ZhiyangQi97 commited on
Commit
9f553e1
·
verified ·
1 Parent(s): 341e9b1

Release merged BF16 CounselTrajLM based on Qwen3.5-9B

Browse files
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 2026 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.
README.md CHANGED
@@ -3,11 +3,12 @@ language:
3
  - zh
4
  base_model: Qwen/Qwen3.5-9B
5
  base_model_relation: finetune
6
- library_name: peft
 
7
  pipeline_tag: text-generation
8
  tags:
9
  - qwen3_5
10
- - lora
11
  - psychological-counseling
12
  - counselor-response-generation
13
  ---
@@ -18,10 +19,12 @@ CounselTrajLM generates the next Chinese counselor response from a counseling
18
  dialogue history. It is fine-tuned on CounselTrajD and is the dialogue model used
19
  in the CounselTraj response-level and dialogue-level experiments.
20
 
21
- This release contains the **LoRA adapter for
22
- [Qwen/Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B)**, its tokenizer, and a
23
- working inference script. The adapter requires the base model, pinned to revision
24
- `c202236235762e1c871ad0ccb60c8ee5ba337b9a`. The base model's license is Apache-2.0.
 
 
25
 
26
  Project, dataset, and evaluation prompts:
27
  [UEC-InabaLab/CounselTraj](https://github.com/UEC-InabaLab/CounselTraj).
@@ -36,6 +39,7 @@ checkpoint.
36
  | Item | Value |
37
  | --- | --- |
38
  | Base model | Qwen/Qwen3.5-9B, text language-model component |
 
39
  | Fine-tuning | QLoRA, NF4 with double quantization and bfloat16 computation |
40
  | Adapter rank / alpha / dropout | 32 / 64 / 0.05 |
41
  | Target modules | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
@@ -53,10 +57,11 @@ checkpoint.
53
  | Random seed | 20260704 |
54
  | Recorded final validation loss | 1.5481888056 |
55
 
56
- The released weights are the final adapter saved after three epochs. The trainer
57
- record contains no best-validation checkpoint selection. The adapter weights are
58
- byte-identical to the source final adapter; the release configuration replaces an
59
- internal base-model path with the official Hub ID and exact revision.
 
60
 
61
  ## Training data and references
62
 
@@ -74,7 +79,7 @@ for the source-specific data terms.
74
  ## Generate a reply
75
 
76
  Authenticate with an account that can access this private model, then download
77
- the adapter and run the included example:
78
 
79
  ```bash
80
  pip install huggingface_hub
@@ -82,19 +87,23 @@ hf auth login
82
  python -c 'from huggingface_hub import snapshot_download; snapshot_download("UEC-InabaLab/CounselTrajLM", local_dir="CounselTrajLM")'
83
  cd CounselTrajLM
84
  pip install -r requirements.txt
85
- python generate_reply.py --adapter . --input-file examples/histories.jsonl
86
  ```
87
 
88
  Or supply a history directly:
89
 
90
  ```bash
91
- python generate_reply.py --adapter . --history 'client:最近工作压力很大,晚上总是睡不好。'
92
  ```
93
 
94
  The script uses the original counselor system prompt and user-history format,
95
- with `enable_thinking=False`. It loads the text language-model component and
96
- adapter with NF4 quantization on CUDA. Use `--no-4bit` on supported hardware with
97
- enough memory to load the base model without 4-bit quantization.
 
 
 
 
98
 
99
  Input JSONL rows contain an optional `id` and a nonempty `history` string. The
100
  script outputs JSON with the generated `response`. The three bundled histories
@@ -127,7 +136,8 @@ provides the prompts, dimensions, and research foundations.
127
 
128
  ## Files
129
 
130
- - `adapter_model.safetensors` and `adapter_config.json`: final LoRA adapter.
 
131
  - `tokenizer.json`, `tokenizer_config.json`, and `chat_template.jinja`: saved tokenizer and chat template.
132
  - `generate_reply.py`: inference entry point.
133
  - `requirements.txt`: inference dependencies.
@@ -135,6 +145,7 @@ provides the prompts, dimensions, and research foundations.
135
  - `examples/histories.jsonl`: synthetic example inputs.
136
  - `validation.json`: release integrity and generation checks.
137
  - `SHA256SUMS`: file checksums.
 
138
 
139
  The model is intended for counseling dialogue research. Its generated responses
140
  have not been validated as clinical care or crisis support.
 
3
  - zh
4
  base_model: Qwen/Qwen3.5-9B
5
  base_model_relation: finetune
6
+ library_name: transformers
7
+ license: apache-2.0
8
  pipeline_tag: text-generation
9
  tags:
10
  - qwen3_5
11
+ - merged
12
  - psychological-counseling
13
  - counselor-response-generation
14
  ---
 
19
  dialogue history. It is fine-tuned on CounselTrajD and is the dialogue model used
20
  in the CounselTraj response-level and dialogue-level experiments.
21
 
22
+ This release contains the **complete merged BF16 text language model**, its
23
+ tokenizer, and a working inference script. It combines the previously released
24
+ three-epoch LoRA adapter with the unquantized
25
+ [Qwen/Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B) base at revision
26
+ `c202236235762e1c871ad0ccb60c8ee5ba337b9a` and loads directly with Transformers.
27
+ The model is released under Apache-2.0; see [LICENSE](LICENSE).
28
 
29
  Project, dataset, and evaluation prompts:
30
  [UEC-InabaLab/CounselTraj](https://github.com/UEC-InabaLab/CounselTraj).
 
39
  | Item | Value |
40
  | --- | --- |
41
  | Base model | Qwen/Qwen3.5-9B, text language-model component |
42
+ | Release format | Complete merged BF16 text language model |
43
  | Fine-tuning | QLoRA, NF4 with double quantization and bfloat16 computation |
44
  | Adapter rank / alpha / dropout | 32 / 64 / 0.05 |
45
  | Target modules | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
 
57
  | Random seed | 20260704 |
58
  | Recorded final validation loss | 1.5481888056 |
59
 
60
+ The merged model uses the final adapter saved after three epochs. Its SHA256 was
61
+ verified against both the source final adapter and the previously published
62
+ adapter. The trainer record contains no best-validation checkpoint selection.
63
+ The [original adapter release](https://huggingface.co/UEC-InabaLab/CounselTrajLM/tree/341e9b198c985ee75c982a43376015d806b15522)
64
+ remains available in the repository history.
65
 
66
  ## Training data and references
67
 
 
79
  ## Generate a reply
80
 
81
  Authenticate with an account that can access this private model, then download
82
+ the complete model and run the included example:
83
 
84
  ```bash
85
  pip install huggingface_hub
 
87
  python -c 'from huggingface_hub import snapshot_download; snapshot_download("UEC-InabaLab/CounselTrajLM", local_dir="CounselTrajLM")'
88
  cd CounselTrajLM
89
  pip install -r requirements.txt
90
+ python generate_reply.py --model . --input-file examples/histories.jsonl
91
  ```
92
 
93
  Or supply a history directly:
94
 
95
  ```bash
96
+ python generate_reply.py --model . --history 'client:最近工作压力很大,晚上总是睡不好。'
97
  ```
98
 
99
  The script uses the original counselor system prompt and user-history format,
100
+ with `enable_thinking=False`. It loads the complete text language model in BF16,
101
+ with additional memory required for the generation cache. A separate base-model
102
+ download and the PEFT library are no longer required.
103
+
104
+ The weights occupy approximately 17.9 GB. The exported model was reloaded on CUDA
105
+ and checked on the three bundled synthetic histories. The merge comparison and
106
+ file integrity results are recorded in [validation.json](validation.json).
107
 
108
  Input JSONL rows contain an optional `id` and a nonempty `history` string. The
109
  script outputs JSON with the generated `response`. The three bundled histories
 
136
 
137
  ## Files
138
 
139
+ - `model-*.safetensors`, `model.safetensors.index.json`, and `config.json`: complete merged text language model.
140
+ - `generation_config.json`: generation defaults.
141
  - `tokenizer.json`, `tokenizer_config.json`, and `chat_template.jinja`: saved tokenizer and chat template.
142
  - `generate_reply.py`: inference entry point.
143
  - `requirements.txt`: inference dependencies.
 
145
  - `examples/histories.jsonl`: synthetic example inputs.
146
  - `validation.json`: release integrity and generation checks.
147
  - `SHA256SUMS`: file checksums.
148
+ - `LICENSE`: Apache-2.0 license.
149
 
150
  The model is intended for counseling dialogue research. Its generated responses
151
  have not been validated as clinical care or crisis support.
SHA256SUMS CHANGED
@@ -1,12 +1,24 @@
1
- 94382b38f0e757e6d632c8d99bf91c991cf216ceaa1b904b318871cb988bb465 README.md
2
- d8476967ec836156efd1b99d571c78ec5d57fde8eba2633d34f24ea8816ce212 adapter_config.json
3
- c4d89ce3b6cfdb53dba2927683331e2182c8767f3380527a78f0533b0f31e5bd adapter_model.safetensors
4
  a4aee8afcf2e0711942cf848899be66016f8d14a889ff9ede07bca099c28f715 chat_template.jinja
 
5
  72b038eb2dd906df9d8efaf5d637a844840fc2547c1a2f2dffa13644d488adac examples/histories.jsonl
6
- 553a41c7ace9b01e76c57a6e3fb789715f89dc42b8461d9de16bf7380d0fec93 examples/replies.jsonl
7
- 7ebe4f0cc7099835e699f6787ee14194fe2cee3dd38b57c9f541fab79c484aee generate_reply.py
8
- 4ad87e0e39e963397f1412a551c1fe04707512eff5dc7f030e1d9ebd7ba747c6 model_details.json
9
- 1bc723f464c8003c711d2eaa47664b028109882d48760fa76f5da12a81a41b2b requirements.txt
 
 
 
 
 
 
 
 
 
 
 
 
10
  f399b3cd12fa270d51457bb749fb30863521e8359b8a27059c71b6c2f7d6dd6c tokenizer.json
11
- bee8eba30f0eb4af73c0fe2cd06d0f89b657d7819941c438157ec42f7c80ea87 tokenizer_config.json
12
- 054f254cbe4a7b5ed11b18a83204542452783da029650e619e7e17c2f6792308 validation.json
 
1
+ bbedc3fda3305820b977265f01b8619d87570a6739de3a5582c3464840f1e57a LICENSE
2
+ 116941973bf068f46c4a7dd952883956c50e0a715dbe89498fabb4e87b39fb04 README.md
 
