Instructions to use modrill/Qwen3-4B-Base-ThinkCode-A-NH025 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use modrill/Qwen3-4B-Base-ThinkCode-A-NH025 with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-4B-Base") model = PeftModel.from_pretrained(base_model, "modrill/Qwen3-4B-Base-ThinkCode-A-NH025") - Transformers
How to use modrill/Qwen3-4B-Base-ThinkCode-A-NH025 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="modrill/Qwen3-4B-Base-ThinkCode-A-NH025")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("modrill/Qwen3-4B-Base-ThinkCode-A-NH025", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use modrill/Qwen3-4B-Base-ThinkCode-A-NH025 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "modrill/Qwen3-4B-Base-ThinkCode-A-NH025" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "modrill/Qwen3-4B-Base-ThinkCode-A-NH025", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/modrill/Qwen3-4B-Base-ThinkCode-A-NH025
- SGLang
How to use modrill/Qwen3-4B-Base-ThinkCode-A-NH025 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "modrill/Qwen3-4B-Base-ThinkCode-A-NH025" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "modrill/Qwen3-4B-Base-ThinkCode-A-NH025", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "modrill/Qwen3-4B-Base-ThinkCode-A-NH025" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "modrill/Qwen3-4B-Base-ThinkCode-A-NH025", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use modrill/Qwen3-4B-Base-ThinkCode-A-NH025 with Docker Model Runner:
docker model run hf.co/modrill/Qwen3-4B-Base-ThinkCode-A-NH025
Add files using upload-large-folder tool
Browse files- ADAPTER_MANIFEST.json +1 -0
- LICENSE +202 -0
- MODULE_SCALE_MANIFEST.json +1 -0
- README.md +131 -0
- VALIDATION.json +1 -0
- adapter_config.json +50 -0
- adapter_model.safetensors +3 -0
ADAPTER_MANIFEST.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"adapter_model":{"bytes":528550328,"path":"adapter_model.safetensors","sha256":"b0c2437fe29dc234871a52202774b76421bfeefdd6467c1770242e9d9e5349ce"},"arm_id":"A-NH025","base_model":"Qwen/Qwen3-4B-Base","base_revision":"906bfd4b4dc7f14ee4320094d8b41684abff8539","contains_base_layer_tensor":false,"contains_base_model_shards":false,"contains_tokenizer_files":false,"created_at_utc":"2026-08-22T18:05:59Z","format":"PEFT_LORA_ADAPTER_ONLY","payload_files":[{"bytes":1519,"path":".gitattributes","sha256":"11ad7efa24975ee4b0c3c3a38ed18737f0658a5f75a0a96787b576a78a023361"},{"bytes":11343,"path":"LICENSE","sha256":"832dd9e00a68dd83b3c3fb9f5588dad7dcf337a0db50f7d9483f310cd292e92e"},{"bytes":82202,"path":"MODULE_SCALE_MANIFEST.json","sha256":"855025049fc7e27a611b5903508f88366a0767f5872e693b481af7c549c6d7db"},{"bytes":4991,"path":"README.md","sha256":"3c8684a7d82461ee02170f061d216865cd0d640a33ac24fb424db5a68c1f026a"},{"bytes":1819,"path":"VALIDATION.json","sha256":"ee3e167d17c8ef25e67fffb0951b665d30587c21404d4a4d10ffa81834b79f1d"},{"bytes":1190,"path":"adapter_config.json","sha256":"a823f881bc5681a80c639d7d874a7a4e98680cebaf776ce569cc80a3a3e14731"},{"bytes":528550328,"path":"adapter_model.safetensors","sha256":"b0c2437fe29dc234871a52202774b76421bfeefdd6467c1770242e9d9e5349ce"}],"payload_manifest_sha256":"552b3af957d6bfcf698a7cad762bf3cb6cdf6f702a3272e8b8d8692b71154845","repo_id":"modrill/Qwen3-4B-Base-ThinkCode-A-NH025","scale_contract":{"body":0.25,"lm_head_and_tied_embedding":0.0},"schema":"HF_PEFT_ADAPTER_RELEASE_MANIFEST_V1","source_artifact_receipt_sha256":"a866ab2be90b8c22e90d0e6f9c373d9152f8788a3ec0be480c5a0b3424a05a83","source_original_adapter_sha256":"2eea0b87617f8b1abdfdb476b1521c70df30402fec638777a694b9d03858ee36","source_phase_a_adapter_recursive_sha256":"e3361b65922de7916dc3ca16f348bae677368f56e304efe94abd38587598cf45","status":"READY"}
|
LICENSE
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
Apache License
|
| 3 |
+
Version 2.0, January 2004
|
| 4 |
+
http://www.apache.org/licenses/
|
| 5 |
+
|
| 6 |
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
| 7 |
+
|
| 8 |
+
1. Definitions.
|
| 9 |
+
|
| 10 |
+
"License" shall mean the terms and conditions for use, reproduction,
|
| 11 |
+
and distribution as defined by Sections 1 through 9 of this document.
|
| 12 |
+
|
| 13 |
+
"Licensor" shall mean the copyright owner or entity authorized by
|
| 14 |
+
the copyright owner that is granting the License.
|
| 15 |
+
|
| 16 |
+
"Legal Entity" shall mean the union of the acting entity and all
|
| 17 |
+
other entities that control, are controlled by, or are under common
|
| 18 |
+
control with that entity. For the purposes of this definition,
|
| 19 |
+
"control" means (i) the power, direct or indirect, to cause the
|
| 20 |
+
direction or management of such entity, whether by contract or
|
| 21 |
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
| 22 |
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
| 23 |
+
|
| 24 |
+
"You" (or "Your") shall mean an individual or Legal Entity
|
| 25 |
+
exercising permissions granted by this License.
