TensorMind commited on
Commit
8ce2c21
·
verified ·
1 Parent(s): 4bc137b

Release TensorMind 1.5 Preview

Browse files

Publish TensorMind 1.5 Preview with its model card, TensorPlay AI poster, benchmark scorecard, and reproducible evaluation assets.

.gitattributes CHANGED
@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ assets/benchmark-matrix.png filter=lfs diff=lfs merge=lfs -text
37
+ assets/benchmark-suite.png filter=lfs diff=lfs merge=lfs -text
38
+ assets/poster.png filter=lfs diff=lfs merge=lfs -text
39
+ assets/tensorplay-ai-logo.png filter=lfs diff=lfs merge=lfs -text
LICENSE ADDED
@@ -0,0 +1,201 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Apache License
2
+ Version 2.0, January 2004
3
+ http://www.apache.org/licenses/
4
+
5
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
6
+
7
+ 1. Definitions.
8
+
9
+ "License" shall mean the terms and conditions for use, reproduction,
10
+ and distribution as defined by Sections 1 through 9 of this document.
11
+
12
+ "Licensor" shall mean the copyright owner or entity authorized by
13
+ the copyright owner that is granting the License.
14
+
15
+ "Legal Entity" shall mean the union of the acting entity and all
16
+ other entities that control, are controlled by, or are under common
17
+ control with that entity. For the purposes of this definition,
18
+ "control" means (i) the power, direct or indirect, to cause the
19
+ direction or management of such entity, whether by contract or
20
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
21
+ outstanding shares, or (iii) beneficial ownership of such entity.
22
+
23
+ "You" (or "Your") shall mean an individual or Legal Entity
24
+ exercising permissions granted by this License.
25
+
26
+ "Source" form shall mean the preferred form for making modifications,
27
+ including but not limited to software source code, documentation
28
+ source, and configuration files.
29
+
30
+ "Object" form shall mean any form resulting from mechanical
31
+ transformation or translation of a Source form, including but
32
+ not limited to compiled object code, generated documentation,
33
+ and conversions to other media types.
34
+
35
+ "Work" shall mean the work of authorship, whether in Source or
36
+ Object form, made available under the License, as indicated by a
37
+ copyright notice that is included in or attached to the work
38
+ (an example is provided in the Appendix below).
39
+
40
+ "Derivative Works" shall mean any work, whether in Source or Object
41
+ form, that is based on (or derived from) the Work and for which the
42
+ editorial revisions, annotations, elaborations, or other modifications
43
+ represent, as a whole, an original work of authorship. For the purposes
44
+ of this License, Derivative Works shall not include works that remain
45
+ separable from, or merely link (or bind by name) to the interfaces of,
46
+ the Work and Derivative Works thereof.
47
+
48
+ "Contribution" shall mean any work of authorship, including
49
+ the original version of the Work and any modifications or additions
50
+ to that Work or Derivative Works thereof, that is intentionally
51
+ submitted to Licensor for inclusion in the Work by the copyright owner
52
+ or by an individual or Legal Entity authorized to submit on behalf of
53
+ the copyright owner. For the purposes of this definition, "submitted"
54
+ means any form of electronic, verbal, or written communication sent
55
+ to the Licensor or its representatives, including but not limited to
56
+ communication on electronic mailing lists, source code control systems,
57
+ and issue tracking systems that are managed by, or on behalf of, the
58
+ Licensor for the purpose of discussing and improving the Work, but
59
+ excluding communication that is conspicuously marked or otherwise
60
+ designated in writing by the copyright owner as "Not a Contribution."
61
+
62
+ "Contributor" shall mean Licensor and any individual or Legal Entity
63
+ on behalf of whom a Contribution has been received by Licensor and
64
+ subsequently incorporated within the Work.
65
+
66
+ 2. Grant of Copyright License. Subject to the terms and conditions of
67
+ this License, each Contributor hereby grants to You a perpetual,
68
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
69
+ copyright license to reproduce, prepare Derivative Works of,
70
+ publicly display, publicly perform, sublicense, and distribute the
71
+ Work and such Derivative Works in Source or Object form.
72
+
73
+ 3. Grant of Patent License. Subject to the terms and conditions of
74
+ this License, each Contributor hereby grants to You a perpetual,
75
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
76
+ (except as stated in this section) patent license to make, have made,
77
+ use, offer to sell, sell, import, and otherwise transfer the Work,
78
+ where such license applies only to those patent claims licensable
79
+ by such Contributor that are necessarily infringed by their
80
+ Contribution(s) alone or by combination of their Contribution(s)
81
+ with the Work to which such Contribution(s) was submitted. If You
82
+ institute patent litigation against any entity (including a
83
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
84
+ or a Contribution incorporated within the Work constitutes direct
85
+ or contributory patent infringement, then any patent licenses
86
+ granted to You under this License for that Work shall terminate
87
+ as of the date such litigation is filed.
