Duplicate from FoundationVision/Infinity

Co-authored-by: Jian Han <hanjian@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,51 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+infinity_vae_d32reg.pth filter=lfs diff=lfs merge=lfs -text
+infinity_2b_reg.pth filter=lfs diff=lfs merge=lfs -text
+infinity_vae_d56_f8_14_patchify.pth filter=lfs diff=lfs merge=lfs -text
+infinity_8b_weights/model-00001-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_weights/model-00002-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_weights/model-00003-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_weights/model-00004-of-00004.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00002-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00005-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00008-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00001-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00003-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00004-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00006-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model-00007-of-00008.safetensors filter=lfs diff=lfs merge=lfs -text
+infinity_8b_512x512_weights/model.safetensors.index.json filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,2 @@
+tmp_scripts
+infinity_8b.pth

README.md

@@ -0,0 +1,46 @@
+---
+license: mit
+language:
+- en
+---
+# Infinity ∞: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis
+
+<div align="center">
+
+[![demo platform](https://img.shields.io/badge/Play%20with%20Infinity%21-Infinity%20demo%20platform-lightblue)](https://opensource.bytedance.com/gmpt/t2i/invite)&nbsp;
+[![arXiv](https://img.shields.io/static/v1?label=Project%20Page&message=Github&color=blue&logo=github-pages)](https://foundationvision.github.io/infinity.project/)&nbsp;
+[![arXiv](https://img.shields.io/badge/arXiv%20paper-2412.04431-b31b1b.svg)](https://arxiv.org/abs/2412.04431)&nbsp;
+[![huggingface weights](https://img.shields.io/badge/%F0%9F%A4%97%20Weights-FoundationVision/Infinity-yellow)](https://huggingface.co/FoundationVision/infinity)&nbsp;
+[![code](https://img.shields.io/badge/%F0%9F%A4%96%20Code-FoundationVision/Infinity-green)](https://github.com/FoundationVision/Infinity)&nbsp;
+
+</div>
+<p align="center" style="font-size: larger;">
+  <a href="https://arxiv.org/abs/2412.04431">Infinity: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis</a>
+</p>
+
+
+
+
+
+
+
+## 📖 Introduction
+We present Infinity, a Bitwise Visual AutoRegressive Modeling capable of generating high-resolution and photorealistic images. Infinity redefines visual autoregressive model under a bitwise token prediction framework with an infinite-vocabulary tokenizer & classifier and bitwise self-correction. Theoretically scaling the tokenizer vocabulary size to infinity and concurrently scaling the transformer size, our method significantly unleashes powerful scaling capabilities. Infinity sets a new record for autoregressive text-to-image models, outperforming top-tier diffusion models like SD3-Medium and SDXL. Notably, Infinity surpasses SD3-Medium by improving the GenEval benchmark score from 0.62 to 0.73 and the ImageReward benchmark score from 0.87 to 0.96, achieving a win rate of 66%. Without extra optimization, Infinity generates a high-quality 1024×1024 image in 0.8 seconds, making it 2.6× faster than SD3-Medium and establishing it as the fastest text-to-image model.
+
+## 📌 Note
+This repo is used for hosting Infinity's checkpoints. For more details, please refer to [![code](https://img.shields.io/badge/%F0%9F%A4%96%20Code-FoundationVision/Infinity-green)](https://github.com/FoundationVision/Infinity)&nbsp;
+
+## 📖 Citation
+If our work assists your research, feel free to give us a star ⭐ or cite us using:
+
+```
+@misc{han2024infinityscalingbitwiseautoregressive,
+    title={Infinity: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis}, 
+    author={Jian Han and Jinlai Liu and Yi Jiang and Bin Yan and Yuqi Zhang and Zehuan Yuan and Bingyue Peng and Xiaobing Liu},
+    year={2024},
+    eprint={2412.04431},
+    archivePrefix={arXiv},
+    primaryClass={cs.CV},
+    url={https://arxiv.org/abs/2412.04431}, 
+}
+```

