean": 1.0002028942108154, + "std": 0.002608145819976926, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.14.4.ff.0.0.weight": { + "min": -0.03244105726480484, + "max": 0.03237903118133545, + "mean": -1.7311865576630225e-06, + "std": 0.018027927726507187, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.14.4.ff.0.0.bias": { + "min": -0.032130636274814606, + "max": 0.03116563893854618, + "mean": -0.0003740063984878361, + "std": 0.01804370991885662, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.14.4.ff.2.weight": { + "min": -0.0012820950942113996, + "max": 0.0011165018659085035, + "mean": -8.955282169154088e-07, + "std": 0.00020968639000784606, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.14.4.ff.2.bias": { + "min": -0.0003423716698307544, + "max": 0.00029734382405877113, + "mean": -3.7682302718167193e-06, + "std": 0.00010476629540789872, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.15.0.weight": { + "min": -0.23448625206947327, + "max": 0.272605836391449, + "mean": 6.777544967917493e-06, + "std": 0.018809372559189796, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.15.1.g": { + "min": 0.32130253314971924, + "max": 0.6949947476387024, + "mean": 0.5816991329193115, + "std": 0.04608374834060669, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.15.2.to_q.weight": { + "min": -0.18193963170051575, + "max": 0.19776132702827454, + "mean": -1.1586925211304333e-05, + "std": 0.033183593302965164, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.15.2.to_q.bias": { + "min": -0.16079005599021912, + "max": 0.12958164513111115, + "mean": -0.0010761492885649204, + "std": 0.03415785729885101, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.15.2.to_k.weight": { + "min": -0.33248019218444824, + "max": 0.31138068437576294, + "mean": -1.0150852176593617e-05, + "std": 0.0322343148291111, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.15.2.to_k.bias": { + "min": -7.811703681945801, + "max": 8.77199935913086, + "mean": 0.09351971745491028, + "std": 1.6208088397979736, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.15.2.to_v.weight": { + "min": -0.23395448923110962, + "max": 0.24196705222129822, + "mean": 4.150588938500732e-05, + "std": 0.04085612669587135, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.15.2.to_v.bias": { + "min": -0.07607380300760269, + "max": 0.06586506962776184, + "mean": 0.0004828007658943534, + "std": 0.01941879838705063, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.15.2.to_out.0.weight": { + "min": -0.24638578295707703, + "max": 0.23463943600654602, + "mean": -3.1122344807954505e-06, + "std": 0.03943074867129326, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.15.2.to_out.0.bias": { + "min": -0.16335651278495789, + "max": 0.16123652458190918, + "mean": 0.001627025194466114, + "std": 0.0652812197804451, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.15.3.g": { + "min": 0.5570001006126404, + "max": 0.9467727541923523, + "mean": 0.7130157351493835, + "std": 0.04052889347076416, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.15.4.ff.0.0.weight": { + "min": -0.22856345772743225, + "max": 0.2556101679801941, + "mean": -4.5706547098234296e-05, + "std": 0.040574584156274796, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.15.4.ff.0.0.bias": { + "min": -0.1351189911365509, + "max": 0.02213732711970806, + "mean": -0.04135933890938759, + "std": 0.018408460542559624, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.15.4.ff.2.weight": { + "min": -0.4225083887577057, + "max": 0.3927571773529053, + "mean": -4.4740827433997765e-06, + "std": 0.04778379574418068, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.15.4.ff.2.bias": { + "min": -0.6081869602203369, + "max": 0.6523037552833557, + "mean": 0.0015862288419157267, + "std": 0.0568697564303875, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.16.0.weight": { + "min": -0.25164270401000977, + "max": 0.32068535685539246, + "mean": -6.094380296417512e-06, + "std": 0.019612763077020645, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.16.1.g": { + "min": 0.3596651554107666, + "max": 0.6836386322975159, + "mean": 0.5707623958587646, + "std": 0.04307318106293678, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.16.2.to_q.weight": { + "min": -0.2204994410276413, + "max": 