88
+
89
+ 4. Redistribution. You may reproduce and distribute copies of the
90
+ Work or Derivative Works thereof in any medium, with or without
91
+ modifications, and in Source or Object form, provided that You
92
+ meet the following conditions:
93
+
94
+ (a) You must give any other recipients of the Work or
95
+ Derivative Works a copy of this License; and
96
+
97
+ (b) You must cause any modified files to carry prominent notices
98
+ stating that You changed the files; and
99
+
100
+ (c) You must retain, in the Source form of any Derivative Works
101
+ that You distribute, all copyright, patent, trademark, and
102
+ attribution notices from the Source form of the Work,
103
+ excluding those notices that do not pertain to any part of
104
+ the Derivative Works; and
105
+
106
+ (d) If the Work includes a "NOTICE" text file as part of its
107
+ distribution, then any Derivative Works that You distribute must
108
+ include a readable copy of the attribution notices contained
109
+ within such NOTICE file, excluding those notices that do not
110
+ pertain to any part of the Derivative Works, in at least one
111
+ of the following places: within a NOTICE text file distributed
112
+ as part of the Derivative Works; within the Source form or
113
+ documentation, if provided along with the Derivative Works; or,
114
+ within a display generated by the Derivative Works, if and
115
+ wherever such third-party notices normally appear. The contents
116
+ of the NOTICE file are for informational purposes only and
117
+ do not modify the License. You may add Your own attribution
118
+ notices within Derivative Works that You distribute, alongside
119
+ or as an addendum to the NOTICE text from the Work, provided
120
+ that such additional attribution notices cannot be construed
121
+ as modifying the License.
122
+
123
+ You may add Your own copyright statement to Your modifications and
124
+ may provide additional or different license terms and conditions
125
+ for use, reproduction, or distribution of Your modifications, or
126
+ for any such Derivative Works as a whole, provided Your use,
127
+ reproduction, and distribution of the Work otherwise complies with
128
+ the conditions stated in this License.
129
+
130
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
131
+ any Contribution intentionally submitted for inclusion in the Work
132
+ by You to the Licensor shall be under the terms and conditions of
133
+ this License, without any additional terms or conditions.
134
+ Notwithstanding the above, nothing herein shall supersede or modify
135
+ the terms of any separate license agreement you may have executed
136
+ with Licensor regarding such Contributions.
137
+
138
+ 6. Trademarks. This License does not grant permission to use the trade
139
+ names, trademarks, service marks, or product names of the Licensor,
140
+ except as required for reasonable and customary use in describing the
141
+ origin of the Work and reproducing the content of the NOTICE file.
142
+
143
+ 7. Disclaimer of Warranty. Unless required by applicable law or
144
+ agreed to in writing, Licensor provides the Work (and each
145
+ Contributor provides its Contributions) on an "AS IS" BASIS,
146
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
147
+ implied, including, without limitation, any warranties or conditions
148
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
149
+ PARTICULAR PURPOSE. You are solely responsible for determining the
150
+ appropriateness of using or redistributing the Work and assume any
151
+ risks associated with Your exercise of permissions under this License.
152
+
153
+ 8. Limitation of Liability. In no event and under no legal theory,
154
+ whether in tort (including negligence), contract, or otherwise,
155
+ unless required by applicable law (such as deliberate and grossly
156
+ negligent acts) or agreed to in writing, shall any Contributor be
157
+ liable to You for damages, including any direct, indirect, special,
158
+ incidental, or consequential damages of any character arising as a
159
+ result of this License or out of the use or inability to use the
160
+ Work (including but not limited to damages for loss of goodwill,
161
+ work stoppage, computer failure or malfunction, or any and all
162
+ other commercial damages or losses), even if such Contributor
163
+ has been advised of the possibility of such damages.
164
+
165
+ 9. Accepting Warranty or Additional Liability. While redistributing
166
+ the Work or Derivative Works thereof, You may choose to offer,
167
+ and charge a fee for, acceptance of support, warranty, indemnity,
168
+ or other liability obligations and/or rights consistent with this
169
+ License. However, in accepting such obligations, You may act only
170
+ on Your own behalf and on Your sole responsibility, not on behalf
171
+ of any other Contributor, and only if You agree to indemnify,
172
+ defend, and hold each Contributor harmless for any liability
173
+ incurred by, or claims asserted against, such Contributor by reason
174
+ of your accepting any such warranty or additional liability.