config.json

@@ -0,0 +1,5 @@
+{
+  "Name": [
+    "Infinity"
+  ],
+}

infinity_125M_256x256.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ace50de09cf8ecdbfe06dc50a7544500e840c6a47dac27a13ead2c0799cd14c9
+size 500795996

infinity_2b_reg.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5958b37e7ecb914230cd2b8c28aae0ec108ce08b315720bc3dc52596cc5b3a7
+size 8819295332

infinity_8b_512x512_weights/model-00001-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75d65aafe35c064a85d093a2dd939e2abccb74480f4d4d2ea2f61bd1d6ae8352
+size 4111791048

infinity_8b_512x512_weights/model-00002-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96b1bd6289dcac698b2bf7263e58cd30d953850515bf8c46e5818652b25ea175
+size 4111791048

infinity_8b_512x512_weights/model-00003-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3826efd9734261583b0edcd5635220b803f9ed9165abff15e8415d51efe5ac33
+size 4111791048

infinity_8b_512x512_weights/model-00004-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e5275269241ea77fb23717582059707993622ea1133387917f5f32e60787c9c
+size 4111791048

infinity_8b_512x512_weights/model-00005-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7bfd368e7a9cf3b95efb6ed72b0c5c755ec452e3b7b6cdedd651b575ba31300
+size 4111791048

infinity_8b_512x512_weights/model-00006-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b5676ea63341d41d36eaf1499fc29813c91149459d03fa6c18fb9056295d3d3a
+size 4111791048

infinity_8b_512x512_weights/model-00007-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ef6ce7d3c33fedd5f9e94bf77b173c8372c56c91ddc91f1bb6b7889a4786bf9
+size 4111791048

infinity_8b_512x512_weights/model-00008-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69de9fb6bc48208f966cf6d1f95d732ab95788e5d539041fd237437b5c0dd79c
+size 4721886800

infinity_8b_512x512_weights/model.safetensors.index.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a1ed885aaafad553d4947311dbc8e490e1804dc000a323009a0dcd46d3ed365
+size 69433

infinity_8b_weights/model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b79c2f4b22b58a686ad490431f724a3d388c144c8e60c23b96eaadc898b5ebfd
+size 8223582120

infinity_8b_weights/model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e304b757cf60a0d3742aac2c5c8f194ee43ad845e637bd2d01bc421254abd8b
+size 8223582120

infinity_8b_weights/model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89f7a89084349ea2f35835ef8cc945fce0809262ccb0b5ac51e27cc21b9f1cc7
+size 8223582120

infinity_8b_weights/model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0247d1534bb17e0816228668eee3db6186110cbc679469217402a43adc7cf049
+size 8833677944