3
  a4aee8afcf2e0711942cf848899be66016f8d14a889ff9ede07bca099c28f715 chat_template.jinja
4
+ 77c837fe7366ccbd656a6cce7ba0f8fcdbb82c505975b9c758bb2e330423173e config.json
5
  72b038eb2dd906df9d8efaf5d637a844840fc2547c1a2f2dffa13644d488adac examples/histories.jsonl
6
+ cd650bc16ac1d4f4a16607d1c3a6345ae53aad1350d847ee96ce781dac816a5f examples/replies.jsonl
7
+ a3b86ea598c81ee3c42c41aa54469906a5700fbbfedb0b1c5ae509d94430c170 generate_reply.py
8
+ 8e29eeb222454c2a417b17547a84aee206c110a033418b7c0342cd75e75d92a0 generation_config.json
9
+ dd63614f1dc80dce2d83be3f3e69af1f8a0ebc8add9b8e0abf3910f563e44344 model-00001-of-00010.safetensors
10
+ 856617f65c8f2d2686c0aec3c6036c9bf53466ed6c90275bee6d0ecaa8263b30 model-00002-of-00010.safetensors
11
+ e956a1f8536baf45076db19dffb96f258a6d94fe0a545f2bb38fed9517ab19a0 model-00003-of-00010.safetensors
12
+ 847db00981824e02e112fa1f57526d1feebfa268d1e85c32c1e24b9d4641afc7 model-00004-of-00010.safetensors
13
+ 5c13aa2a1fe366710e1dcd1484901e7725afdde55b2c717a2a25797d93b75058 model-00005-of-00010.safetensors
14
+ 6c65097043a90de2d46d0d01d698ef1a32bde00a3baf458434ffbf0906254a31 model-00006-of-00010.safetensors
15
+ 7de8c454f2661fd3f6e6a7c0a46dc8e6915a34caf3fac961b895b306d64b89b9 model-00007-of-00010.safetensors
16
+ 293d1a770da768e17860c9915a55ea26d178067d118cfdbdc2c285577dc21b87 model-00008-of-00010.safetensors
17
+ 9d99b5f3e4bff23afe10799d21ca22eb6e866cbd17a78428e6e4da6fb8466211 model-00009-of-00010.safetensors
18
+ d420fe217cc622433d5ec2bb1620ea34eaf4a4561e3a55b69cbe8856b17705dc model-00010-of-00010.safetensors
19
+ 2e1e25b0bacb71b1a4d670f12ed93c8790619bf5bef77a5739a017b1e3940f95 model.safetensors.index.json
20
+ d733550091085c313f62d2af14323484d3183e78b3e1fdb41de91dcbe5232cd7 model_details.json
21
+ 338ccf69cc3501c68ac17595e7206fbc8c53ab700bd8f8ab051ccd1ed879786b requirements.txt
22
  f399b3cd12fa270d51457bb749fb30863521e8359b8a27059c71b6c2f7d6dd6c tokenizer.json
23
+ 91a08f825d370d085d692e04cf117cdd7faad7bf18e996f1e6031b6dab03db72 tokenizer_config.json
24
+ afae503c250b36b28b194f811664147dac18f38ae59a2e463c9c4eaca9269361 validation.json
adapter_config.json DELETED
@@ -1,48 +0,0 @@
1
- {
2
- "alora_invocation_tokens": null,
3
- "alpha_pattern": {},
4
- "arrow_config": null,
5
- "auto_mapping": null,
6
- "base_model_name_or_path": "Qwen/Qwen3.5-9B",
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": 64,
20
- "lora_bias": false,
21
- "lora_dropout": 0.05,
22
- "lora_ga_config": null,
23
- "megatron_config": null,
24
- "megatron_core": "megatron.core",
25
- "modules_to_save": null,
26
- "peft_type": "LORA",
27
- "peft_version": "0.19.1",
28
- "qalora_group_size": 16,
29
- "r": 32,
30
- "rank_pattern": {},
31
- "revision": "c202236235762e1c871ad0ccb60c8ee5ba337b9a",
32
- "target_modules": [
33
- "k_proj",
34
- "v_proj",
35
- "gate_proj",
36
- "q_proj",
37
- "o_proj",
38
- "up_proj",
39
- "down_proj"
40
- ],
41
- "target_parameters": null,
42
- "task_type": "CAUSAL_LM",
43
- "trainable_token_indices": null,
44
- "use_bdlora": null,
45
- "use_dora": false,
46
- "use_qalora": false,
47
- "use_rslora": false
48
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
config.json ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3_5ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "attn_output_gate": true,
8
+ "bos_token_id": null,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 248044,
11
+ "full_attention_interval": 4,
12
+ "head_dim": 256,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 4096,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 12288,
17
+ "layer_types": [
18
+ "linear_attention",
19
+ "linear_attention",
20
+ "linear_attention",
21
+ "full_attention",
22
+ "linear_attention",
23
+ "linear_attention",
24
+ "linear_attention",
25
+ "full_attention",
26
+ "linear_attention",
27
+ "linear_attention",
28
+ "linear_attention",
29
+ "full_attention",
30
+ "linear_attention",
31
+ "linear_attention",
32
+ "linear_attention",
33
+ "full_attention",
34
+ "linear_attention",
35
+ "linear_attention",
36
+ "linear_attention",
37
+ "full_attention",
38
+ "linear_attention",
39
+ "linear_attention",
40
+ "linear_attention",
41
+ "full_attention",
42
+ "linear_attention",
43
+ "linear_attention",
44
+ "linear_attention",
45
+ "full_attention",
46
+ "linear_attention",
47
+ "linear_attention",
48
+ "linear_attention",
49
+ "full_attention"
50
+ ],
51
+ "linear_conv_kernel_dim": 4,
52
+ "linear_key_head_dim": 128,
53
+ "linear_num_key_heads": 16,
54
+ "linear_num_value_heads": 32,
55
+ "linear_value_head_dim": 128,
56
+ "mamba_ssm_dtype": "float32",
57
+ "max_position_embeddings": 262144,
58
+ "mlp_only_layers": [],
59
+ "model_type": "qwen3_5_text",
60
+ "mtp_num_hidden_layers": 1,
61
+ "mtp_use_dedicated_embeddings": false,
62
+ "num_attention_heads": 16,
63
+ "num_hidden_layers": 32,
64
+ "num_key_value_heads": 4,
65
+ "pad_token_id": null,
66
+ "partial_rotary_factor": 0.25,
67
+ "rms_norm_eps": 1e-06,
68
+ "rope_parameters": {
69
+ "mrope_interleaved": true,
70
+ "mrope_section": [
71
+ 11,
72
+ 11,
73
+ 10
74
+ ],
75
+ "partial_rotary_factor": 0.25,
76
+ "rope_theta": 10000000,
77
+ "rope_type": "default"
78
+ },
79
+ "tie_word_embeddings": false,
80
+ "transformers_version": "5.13.0",
81
+ "use_cache": true,
82
+ "vocab_size": 248320
83
+ }
examples/replies.jsonl CHANGED
@@ -1,3 +1,3 @@
1
- {"id": "synthetic-anxiety", "history": "client:最近快考试了,我总担心考不好,晚上也睡不着。\ncounselor:考试的压力似乎让你一直很紧张。\nclient:是的,一躺下就忍不住想最坏的结果。", "response": "你能把这种“越想睡越清醒”的过程说出来,已经是在认真观察自己的状态了。"}
2
- {"id": "synthetic-interpersonal", "history": "client:最近朋友很少主动联系我,我总觉得自己不重要。\ncounselor:你似乎很在意自己在这段友谊中的位置。\nclient:嗯,但我也不敢问,怕别人觉得我太黏人。", "response": "你能把这种担心说出来,其实已经很不容易了。你不是简单地“想太多”,而是在努力理解关系里发生了什么。"}
3
- {"id": "synthetic-work", "history": "client:刚换了一份工作,我觉得自己什么都做不好。\ncounselor:面对新的工作要求,你好像对自己有很多怀疑。\nclient:是的,每次遇到不会的事情,我就觉得他们一定后悔招了我。", "response": "你能把这种“被否定”的感觉说出来,其实很不容易,尤其是在刚进入新环境的时候。"}
 
1
+ {"id": "synthetic-anxiety", "history": "client:最近快考试了,我总担心考不好,晚上也睡不着。\ncounselor:考试的压力似乎让你一直很紧张。\nclient:是的,一躺下就忍不住想最坏的结果。", "response": "你能把这种“最坏结果”的想法说出来,已经是在尝试靠近它,而不是完全被它带着走。"}
2
+ {"id": "synthetic-interpersonal", "history": "client:最近朋友很少主动联系我,我总觉得自己不重要。\ncounselor:你似乎很在意自己在这段友谊中的位置。\nclient:嗯,但我也不敢问,怕别人觉得我太黏人。", "response": "你能觉察到“想靠近”和“怕被嫌弃”同时存在,这其实很不容易。"}
3
+ {"id": "synthetic-work", "history": "client:刚换了一份工作,我觉得自己什么都做不好。\ncounselor:面对新的工作要求,你好像对自己有很多怀疑。\nclient:是的,每次遇到不会的事情,我就觉得他们一定后悔招了我。", "response": "你能把这种“刚进入新环境时的不确定”说出来,其实已经是在认真面对它了。"}
generate_reply.py CHANGED
@@ -1,17 +1,14 @@
1
- """Generate the next Chinese counselor reply with the CounselTrajLM adapter."""
2
 
3
  import argparse
4
  import json
5
  from pathlib import Path
6
 
7
  import torch
8
- from peft import PeftModel
9
- from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
10
 
11
 
12
  REPO_ID = "UEC-InabaLab/CounselTrajLM"
13
- BASE_MODEL = "Qwen/Qwen3.5-9B"
14
- BASE_REVISION = "c202236235762e1c871ad0ccb60c8ee5ba337b9a"
15
  SYSTEM_PROMPT = (
16
  "你是一名专业、温和、自然的线上心理咨询师。"
17
  "请根据对话历史生成下一句咨询师回复。"
@@ -35,8 +32,8 @@ def build_prompt(tokenizer, history):
35
  )
36
 