0.17691564559936523, + "mean": -3.469674993539229e-05, + "std": 0.034298643469810486, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.16.2.to_q.bias": { + "min": -0.16342805325984955, + "max": 0.23329652845859528, + "mean": 0.0003627383557613939, + "std": 0.03284167870879173, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.16.2.to_k.weight": { + "min": -0.2643727660179138, + "max": 0.2404099404811859, + "mean": -5.280954064801335e-05, + "std": 0.03389745578169823, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.16.2.to_k.bias": { + "min": -4.859966278076172, + "max": 5.0964674949646, + "mean": 0.04393793269991875, + "std": 1.230094075202942, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.16.2.to_v.weight": { + "min": -0.2466304451227188, + "max": 0.25078442692756653, + "mean": 7.233464566525072e-05, + "std": 0.04398677870631218, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.16.2.to_v.bias": { + "min": -0.06267692148685455, + "max": 0.054532695561647415, + "mean": 0.000642440456431359, + "std": 0.017191536724567413, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.16.2.to_out.0.weight": { + "min": -0.28690966963768005, + "max": 0.27239924669265747, + "mean": -5.01475042256061e-05, + "std": 0.04298488423228264, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.16.2.to_out.0.bias": { + "min": -0.1612030565738678, + "max": 0.1705736219882965, + "mean": -0.0028862706385552883, + "std": 0.05929599329829216, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.16.3.g": { + "min": 0.519792377948761, + "max": 0.9359998106956482, + "mean": 0.7136070132255554, + "std": 0.03880562260746956, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.16.4.ff.0.0.weight": { + "min": -0.23828113079071045, + "max": 0.24893540143966675, + "mean": 0.0004648254835046828, + "std": 0.040453579276800156, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.16.4.ff.0.0.bias": { + "min": -0.14526499807834625, + "max": 0.041103385388851166, + "mean": -0.03970393165946007, + "std": 0.02056412398815155, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.16.4.ff.2.weight": { + "min": -0.5338290929794312, + "max": 0.5837586522102356, + "mean": 5.762096407124773e-06, + "std": 0.04885942488908768, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.16.4.ff.2.bias": { + "min": -0.5195844769477844, + "max": 0.4939325749874115, + "mean": 0.002366485306993127, + "std": 0.05347662419080734, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.17.0.weight": { + "min": -0.273802787065506, + "max": 0.3155968487262726, + "mean": 2.01077523342974e-06, + "std": 0.02004941552877426, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.17.1.g": { + "min": 0.36614885926246643, + "max": 0.7128685116767883, + "mean": 0.5932222604751587, + "std": 0.04609934985637665, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.17.2.to_q.weight": { + "min": -0.21129509806632996, + "max": 0.19956757128238678, + "mean": 3.06197653117124e-05, + "std": 0.034865960478782654, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.17.2.to_q.bias": { + "min": -0.18738499283790588, + "max": 0.20401518046855927, + "mean": 0.0009546762448735535, + "std": 0.031527843326330185, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.17.2.to_k.weight": { + "min": -0.2900747060775757, + "max": 0.3402419686317444, + "mean": -4.711254223366268e-05, + "std": 0.03458685800433159, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.17.2.to_k.bias": { + "min": -3.881408214569092, + "max": 3.3909339904785156, + "mean": 0.014485932886600494, + "std": 0.8588526248931885, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.17.2.to_v.weight": { + "min": -0.22496825456619263, + "max": 0.2504532039165497, + "mean": -3.7677732507290784e-06, + "std": 0.04222949594259262, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.17.2.to_v.bias": { + "min": -0.05539275333285332, + "max": 0.046729691326618195, + "mean": -1.6585952835157514e-05, + "std": 0.01585092395544052, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.17.2.to_out.0.weight": { + "min": -0.29304519295692444, + "max": 0.2904603183269501, + "mean": -7.356060450547375e-06, + "std": 