infinity_8b_weights/model.safetensors.index.json

@@ -0,0 +1,869 @@
+{
+  "metadata": {
+    "total_size": 33504330240
+  },
+  "weight_map": {
+    "block_chunks.0.module.0.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.0.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.1.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.2.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.3.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.0.module.4.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.0.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.1.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.2.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.3.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ada_gss": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.sa.scale_mul_1H11": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.sa.q_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.sa.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.sa.mat_qkv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.sa.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.sa.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca.v_bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca.mat_q.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca.mat_q.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca.mat_kv.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca.proj.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca.proj.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ffn.fc1.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ffn.fc1.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ffn.fc2.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ffn.fc2.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca_norm.weight": "model-00001-of-00004.safetensors",
+    "block_chunks.1.module.4.ca_norm.bias": "model-00001-of-00004.safetensors",
+    "block_chunks.2.module.0.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.0.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.1.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.2.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.3.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.2.module.4.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.0.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.1.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.2.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.3.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ada_gss": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.sa.scale_mul_1H11": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.sa.q_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.sa.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.sa.mat_qkv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.sa.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.sa.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca.v_bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca.mat_q.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca.mat_q.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca.mat_kv.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca.proj.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca.proj.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ffn.fc1.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ffn.fc1.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ffn.fc2.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ffn.fc2.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca_norm.weight": "model-00002-of-00004.safetensors",
+    "block_chunks.3.module.4.ca_norm.bias": "model-00002-of-00004.safetensors",
+    "block_chunks.4.module.0.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.0.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.1.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.2.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.3.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.4.module.4.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.0.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.1.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.2.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.3.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ada_gss": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.sa.scale_mul_1H11": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.sa.q_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.sa.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.sa.mat_qkv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.sa.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.sa.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca.v_bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca.mat_q.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca.mat_q.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca.mat_kv.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca.proj.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca.proj.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ffn.fc1.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ffn.fc1.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ffn.fc2.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ffn.fc2.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca_norm.weight": "model-00003-of-00004.safetensors",
+    "block_chunks.5.module.4.ca_norm.bias": "model-00003-of-00004.safetensors",
+    "block_chunks.6.module.0.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ada_gss": "model-00004-of-00004.safetensors",
+    "text_norm.weight": "model-00004-of-00004.safetensors",
+    "text_proj_for_sos.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.0.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.1.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.2.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.3.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.4.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.5.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.sa.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca.zero_k_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.0.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.1.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.2.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.3.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.6.module.4.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.0.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.1.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.2.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.3.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ada_gss": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.sa.scale_mul_1H11": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.sa.q_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.sa.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.sa.mat_qkv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.sa.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.sa.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca.v_bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca.mat_q.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca.mat_q.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ffn.fc1.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ffn.fc1.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ffn.fc2.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ffn.fc2.bias": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca_norm.weight": "model-00004-of-00004.safetensors",
+    "block_chunks.7.module.4.ca_norm.bias": "model-00004-of-00004.safetensors",
+    "text_proj_for_sos.ca.mat_q": "model-00004-of-00004.safetensors",
+    "text_proj_for_sos.ca.v_bias": "model-00004-of-00004.safetensors",
+    "text_proj_for_sos.ca.mat_kv.weight": "model-00004-of-00004.safetensors",
+    "text_proj_for_sos.ca.proj.weight": "model-00004-of-00004.safetensors",
+    "text_proj_for_sos.ca.proj.bias": "model-00004-of-00004.safetensors",
+    "text_proj_for_ca.0.weight": "model-00004-of-00004.safetensors",
+    "text_proj_for_ca.0.bias": "model-00004-of-00004.safetensors",
+    "text_proj_for_ca.2.weight": "model-00004-of-00004.safetensors",
+    "text_proj_for_ca.2.bias": "model-00004-of-00004.safetensors",
+    "lvl_embed.weight": "model-00004-of-00004.safetensors",
+    "word_embed.weight": "model-00004-of-00004.safetensors",
+    "word_embed.bias": "model-00004-of-00004.safetensors",
+    "shared_ada_lin.1.weight": "model-00004-of-00004.safetensors",
+    "shared_ada_lin.1.bias": "model-00004-of-00004.safetensors",
+    "head_nm.ada_lin.1.weight": "model-00004-of-00004.safetensors",
+    "head_nm.ada_lin.1.bias": "model-00004-of-00004.safetensors",
+    "head.weight": "model-00004-of-00004.safetensors",
+    "head.bias": "model-00004-of-00004.safetensors",
+    "cfg_uncond": "model-00004-of-00004.safetensors",
+    "pos_start": "model-00004-of-00004.safetensors"
+  }
+}

infinity_vae_d16.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8b866165f7efbc0ae15667c618bc78c6f37b5e71078a948b78d69ff2348d8a1
+size 1555555549

infinity_vae_d24.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11afd37acbf6b02ae5f176add388ff4d264203e9ee40d1d4f0db398525a4c033
+size 1556439517

infinity_vae_d32.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a669bca347d46dc1b10ca601ab67d4f231ad9bad07339e80b22f807ee36ad3e4
+size 1557325341

infinity_vae_d32reg.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a37fa3ea1b2a1ebd23de61d91a5e68202825e5a67edaef4b7c55f5fd5b9cf26
+size 1557324701

infinity_vae_d56_f8_14_patchify.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4c695350125593dde7ec7120e6dff7039a9adea01516f6e14758eb94e3ec6c2
+size 1215341746

infinity_vae_d64.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab1d29c91d52aece43bae1e9c3f11ae993427a315db0be7c26780c5d1e2dbe28
+size 1560864925