37
 
38
- def load_model(adapter, base_model=BASE_MODEL, local_files_only=False, use_4bit=True):
39
- tokenizer = AutoTokenizer.from_pretrained(adapter, local_files_only=local_files_only)
40
  if tokenizer.pad_token_id is None:
41
  tokenizer.pad_token = tokenizer.eos_token
42
  kwargs = {
@@ -45,17 +42,7 @@ def load_model(adapter, base_model=BASE_MODEL, local_files_only=False, use_4bit=
45
  "attn_implementation": "eager",
46
  "local_files_only": local_files_only,
47
  }
48
- if base_model == BASE_MODEL:
49
- kwargs["revision"] = BASE_REVISION
50
- if use_4bit:
51
- kwargs["quantization_config"] = BitsAndBytesConfig(
52
- load_in_4bit=True,
53
- bnb_4bit_use_double_quant=True,
54
- bnb_4bit_quant_type="nf4",
55
- bnb_4bit_compute_dtype=torch.bfloat16,
56
- )
57
- base = AutoModelForCausalLM.from_pretrained(base_model, **kwargs)
58
- model = PeftModel.from_pretrained(base, adapter, local_files_only=local_files_only)
59
  model.eval()
60
  return tokenizer, model
61
 
@@ -85,14 +72,12 @@ def generate_reply(tokenizer, model, history, max_new_tokens=256):
85
 
86
  def main():
87
  parser = argparse.ArgumentParser(description=__doc__)
88
- parser.add_argument("--adapter", default=REPO_ID)
89
- parser.add_argument("--base-model", default=BASE_MODEL)
90
  parser.add_argument("--history")
91
  parser.add_argument("--input-file", type=Path)
92
  parser.add_argument("--output-file", type=Path)
93
  parser.add_argument("--max-new-tokens", type=int, default=256)
94
  parser.add_argument("--local-files-only", action="store_true")
95
- parser.add_argument("--no-4bit", action="store_true")
96
  args = parser.parse_args()
97
  if bool(args.history) == bool(args.input_file):
98
  parser.error("Provide exactly one of --history or --input-file")
@@ -104,10 +89,8 @@ def main():
104
  if not rows or any(not isinstance(row.get("history"), str) or not row["history"].strip() for row in rows):
105
  parser.error("Every input row must contain a nonempty history string")
106
  tokenizer, model = load_model(
107
- args.adapter,
108
- base_model=args.base_model,
109
  local_files_only=args.local_files_only,
110
- use_4bit=not args.no_4bit,
111
  )
112
  results = []
113
  for row in rows:
 
1
+ """Generate the next Chinese counselor reply with the complete CounselTrajLM model."""
2
 
3
  import argparse
4
  import json
5
  from pathlib import Path
6
 
7
  import torch
8
+ from transformers import AutoModelForCausalLM, AutoTokenizer
 
9
 
10
 
11
  REPO_ID = "UEC-InabaLab/CounselTrajLM"
 
 
12
  SYSTEM_PROMPT = (
13
  "你是一名专业、温和、自然的线上心理咨询师。"
14
  "请根据对话历史生成下一句咨询师回复。"
 
32
  )
33
 
34
 
35
+ def load_model(model_id=REPO_ID, local_files_only=False):
36
+ tokenizer = AutoTokenizer.from_pretrained(model_id, local_files_only=local_files_only)
37
  if tokenizer.pad_token_id is None:
38
  tokenizer.pad_token = tokenizer.eos_token
39
  kwargs = {
 
42
  "attn_implementation": "eager",
43
  "local_files_only": local_files_only,
44
  }
45
+ model = AutoModelForCausalLM.from_pretrained(model_id, **kwargs)
 
 
 
 
 
 
 
 
 
 
46
  model.eval()
47
  return tokenizer, model
48
 
 
72
 
73
  def main():
74
  parser = argparse.ArgumentParser(description=__doc__)
75
+ parser.add_argument("--model", "--adapter", dest="model", default=REPO_ID)
 
76
  parser.add_argument("--history")
77
  parser.add_argument("--input-file", type=Path)
78
  parser.add_argument("--output-file", type=Path)
79
  parser.add_argument("--max-new-tokens", type=int, default=256)
80
  parser.add_argument("--local-files-only", action="store_true")
 
81
  args = parser.parse_args()
82
  if bool(args.history) == bool(args.input_file):
83
  parser.error("Provide exactly one of --history or --input-file")
 
89
  if not rows or any(not isinstance(row.get("history"), str) or not row["history"].strip() for row in rows):
90
  parser.error("Every input row must contain a nonempty history string")
91
  tokenizer, model = load_model(
92
+ args.model,
 