0.04194435849785805, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.17.2.to_out.0.bias": { + "min": -0.12514667212963104, + "max": 0.25951117277145386, + "mean": -0.003241210710257292, + "std": 0.05318121612071991, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.17.3.g": { + "min": 0.4564977288246155, + "max": 0.847152590751648, + "mean": 0.7056270837783813, + "std": 0.03555477410554886, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.17.4.ff.0.0.weight": { + "min": -0.5119993090629578, + "max": 0.3481258749961853, + "mean": 0.0003428043273743242, + "std": 0.04019870236515999, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.17.4.ff.0.0.bias": { + "min": -0.18613915145397186, + "max": 0.03958306089043617, + "mean": -0.03939869999885559, + "std": 0.021371137350797653, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.17.4.ff.2.weight": { + "min": -0.5454614162445068, + "max": 0.5573456287384033, + "mean": -7.15605856385082e-05, + "std": 0.05073413625359535, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.17.4.ff.2.bias": { + "min": -0.5122924447059631, + "max": 0.6649084091186523, + "mean": 0.002443553414195776, + "std": 0.04954148083925247, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.18.0.weight": { + "min": -0.3326307237148285, + "max": 0.2655903100967407, + "mean": 3.417561856622342e-06, + "std": 0.01938662678003311, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.18.1.g": { + "min": 0.32189854979515076, + "max": 0.7676428556442261, + "mean": 0.6510834097862244, + "std": 0.045412834733724594, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.18.2.to_q.weight": { + "min": -0.24963903427124023, + "max": 0.21975325047969818, + "mean": -2.1360538084991276e-06, + "std": 0.03650053218007088, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.18.2.to_q.bias": { + "min": -0.3272336423397064, + "max": 0.2872598171234131, + "mean": -0.000690902175847441, + "std": 0.038575589656829834, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.18.2.to_k.weight": { + "min": -0.3108454644680023, + "max": 0.3709103763103485, + "mean": 6.501353345811367e-05, + "std": 0.036241017282009125, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.18.2.to_k.bias": { + "min": -4.7285308837890625, + "max": 5.821481227874756, + "mean": 0.03798262029886246, + "std": 1.4149147272109985, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.18.2.to_v.weight": { + "min": -0.22184839844703674, + "max": 0.20582044124603271, + "mean": -7.514897151850164e-05, + "std": 0.04248502478003502, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.18.2.to_v.bias": { + "min": -0.07779642939567566, + "max": 0.05152571201324463, + "mean": -0.0009286667918786407, + "std": 0.016416585072875023, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.18.2.to_out.0.weight": { + "min": -0.33085039258003235, + "max": 0.3292792737483978, + "mean": -4.624932898877887e-06, + "std": 0.04279141500592232, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.18.2.to_out.0.bias": { + "min": -0.2853098511695862, + "max": 0.11214260756969452, + "mean": -0.001206133747473359, + "std": 0.0470227487385273, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.18.3.g": { + "min": 0.48610714077949524, + "max": 0.8880516886711121, + "mean": 0.7374852299690247, + "std": 0.038454823195934296, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.18.4.ff.0.0.weight": { + "min": -0.3625198006629944, + "max": 0.274814248085022, + "mean": 5.1260511099826545e-05, + "std": 0.040644217282533646, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.18.4.ff.0.0.bias": { + "min": -0.24789389967918396, + "max": 0.046399183571338654, + "mean": -0.0392770953476429, + "std": 0.023303059861063957, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.18.4.ff.2.weight": { + "min": -0.6276291012763977, + "max": 0.5983994007110596, + "mean": -6.147650128696114e-05, + "std": 0.053116291761398315, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.18.4.ff.2.bias": { + "min": -0.7109575271606445, + "max": 0.2664211392402649, + "mean": 0.0009173410944640636, + "std": 0.05126515030860901, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.19.0.weight": { + "min": -0.3433755040168762, + "max": 0.30368152260780334, + "mean": 