|
| 26 |
+
|
| 27 |
+
"Source" form shall mean the preferred form for making modifications,
|
| 28 |
+
including but not limited to software source code, documentation
|
| 29 |
+
source, and configuration files.
|
| 30 |
+
|
| 31 |
+
"Object" form shall mean any form resulting from mechanical
|
| 32 |
+
transformation or translation of a Source form, including but
|
| 33 |
+
not limited to compiled object code, generated documentation,
|
| 34 |
+
and conversions to other media types.
|
| 35 |
+
|
| 36 |
+
"Work" shall mean the work of authorship, whether in Source or
|
| 37 |
+
Object form, made available under the License, as indicated by a
|
| 38 |
+
copyright notice that is included in or attached to the work
|
| 39 |
+
(an example is provided in the Appendix below).
|
| 40 |
+
|
| 41 |
+
"Derivative Works" shall mean any work, whether in Source or Object
|
| 42 |
+
form, that is based on (or derived from) the Work and for which the
|
| 43 |
+
editorial revisions, annotations, elaborations, or other modifications
|
| 44 |
+
represent, as a whole, an original work of authorship. For the purposes
|
| 45 |
+
of this License, Derivative Works shall not include works that remain
|
| 46 |
+
separable from, or merely link (or bind by name) to the interfaces of,
|
| 47 |
+
the Work and Derivative Works thereof.
|
| 48 |
+
|
| 49 |
+
"Contribution" shall mean any work of authorship, including
|
| 50 |
+
the original version of the Work and any modifications or additions
|
| 51 |
+
to that Work or Derivative Works thereof, that is intentionally
|
| 52 |
+
submitted to Licensor for inclusion in the Work by the copyright owner
|
| 53 |
+
or by an individual or Legal Entity authorized to submit on behalf of
|
| 54 |
+
the copyright owner. For the purposes of this definition, "submitted"
|
| 55 |
+
means any form of electronic, verbal, or written communication sent
|
| 56 |
+
to the Licensor or its representatives, including but not limited to
|
| 57 |
+
communication on electronic mailing lists, source code control systems,
|
| 58 |
+
and issue tracking systems that are managed by, or on behalf of, the
|
| 59 |
+
Licensor for the purpose of discussing and improving the Work, but
|
| 60 |
+
excluding communication that is conspicuously marked or otherwise
|
| 61 |
+
designated in writing by the copyright owner as "Not a Contribution."
|
| 62 |
+
|
| 63 |
+
"Contributor" shall mean Licensor and any individual or Legal Entity
|
| 64 |
+
on behalf of whom a Contribution has been received by Licensor and
|
| 65 |
+
subsequently incorporated within the Work.
|
| 66 |
+
|
| 67 |
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
| 68 |
+
this License, each Contributor hereby grants to You a perpetual,
|
| 69 |
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
| 70 |
+
copyright license to reproduce, prepare Derivative Works of,
|
| 71 |
+
publicly display, publicly perform, sublicense, and distribute the
|
| 72 |
+
Work and such Derivative Works in Source or Object form.
|
| 73 |
+
|
| 74 |
+
3. Grant of Patent License. Subject to the terms and conditions of
|
| 75 |
+
this License, each Contributor hereby grants to You a perpetual,
|
| 76 |
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
| 77 |
+
(except as stated in this section) patent license to make, have made,
|
| 78 |
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
| 79 |
+
where such license applies only to those patent claims licensable
|
| 80 |
+
by such Contributor that are necessarily infringed by their
|
| 81 |
+
Contribution(s) alone or by combination of their Contribution(s)
|
| 82 |
+
with the Work to which such Contribution(s) was submitted. If You
|
| 83 |
+
institute patent litigation against any entity (including a
|
| 84 |
+
cross-claim or counterclaim in a lawsuit) alleging that the Work
|
| 85 |
+
or a Contribution incorporated within the Work constitutes direct
|
| 86 |
+
or contributory patent infringement, then any patent licenses
|
| 87 |
+
granted to You under this License for that Work shall terminate
|
| 88 |
+
as of the date such litigation is filed.
|
| 89 |
+
|
| 90 |
+
4. Redistribution. You may reproduce and distribute copies of the
|
| 91 |
+
Work or Derivative Works thereof in any medium, with or without
|
| 92 |
+
modifications, and in Source or Object form, provided that You
|
| 93 |
+
meet the following conditions:
|
| 94 |
+
|
| 95 |
+
(a) You must give any other recipients of the Work or
|
| 96 |
+
Derivative Works a copy of this License; and
|
| 97 |
+
|
| 98 |
+
(b) You must cause any modified files to carry prominent notices
|
| 99 |
+
stating that You changed the files; and
|
| 100 |
+
|
| 101 |
+
(c) You must retain, in the Source form of any Derivative Works
|
| 102 |
+
that You distribute, all copyright, patent, trademark, and
|
| 103 |
+
attribution notices from the Source form of the Work,
|
| 104 |
+
excluding those notices that do not pertain to any part of
|
| 105 |
+
the Derivative Works; and
|
| 106 |
+
|
| 107 |
+
(d) If the Work includes a "NOTICE" text file as part of its
|
| 108 |
+
distribution, then any Derivative Works that You distribute must
|
| 109 |
+
include a readable copy of the attribution notices contained
|
| 110 |
+
within such NOTICE file, excluding those notices that do not
|
| 111 |
+
pertain to any part of the Derivative Works, in at least one
|
| 112 |
+
of the following places: within a NOTICE text file distributed
|
| 113 |
+
as part of the Derivative Works; within the Source form or
|
| 114 |
+
documentation, if provided along with the Derivative Works; or,
|
| 115 |
+
within a display generated by the Derivative Works, if and
|
| 116 |
+
wherever such third-party notices normally appear. The contents
|
| 117 |
+
of the NOTICE file are for informational purposes only and
|
| 118 |
+
do not modify the License. You may add Your own attribution
|
| 119 |
+
notices within Derivative Works that You distribute, alongside
|
| 120 |
+
or as an addendum to the NOTICE text from the Work, provided
|
| 121 |
+
that such additional attribution notices cannot be construed
|
| 122 |
+
as modifying the License.