93
  local_files_only=args.local_files_only,
 
94
  )
95
  results = []
96
  for row in rows:
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "eos_token_id": 248044,
4
+ "transformers_version": "5.13.0",
5
+ "use_cache": true
6
+ }
model-00001-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd63614f1dc80dce2d83be3f3e69af1f8a0ebc8add9b8e0abf3910f563e44344
3
+ size 2034237568
model-00002-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:856617f65c8f2d2686c0aec3c6036c9bf53466ed6c90275bee6d0ecaa8263b30
3
+ size 2034237592
model-00003-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e956a1f8536baf45076db19dffb96f258a6d94fe0a545f2bb38fed9517ab19a0
3
+ size 1965378120
model-00004-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:847db00981824e02e112fa1f57526d1feebfa268d1e85c32c1e24b9d4641afc7
3
+ size 1998940424
model-00005-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5c13aa2a1fe366710e1dcd1484901e7725afdde55b2c717a2a25797d93b75058
3
+ size 1998333688
model-00006-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c65097043a90de2d46d0d01d698ef1a32bde00a3baf458434ffbf0906254a31
3
+ size 1965386568
model-00007-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7de8c454f2661fd3f6e6a7c0a46dc8e6915a34caf3fac961b895b306d64b89b9
3
+ size 1931224240
model-00008-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:293d1a770da768e17860c9915a55ea26d178067d118cfdbdc2c285577dc21b87
3
+ size 1931240872
model-00009-of-00010.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d99b5f3e4bff23afe10799d21ca22eb6e866cbd17a78428e6e4da6fb8466211
3
+ size 1973177136
adapter_model.safetensors → model-00010-of-00010.safetensors RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c4d89ce3b6cfdb53dba2927683331e2182c8767f3380527a78f0533b0f31e5bd
3
- size 232818064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d420fe217cc622433d5ec2bb1620ea34eaf4a4561e3a55b69cbe8856b17705dc
3
+ size 75506064
model.safetensors.index.json ADDED
@@ -0,0 +1,435 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_parameters": 8953803264,
4
+ "total_size": 17907606528
5
+ },
6
+ "weight_map": {
7
+ "lm_head.weight": "model-00001-of-00010.safetensors",
8
+ "model.language_model.embed_tokens.weight": "model-00002-of-00010.safetensors",
9
+ "model.language_model.layers.0.input_layernorm.weight": "model-00003-of-00010.safetensors",
10
+ "model.language_model.layers.0.linear_attn.A_log": "model-00003-of-00010.safetensors",
11
+ "model.language_model.layers.0.linear_attn.conv1d.weight": "model-00003-of-00010.safetensors",
12
+ "model.language_model.layers.0.linear_attn.dt_bias": "model-00003-of-00010.safetensors",
13
+ "model.language_model.layers.0.linear_attn.in_proj_a.weight": "model-00003-of-00010.safetensors",
14
+ "model.language_model.layers.0.linear_attn.in_proj_b.weight": "model-00003-of-00010.safetensors",
15
+ "model.language_model.layers.0.linear_attn.in_proj_qkv.weight": "model-00003-of-00010.safetensors",
16
+ "model.language_model.layers.0.linear_attn.in_proj_z.weight": "model-00003-of-00010.safetensors",
17
+ "model.language_model.layers.0.linear_attn.norm.weight": "model-00003-of-00010.safetensors",
18
+ "model.language_model.layers.0.linear_attn.out_proj.weight": "model-00003-of-00010.safetensors",
19
+ "model.language_model.layers.0.mlp.down_proj.weight": "model-00003-of-00010.safetensors",
20
+ "model.language_model.layers.0.mlp.gate_proj.weight": "model-00003-of-00010.safetensors",
21
+ "model.language_model.layers.0.mlp.up_proj.weight": "model-00003-of-00010.safetensors",
22
+ "model.language_model.layers.0.post_attention_layernorm.weight": "model-00003-of-00010.safetensors",
23
+ "model.language_model.layers.1.input_layernorm.weight": "model-00003-of-00010.safetensors",
24
+ "model.language_model.layers.1.linear_attn.A_log": "model-00003-of-00010.safetensors",
25
+ "model.language_model.layers.1.linear_attn.conv1d.weight": "model-00003-of-00010.safetensors",
26
+ "model.language_model.layers.1.linear_attn.dt_bias": "model-00003-of-00010.safetensors",
27
+ "model.language_model.layers.1.linear_attn.in_proj_a.weight": "model-00003-of-00010.safetensors",
28
+ "model.language_model.layers.1.linear_attn.in_proj_b.weight": "model-00003-of-00010.safetensors",
29
+ "model.language_model.layers.1.linear_attn.in_proj_qkv.weight": "model-00003-of-00010.safetensors",
30
+ "model.language_model.layers.1.linear_attn.in_proj_z.weight": "model-00003-of-00010.safetensors",
31
+ "model.language_model.layers.1.linear_attn.norm.weight": "model-00003-of-00010.safetensors",
32
+ "model.language_model.layers.1.linear_attn.out_proj.weight": "model-00003-of-00010.safetensors",
33
+ "model.language_model.layers.1.mlp.down_proj.weight": "model-00003-of-00010.safetensors",
34
+ "model.language_model.layers.1.mlp.gate_proj.weight": "model-00003-of-00010.safetensors",
35
+ "model.language_model.layers.1.mlp.up_proj.weight": "model-00003-of-00010.safetensors",
36
+ "model.language_model.layers.1.post_attention_layernorm.weight": "model-00003-of-00010.safetensors",
37
+ "model.language_model.layers.10.input_layernorm.weight": "model-00005-of-00010.safetensors",
38
+ "model.language_model.layers.10.linear_attn.A_log": "model-00005-of-00010.safetensors",
39
+ "model.language_model.layers.10.linear_attn.conv1d.weight": "model-00005-of-00010.safetensors",
40
+ "model.language_model.layers.10.linear_attn.dt_bias": "model-00005-of-00010.safetensors",
41
+ "model.language_model.layers.10.linear_attn.in_proj_a.weight": "model-00005-of-00010.safetensors",
42
+ "model.language_model.layers.10.linear_attn.in_proj_b.weight": "model-00005-of-00010.safetensors",
43
+ "model.language_model.layers.10.linear_attn.in_proj_qkv.weight": "model-00005-of-00010.safetensors",
44
+ "model.language_model.layers.10.linear_attn.in_proj_z.weight": "model-00005-of-00010.safetensors",
45
+ "model.language_model.layers.10.linear_attn.norm.weight": "model-00005-of-00010.safetensors",
46
+ "model.language_model.layers.10.linear_attn.out_proj.weight": "model-00005-of-00010.safetensors",
47
+ "model.language_model.layers.10.mlp.down_proj.weight": "model-00005-of-00010.safetensors",
48
+ "model.language_model.layers.10.mlp.gate_proj.weight": "model-00005-of-00010.safetensors",
49
+ "model.language_model.layers.10.mlp.up_proj.weight": "model-00005-of-00010.safetensors",
50
+ "model.language_model.layers.10.post_attention_layernorm.weight": "model-00005-of-00010.safetensors",
51
+ "model.language_model.layers.11.input_layernorm.weight": "model-00005-of-00010.safetensors",
52
+ "model.language_model.layers.11.mlp.down_proj.weight": "model-00005-of-00010.safetensors",
53
+ "model.language_model.layers.11.mlp.gate_proj.weight": "model-00005-of-00010.safetensors",
54
+ "model.language_model.layers.11.mlp.up_proj.weight": "model-00005-of-00010.safetensors",
55
+ "model.language_model.layers.11.post_attention_layernorm.weight": "model-00005-of-00010.safetensors",
56
+ "model.language_model.layers.11.self_attn.k_norm.weight": "model-00005-of-00010.safetensors",
57
+ "model.language_model.layers.11.self_attn.k_proj.weight": "model-00005-of-00010.safetensors",
58
+ "model.language_model.layers.11.self_attn.o_proj.weight": "model-00005-of-00010.safetensors",
59
+ "model.language_model.layers.11.self_attn.q_norm.weight": "model-00005-of-00010.safetensors",
60
+ "model.language_model.layers.11.self_attn.q_proj.weight": "model-00005-of-00010.safetensors",
61
+ "model.language_model.layers.11.self_attn.v_proj.weight": "model-00005-of-00010.safetensors",
62
+ "model.language_model.layers.12.input_layernorm.weight": "model-00005-of-00010.safetensors",
63
+ "model.language_model.layers.12.linear_attn.A_log": "model-00005-of-00010.safetensors",
64
+ "model.language_model.layers.12.linear_attn.conv1d.weight": "model-00005-of-00010.safetensors",
65
+ "model.language_model.layers.12.linear_attn.dt_bias": "model-00005-of-00010.safetensors",
66
+ "model.language_model.layers.12.linear_attn.in_proj_a.weight": "model-00005-of-00010.safetensors",
67
+ "model.language_model.layers.12.linear_attn.in_proj_b.weight": "model-00005-of-00010.safetensors",
68
+ "model.language_model.layers.12.linear_attn.in_proj_qkv.weight": "model-00005-of-00010.safetensors",
69
+ "model.language_model.layers.12.linear_attn.in_proj_z.weight": "model-00005-of-00010.safetensors",
70
+ "model.language_model.layers.12.linear_attn.norm.weight": "model-00005-of-00010.safetensors",
71
+ "model.language_model.layers.12.linear_attn.out_proj.weight": "model-00005-of-00010.safetensors",
72
+ "model.language_model.layers.12.mlp.down_proj.weight": "model-00005-of-00010.safetensors",
73
+ "model.language_model.layers.12.mlp.gate_proj.weight": "model-00005-of-00010.safetensors",
74
+ "model.language_model.layers.12.mlp.up_proj.weight": "model-00005-of-00010.safetensors",
75
+ "model.language_model.layers.12.post_attention_layernorm.weight": "model-00005-of-00010.safetensors",
76
+ "model.language_model.layers.13.input_layernorm.weight": "model-00005-of-00010.safetensors",
77
+ "model.language_model.layers.13.linear_attn.A_log": "model-00005-of-00010.safetensors",
78
+ "model.language_model.layers.13.linear_attn.conv1d.weight": "model-00005-of-00010.safetensors",
79
+ "model.language_model.layers.13.linear_attn.dt_bias": "model-00005-of-00010.safetensors",
80
+ "model.language_model.layers.13.linear_attn.in_proj_a.weight": "model-00005-of-00010.safetensors",
81
+ "model.language_model.layers.13.linear_attn.in_proj_b.weight": "model-00005-of-00010.safetensors",
82
+ "model.language_model.layers.13.linear_attn.in_proj_qkv.weight": "model-00005-of-00010.safetensors",
83
+ "model.language_model.layers.13.linear_attn.in_proj_z.weight": "model-00005-of-00010.safetensors",
84
+ "model.language_model.layers.13.linear_attn.norm.weight": "model-00005-of-00010.safetensors",
85
+ "model.language_model.layers.13.linear_attn.out_proj.weight": "model-00005-of-00010.safetensors",
86
+ "model.language_model.layers.13.mlp.down_proj.weight": "model-00005-of-00010.safetensors",
87
+ "model.language_model.layers.13.mlp.gate_proj.weight": "model-00005-of-00010.safetensors",
88
+ "model.language_model.layers.13.mlp.up_proj.weight": "model-00006-of-00010.safetensors",
89
+ "model.language_model.layers.13.post_attention_layernorm.weight": "model-00006-of-00010.safetensors",
90
+ "model.language_model.layers.14.input_layernorm.weight": "model-00006-of-00010.safetensors",
91
+ "model.language_model.layers.14.linear_attn.A_log": "model-00006-of-00010.safetensors",
92
+ "model.language_model.layers.14.linear_attn.conv1d.weight": "model-00006-of-00010.safetensors",
93
+ "model.language_model.layers.14.linear_attn.dt_bias": "model-00006-of-00010.safetensors",
94
+ "model.language_model.layers.14.linear_attn.in_proj_a.weight": "model-00006-of-00010.safetensors",
95
+ "model.language_model.layers.14.linear_attn.in_proj_b.weight": "model-00006-of-00010.safetensors",
96
+ "model.language_model.layers.14.linear_attn.in_proj_qkv.weight": "model-00006-of-00010.safetensors",
97
+ "model.language_model.layers.14.linear_attn.in_proj_z.weight": "model-00006-of-00010.safetensors",
98
+ "model.language_model.layers.14.linear_attn.norm.weight": "model-00006-of-00010.safetensors",
99
+ "model.language_model.layers.14.linear_attn.out_proj.weight": "model-00006-of-00010.safetensors",
100