1.5963701116561424e-07, + "std": 0.01913503371179104, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.19.1.g": { + "min": 0.34989097714424133, + "max": 0.7839252948760986, + "mean": 0.6388714909553528, + "std": 0.04933994635939598, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.19.2.to_q.weight": { + "min": -0.20559623837471008, + "max": 0.20719166100025177, + "mean": -5.992387013975531e-05, + "std": 0.03769540786743164, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.19.2.to_q.bias": { + "min": -0.25907090306282043, + "max": 0.2685673236846924, + "mean": -0.00039763032691553235, + "std": 0.04464223235845566, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.19.2.to_k.weight": { + "min": -0.35467270016670227, + "max": 0.3229817748069763, + "mean": -6.9561183408950455e-06, + "std": 0.03720381483435631, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.19.2.to_k.bias": { + "min": -5.267129421234131, + "max": 4.20892858505249, + "mean": -0.02641383744776249, + "std": 1.0074299573898315, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.19.2.to_v.weight": { + "min": -0.23873740434646606, + "max": 0.24359266459941864, + "mean": -2.525941454223357e-05, + "std": 0.04320967569947243, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.19.2.to_v.bias": { + "min": -0.06238892674446106, + "max": 0.056785948574543, + "mean": 0.0003448878414928913, + "std": 0.014156854711472988, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.19.2.to_out.0.weight": { + "min": -0.43733710050582886, + "max": 0.3737178444862366, + "mean": 1.443843029846903e-05, + "std": 0.044121142476797104, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.19.2.to_out.0.bias": { + "min": -0.09657814353704453, + "max": 0.1761663407087326, + "mean": -0.0006602209759876132, + "std": 0.03516199812293053, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.19.3.g": { + "min": 0.4218268096446991, + "max": 1.070821762084961, + "mean": 0.7484229803085327, + "std": 0.042183347046375275, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.19.4.ff.0.0.weight": { + "min": -0.2668064832687378, + "max": 0.2973981201648712, + "mean": -7.947084668558091e-05, + "std": 0.0408041812479496, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.19.4.ff.0.0.bias": { + "min": -0.18554465472698212, + "max": 0.04366818815469742, + "mean": -0.03683188557624817, + "std": 0.025637373328208923, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.19.4.ff.2.weight": { + "min": -0.4576941728591919, + "max": 0.4877614378929138, + "mean": 4.342636384535581e-05, + "std": 0.05420947074890137, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.19.4.ff.2.bias": { + "min": -0.28702330589294434, + "max": 0.5525704622268677, + "mean": -0.0008802832453511655, + "std": 0.04786703363060951, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.20.0.weight": { + "min": -0.2927917540073395, + "max": 0.32283690571784973, + "mean": 6.15146973359515e-06, + "std": 0.019968591630458832, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.20.1.g": { + "min": 0.2908625304698944, + "max": 0.7625526785850525, + "mean": 0.650852382183075, + "std": 0.052188921719789505, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.20.2.to_q.weight": { + "min": -0.24394108355045319, + "max": 0.2618102431297302, + "mean": -5.981732101645321e-06, + "std": 0.0396115742623806, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.20.2.to_q.bias": { + "min": -0.267729252576828, + "max": 0.20025481283664703, + "mean": -0.0008811865700408816, + "std": 0.05178782343864441, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.20.2.to_k.weight": { + "min": -0.27242225408554077, + "max": 0.25395235419273376, + "mean": 4.551842721411958e-06, + "std": 0.03870858997106552, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.20.2.to_k.bias": { + "min": -12.978915214538574, + "max": 15.964410781860352, + "mean": 0.033282238990068436, + "std": 1.9907665252685547, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.20.2.to_v.weight": { + "min": -0.20757102966308594, + "max": 0.2263997346162796, + "mean": -7.214213110273704e-05, + "std": 0.04055347666144371, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.20.2.to_v.bias": { + "min": -0.06938357651233673, + "max": 