|
| 123 |
+
|
| 124 |
+
You may add Your own copyright statement to Your modifications and
|
| 125 |
+
may provide additional or different license terms and conditions
|
| 126 |
+
for use, reproduction, or distribution of Your modifications, or
|
| 127 |
+
for any such Derivative Works as a whole, provided Your use,
|
| 128 |
+
reproduction, and distribution of the Work otherwise complies with
|
| 129 |
+
the conditions stated in this License.
|
| 130 |
+
|
| 131 |
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
| 132 |
+
any Contribution intentionally submitted for inclusion in the Work
|
| 133 |
+
by You to the Licensor shall be under the terms and conditions of
|
| 134 |
+
this License, without any additional terms or conditions.
|
| 135 |
+
Notwithstanding the above, nothing herein shall supersede or modify
|
| 136 |
+
the terms of any separate license agreement you may have executed
|
| 137 |
+
with Licensor regarding such Contributions.
|
| 138 |
+
|
| 139 |
+
6. Trademarks. This License does not grant permission to use the trade
|
| 140 |
+
names, trademarks, service marks, or product names of the Licensor,
|
| 141 |
+
except as required for reasonable and customary use in describing the
|
| 142 |
+
origin of the Work and reproducing the content of the NOTICE file.
|
| 143 |
+
|
| 144 |
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
| 145 |
+
agreed to in writing, Licensor provides the Work (and each
|
| 146 |
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
| 147 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
| 148 |
+
implied, including, without limitation, any warranties or conditions
|
| 149 |
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
| 150 |
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
| 151 |
+
appropriateness of using or redistributing the Work and assume any
|
| 152 |
+
risks associated with Your exercise of permissions under this License.
|
| 153 |
+
|
| 154 |
+
8. Limitation of Liability. In no event and under no legal theory,
|
| 155 |
+
whether in tort (including negligence), contract, or otherwise,
|
| 156 |
+
unless required by applicable law (such as deliberate and grossly
|
| 157 |
+
negligent acts) or agreed to in writing, shall any Contributor be
|
| 158 |
+
liable to You for damages, including any direct, indirect, special,
|
| 159 |
+
incidental, or consequential damages of any character arising as a
|
| 160 |
+
result of this License or out of the use or inability to use the
|
| 161 |
+
Work (including but not limited to damages for loss of goodwill,
|
| 162 |
+
work stoppage, computer failure or malfunction, or any and all
|
| 163 |
+
other commercial damages or losses), even if such Contributor
|
| 164 |
+
has been advised of the possibility of such damages.
|
| 165 |
+
|
| 166 |
+
9. Accepting Warranty or Additional Liability. While redistributing
|
| 167 |
+
the Work or Derivative Works thereof, You may choose to offer,
|
| 168 |
+
and charge a fee for, acceptance of support, warranty, indemnity,
|
| 169 |
+
or other liability obligations and/or rights consistent with this
|
| 170 |
+
License. However, in accepting such obligations, You may act only
|
| 171 |
+
on Your own behalf and on Your sole responsibility, not on behalf
|
| 172 |
+
of any other Contributor, and only if You agree to indemnify,
|
| 173 |
+
defend, and hold each Contributor harmless for any liability
|
| 174 |
+
incurred by, or claims asserted against, such Contributor by reason
|
| 175 |
+
of your accepting any such warranty or additional liability.
|
| 176 |
+
|
| 177 |
+
END OF TERMS AND CONDITIONS
|
| 178 |
+
|
| 179 |
+
APPENDIX: How to apply the Apache License to your work.
|
| 180 |
+
|
| 181 |
+
To apply the Apache License to your work, attach the following
|
| 182 |
+
boilerplate notice, with the fields enclosed by brackets "[]"
|
| 183 |
+
replaced with your own identifying information. (Don't include
|
| 184 |
+
the brackets!) The text should be enclosed in the appropriate
|
| 185 |
+
comment syntax for the file format. We also recommend that a
|
| 186 |
+
file or class name and description of purpose be included on the
|
| 187 |
+
same "printed page" as the copyright notice for easier
|
| 188 |
+
identification within third-party archives.
|
| 189 |
+
|
| 190 |
+
Copyright 2024 Alibaba Cloud
|
| 191 |
+
|
| 192 |
+
Licensed under the Apache License, Version 2.0 (the "License");
|
| 193 |
+
you may not use this file except in compliance with the License.
|
| 194 |
+
You may obtain a copy of the License at
|
| 195 |
+
|
| 196 |
+
http://www.apache.org/licenses/LICENSE-2.0
|
| 197 |
+
|
| 198 |
+
Unless required by applicable law or agreed to in writing, software
|
| 199 |
+
distributed under the License is distributed on an "AS IS" BASIS,
|
| 200 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 201 |
+
See the License for the specific language governing permissions and
|
| 202 |
+
limitations under the License.