+ "model.language_model.layers.14.mlp.down_proj.weight": "model-00006-of-00010.safetensors",
101
+ "model.language_model.layers.14.mlp.gate_proj.weight": "model-00006-of-00010.safetensors",
102
+ "model.language_model.layers.14.mlp.up_proj.weight": "model-00006-of-00010.safetensors",
103
+ "model.language_model.layers.14.post_attention_layernorm.weight": "model-00006-of-00010.safetensors",
104
+ "model.language_model.layers.15.input_layernorm.weight": "model-00006-of-00010.safetensors",
105
+ "model.language_model.layers.15.mlp.down_proj.weight": "model-00006-of-00010.safetensors",
106
+ "model.language_model.layers.15.mlp.gate_proj.weight": "model-00006-of-00010.safetensors",
107
+ "model.language_model.layers.15.mlp.up_proj.weight": "model-00006-of-00010.safetensors",
108
+ "model.language_model.layers.15.post_attention_layernorm.weight": "model-00006-of-00010.safetensors",
109
+ "model.language_model.layers.15.self_attn.k_norm.weight": "model-00006-of-00010.safetensors",
110
+ "model.language_model.layers.15.self_attn.k_proj.weight": "model-00006-of-00010.safetensors",
111
+ "model.language_model.layers.15.self_attn.o_proj.weight": "model-00006-of-00010.safetensors",
112
+ "model.language_model.layers.15.self_attn.q_norm.weight": "model-00006-of-00010.safetensors",
113
+ "model.language_model.layers.15.self_attn.q_proj.weight": "model-00006-of-00010.safetensors",
114
+ "model.language_model.layers.15.self_attn.v_proj.weight": "model-00006-of-00010.safetensors",
115
+ "model.language_model.layers.16.input_layernorm.weight": "model-00006-of-00010.safetensors",
116
+ "model.language_model.layers.16.linear_attn.A_log": "model-00006-of-00010.safetensors",
117
+ "model.language_model.layers.16.linear_attn.conv1d.weight": "model-00006-of-00010.safetensors",
118
+ "model.language_model.layers.16.linear_attn.dt_bias": "model-00006-of-00010.safetensors",
119
+ "model.language_model.layers.16.linear_attn.in_proj_a.weight": "model-00006-of-00010.safetensors",
120
+ "model.language_model.layers.16.linear_attn.in_proj_b.weight": "model-00006-of-00010.safetensors",
121
+ "model.language_model.layers.16.linear_attn.in_proj_qkv.weight": "model-00006-of-00010.safetensors",
122
+ "model.language_model.layers.16.linear_attn.in_proj_z.weight": "model-00006-of-00010.safetensors",
123
+ "model.language_model.layers.16.linear_attn.norm.weight": "model-00006-of-00010.safetensors",
124
+ "model.language_model.layers.16.linear_attn.out_proj.weight": "model-00006-of-00010.safetensors",
125
+ "model.language_model.layers.16.mlp.down_proj.weight": "model-00006-of-00010.safetensors",
126
+ "model.language_model.layers.16.mlp.gate_proj.weight": "model-00006-of-00010.safetensors",
127
+ "model.language_model.layers.16.mlp.up_proj.weight": "model-00006-of-00010.safetensors",
128
+ "model.language_model.layers.16.post_attention_layernorm.weight": "model-00006-of-00010.safetensors",
129
+ "model.language_model.layers.17.input_layernorm.weight": "model-00006-of-00010.safetensors",
130
+ "model.language_model.layers.17.linear_attn.A_log": "model-00006-of-00010.safetensors",
131
+ "model.language_model.layers.17.linear_attn.conv1d.weight": "model-00006-of-00010.safetensors",
132
+ "model.language_model.layers.17.linear_attn.dt_bias": "model-00006-of-00010.safetensors",
133
+ "model.language_model.layers.17.linear_attn.in_proj_a.weight": "model-00006-of-00010.safetensors",
134
+ "model.language_model.layers.17.linear_attn.in_proj_b.weight": "model-00006-of-00010.safetensors",
135
+ "model.language_model.layers.17.linear_attn.in_proj_qkv.weight": "model-00006-of-00010.safetensors",
136
+ "model.language_model.layers.17.linear_attn.in_proj_z.weight": "model-00006-of-00010.safetensors",
137
+ "model.language_model.layers.17.linear_attn.norm.weight": "model-00006-of-00010.safetensors",
138
+ "model.language_model.layers.17.linear_attn.out_proj.weight": "model-00006-of-00010.safetensors",
139
+ "model.language_model.layers.17.mlp.down_proj.weight": "model-00006-of-00010.safetensors",
140
+ "model.language_model.layers.17.mlp.gate_proj.weight": "model-00006-of-00010.safetensors",
141
+ "model.language_model.layers.17.mlp.up_proj.weight": "model-00006-of-00010.safetensors",
142
+ "model.language_model.layers.17.post_attention_layernorm.weight": "model-00006-of-00010.safetensors",
143
+ "model.language_model.layers.18.input_layernorm.weight": "model-00006-of-00010.safetensors",
144
+ "model.language_model.layers.18.linear_attn.A_log": "model-00006-of-00010.safetensors",
145
+ "model.language_model.layers.18.linear_attn.conv1d.weight": "model-00006-of-00010.safetensors",
146
+ "model.language_model.layers.18.linear_attn.dt_bias": "model-00006-of-00010.safetensors",
147
+ "model.language_model.layers.18.linear_attn.in_proj_a.weight": "model-00006-of-00010.safetensors",
148
+ "model.language_model.layers.18.linear_attn.in_proj_b.weight": "model-00006-of-00010.safetensors",
149
+ "model.language_model.layers.18.linear_attn.in_proj_qkv.weight": "model-00006-of-00010.safetensors",
150
+ "model.language_model.layers.18.linear_attn.in_proj_z.weight": "model-00006-of-00010.safetensors",
151
+ "model.language_model.layers.18.linear_attn.norm.weight": "model-00006-of-00010.safetensors",
152
+ "model.language_model.layers.18.linear_attn.out_proj.weight": "model-00006-of-00010.safetensors",
153
+ "model.language_model.layers.18.mlp.down_proj.weight": "model-00007-of-00010.safetensors",
154
+ "model.language_model.layers.18.mlp.gate_proj.weight": "model-00007-of-00010.safetensors",
155
+ "model.language_model.layers.18.mlp.up_proj.weight": "model-00007-of-00010.safetensors",
156
+ "model.language_model.layers.18.post_attention_layernorm.weight": "model-00007-of-00010.safetensors",
157
+ "model.language_model.layers.19.input_layernorm.weight": "model-00007-of-00010.safetensors",
158
+ "model.language_model.layers.19.mlp.down_proj.weight": "model-00007-of-00010.safetensors",
159
+ "model.language_model.layers.19.mlp.gate_proj.weight": "model-00007-of-00010.safetensors",
160
+ "model.language_model.layers.19.mlp.up_proj.weight": "model-00007-of-00010.safetensors",
161
+ "model.language_model.layers.19.post_attention_layernorm.weight": "model-00007-of-00010.safetensors",
162
+ "model.language_model.layers.19.self_attn.k_norm.weight": "model-00007-of-00010.safetensors",
163
+ "model.language_model.layers.19.self_attn.k_proj.weight": "model-00007-of-00010.safetensors",
164
+ "model.language_model.layers.19.self_attn.o_proj.weight": "model-00007-of-00010.safetensors",
165
+ "model.language_model.layers.19.self_attn.q_norm.weight": "model-00007-of-00010.safetensors",
166
+ "model.language_model.layers.19.self_attn.q_proj.weight": "model-00007-of-00010.safetensors",
167
+ "model.language_model.layers.19.self_attn.v_proj.weight": "model-00007-of-00010.safetensors",
168
+ "model.language_model.layers.2.input_layernorm.weight": "model-00003-of-00010.safetensors",
169
+ "model.language_model.layers.2.linear_attn.A_log": "model-00003-of-00010.safetensors",
170
+ "model.language_model.layers.2.linear_attn.conv1d.weight": "model-00003-of-00010.safetensors",
171
+ "model.language_model.layers.2.linear_attn.dt_bias": "model-00003-of-00010.safetensors",
172
+ "model.language_model.layers.2.linear_attn.in_proj_a.weight": "model-00003-of-00010.safetensors",
173
+ "model.language_model.layers.2.linear_attn.in_proj_b.weight": "model-00003-of-00010.safetensors",
174
+ "model.language_model.layers.2.linear_attn.in_proj_qkv.weight": "model-00003-of-00010.safetensors",
175
+ "model.language_model.layers.2.linear_attn.in_proj_z.weight": "model-00003-of-00010.safetensors",
176
+ "model.language_model.layers.2.linear_attn.norm.weight": "model-00003-of-00010.safetensors",
177
+ "model.language_model.layers.2.linear_attn.out_proj.weight": "model-00003-of-00010.safetensors",
178
+ "model.language_model.layers.2.mlp.down_proj.weight": "model-00003-of-00010.safetensors",
179
+ "model.language_model.layers.2.mlp.gate_proj.weight": "model-00003-of-00010.safetensors",
180
+ "model.language_model.layers.2.mlp.up_proj.weight": "model-00003-of-00010.safetensors",
181
+ "model.language_model.layers.2.post_attention_layernorm.weight": "model-00003-of-00010.safetensors",
182
+ "model.language_model.layers.20.input_layernorm.weight": "model-00007-of-00010.safetensors",
183
+ "model.language_model.layers.20.linear_attn.A_log": "model-00007-of-00010.safetensors",
184
+ "model.language_model.layers.20.linear_attn.conv1d.weight": "model-00007-of-00010.safetensors",
185
+ "model.language_model.layers.20.linear_attn.dt_bias": "model-00007-of-00010.safetensors",
186
+ "model.language_model.layers.20.linear_attn.in_proj_a.weight": "model-00007-of-00010.safetensors",
187
+ "model.language_model.layers.20.linear_attn.in_proj_b.weight": "model-00007-of-00010.safetensors",
188
+ "model.language_model.layers.20.linear_attn.in_proj_qkv.weight": "model-00007-of-00010.safetensors",
189
+ "model.language_model.layers.20.linear_attn.in_proj_z.weight": "model-00007-of-00010.safetensors",
190
+ "model.language_model.layers.20.linear_attn.norm.weight": "model-00007-of-00010.safetensors",
191
+ "model.language_model.layers.20.linear_attn.out_proj.weight": "model-00007-of-00010.safetensors",
192
+ "model.language_model.layers.20.mlp.down_proj.weight": "model-00007-of-00010.safetensors",
193
+ "model.language_model.layers.20.mlp.gate_proj.weight": "model-00007-of-00010.safetensors",
194
+ "model.language_model.layers.20.mlp.up_proj.weight": "model-00007-of-00010.safetensors",
195
+ "model.language_model.layers.20.post_attention_layernorm.weight": "model-00007-of-00010.safetensors",
196
+ "model.language_model.layers.21.input_layernorm.weight": "model-00007-of-00010.safetensors",
197
+ "model.language_model.layers.21.linear_attn.A_log": "model-00007-of-00010.safetensors",
198
+ "model.language_model.layers.21.linear_attn.conv1d.weight": "model-00007-of-00010.safetensors",
199
+ "model.language_model.layers.21.linear_attn.dt_bias": "model-00007-of-00010.safetensors",
200
+ "model.language_model.layers.21.linear_attn.in_proj_a.weight": "model-00007-of-00010.safetensors",
201
+ "model.language_model.layers.21.linear_attn.in_proj_b.weight": "model-00007-of-00010.safetensors",
202
+ "model.language_model.layers.21.linear_attn.in_proj_qkv.weight": "model-00007-of-00010.safetensors",
203
+ "model.language_model.layers.21.linear_attn.in_proj_z.weight": "model-00007-of-00010.safetensors",
204
+ "model.language_model.layers.21.linear_attn.norm.weight": "model-00007-of-00010.safetensors",
205
+ "model.language_model.layers.21.linear_attn.out_proj.weight": "model-00007-of-00010.safetensors",
206
+ "model.language_model.layers.21.mlp.down_proj.weight": "model-00007-of-00010.safetensors",
207
+ "model.language_model.layers.21.mlp.gate_proj.weight": "model-00007-of-00010.safetensors",
208
+ "model.language_model.layers.21.mlp.up_proj.weight": "model-00007-of-00010.safetensors",
209
+ "model.language_model.layers.21.post_attention_layernorm.weight": "model-00007-of-00010.safetensors",
210
+ "model.language_model.layers.22.input_layernorm.weight": "model-00007-of-00010.safetensors",
211
+ "model.language_model.layers.22.linear_attn.A_log": "model-00007-of-00010.safetensors",
212
+ "model.language_model.layers.22.linear_attn.conv1d.weight": "model-00007-of-00010.safetensors",
213
+ "model.language_model.layers.22.linear_attn.dt_bias": "model-00007-of-00010.safetensors",