0.06327643245458603, + "mean": 0.00015629694098606706, + "std": 0.014746708795428276, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.20.2.to_out.0.weight": { + "min": -0.46517401933670044, + "max": 0.320604145526886, + "mean": 1.968832475540694e-05, + "std": 0.040588606148958206, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.20.2.to_out.0.bias": { + "min": -0.06422771513462067, + "max": 0.11537671089172363, + "mean": 0.0011921785771846771, + "std": 0.024717185646295547, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.20.3.g": { + "min": 0.3747407793998718, + "max": 0.935266375541687, + "mean": 0.750953733921051, + "std": 0.040338218212127686, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.20.4.ff.0.0.weight": { + "min": -0.2802576720714569, + "max": 0.2736368775367737, + "mean": -0.00016840582247823477, + "std": 0.04099476709961891, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.20.4.ff.0.0.bias": { + "min": -0.19888785481452942, + "max": 0.05115103721618652, + "mean": -0.0320354662835598, + "std": 0.025122012943029404, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.20.4.ff.2.weight": { + "min": -0.6593934297561646, + "max": 0.5366666913032532, + "mean": -4.888622788712382e-05, + "std": 0.05284604802727699, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.20.4.ff.2.bias": { + "min": -0.19323143362998962, + "max": 0.5829473733901978, + "mean": -0.0005128738121129572, + "std": 0.041099581867456436, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.21.0.weight": { + "min": -0.41776803135871887, + "max": 0.3719577491283417, + "mean": 6.155986739031505e-06, + "std": 0.02162076160311699, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.21.1.g": { + "min": 0.21425622701644897, + "max": 0.7496172189712524, + "mean": 0.6495488882064819, + "std": 0.054406262934207916, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.21.2.to_q.weight": { + "min": -0.20992936193943024, + "max": 0.1961071640253067, + "mean": 4.025327507406473e-05, + "std": 0.039461661130189896, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.21.2.to_q.bias": { + "min": -0.3297625780105591, + "max": 0.25971850752830505, + "mean": -0.003232162445783615, + "std": 0.05629448592662811, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.21.2.to_k.weight": { + "min": -0.2059866487979889, + "max": 0.25485166907310486, + "mean": 5.424032860901207e-05, + "std": 0.0385642871260643, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.21.2.to_k.bias": { + "min": -6.252347469329834, + "max": 6.942240238189697, + "mean": 0.0483565516769886, + "std": 1.3863071203231812, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.21.2.to_v.weight": { + "min": -0.21015214920043945, + "max": 0.2306891679763794, + "mean": -5.141047040524427e-06, + "std": 0.041312482208013535, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.21.2.to_v.bias": { + "min": -0.04387219622731209, + "max": 0.036041487008333206, + "mean": 6.907794158905745e-07, + "std": 0.012801294215023518, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.21.2.to_out.0.weight": { + "min": -0.3976612091064453, + "max": 0.3448401689529419, + "mean": -5.557302574743517e-05, + "std": 0.04238886013627052, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.21.2.to_out.0.bias": { + "min": -0.055147796869277954, + "max": 0.06285040080547333, + "mean": 0.00036463249125517905, + "std": 0.018676765263080597, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.21.3.g": { + "min": 0.3504812717437744, + "max": 1.0465654134750366, + "mean": 0.7894250154495239, + "std": 0.048819400370121, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.21.4.ff.0.0.weight": { + "min": -0.3337032198905945, + "max": 0.3862806558609009, + "mean": -0.00016953393060248345, + "std": 0.04147983714938164, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.21.4.ff.0.0.bias": { + "min": -0.1576613336801529, + "max": 0.0590929239988327, + "mean": -0.03184548765420914, + "std": 0.02515709400177002, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.21.4.ff.2.weight": { + "min": -0.6968328952789307, + "max": 0.469901978969574, + "mean": -8.902316039893776e-05, + "std": 