|
MODULE_SCALE_MANIFEST.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"arm_id":"A-NH025","construction":"FP32_SCALE_LORA_B_THEN_FP32_B_AT_A_ALPHA_OVER_R","created_at_utc":"2026-08-22T05:54:14Z","layer_contract":{"layers_0_23":0.25,"layers_24_35":0.25,"lm_head_and_tied_shared_embedding":0.0},"module_count":253,"modules":[{"base_physical_weight":"model.layers.0.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.0.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.0.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.0.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.0.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.0.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.0.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.0.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.0.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.0.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.0.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.0.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.0.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.0.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.0.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.0.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.0.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.0.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.0.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.0.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.0.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.0.mlp.up_proj","lora_A_key":"base_model.model.model.layers.0.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.0.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.0.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.0.mlp.down_proj","lora_A_key":"base_model.model.model.layers.0.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.0.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.1.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.1.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.1.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.1.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.1.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.1.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.1.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.1.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.1.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.1.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.1.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.1.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.1.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.1.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.1.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.1.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.1.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.1.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.1.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.1.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.1.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.1.mlp.up_proj","lora_A_key":"base_model.model.model.layers.1.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.1.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.1.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.1.mlp.down_proj","lora_A_key":"base_model.model.model.layers.1.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.1.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.2.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.2.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.2.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.2.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.2.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.2.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.2.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.2.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.2.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.2.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.2.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.2.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.2.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.2.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.2.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.2.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.2.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.2.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.2.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.2.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.2.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.2.mlp.up_proj","lora_A_key":"base_model.model.model.layers.2.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.2.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.2.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.2.mlp.down_proj","lora_A_key":"base_model.model.model.layers.2.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.2.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.3.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.3.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.3.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.3.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.3.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.3.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.3.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.3.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.3.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.3.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.3.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.3.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.3.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.3.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.3.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.3.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.3.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.3.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.3.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.3.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.3.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.3.mlp.up_proj","lora_A_key":"base_model.model.model.layers.3.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.3.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.3.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.3.mlp.down_proj","lora_A_key":"base_model.model.model.layers.3.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.3.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.4.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.4.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.4.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.4.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.4.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.4.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.4.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.4.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.4.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.4.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.4.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.4.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.4.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.4.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.4.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.4.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.4.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.4.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.4.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.4.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.4.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.4.mlp.up_proj","lora_A_key":"base_model.model.model.layers.4.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.4.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.4.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.4.mlp.down_proj","lora_A_key":"base_model.model.model.layers.4.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.4.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.5.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.5.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.5.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.5.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.5.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.5.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.5.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.5.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.5.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.5.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.5.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.5.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.5.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.5.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.5.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.5.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.5.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.5.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.5.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.5.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.5.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.5.mlp.up_proj","lora_A_key":"base_model.model.model.layers.5.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.5.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.5.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.5.mlp.down_proj","lora_A_key":"base_model.model.model.layers.5.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.5.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.6.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.6.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.6.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.6.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.6.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.6.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.6.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.6.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.6.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.6.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.6.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.6.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.6.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.6.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.6.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.6.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.6.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.6.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.6.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.6.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.6.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.6.mlp.up_proj","lora_A_key":"base_model.model.model.layers.6.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.6.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.6.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.6.mlp.down_proj","lora_A_key":"base_model.model.model.layers.6.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.6.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.7.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.7.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.7.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.7.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.7.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.7.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.7.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.7.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.7.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.7.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.7.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.7.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.7.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.7.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.7.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.7.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.7.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.7.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.7.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.7.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.7.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.7.mlp.up_proj","lora_A_key":"base_model.model.model.layers.7.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.7.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.7.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.7.mlp.down_proj","lora_A_key":"base_model.model.model.layers.7.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.7.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.8.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.8.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.8.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.8.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.8.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.8.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.8.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.8.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.8.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.8.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.8.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.8.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.8.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.8.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.8.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.8.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.8.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.8.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.8.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.8.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.8.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.8.mlp.up_proj","lora_A_key":"base_model.model.model.layers.8.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.8.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.8.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.8.mlp.down_proj","lora_A_key":"base_model.model.model.layers.8.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.8.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.9.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.9.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.9.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.9.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.9.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.9.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.9.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.9.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.9.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.9.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.9.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.9.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.9.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.9.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.9.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.9.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.9.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.9.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.9.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.9.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.9.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.9.mlp.up_proj","lora_A_key":"base_model.model.model.layers.9.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.9.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.9.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.9.mlp.down_proj","lora_A_key":"base_model.model.model.layers.9.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.9.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.10.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.10.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.10.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.10.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.10.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.10.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.10.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.10.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.10.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.10.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.10.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.10.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.10.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.10.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.10.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.10.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.10.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.10.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.10.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.10.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.10.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.10.mlp.up_proj","lora_A_key":"base_model.model.model.layers.10.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.10.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.10.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.10.mlp.down_proj","lora_A_key":"base_model.model.model.layers.10.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.10.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.11.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.11.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.11.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.11.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.11.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.11.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.11.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.11.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.11.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.11.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.11.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.11.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.11.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.11.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.11.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.11.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.11.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.11.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.11.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.11.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.11.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.11.mlp.up_proj","lora_A_key":"base_model.model.model.layers.11.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.11.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.11.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.11.mlp.down_proj","lora_A_key":"base_model.model.model.layers.11.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.11.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.12.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.12.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.12.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.12.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.12.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.12.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.12.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.12.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.12.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.12.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.12.