214
+ "model.language_model.layers.22.linear_attn.in_proj_a.weight": "model-00007-of-00010.safetensors",
215
+ "model.language_model.layers.22.linear_attn.in_proj_b.weight": "model-00007-of-00010.safetensors",
216
+ "model.language_model.layers.22.linear_attn.in_proj_qkv.weight": "model-00007-of-00010.safetensors",
217
+ "model.language_model.layers.22.linear_attn.in_proj_z.weight": "model-00007-of-00010.safetensors",
218
+ "model.language_model.layers.22.linear_attn.norm.weight": "model-00007-of-00010.safetensors",
219
+ "model.language_model.layers.22.linear_attn.out_proj.weight": "model-00007-of-00010.safetensors",
220
+ "model.language_model.layers.22.mlp.down_proj.weight": "model-00007-of-00010.safetensors",
221
+ "model.language_model.layers.22.mlp.gate_proj.weight": "model-00007-of-00010.safetensors",
222
+ "model.language_model.layers.22.mlp.up_proj.weight": "model-00008-of-00010.safetensors",
223
+ "model.language_model.layers.22.post_attention_layernorm.weight": "model-00008-of-00010.safetensors",
224
+ "model.language_model.layers.23.input_layernorm.weight": "model-00008-of-00010.safetensors",
225
+ "model.language_model.layers.23.mlp.down_proj.weight": "model-00008-of-00010.safetensors",
226
+ "model.language_model.layers.23.mlp.gate_proj.weight": "model-00008-of-00010.safetensors",
227
+ "model.language_model.layers.23.mlp.up_proj.weight": "model-00008-of-00010.safetensors",
228
+ "model.language_model.layers.23.post_attention_layernorm.weight": "model-00008-of-00010.safetensors",
229
+ "model.language_model.layers.23.self_attn.k_norm.weight": "model-00008-of-00010.safetensors",
230
+ "model.language_model.layers.23.self_attn.k_proj.weight": "model-00008-of-00010.safetensors",
231
+ "model.language_model.layers.23.self_attn.o_proj.weight": "model-00008-of-00010.safetensors",
232
+ "model.language_model.layers.23.self_attn.q_norm.weight": "model-00008-of-00010.safetensors",
233
+ "model.language_model.layers.23.self_attn.q_proj.weight": "model-00008-of-00010.safetensors",
234
+ "model.language_model.layers.23.self_attn.v_proj.weight": "model-00008-of-00010.safetensors",
235
+ "model.language_model.layers.24.input_layernorm.weight": "model-00008-of-00010.safetensors",
236
+ "model.language_model.layers.24.linear_attn.A_log": "model-00008-of-00010.safetensors",
237
+ "model.language_model.layers.24.linear_attn.conv1d.weight": "model-00008-of-00010.safetensors",
238
+ "model.language_model.layers.24.linear_attn.dt_bias": "model-00008-of-00010.safetensors",
239
+ "model.language_model.layers.24.linear_attn.in_proj_a.weight": "model-00008-of-00010.safetensors",
240
+ "model.language_model.layers.24.linear_attn.in_proj_b.weight": "model-00008-of-00010.safetensors",
241
+ "model.language_model.layers.24.linear_attn.in_proj_qkv.weight": "model-00008-of-00010.safetensors",
242
+ "model.language_model.layers.24.linear_attn.in_proj_z.weight": "model-00008-of-00010.safetensors",
243
+ "model.language_model.layers.24.linear_attn.norm.weight": "model-00008-of-00010.safetensors",
244
+ "model.language_model.layers.24.linear_attn.out_proj.weight": "model-00008-of-00010.safetensors",
245
+ "model.language_model.layers.24.mlp.down_proj.weight": "model-00008-of-00010.safetensors",
246
+ "model.language_model.layers.24.mlp.gate_proj.weight": "model-00008-of-00010.safetensors",
247
+ "model.language_model.layers.24.mlp.up_proj.weight": "model-00008-of-00010.safetensors",
248
+ "model.language_model.layers.24.post_attention_layernorm.weight": "model-00008-of-00010.safetensors",
249
+ "model.language_model.layers.25.input_layernorm.weight": "model-00008-of-00010.safetensors",
250
+ "model.language_model.layers.25.linear_attn.A_log": "model-00008-of-00010.safetensors",
251
+ "model.language_model.layers.25.linear_attn.conv1d.weight": "model-00008-of-00010.safetensors",
252
+ "model.language_model.layers.25.linear_attn.dt_bias": "model-00008-of-00010.safetensors",
253
+ "model.language_model.layers.25.linear_attn.in_proj_a.weight": "model-00008-of-00010.safetensors",
254
+ "model.language_model.layers.25.linear_attn.in_proj_b.weight": "model-00008-of-00010.safetensors",
255
+ "model.language_model.layers.25.linear_attn.in_proj_qkv.weight": "model-00008-of-00010.safetensors",
256
+ "model.language_model.layers.25.linear_attn.in_proj_z.weight": "model-00008-of-00010.safetensors",
257
+ "model.language_model.layers.25.linear_attn.norm.weight": "model-00008-of-00010.safetensors",
258
+ "model.language_model.layers.25.linear_attn.out_proj.weight": "model-00008-of-00010.safetensors",
259
+ "model.language_model.layers.25.mlp.down_proj.weight": "model-00008-of-00010.safetensors",
260
+ "model.language_model.layers.25.mlp.gate_proj.weight": "model-00008-of-00010.safetensors",
261
+ "model.language_model.layers.25.mlp.up_proj.weight": "model-00008-of-00010.safetensors",
262
+ "model.language_model.layers.25.post_attention_layernorm.weight": "model-00008-of-00010.safetensors",
263
+ "model.language_model.layers.26.input_layernorm.weight": "model-00008-of-00010.safetensors",
264
+ "model.language_model.layers.26.linear_attn.A_log": "model-00008-of-00010.safetensors",
265
+ "model.language_model.layers.26.linear_attn.conv1d.weight": "model-00008-of-00010.safetensors",
266
+ "model.language_model.layers.26.linear_attn.dt_bias": "model-00008-of-00010.safetensors",
267
+ "model.language_model.layers.26.linear_attn.in_proj_a.weight": "model-00008-of-00010.safetensors",
268
+ "model.language_model.layers.26.linear_attn.in_proj_b.weight": "model-00008-of-00010.safetensors",
269
+ "model.language_model.layers.26.linear_attn.in_proj_qkv.weight": "model-00008-of-00010.safetensors",
270
+ "model.language_model.layers.26.linear_attn.in_proj_z.weight": "model-00008-of-00010.safetensors",
271
+ "model.language_model.layers.26.linear_attn.norm.weight": "model-00008-of-00010.safetensors",
272
+ "model.language_model.layers.26.linear_attn.out_proj.weight": "model-00008-of-00010.safetensors",
273
+ "model.language_model.layers.26.mlp.down_proj.weight": "model-00008-of-00010.safetensors",
274
+ "model.language_model.layers.26.mlp.gate_proj.weight": "model-00008-of-00010.safetensors",
275
+ "model.language_model.layers.26.mlp.up_proj.weight": "model-00008-of-00010.safetensors",
276
+ "model.language_model.layers.26.post_attention_layernorm.weight": "model-00008-of-00010.safetensors",
277
+ "model.language_model.layers.27.input_layernorm.weight": "model-00008-of-00010.safetensors",
278
+ "model.language_model.layers.27.mlp.down_proj.weight": "model-00008-of-00010.safetensors",
279
+ "model.language_model.layers.27.mlp.gate_proj.weight": "model-00009-of-00010.safetensors",
280
+ "model.language_model.layers.27.mlp.up_proj.weight": "model-00009-of-00010.safetensors",
281
+ "model.language_model.layers.27.post_attention_layernorm.weight": "model-00009-of-00010.safetensors",
282
+ "model.language_model.layers.27.self_attn.k_norm.weight": "model-00009-of-00010.safetensors",
283
+ "model.language_model.layers.27.self_attn.k_proj.weight": "model-00009-of-00010.safetensors",
284
+ "model.language_model.layers.27.self_attn.o_proj.weight": "model-00009-of-00010.safetensors",
285
+ "model.language_model.layers.27.self_attn.q_norm.weight": "model-00009-of-00010.safetensors",
286
+ "model.language_model.layers.27.self_attn.q_proj.weight": "model-00009-of-00010.safetensors",
287
+ "model.language_model.layers.27.self_attn.v_proj.weight": "model-00009-of-00010.safetensors",
288
+ "model.language_model.layers.28.input_layernorm.weight": "model-00009-of-00010.safetensors",
289
+ "model.language_model.layers.28.linear_attn.A_log": "model-00009-of-00010.safetensors",
290
+ "model.language_model.layers.28.linear_attn.conv1d.weight": "model-00009-of-00010.safetensors",
291
+ "model.language_model.layers.28.linear_attn.dt_bias": "model-00009-of-00010.safetensors",
292
+ "model.language_model.layers.28.linear_attn.in_proj_a.weight": "model-00009-of-00010.safetensors",
293
+ "model.language_model.layers.28.linear_attn.in_proj_b.weight": "model-00009-of-00010.safetensors",
294
+ "model.language_model.layers.28.linear_attn.in_proj_qkv.weight": "model-00009-of-00010.safetensors",
295
+ "model.language_model.layers.28.linear_attn.in_proj_z.weight": "model-00009-of-00010.safetensors",
296
+ "model.language_model.layers.28.linear_attn.norm.weight": "model-00009-of-00010.safetensors",
297
+ "model.language_model.layers.28.linear_attn.out_proj.weight": "model-00009-of-00010.safetensors",
298
+ "model.language_model.layers.28.mlp.down_proj.weight": "model-00009-of-00010.safetensors",
299
+ "model.language_model.layers.28.mlp.gate_proj.weight": "model-00009-of-00010.safetensors",
300
+ "model.language_model.layers.28.mlp.up_proj.weight": "model-00009-of-00010.safetensors",
301
+ "model.language_model.layers.28.post_attention_layernorm.weight": "model-00009-of-00010.safetensors",
302
+ "model.language_model.layers.29.input_layernorm.weight": "model-00009-of-00010.safetensors",
303
+ "model.language_model.layers.29.linear_attn.A_log": "model-00009-of-00010.safetensors",
304
+ "model.language_model.layers.29.linear_attn.conv1d.weight": "model-00009-of-00010.safetensors",
305
+ "model.language_model.layers.29.linear_attn.dt_bias": "model-00009-of-00010.safetensors",
306
+ "model.language_model.layers.29.linear_attn.in_proj_a.weight": "model-00009-of-00010.safetensors",
307
+ "model.language_model.layers.29.linear_attn.in_proj_b.weight": "model-00009-of-00010.safetensors",
308
+ "model.language_model.layers.29.linear_attn.in_proj_qkv.weight": "model-00009-of-00010.safetensors",
309
+ "model.language_model.layers.29.linear_attn.in_proj_z.weight": "model-00009-of-00010.safetensors",
310
+ "model.language_model.layers.29.linear_attn.norm.weight": "model-00009-of-00010.safetensors",
311
+ "model.language_model.layers.29.linear_attn.out_proj.weight": "model-00009-of-00010.safetensors",
312
+ "model.language_model.layers.29.mlp.down_proj.weight": "model-00009-of-00010.safetensors",
313
+ "model.language_model.layers.29.mlp.gate_proj.weight": "model-00009-of-00010.safetensors",
314
+ "model.language_model.layers.29.mlp.up_proj.weight": "model-00009-of-00010.safetensors",
315
+ "model.language_model.layers.29.post_attention_layernorm.weight": "model-00009-of-00010.safetensors",
316
+ "model.language_model.layers.3.input_layernorm.weight": "model-00003-of-00010.safetensors",
317
+ "model.language_model.layers.3.mlp.down_proj.weight": "model-00003-of-00010.safetensors",
318
+ "model.language_model.layers.3.mlp.gate_proj.weight": "model-00003-of-00010.safetensors",
319
+ "model.language_model.layers.3.mlp.up_proj.weight": "model-00003-of-00010.safetensors",
320
+ "model.language_model.layers.3.post_attention_layernorm.weight": "model-00003-of-00010.safetensors",
321
+ "model.language_model.layers.3.self_attn.k_norm.weight": "model-00003-of-00010.safetensors",
322
+ "model.language_model.layers.3.self_attn.k_proj.weight": "model-00003-of-00010.safetensors",
323
+ "model.language_model.layers.3.self_attn.o_proj.weight": "model-00003-of-00010.safetensors",
324
+ "model.language_model.layers.3.self_attn.q_norm.weight": "model-00003-of-00010.safetensors",
325
+ "model.language_model.layers.3.self_attn.q_proj.weight": "model-00003-of-00010.safetensors",
326
+ "model.language_model.layers.3.self_attn.v_proj.weight": "model-00003-of-00010.safetensors",