0.051792412996292114, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.21.4.ff.2.bias": { + "min": -0.24871186912059784, + "max": 0.32932594418525696, + "mean": -0.0002525809977669269, + "std": 0.04146667197346687, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.22.0.weight": { + "min": -0.2871420085430145, + "max": 0.35027819871902466, + "mean": -2.14410374610452e-06, + "std": 0.024236002936959267, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.22.1.g": { + "min": 0.19654829800128937, + "max": 0.7817674279212952, + "mean": 0.6702600121498108, + "std": 0.058710552752017975, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.22.2.to_q.weight": { + "min": -0.22942325472831726, + "max": 0.2315986454486847, + "mean": -1.993781370401848e-05, + "std": 0.04043741896748543, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.22.2.to_q.bias": { + "min": -0.22012899816036224, + "max": 0.24119356274604797, + "mean": 0.0007787380600348115, + "std": 0.0558554045855999, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.22.2.to_k.weight": { + "min": -0.21590574085712433, + "max": 0.22671166062355042, + "mean": -7.169770105974749e-05, + "std": 0.03937356546521187, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.22.2.to_k.bias": { + "min": -8.917876243591309, + "max": 9.080994606018066, + "mean": -0.001221940852701664, + "std": 1.850203514099121, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.22.2.to_v.weight": { + "min": -0.2697039842605591, + "max": 0.2592160999774933, + "mean": 4.3639320210786536e-05, + "std": 0.03840581700205803, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.22.2.to_v.bias": { + "min": -0.057751722633838654, + "max": 0.05785238742828369, + "mean": 0.0003506582579575479, + "std": 0.014723116531968117, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.22.2.to_out.0.weight": { + "min": -0.26493385434150696, + "max": 0.28856679797172546, + "mean": -6.166309321997687e-05, + "std": 0.0390719398856163, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.22.2.to_out.0.bias": { + "min": -0.04392173886299133, + "max": 0.037354789674282074, + "mean": -9.023403254104778e-05, + "std": 0.013362305238842964, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.22.3.g": { + "min": 0.339423805475235, + "max": 1.0940691232681274, + "mean": 0.8637771010398865, + "std": 0.06392761319875717, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.22.4.ff.0.0.weight": { + "min": -0.42344674468040466, + "max": 0.419131875038147, + "mean": 0.0003126289520878345, + "std": 0.04350034520030022, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.22.4.ff.0.0.bias": { + "min": -0.2149772197008133, + "max": 0.1709900051355362, + "mean": -0.02949333004653454, + "std": 0.03195162117481232, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.22.4.ff.2.weight": { + "min": -0.6006098985671997, + "max": 0.5608966946601868, + "mean": -0.00015077056013979018, + "std": 0.05344511568546295, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.22.4.ff.2.bias": { + "min": -0.17873013019561768, + "max": 0.3772476017475128, + "mean": 0.001360590336844325, + "std": 0.03732540085911751, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.23.0.weight": { + "min": -0.3945685923099518, + "max": 0.3692152500152588, + "mean": 3.696953717735596e-05, + "std": 0.02861735410988331, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.23.1.g": { + "min": 0.2903454899787903, + "max": 0.8293581604957581, + "mean": 0.7055460214614868, + "std": 0.0678996667265892, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.23.2.to_q.weight": { + "min": -0.9263197779655457, + "max": 1.0265021324157715, + "mean": -2.6120340407942422e-05, + "std": 0.04762475937604904, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.23.2.to_q.bias": { + "min": -0.8807425498962402, + "max": 0.8172140717506409, + "mean": -0.00030884621082805097, + "std": 0.09569496661424637, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.23.2.to_k.weight": { + "min": -0.2697499990463257, + "max": 0.24099533259868622, + "mean": -2.2782449377700686e-05, + "std": 0.03895165026187897, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.23.2.to_k.bias": { + "min": -23.798847198486328, + "max": 22.90509796142578, + "mean": -0.0919695645570755, + "std": 