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.12.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.12.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.12.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.12.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.12.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.12.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.12.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.12.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.12.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.12.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.12.mlp.up_proj","lora_A_key":"base_model.model.model.layers.12.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.12.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.12.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.12.mlp.down_proj","lora_A_key":"base_model.model.model.layers.12.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.12.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.13.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.13.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.13.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.13.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.13.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.13.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.13.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.13.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.13.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.13.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.13.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.13.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.13.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.13.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.13.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.13.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.13.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.13.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.13.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.13.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.13.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.13.mlp.up_proj","lora_A_key":"base_model.model.model.layers.13.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.13.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.13.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.13.mlp.down_proj","lora_A_key":"base_model.model.model.layers.13.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.13.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.14.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.14.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.14.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.14.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.14.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.14.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.14.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.14.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.14.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.14.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.14.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.14.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.14.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.14.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.14.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.14.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.14.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.14.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.14.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.14.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.14.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.14.mlp.up_proj","lora_A_key":"base_model.model.model.layers.14.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.14.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.14.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.14.mlp.down_proj","lora_A_key":"base_model.model.model.layers.14.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.14.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.15.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.15.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.15.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.15.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.15.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.15.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.15.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.15.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.15.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.15.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.15.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.15.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.15.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.15.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.15.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.15.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.15.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.15.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.15.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.15.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.15.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.15.mlp.up_proj","lora_A_key":"base_model.model.model.layers.15.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.15.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.15.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.15.mlp.down_proj","lora_A_key":"base_model.model.model.layers.15.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.15.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.16.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.16.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.16.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.16.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.16.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.16.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.16.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.16.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.16.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.16.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.16.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.16.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.16.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.16.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.16.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.16.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.16.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.16.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.16.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.16.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.16.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.16.mlp.up_proj","lora_A_key":"base_model.model.model.layers.16.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.16.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.16.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.16.mlp.down_proj","lora_A_key":"base_model.model.model.layers.16.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.16.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.17.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.17.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.17.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.17.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.17.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.17.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.17.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.17.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.17.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.17.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.17.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.17.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.17.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.17.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.17.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.17.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.17.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.17.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.17.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.17.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.17.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.17.mlp.up_proj","lora_A_key":"base_model.model.model.layers.17.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.17.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.17.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.17.mlp.down_proj","lora_A_key":"base_model.model.model.layers.17.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.17.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.18.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.18.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.18.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.18.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.18.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.18.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.18.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.18.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.18.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.18.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.18.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.18.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.18.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.18.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.18.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.18.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.18.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.18.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.18.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.18.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.18.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.18.mlp.up_proj","lora_A_key":"base_model.model.model.layers.18.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.18.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.18.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.18.mlp.down_proj","lora_A_key":"base_model.model.model.layers.18.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.18.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.19.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.19.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.19.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.19.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.19.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.19.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.19.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.19.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.19.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.19.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.19.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.19.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.19.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.19.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.19.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.19.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.19.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.19.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.19.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.19.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.19.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.19.mlp.up_proj","lora_A_key":"base_model.model.model.layers.19.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.19.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.19.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.19.mlp.down_proj","lora_A_key":"base_model.model.model.layers.19.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.19.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.20.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.20.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.20.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.20.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.20.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.20.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.20.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.20.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.20.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.20.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.20.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.20.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.20.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.20.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.20.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.20.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.20.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.20.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.20.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.20.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.20.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.20.mlp.up_proj","lora_A_key":"base_model.model.model.layers.20.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.20.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.20.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.20.mlp.down_proj","lora_A_key":"base_model.model.model.layers.20.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.20.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.21.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.21.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.21.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.21.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.21.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.21.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.21.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.21.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.21.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.21.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.21.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.21.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.21.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.21.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.21.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.21.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.21.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.21.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.21.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.21.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.21.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.21.mlp.up_proj","lora_A_key":"base_model.model.model.layers.21.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.21.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.21.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.21.mlp.down_proj","lora_A_key":"base_model.model.model.layers.21.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.21.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.22.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.22.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.22.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.22.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.22.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.22.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.22.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.22.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.22.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.22.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.22.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.22.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.22.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.22.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.22.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.22.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.22.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.22.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.22.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.22.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.22.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.22.mlp.up_proj","lora_A_key":"base_model.model.model.layers.22.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.22.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.22.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.22.mlp.down_proj","lora_A_key":"base_model.model.model.layers.22.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.22.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.23.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.23.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.23.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.23.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.23.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.23.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.23.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.23.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.23.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.23.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.23.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.23.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.23.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.23.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.23.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.23.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.23.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.23.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.23.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.23.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.23.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.23.mlp.up_proj","lora_A_key":"base_model.model.model.layers.23.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.23.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.23.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.23.mlp.down_proj","lora_A_key":"base_model.model.model.layers.23.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.23.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.24.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.24.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.24.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.24.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.24.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.24.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.24.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.24.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.24.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.24.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.24.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.24.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.24.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.24.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.24.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.24.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.24.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.24.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.24.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.24.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.24.