327
+ "model.language_model.layers.30.input_layernorm.weight": "model-00009-of-00010.safetensors",
328
+ "model.language_model.layers.30.linear_attn.A_log": "model-00009-of-00010.safetensors",
329
+ "model.language_model.layers.30.linear_attn.conv1d.weight": "model-00009-of-00010.safetensors",
330
+ "model.language_model.layers.30.linear_attn.dt_bias": "model-00009-of-00010.safetensors",
331
+ "model.language_model.layers.30.linear_attn.in_proj_a.weight": "model-00009-of-00010.safetensors",
332
+ "model.language_model.layers.30.linear_attn.in_proj_b.weight": "model-00009-of-00010.safetensors",
333
+ "model.language_model.layers.30.linear_attn.in_proj_qkv.weight": "model-00009-of-00010.safetensors",
334
+ "model.language_model.layers.30.linear_attn.in_proj_z.weight": "model-00009-of-00010.safetensors",
335
+ "model.language_model.layers.30.linear_attn.norm.weight": "model-00009-of-00010.safetensors",
336
+ "model.language_model.layers.30.linear_attn.out_proj.weight": "model-00009-of-00010.safetensors",
337
+ "model.language_model.layers.30.mlp.down_proj.weight": "model-00009-of-00010.safetensors",
338
+ "model.language_model.layers.30.mlp.gate_proj.weight": "model-00009-of-00010.safetensors",
339
+ "model.language_model.layers.30.mlp.up_proj.weight": "model-00009-of-00010.safetensors",
340
+ "model.language_model.layers.30.post_attention_layernorm.weight": "model-00009-of-00010.safetensors",
341
+ "model.language_model.layers.31.input_layernorm.weight": "model-00009-of-00010.safetensors",
342
+ "model.language_model.layers.31.mlp.down_proj.weight": "model-00009-of-00010.safetensors",
343
+ "model.language_model.layers.31.mlp.gate_proj.weight": "model-00009-of-00010.safetensors",
344
+ "model.language_model.layers.31.mlp.up_proj.weight": "model-00009-of-00010.safetensors",
345
+ "model.language_model.layers.31.post_attention_layernorm.weight": "model-00009-of-00010.safetensors",
346
+ "model.language_model.layers.31.self_attn.k_norm.weight": "model-00009-of-00010.safetensors",
347
+ "model.language_model.layers.31.self_attn.k_proj.weight": "model-00009-of-00010.safetensors",
348
+ "model.language_model.layers.31.self_attn.o_proj.weight": "model-00009-of-00010.safetensors",
349
+ "model.language_model.layers.31.self_attn.q_norm.weight": "model-00009-of-00010.safetensors",
350
+ "model.language_model.layers.31.self_attn.q_proj.weight": "model-00010-of-00010.safetensors",
351
+ "model.language_model.layers.31.self_attn.v_proj.weight": "model-00010-of-00010.safetensors",
352
+ "model.language_model.layers.4.input_layernorm.weight": "model-00003-of-00010.safetensors",
353
+ "model.language_model.layers.4.linear_attn.A_log": "model-00003-of-00010.safetensors",
354
+ "model.language_model.layers.4.linear_attn.conv1d.weight": "model-00003-of-00010.safetensors",
355
+ "model.language_model.layers.4.linear_attn.dt_bias": "model-00003-of-00010.safetensors",
356
+ "model.language_model.layers.4.linear_attn.in_proj_a.weight": "model-00003-of-00010.safetensors",
357
+ "model.language_model.layers.4.linear_attn.in_proj_b.weight": "model-00003-of-00010.safetensors",
358
+ "model.language_model.layers.4.linear_attn.in_proj_qkv.weight": "model-00003-of-00010.safetensors",
359
+ "model.language_model.layers.4.linear_attn.in_proj_z.weight": "model-00003-of-00010.safetensors",
360
+ "model.language_model.layers.4.linear_attn.norm.weight": "model-00003-of-00010.safetensors",
361
+ "model.language_model.layers.4.linear_attn.out_proj.weight": "model-00003-of-00010.safetensors",
362
+ "model.language_model.layers.4.mlp.down_proj.weight": "model-00003-of-00010.safetensors",
363
+ "model.language_model.layers.4.mlp.gate_proj.weight": "model-00004-of-00010.safetensors",
364
+ "model.language_model.layers.4.mlp.up_proj.weight": "model-00004-of-00010.safetensors",
365
+ "model.language_model.layers.4.post_attention_layernorm.weight": "model-00004-of-00010.safetensors",
366
+ "model.language_model.layers.5.input_layernorm.weight": "model-00004-of-00010.safetensors",
367
+ "model.language_model.layers.5.linear_attn.A_log": "model-00004-of-00010.safetensors",
368
+ "model.language_model.layers.5.linear_attn.conv1d.weight": "model-00004-of-00010.safetensors",
369
+ "model.language_model.layers.5.linear_attn.dt_bias": "model-00004-of-00010.safetensors",
370
+ "model.language_model.layers.5.linear_attn.in_proj_a.weight": "model-00004-of-00010.safetensors",
371
+ "model.language_model.layers.5.linear_attn.in_proj_b.weight": "model-00004-of-00010.safetensors",
372
+ "model.language_model.layers.5.linear_attn.in_proj_qkv.weight": "model-00004-of-00010.safetensors",
373
+ "model.language_model.layers.5.linear_attn.in_proj_z.weight": "model-00004-of-00010.safetensors",
374
+ "model.language_model.layers.5.linear_attn.norm.weight": "model-00004-of-00010.safetensors",
375
+ "model.language_model.layers.5.linear_attn.out_proj.weight": "model-00004-of-00010.safetensors",
376
+ "model.language_model.layers.5.mlp.down_proj.weight": "model-00004-of-00010.safetensors",
377
+ "model.language_model.layers.5.mlp.gate_proj.weight": "model-00004-of-00010.safetensors",
378
+ "model.language_model.layers.5.mlp.up_proj.weight": "model-00004-of-00010.safetensors",
379
+ "model.language_model.layers.5.post_attention_layernorm.weight": "model-00004-of-00010.safetensors",
380
+ "model.language_model.layers.6.input_layernorm.weight": "model-00004-of-00010.safetensors",
381
+ "model.language_model.layers.6.linear_attn.A_log": "model-00004-of-00010.safetensors",
382
+ "model.language_model.layers.6.linear_attn.conv1d.weight": "model-00004-of-00010.safetensors",
383
+ "model.language_model.layers.6.linear_attn.dt_bias": "model-00004-of-00010.safetensors",
384
+ "model.language_model.layers.6.linear_attn.in_proj_a.weight": "model-00004-of-00010.safetensors",
385
+ "model.language_model.layers.6.linear_attn.in_proj_b.weight": "model-00004-of-00010.safetensors",
386
+ "model.language_model.layers.6.linear_attn.in_proj_qkv.weight": "model-00004-of-00010.safetensors",
387
+ "model.language_model.layers.6.linear_attn.in_proj_z.weight": "model-00004-of-00010.safetensors",
388
+ "model.language_model.layers.6.linear_attn.norm.weight": "model-00004-of-00010.safetensors",
389
+ "model.language_model.layers.6.linear_attn.out_proj.weight": "model-00004-of-00010.safetensors",
390
+ "model.language_model.layers.6.mlp.down_proj.weight": "model-00004-of-00010.safetensors",
391
+ "model.language_model.layers.6.mlp.gate_proj.weight": "model-00004-of-00010.safetensors",
392
+ "model.language_model.layers.6.mlp.up_proj.weight": "model-00004-of-00010.safetensors",
393
+ "model.language_model.layers.6.post_attention_layernorm.weight": "model-00004-of-00010.safetensors",
394
+ "model.language_model.layers.7.input_layernorm.weight": "model-00004-of-00010.safetensors",
395
+ "model.language_model.layers.7.mlp.down_proj.weight": "model-00004-of-00010.safetensors",
396
+ "model.language_model.layers.7.mlp.gate_proj.weight": "model-00004-of-00010.safetensors",
397
+ "model.language_model.layers.7.mlp.up_proj.weight": "model-00004-of-00010.safetensors",
398
+ "model.language_model.layers.7.post_attention_layernorm.weight": "model-00004-of-00010.safetensors",
399
+ "model.language_model.layers.7.self_attn.k_norm.weight": "model-00004-of-00010.safetensors",
400
+ "model.language_model.layers.7.self_attn.k_proj.weight": "model-00004-of-00010.safetensors",
401
+ "model.language_model.layers.7.self_attn.o_proj.weight": "model-00004-of-00010.safetensors",
402
+ "model.language_model.layers.7.self_attn.q_norm.weight": "model-00004-of-00010.safetensors",
403
+ "model.language_model.layers.7.self_attn.q_proj.weight": "model-00004-of-00010.safetensors",
404
+ "model.language_model.layers.7.self_attn.v_proj.weight": "model-00004-of-00010.safetensors",
405
+ "model.language_model.layers.8.input_layernorm.weight": "model-00004-of-00010.safetensors",
406
+ "model.language_model.layers.8.linear_attn.A_log": "model-00004-of-00010.safetensors",
407
+ "model.language_model.layers.8.linear_attn.conv1d.weight": "model-00004-of-00010.safetensors",
408
+ "model.language_model.layers.8.linear_attn.dt_bias": "model-00004-of-00010.safetensors",
409
+ "model.language_model.layers.8.linear_attn.in_proj_a.weight": "model-00004-of-00010.safetensors",
410
+ "model.language_model.layers.8.linear_attn.in_proj_b.weight": "model-00004-of-00010.safetensors",
411
+ "model.language_model.layers.8.linear_attn.in_proj_qkv.weight": "model-00004-of-00010.safetensors",
412
+ "model.language_model.layers.8.linear_attn.in_proj_z.weight": "model-00004-of-00010.safetensors",
413
+ "model.language_model.layers.8.linear_attn.norm.weight": "model-00004-of-00010.safetensors",
414
+ "model.language_model.layers.8.linear_attn.out_proj.weight": "model-00004-of-00010.safetensors",
415
+ "model.language_model.layers.8.mlp.down_proj.weight": "model-00004-of-00010.safetensors",
416
+ "model.language_model.layers.8.mlp.gate_proj.weight": "model-00004-of-00010.safetensors",
417
+ "model.language_model.layers.8.mlp.up_proj.weight": "model-00004-of-00010.safetensors",
418
+ "model.language_model.layers.8.post_attention_layernorm.weight": "model-00004-of-00010.safetensors",
419
+ "model.language_model.layers.9.input_layernorm.weight": "model-00004-of-00010.safetensors",
420
+ "model.language_model.layers.9.linear_attn.A_log": "model-00004-of-00010.safetensors",
421
+ "model.language_model.layers.9.linear_attn.conv1d.weight": "model-00004-of-00010.safetensors",
422
+ "model.language_model.layers.9.linear_attn.dt_bias": "model-00004-of-00010.safetensors",
423
+ "model.language_model.layers.9.linear_attn.in_proj_a.weight": "model-00004-of-00010.safetensors",
424
+ "model.language_model.layers.9.linear_attn.in_proj_b.weight": "model-00004-of-00010.safetensors",
425
+ "model.language_model.layers.9.linear_attn.in_proj_qkv.weight": "model-00004-of-00010.safetensors",
426
+ "model.language_model.layers.9.linear_attn.in_proj_z.weight": "model-00005-of-00010.safetensors",
427
+ "model.language_model.layers.9.linear_attn.norm.weight": "model-00005-of-00010.safetensors",
428
+ "model.language_model.layers.9.linear_attn.out_proj.weight": "model-00005-of-00010.safetensors",
429
+ "model.language_model.layers.9.mlp.down_proj.weight": "model-00005-of-00010.safetensors",
430
+ "model.language_model.layers.9.mlp.gate_proj.weight": "model-00005-of-00010.safetensors",
431
+ "model.language_model.layers.9.mlp.up_proj.weight": "model-00005-of-00010.safetensors",
432
+ "model.language_model.layers.9.post_attention_layernorm.weight": "model-00005-of-00010.safetensors",
433
+ "model.language_model.norm.weight": "model-00010-of-00010.safetensors"
434
+ }
435
+ }
model_details.json CHANGED
@@ -1,12 +1,10 @@
1
  {
2
  "base_model": "Qwen/Qwen3.5-9B",
3
  "base_revision": "c202236235762e1c871ad0ccb60c8ee5ba337b9a",
4
- "adapter_type": "LoRA",
5
  "rank": 32,
6
  "alpha": 64,
7
  "dropout": 0.05,
8
  "adapter_parameters": 58195968,
9
- "selected_checkpoint": "final_adapter",
10
  "selection": "Final checkpoint after three epochs; no best-validation checkpoint reload.",
11
  "epoch": 3.0,
12
  "global_step": 79815,
@@ -27,6 +25,12 @@
27
  "seed": 20260704
28
  },
29
  "release_changes": [
30
- "Replace the internal base-model path with its official Hub ID and pin the exact base revision."
31
- ]
 