4.078832626342773, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.23.2.to_v.weight": { + "min": -0.22787398099899292, + "max": 0.24508967995643616, + "mean": -2.5707324311952107e-05, + "std": 0.038637157529592514, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.23.2.to_v.bias": { + "min": -0.06037602946162224, + "max": 0.04592515528202057, + "mean": -0.00014296159497462213, + "std": 0.01469582598656416, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.23.2.to_out.0.weight": { + "min": -0.33830153942108154, + "max": 0.3749238848686218, + "mean": 7.406164513668045e-06, + "std": 0.04081294313073158, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.23.2.to_out.0.bias": { + "min": -0.04650312289595604, + "max": 0.19583187997341156, + "mean": 0.00027365636196918786, + "std": 0.01356838084757328, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.23.3.g": { + "min": 0.37443819642066956, + "max": 1.133804440498352, + "mean": 0.8900732398033142, + "std": 0.06407663971185684, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.23.4.ff.0.0.weight": { + "min": -0.44803717732429504, + "max": 0.5431130528450012, + "mean": 2.468598904670216e-05, + "std": 0.045565586537122726, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.23.4.ff.0.0.bias": { + "min": -0.22437457740306854, + "max": 0.08822718262672424, + "mean": -0.03203187137842178, + "std": 0.037792954593896866, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.23.4.ff.2.weight": { + "min": -0.7267696261405945, + "max": 0.6905267834663391, + "mean": 3.431630102568306e-05, + "std": 0.05177779868245125, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.23.4.ff.2.bias": { + "min": -0.17477792501449585, + "max": 0.2187574803829193, + "mean": 4.145095590502024e-05, + "std": 0.03179146349430084, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.24.0.weight": { + "min": -0.34067779779434204, + "max": 0.37430673837661743, + "mean": 4.298752173781395e-05, + "std": 0.034139689058065414, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.24.1.g": { + "min": 0.3176548182964325, + "max": 1.2885946035385132, + "mean": 0.6015164256095886, + "std": 0.08361472934484482, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.24.2.to_q.weight": { + "min": -0.2833174467086792, + "max": 0.2604674696922302, + "mean": -2.836968405972584e-06, + "std": 0.0359807163476944, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.24.2.to_q.bias": { + "min": -0.23581622540950775, + "max": 0.20569506287574768, + "mean": 0.00023786764359101653, + "std": 0.05603973567485809, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.24.2.to_k.weight": { + "min": -0.4358341097831726, + "max": 0.3255886137485504, + "mean": 2.4293056412716396e-05, + "std": 0.03413134440779686, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.24.2.to_k.bias": { + "min": -5.55698823928833, + "max": 7.328329086303711, + "mean": -0.007412843406200409, + "std": 0.7006030082702637, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.24.2.to_v.weight": { + "min": -0.3444630801677704, + "max": 0.36411502957344055, + "mean": 0.00010332845704397187, + "std": 0.04782791808247566, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.24.2.to_v.bias": { + "min": -0.07393012195825577, + "max": 0.06051904335618019, + "mean": 0.0009339260286651552, + "std": 0.014950446784496307, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.24.2.to_out.0.weight": { + "min": -0.2559783458709717, + "max": 0.2868276536464691, + "mean": 4.447174433153123e-06, + "std": 0.041554734110832214, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.24.2.to_out.0.bias": { + "min": -0.05538920685648918, + "max": 0.06289947777986526, + "mean": 0.0001379675231873989, + "std": 0.007169328164309263, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.24.3.g": { + "min": 0.4936121106147766, + "max": 1.2250889539718628, + "mean": 1.0134532451629639, + "std": 0.11746872216463089, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.24.4.ff.0.0.weight": { + "min": -1.0937550067901611, + "max": 1.0471408367156982, + "mean": -4.919863567920402e-05, + "std": 0.05240846797823906, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.24.4.ff.0.0.bias": { + "min": -0.22377891838550568, + "max": 