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.24.mlp.up_proj","lora_A_key":"base_model.model.model.layers.24.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.24.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.24.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.24.mlp.down_proj","lora_A_key":"base_model.model.model.layers.24.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.24.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.25.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.25.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.25.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.25.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.25.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.25.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.25.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.25.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.25.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.25.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.25.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.25.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.25.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.25.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.25.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.25.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.25.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.25.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.25.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.25.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.25.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.25.mlp.up_proj","lora_A_key":"base_model.model.model.layers.25.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.25.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.25.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.25.mlp.down_proj","lora_A_key":"base_model.model.model.layers.25.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.25.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.26.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.26.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.26.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.26.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.26.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.26.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.26.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.26.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.26.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.26.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.26.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.26.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.26.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.26.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.26.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.26.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.26.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.26.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.26.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.26.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.26.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.26.mlp.up_proj","lora_A_key":"base_model.model.model.layers.26.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.26.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.26.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.26.mlp.down_proj","lora_A_key":"base_model.model.model.layers.26.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.26.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.27.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.27.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.27.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.27.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.27.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.27.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.27.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.27.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.27.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.27.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.27.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.27.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.27.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.27.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.27.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.27.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.27.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.27.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.27.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.27.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.27.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.27.mlp.up_proj","lora_A_key":"base_model.model.model.layers.27.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.27.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.27.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.27.mlp.down_proj","lora_A_key":"base_model.model.model.layers.27.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.27.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.28.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.28.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.28.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.28.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.28.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.28.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.28.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.28.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.28.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.28.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.28.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.28.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.28.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.28.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.28.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.28.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.28.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.28.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.28.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.28.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.28.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.28.mlp.up_proj","lora_A_key":"base_model.model.model.layers.28.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.28.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.28.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.28.mlp.down_proj","lora_A_key":"base_model.model.model.layers.28.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.28.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.29.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.29.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.29.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.29.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.29.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.29.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.29.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.29.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.29.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.29.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.29.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.29.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.29.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.29.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.29.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.29.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.29.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.29.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.29.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.29.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.29.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.29.mlp.up_proj","lora_A_key":"base_model.model.model.layers.29.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.29.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.29.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.29.mlp.down_proj","lora_A_key":"base_model.model.model.layers.29.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.29.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.30.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.30.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.30.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.30.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.30.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.30.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.30.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.30.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.30.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.30.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.30.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.30.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.30.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.30.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.30.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.30.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.30.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.30.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.30.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.30.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.30.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.30.mlp.up_proj","lora_A_key":"base_model.model.model.layers.30.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.30.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.30.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.30.mlp.down_proj","lora_A_key":"base_model.model.model.layers.30.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.30.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.31.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.31.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.31.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.31.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.31.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.31.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.31.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.31.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.31.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.31.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.31.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.31.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.31.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.31.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.31.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.31.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.31.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.31.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.31.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.31.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.31.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.31.mlp.up_proj","lora_A_key":"base_model.model.model.layers.31.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.31.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.31.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.31.mlp.down_proj","lora_A_key":"base_model.model.model.layers.31.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.31.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.32.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.32.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.32.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.32.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.32.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.32.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.32.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.32.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.32.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.32.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.32.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.32.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.32.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.32.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.32.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.32.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.32.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.32.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.32.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.32.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.32.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.32.mlp.up_proj","lora_A_key":"base_model.model.model.layers.32.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.32.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.32.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.32.mlp.down_proj","lora_A_key":"base_model.model.model.layers.32.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.32.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.33.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.33.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.33.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.33.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.33.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.33.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.33.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.33.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.33.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.33.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.33.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.33.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.33.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.33.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.33.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.33.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.33.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.33.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.33.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.33.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.33.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.33.mlp.up_proj","lora_A_key":"base_model.model.model.layers.33.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.33.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.33.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.33.mlp.down_proj","lora_A_key":"base_model.model.model.layers.33.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.33.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.34.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.34.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.34.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.34.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.34.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.34.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.34.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.34.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.34.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.34.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.34.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.34.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.34.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.34.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.34.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.34.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.34.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.34.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.34.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.34.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.34.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.34.mlp.up_proj","lora_A_key":"base_model.model.model.layers.34.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.34.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.34.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.34.mlp.down_proj","lora_A_key":"base_model.model.model.layers.34.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.34.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.35.self_attn.q_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.35.self_attn.q_proj","lora_A_key":"base_model.model.model.layers.35.self_attn.q_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.35.self_attn.q_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.35.self_attn.k_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.35.self_attn.k_proj","lora_A_key":"base_model.model.model.layers.35.self_attn.k_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.35.self_attn.k_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.35.self_attn.v_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.35.self_attn.v_proj","lora_A_key":"base_model.model.model.layers.35.self_attn.v_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.35.self_attn.v_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.35.self_attn.o_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.35.self_attn.o_proj","lora_A_key":"base_model.model.model.layers.35.self_attn.o_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.35.self_attn.o_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.35.mlp.gate_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.35.mlp.gate_proj","lora_A_key":"base_model.model.model.layers.35.mlp.gate_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.35.mlp.gate_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.35.mlp.up_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.35.mlp.up_proj","lora_A_key":"base_model.model.model.layers.35.mlp.up_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.35.mlp.up_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.layers.35.mlp.down_proj.weight","head_shared_embedding_delta_zero":false,"logical_module":"model.layers.35.mlp.down_proj","lora_A_key":"base_model.model.model.layers.35.mlp.down_proj.lora_A.weight","lora_B_key":"base_model.model.model.layers.35.mlp.down_proj.lora_B.weight","scale":0.25},{"base_physical_weight":"model.embed_tokens.weight","head_shared_embedding_delta_zero":true,"logical_module":"lm_head","lora_A_key":"base_model.model.lm_head.lora_A.weight","lora_B_key":"base_model.model.lm_head.lora_B.weight","scale":0.0}],"original_adapter_sha256":"2eea0b87617f8b1abdfdb476b1521c70df30402fec638777a694b9d03858ee36","schema":"CTX_PHASE_A_MODULE_SCALE_MANIFEST_V1","status":"FROZEN"}
|
README.md
ADDED
|
@@ -0,0 +1,131 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: peft
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
base_model: Qwen/Qwen3-4B-Base
|
| 6 |
+
base_model_relation: adapter
|
| 7 |
+
tags:
|
| 8 |
+
- peft
|
| 9 |
+
- lora
|
| 10 |
+
- transformers
|
| 11 |
+
- safetensors
|
| 12 |
+
- qwen3
|
| 13 |
+
- code
|
| 14 |
+
- text-generation
|
| 15 |
+
model-index:
|
| 16 |
+
- name: Qwen3-4B-Base-ThinkCode-A-NH025 PEFT Adapter
|
| 17 |
+
results:
|
| 18 |
+
- task:
|
| 19 |
+
type: text-generation
|
| 20 |
+
name: Code Generation
|
| 21 |
+
dataset:
|
| 22 |
+
name: EvalScope Full1055 corrected (development-only)
|
| 23 |
+
type: evalscope-full1055-corrected-development
|
| 24 |
+
metrics:
|
| 25 |
+
- type: pass@1
|
| 26 |
+
name: resolved aggregate code_only pass@1 (3 seeds)
|
| 27 |
+
value: 25.09
|
| 28 |
+
---
|
| 29 |
+
|
| 30 |
+
# Qwen3-4B-Base-ThinkCode-A-NH025 — PEFT Adapter
|
| 31 |
+
|
| 32 |
+
This repository contains a **PEFT LoRA adapter only**. It does not contain the
|
| 33 |
+
Qwen3 base-model weights and cannot be loaded as a standalone causal language
|
| 34 |
+
model.