 
 
 
 
 
32
  }
 
1
  {
2
  "base_model": "Qwen/Qwen3.5-9B",
3
  "base_revision": "c202236235762e1c871ad0ccb60c8ee5ba337b9a",
 
4
  "rank": 32,
5
  "alpha": 64,
6
  "dropout": 0.05,
7
  "adapter_parameters": 58195968,
 
8
  "selection": "Final checkpoint after three epochs; no best-validation checkpoint reload.",
9
  "epoch": 3.0,
10
  "global_step": 79815,
 
25
  "seed": 20260704
26
  },
27
  "release_changes": [
28
+ "Replace the internal base-model path with its official Hub ID and pin the exact base revision.",
29
+ "Merge the published final adapter into the unquantized pinned base text language model."
30
+ ],
31
+ "training_adapter_type": "LoRA",
32
+ "training_checkpoint": "final_adapter",
33
+ "release_format": "merged BF16 text language model",
34
+ "source_adapter_sha256": "c4d89ce3b6cfdb53dba2927683331e2182c8767f3380527a78f0533b0f31e5bd",
35
+ "source_adapter_hf_revision": "341e9b198c985ee75c982a43376015d806b15522"
36
  }
requirements.txt CHANGED
@@ -1,7 +1,5 @@
1
- torch>=2.5
2
  transformers==5.13.0
3
- peft==0.19.1
4
  accelerate>=1.13
5
- bitsandbytes>=0.49.2
6
  huggingface_hub>=0.34
7
- safetensors>=0.5
 
1
+ torch>=2.5.1
2
  transformers==5.13.0
 
3
  accelerate>=1.13
 
4
  huggingface_hub>=0.34
5
+ safetensors>=0.8
tokenizer_config.json CHANGED
@@ -9,8 +9,6 @@
9
  "eos_token": "<|im_end|>",
10
  "errors": "replace",
11
  "image_token": "<|image_pad|>",
12
- "is_local": true,
13
- "local_files_only": true,
14
  "model_max_length": 262144,
15
  "model_specific_special_tokens": {
16
  "audio_bos_token": "<|audio_start|>",
 
9
  "eos_token": "<|im_end|>",
10
  "errors": "replace",
11
  "image_token": "<|image_pad|>",
 
 
12
  "model_max_length": 262144,
13
  "model_specific_special_tokens": {
14
  "audio_bos_token": "<|audio_start|>",
validation.json CHANGED
@@ -1,26 +1,61 @@
1
  {
2
- "adapter_sha256": "c4d89ce3b6cfdb53dba2927683331e2182c8767f3380527a78f0533b0f31e5bd",
 
 
 
3
  "matches_source_final_adapter": true,
4
- "adapter_tensors": 256,
5
- "adapter_parameters": 58195968,
6
- "all_tensors_finite": true,
7
  "base_revision": "c202236235762e1c871ad0ccb60c8ee5ba337b9a",
8
- "base_revision_verified_on_hub": true,
9
- "model_card_and_python_syntax_validated": true,
10
- "original_system_prompt_verified": true,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
11
  "gpu_smoke_test": {
12
  "passed": true,
13
  "examples": 3,
14
- "input": "examples/histories.jsonl",
15
- "output": "examples/replies.jsonl",
16
- "nonempty_responses": 3,
17
  "thinking_enabled": false,
18
- "quantization": "NF4 double quantization",
19
- "max_new_tokens": 128,
20
- "torch": "2.5.1+cu121",
21
  "transformers": "5.13.0",
22
- "peft": "0.19.1",
23
- "accelerate": "1.13.0",
24
- "bitsandbytes": "0.49.2"
25
- }
26
  }
 
1
  {
2
+ "status": "passed",
3
+ "format": "merged BF16 text language model",
4
+ "source_adapter_revision": "341e9b198c985ee75c982a43376015d806b15522",
5
+ "source_adapter_sha256": "c4d89ce3b6cfdb53dba2927683331e2182c8767f3380527a78f0533b0f31e5bd",
6
  "matches_source_final_adapter": true,
 
 
 
7
  "base_revision": "c202236235762e1c871ad0ccb60c8ee5ba337b9a",
8
+ "all_parameters_finite": true,
9
+ "parameter_count": 8953803264,
10
+ "parameter_dtypes": {
11
+ "torch.bfloat16": 8953803264
12
+ },
13
+ "tensor_count": 427,
14
+ "weight_shards": 10,
15
+ "weight_file_bytes": 17907662272,
16
+ "weight_sha256": {
17
+ "model-00001-of-00010.safetensors": "dd63614f1dc80dce2d83be3f3e69af1f8a0ebc8add9b8e0abf3910f563e44344",
18
+ "model-00002-of-00010.safetensors": "856617f65c8f2d2686c0aec3c6036c9bf53466ed6c90275bee6d0ecaa8263b30",
19
+ "model-00003-of-00010.safetensors": "e956a1f8536baf45076db19dffb96f258a6d94fe0a545f2bb38fed9517ab19a0",
20
+ "model-00004-of-00010.safetensors": "847db00981824e02e112fa1f57526d1feebfa268d1e85c32c1e24b9d4641afc7",
21
+ "model-00005-of-00010.safetensors": "5c13aa2a1fe366710e1dcd1484901e7725afdde55b2c717a2a25797d93b75058",
22
+ "model-00006-of-00010.safetensors": "6c65097043a90de2d46d0d01d698ef1a32bde00a3baf458434ffbf0906254a31",
23
+ "model-00007-of-00010.safetensors": "7de8c454f2661fd3f6e6a7c0a46dc8e6915a34caf3fac961b895b306d64b89b9",
24
+ "model-00008-of-00010.safetensors": "293d1a770da768e17860c9915a55ea26d178067d118cfdbdc2c285577dc21b87",
25
+ "model-00009-of-00010.safetensors": "9d99b5f3e4bff23afe10799d21ca22eb6e866cbd17a78428e6e4da6fb8466211",
26
+ "model-00010-of-00010.safetensors": "d420fe217cc622433d5ec2bb1620ea34eaf4a4561e3a55b69cbe8856b17705dc"
27
+ },
28
+ "merge_comparison": [
29
+ {
30
+ "id": "synthetic-anxiety",
31
+ "logit_cosine_similarity": 0.9999359846115112,
32
+ "mean_absolute_logit_difference": 0.01997707225382328,
33
+ "max_absolute_logit_difference": 0.125,
34
+ "same_next_token": true
35
+ },
36
+ {
37
+ "id": "synthetic-interpersonal",
38
+ "logit_cosine_similarity": 0.9999428391456604,
39
+ "mean_absolute_logit_difference": 0.018034368753433228,
40
+ "max_absolute_logit_difference": 0.125,
41
+ "same_next_token": true
42
+ },
43
+ {
44
+ "id": "synthetic-work",
45
+ "logit_cosine_similarity": 0.9999374151229858,
46
+ "mean_absolute_logit_difference": 0.02002718113362789,
47
+ "max_absolute_logit_difference": 0.125,
48
+ "same_next_token": true
49
+ }
50
+ ],
51
  "gpu_smoke_test": {
52
  "passed": true,
53
  "examples": 3,
 
 
 
54
  "thinking_enabled": false,
55
+ "quantization": "none",
56
+ "output": "examples/replies.jsonl",
 
57
  "transformers": "5.13.0",
58
+ "torch": "2.5.1+cu121"
59
+ },
60
+ "removed_local_config_fields": []
 
61
  }