0.1730729043483734, + "mean": -0.0272611565887928, + "std": 0.036391731351614, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.24.4.ff.2.weight": { + "min": -0.8862237930297852, + "max": 0.9243613481521606, + "mean": -0.000145945290569216, + "std": 0.05328156799077988, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.24.4.ff.2.bias": { + "min": -0.1714460700750351, + "max": 0.38068291544914246, + "mean": 0.0033734007738530636, + "std": 0.03993367776274681, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.25.0.weight": { + "min": -0.7782041430473328, + "max": 0.7242955565452576, + "mean": 1.8867685867007822e-05, + "std": 0.046160779893398285, + "sparsity": 0.0, + "shape": [ + 1024, + 2048 + ] + }, + "transformer.layers.25.1.g": { + "min": 0.338652104139328, + "max": 1.4327832460403442, + "mean": 0.9483770728111267, + "std": 0.20681361854076385, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.25.2.to_q.weight": { + "min": -1.7457636594772339, + "max": 1.704433560371399, + "mean": 0.00022719459957443178, + "std": 0.15868502855300903, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.25.2.to_q.bias": { + "min": -1.2020785808563232, + "max": 1.102237343788147, + "mean": -0.009557764045894146, + "std": 0.20423445105552673, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.25.2.to_k.weight": { + "min": -0.4211972653865814, + "max": 0.42695388197898865, + "mean": 6.460870645241812e-05, + "std": 0.04801572859287262, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.25.2.to_k.bias": { + "min": -19.786317825317383, + "max": 19.58098602294922, + "mean": -0.24868716299533844, + "std": 4.785643100738525, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.25.2.to_v.weight": { + "min": -0.32400673627853394, + "max": 0.4385600686073303, + "mean": -1.1902460755663924e-05, + "std": 0.046161260455846786, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.25.2.to_v.bias": { + "min": -0.03411344811320305, + "max": 0.03715973347425461, + "mean": 0.000642350991256535, + "std": 0.012920677661895752, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.25.2.to_out.0.weight": { + "min": -0.7036018371582031, + "max": 0.6655198335647583, + "mean": 4.3310083128744736e-05, + "std": 0.057881489396095276, + "sparsity": 0.0, + "shape": [ + 1024, + 1024 + ] + }, + "transformer.layers.25.2.to_out.0.bias": { + "min": -0.07239808887243271, + "max": 0.06768179684877396, + "mean": -0.0001333777909167111, + "std": 0.012929531745612621, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.25.3.g": { + "min": 0.38025927543640137, + "max": 1.3922340869903564, + "mean": 1.0665740966796875, + "std": 0.21970504522323608, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.layers.25.4.ff.0.0.weight": { + "min": -0.6167835593223572, + "max": 0.7178800106048584, + "mean": 0.00011188755161128938, + "std": 0.05802030861377716, + "sparsity": 0.0, + "shape": [ + 4096, + 1024 + ] + }, + "transformer.layers.25.4.ff.0.0.bias": { + "min": -0.21991026401519775, + "max": 0.22539444267749786, + "mean": 0.006232057698071003, + "std": 0.049761686474084854, + "sparsity": 0.0, + "shape": [ + 4096 + ] + }, + "transformer.layers.25.4.ff.2.weight": { + "min": -0.6298967599868774, + "max": 0.8895401954650879, + "mean": 1.17591189336963e-05, + "std": 0.023527691140770912, + "sparsity": 0.0, + "shape": [ + 1024, + 4096 + ] + }, + "transformer.layers.25.4.ff.2.bias": { + "min": -0.5078860521316528, + "max": 0.47492364048957825, + "mean": -0.0030241229105740786, + "std": 0.0694146603345871, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.norm_out.g": { + "min": 0.5377801060676575, + "max": 1.1812876462936401, + "mean": 0.7827885746955872, + "std": 0.09896031767129898, + "sparsity": 0.0, + "shape": [ + 1024 + ] + }, + "transformer.proj_out.weight": { + "min": -0.2672213613986969, + "max": 0.21292650699615479, + "mean": -0.00022339042334351689, + "std": 0.05399598926305771, + "sparsity": 0.0, + "shape": [ + 100, + 1024 + ] + }, + "transformer.proj_out.bias": { + "min": -0.23856915533542633, + "max": 0.014836194925010204, + "mean": -0.043973349034786224, + "std": 0.03437991812825203, + "sparsity": 0.0, + "shape": [ + 100 + ] + } + } +} \ No newline at end of file