|
| 35 |
+
|
| 36 |
+
The required base is
|
| 37 |
+
[`Qwen/Qwen3-4B-Base`](https://huggingface.co/Qwen/Qwen3-4B-Base) at the fixed
|
| 38 |
+
revision `906bfd4b4dc7f14ee4320094d8b41684abff8539`.
|
| 39 |
+
|
| 40 |
+
## Adapter construction
|
| 41 |
+
|
| 42 |
+
`A-NH025` is the Phase A no-head arm. Starting from the completed source LoRA,
|
| 43 |
+
every selected transformer-body LoRA `B` tensor is multiplied by `0.25` in
|
| 44 |
+
FP32, while the `lm_head` LoRA `B` tensor is multiplied by `0`, making its
|
| 45 |
+
effective head/shared-embedding delta exactly zero. LoRA `A` tensors are
|
| 46 |
+
unchanged. With `lora_alpha=128` and `r=64`, PEFT applies the intended body
|
| 47 |
+
delta without a language-model-head delta across the 253 declared modules.
|
| 48 |
+
|
| 49 |
+
The effective-zero `lm_head` adapter is omitted from the release state and
|
| 50 |
+
target list; this is exactly equivalent to its validated zero delta and avoids
|
| 51 |
+
packaging any base-layer tensor. `MODULE_SCALE_MANIFEST.json` retains the
|
| 52 |
+
explicit zero-head contract and records every logical module, source tensor
|
| 53 |
+
key, physical base weight, and scale. This release is from the completed Phase
|
| 54 |
+
A delta-scaling line; it is **not** the later failed NEXTGEN route and does not
|
| 55 |
+
include subsequent protocol-repair experiments.
|
| 56 |
+
|
| 57 |
+
## Loading with PEFT
|
| 58 |
+
|
| 59 |
+
Use recent `transformers` and `peft` versions. Load the fixed base first, then
|
| 60 |
+
attach this adapter:
|
| 61 |
+
|
| 62 |
+
```python
|
| 63 |
+
from peft import PeftModel
|
| 64 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 65 |
+
|
| 66 |
+
base_id = "Qwen/Qwen3-4B-Base"
|
| 67 |
+
base_revision = "906bfd4b4dc7f14ee4320094d8b41684abff8539"
|
| 68 |
+
adapter_id = "modrill/Qwen3-4B-Base-ThinkCode-A-NH025"
|
| 69 |
+
|
| 70 |
+
tokenizer = AutoTokenizer.from_pretrained(base_id, revision=base_revision)
|
| 71 |
+
base = AutoModelForCausalLM.from_pretrained(
|
| 72 |
+
base_id,
|
| 73 |
+
revision=base_revision,
|
| 74 |
+
torch_dtype="auto",
|
| 75 |
+
device_map="auto",
|
| 76 |
+
)
|
| 77 |
+
model = PeftModel.from_pretrained(base, adapter_id)
|
| 78 |
+
|
| 79 |
+
messages = [{"role": "user", "content": "Write a Python function that checks whether a number is prime."}]
|
| 80 |
+
prompt = tokenizer.apply_chat_template(
|
| 81 |
+
messages,
|
| 82 |
+
tokenize=False,
|
| 83 |
+
add_generation_prompt=True,
|
| 84 |
+
enable_thinking=False,
|
| 85 |
+
)
|
| 86 |
+
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
| 87 |
+
eos_ids = [
|
| 88 |
+
tokenizer.eos_token_id,
|
| 89 |
+
tokenizer.convert_tokens_to_ids("<|im_end|>"),
|
| 90 |
+
]
|
| 91 |
+
outputs = model.generate(**inputs, max_new_tokens=2048, eos_token_id=eos_ids)
|
| 92 |
+
print(tokenizer.decode(outputs[0][inputs.input_ids.shape[-1]:], skip_special_tokens=True))
|
| 93 |
+
```
|
| 94 |
+
|
| 95 |
+
The base tokenizer's chat template supports `enable_thinking`. Disable it for
|
| 96 |
+
direct code generation matching the concise screening style, or enable it when
|
| 97 |
+
explicit reasoning is desired. Pass both `<|endoftext|>` and `<|im_end|>` as
|
| 98 |
+
EOS IDs. Keep the combined prompt and generated sequence within **32K tokens**,
|
| 99 |
+
the fixed base model configuration limit, unless a separate long-context
|
| 100 |
+
extension is validated.
