Upload 5 files

config.json

@@ -0,0 +1,91 @@
+{
+  "_commit_hash": "9562f3fdc38f09b92594864c5e98264f1bfbca33",
+  "_name_or_path": "togethercomputer/evo-1-131k-base",
+  "architectures": [
+    "StripedHyenaModelForCausalLM"
+  ],
+  "attn_layer_idxs": [
+    8,
+    16,
+    24
+  ],
+  "auto_map": {
+    "AutoConfig": "togethercomputer/evo-1-131k-base--configuration_hyena.StripedHyenaConfig",
+    "AutoModelForCausalLM": "togethercomputer/evo-1-131k-base--modeling_hyena.StripedHyenaModelForCausalLM",
+    "AutoTokenizer": [
+      "togethercomputer/evo-1-131k-base--tokenizer.ByteTokenizer",
+      null
+    ]
+  },
+  "column_split": false,
+  "column_split_hyena": true,
+  "eps": 1e-06,
+  "final_norm": true,
+  "hidden_size": 4096,
+  "hyena_filter_groups": 1,
+  "hyena_layer_idxs": [
+    0,
+    1,
+    2,
+    3,
+    4,
+    5,
+    6,
+    7,
+    9,
+    10,
+    11,
+    12,
+    13,
+    14,
+    15,
+    17,
+    18,
+    19,
+    20,
+    21,
+    22,
+    23,
+    25,
+    26,
+    27,
+    28,
+    29,
+    30,
+    31
+  ],
+  "inference_mode": false,
+  "inner_mlp_size": 10928,
+  "log_intermediate_values": false,
+  "make_vocab_size_divisible_by": 8,
+  "max_seqlen": 131072,
+  "mha_out_proj_bias": true,
+  "mlp_activation": "gelu",
+  "model_parallel_size": 1,
+  "model_type": "stripedhyena",
+  "num_attention_heads": 32,
+  "num_filters": 4096,
+  "num_layers": 32,
+  "pipe_parallel_size": 1,
+  "prefill_style": "fft",
+  "proj_groups": 1,
+  "qkv_proj_bias": true,
+  "rotary_emb_base": 10000,
+  "rotary_emb_scaling_factor": 16,
+  "short_filter_bias": true,
+  "short_filter_length": 3,
+  "smeared_gqa": false,
+  "split_k0": true,
+  "state_size": 8,
+  "tie_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": null,
+  "use_cache": true,
+  "use_flash_attention_2": true,
+  "use_flash_attn": true,
+  "use_flash_depthwise": false,
+  "use_flash_rmsnorm": false,
+  "use_flashfft": false,
+  "use_interpolated_rotary_pos_emb": true,
+  "vocab_size": 512
+}

generation_config.json

@@ -0,0 +1,4 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.40.1"
+}

pytorch_model-00001-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ae71d1a896e490da51751ef0ef2ed00879f66824259653538a6a496e3cd3243
+size 4980102165

pytorch_model-00003-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2591139f7373fc6be3fd9712cfb1c2ad9364bbb3ee5edf9331e02188d1598272
+size 3003328882

pytorch_model.bin.index.json

@@ -0,0 +1,446 @@
+{
+  "metadata": {
+    "total_size": 12913164672
+  },
+  "weight_map": {
+    "backbone.blocks.0.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.0.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.1.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.10.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.11.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.12.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.12.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.12.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.12.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.13.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.13.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.14.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.15.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.inner_mha_cls.Wqkv.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.inner_mha_cls.Wqkv.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.inner_mha_cls.out_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.inner_mha_cls.out_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.inner_mha_cls.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.16.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.17.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.18.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.19.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.2.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.2.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.20.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.20.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.21.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.22.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.filter.D": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.filter.poles": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.filter.residues": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.filter.short_filter_bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.filter.short_filter_weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.mlp.l2.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.mlp.l3.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.out_filter_dense.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.out_filter_dense.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.projections.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.23.projections.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.inner_mha_cls.Wqkv.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.inner_mha_cls.Wqkv.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.inner_mha_cls.out_proj.bias": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.inner_mha_cls.out_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.inner_mha_cls.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.mlp.l1.weight": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.24.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.24.post_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.24.pre_norm.scale": "pytorch_model-00002-of-00003.bin",
+    "backbone.blocks.25.filter.D": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.filter.poles": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.filter.residues": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.filter.short_filter_bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.filter.short_filter_weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.mlp.l1.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.out_filter_dense.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.out_filter_dense.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.post_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.pre_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.projections.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.25.projections.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.filter.D": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.filter.poles": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.filter.residues": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.filter.short_filter_bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.filter.short_filter_weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.mlp.l1.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.out_filter_dense.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.out_filter_dense.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.post_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.pre_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.projections.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.26.projections.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.filter.D": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.filter.poles": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.filter.residues": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.filter.short_filter_bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.filter.short_filter_weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.mlp.l1.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.out_filter_dense.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.out_filter_dense.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.post_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.pre_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.projections.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.27.projections.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.filter.D": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.filter.poles": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.filter.residues": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.filter.short_filter_bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.filter.short_filter_weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.mlp.l1.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.out_filter_dense.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.out_filter_dense.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.post_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.pre_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.projections.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.28.projections.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.filter.D": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.filter.poles": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.filter.residues": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.filter.short_filter_bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.filter.short_filter_weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.mlp.l1.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.out_filter_dense.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.out_filter_dense.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.post_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.pre_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.projections.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.29.projections.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.3.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.3.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.30.filter.D": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.filter.poles": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.filter.residues": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.filter.short_filter_bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.filter.short_filter_weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.mlp.l1.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.out_filter_dense.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.out_filter_dense.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.post_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.pre_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.projections.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.30.projections.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.filter.D": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.filter.poles": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.filter.residues": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.filter.short_filter_bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.filter.short_filter_weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.mlp.l1.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.mlp.l2.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.mlp.l3.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.out_filter_dense.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.out_filter_dense.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.post_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.pre_norm.scale": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.projections.bias": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.31.projections.weight": "pytorch_model-00003-of-00003.bin",
+    "backbone.blocks.4.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.4.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.5.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.6.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.7.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.inner_mha_cls.Wqkv.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.inner_mha_cls.Wqkv.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.inner_mha_cls.out_proj.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.inner_mha_cls.out_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.inner_mha_cls.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.8.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.filter.D": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.filter.poles": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.filter.residues": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.filter.short_filter_bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.filter.short_filter_weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.mlp.l1.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.mlp.l2.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.mlp.l3.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.out_filter_dense.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.out_filter_dense.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.post_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.pre_norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.projections.bias": "pytorch_model-00001-of-00003.bin",
+    "backbone.blocks.9.projections.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.embedding_layer.weight": "pytorch_model-00001-of-00003.bin",
+    "backbone.norm.scale": "pytorch_model-00001-of-00003.bin",
+    "backbone.unembed.weight": "pytorch_model-00001-of-00003.bin"
+  }
+}