|
| 101 |
+
|
| 102 |
+
## Development evaluation
|
| 103 |
+
|
| 104 |
+
Across three preregistered seeds on the corrected EvalScope Full1055
|
| 105 |
+
development suite, the resolved `code_only` aggregate was
|
| 106 |
+
**794/3165 = 25.09%**. Relative to the fixed BASE, the estimated change was
|
| 107 |
+
approximately **+0.98 percentage points**, with an approximate 95% confidence
|
| 108 |
+
interval of **[+0.095, +1.833] percentage points**. The Holm-adjusted
|
| 109 |
+
**p-value was 0.489**.
|
| 110 |
+
|
| 111 |
+
These results are development-only, not a held-out formal claim. In the
|
| 112 |
+
original bidirectional scoring for `seed=3407`, some outcomes flipped between
|
| 113 |
+
PASS and TLE because of the execution environment. Those cases were resolved
|
| 114 |
+
by fixed single-CPU serial rejudgment, which does not eliminate all scorer,
|
| 115 |
+
timing, or environment uncertainty.
|
| 116 |
+
|
| 117 |
+
## Limitations
|
| 118 |
+
|
| 119 |
+
- This adapter requires the exact base model and should not be loaded alone.
|
| 120 |
+
- The evidence is development-only and includes scorer-environment uncertainty.
|
| 121 |
+
- Generated code can be incorrect, insecure, or non-compiling; sandbox and
|
| 122 |
+
test it independently.
|
| 123 |
+
- No production safety, security, or suitability certification is implied.
|
| 124 |
+
|
| 125 |
+
## License
|
| 126 |
+
|
| 127 |
+
The fixed base card and included license identify Apache-2.0. This adapter
|
| 128 |
+
preserves that license text and metadata. Users should independently verify the
|
| 129 |
+
upstream Qwen3 license, notices, training-data terms, and applicability to their
|
| 130 |
+
use case.
|
| 131 |
+
|
VALIDATION.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"arm_id":"A-NH025","base_model":"Qwen/Qwen3-4B-Base","base_revision":"906bfd4b4dc7f14ee4320094d8b41684abff8539","fixed_prompt_fp32_quantization_isolated_logits":[{"max_abs":0.11317825317382812,"mean_abs":0.015323368832468987,"p99_abs":0.05210685729980469,"prompt_index":0,"top10_overlap":10,"top1_equal":true},{"max_abs":0.11573135852813721,"mean_abs":0.03259986639022827,"p99_abs":0.07636938244104385,"prompt_index":1,"top10_overlap":10,"top1_equal":true},{"max_abs":0.08328914642333984,"mean_abs":0.02721281535923481,"p99_abs":0.05772271752357483,"prompt_index":2,"top10_overlap":10,"top1_equal":true}],"logit_gate":{"max_abs_le":0.25,"mean_abs_le":0.05,"top10_overlap_min":9,"top1_all_equal":true},"phase_a_canary_status":"PASS","representative_bf16_weight_parity":[{"bf16_exact_samples":6,"logical_module":"model.layers.0.self_attn.q_proj","max_abs":0.0,"physical_weight":"model.layers.0.self_attn.q_proj.weight","sample_count":6},{"bf16_exact_samples":6,"logical_module":"model.layers.23.mlp.down_proj","max_abs":0.0,"physical_weight":"model.layers.23.mlp.down_proj.weight","sample_count":6},{"bf16_exact_samples":6,"logical_module":"model.layers.24.mlp.up_proj","max_abs":0.0,"physical_weight":"model.layers.24.mlp.up_proj.weight","sample_count":6},{"bf16_exact_samples":6,"logical_module":"model.layers.35.self_attn.o_proj","max_abs":0.0,"physical_weight":"model.layers.35.self_attn.o_proj.weight","sample_count":6},{"bf16_exact_samples":6,"logical_module":"lm_head","max_abs":0.0,"physical_weight":"model.embed_tokens.weight","sample_count":6}],"schema":"HF_PEFT_ADAPTER_VALIDATION_V1","status":"PASS","tensor_contract":{"base_layer_tensor_absent":true,"body_scale":0.25,"lm_head_scale":0.0,"peft_tied_adapter_storage_verified":false,"source_scaling_all_exact":true},"validated_at_utc":"2026-08-22T18:05:45Z"}
|
adapter_config.json
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3-4B-Base",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": false,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 128,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.0,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"monteclora_config": null,
|
| 27 |
+
"peft_type": "LORA",
|
| 28 |
+
"peft_version": "0.20.0",
|
| 29 |
+
"qalora_group_size": 16,
|
| 30 |
+
"r": 64,
|
| 31 |
+
"rank_pattern": {},
|
| 32 |
+
"revision": "906bfd4b4dc7f14ee4320094d8b41684abff8539",
|
| 33 |
+
"target_modules": [
|
| 34 |
+
"down_proj",
|
| 35 |
+
"gate_proj",
|
| 36 |
+
"k_proj",
|
| 37 |
+
"o_proj",
|
| 38 |
+
"q_proj",
|
| 39 |
+
"up_proj",
|
| 40 |
+
"v_proj"
|
| 41 |
+
],
|
| 42 |
+
"target_parameters": null,
|
| 43 |
+
"task_type": "CAUSAL_LM",
|
| 44 |
+
"trainable_token_indices": null,
|
| 45 |
+
"use_bdlora": null,
|
| 46 |
+
"use_dora": false,
|
| 47 |
+
"use_qalora": false,
|
| 48 |
+
"use_rslora": false,
|
| 49 |
+
"velora_config": null
|
| 50 |
+
}
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b0c2437fe29dc234871a52202774b76421bfeefdd6467c1770242e9d9e5349ce
|
| 3 |
+
size 528550328
|