175
+
176
+ END OF TERMS AND CONDITIONS
177
+
178
+ APPENDIX: How to apply the Apache License to your work.
179
+
180
+ To apply the Apache License to your work, attach the following
181
+ boilerplate notice, with the fields enclosed by brackets "[]"
182
+ replaced with your own identifying information. (Don't include
183
+ the brackets!) The text should be enclosed in the appropriate
184
+ comment syntax for the file format. We also recommend that a
185
+ file or class name and description of purpose be included on the
186
+ same "printed page" as the copyright notice for easier
187
+ identification within third-party archives.
188
+
189
+ Copyright [yyyy] [name of copyright owner]
190
+
191
+ Licensed under the Apache License, Version 2.0 (the "License");
192
+ you may not use this file except in compliance with the License.
193
+ You may obtain a copy of the License at
194
+
195
+ http://www.apache.org/licenses/LICENSE-2.0
196
+
197
+ Unless required by applicable law or agreed to in writing, software
198
+ distributed under the License is distributed on an "AS IS" BASIS,
199
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
200
+ See the License for the specific language governing permissions and
201
+ limitations under the License.
README.md ADDED
@@ -0,0 +1,175 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ library_name: transformers
4
+ pipeline_tag: text-generation
5
+ language:
6
+ - zh
7
+ - en
8
+ tags:
9
+ - qwen3
10
+ - conversational
11
+ - text-generation
12
+ - safetensors
13
+ - tensorplay
14
+ - tensormind
15
+ - preview
16
+ inference: true
17
+ ---
18
+
19
+ <div align="center">
20
+ <img src="./assets/poster.png" alt="TensorMind 1.5 Preview" width="100%" />
21
+ </div>
22
+
23
+ <h1 align="center">TensorMind 1.5 Preview</h1>
24
+
25
+ <p align="center">
26
+ A compact, open-weight bilingual language model from <strong>TensorPlay AI</strong>.
27
+ </p>
28
+
29
+ <p align="center">
30
+ <code>536.9M parameters</code> · <code>BF16</code> · <code>Qwen3 architecture</code> · <code>Apache-2.0</code>
31
+ </p>
32
+
33
+ ## Overview
34
+
35
+ TensorMind 1.5 Preview is a lightweight Chinese/English conversational model for local experimentation, instruction-following research, and continued fine-tuning. It ships as a standard Transformers package with Safetensors weights, tokenizer, generation config, and chat template—no custom model code is required.
36
+
37
+ ### Highlights
38
+
39
+ - **Compact deployment:** 536,941,568 parameters in a 1.00 GiB BF16 weight file.
40
+ - **Standard runtime:** native `Qwen3ForCausalLM` support in Transformers.
41
+ - **Chat ready:** bundled bilingual chat template with optional non-thinking generation.
42
+ - **Reproducible evaluation:** exact score data, vector figures, and rendering source are included.
43
+ - **Open release:** Apache-2.0 licensed weights for research and application prototyping.
44
+
45
+ ## Quick start
46
+
47
+ ```bash
48
+ pip install "transformers>=4.51.0" accelerate torch
49
+ ```
50
+
51
+ ```python
52
+ import torch
53
+ from transformers import AutoModelForCausalLM, AutoTokenizer
54
+
55
+ model_id = "AATensorPlay/TensorMind-1.5-preview"
56
+
57
+ tokenizer = AutoTokenizer.from_pretrained(model_id)
58
+ model = AutoModelForCausalLM.from_pretrained(
59
+ model_id,
60
+ torch_dtype="auto",
61
+ device_map="auto",
62
+ )
63
+
64
+ messages = [
65
+ {"role": "system", "content": "You are a helpful assistant."},
66
+ {"role": "user", "content": "用三句话解释什么是强化学习。"},
67
+ ]
68
+
69
+ prompt = tokenizer.apply_chat_template(
70
+ messages,
71
+ tokenize=False,
72
+ add_generation_prompt=True,
73
+ enable_thinking=False,
74
+ )
75
+ inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
76
+
77
+ with torch.inference_mode():
78
+ output = model.generate(
79
+ **inputs,
80
+ max_new_tokens=256,
81
+ do_sample=True,
82
+ temperature=0.7,
83
+ top_p=0.8,
84
+ )
85
+
86
+ new_tokens = output[0, inputs.input_ids.shape[1]:]
87
+ print(tokenizer.decode(new_tokens, skip_special_tokens=True))
88
+ ```
89
+
90
+ For deterministic decoding, set `do_sample=False` and omit `temperature` and `top_p`.
91
+
92
+ ## Model details
93
+
94
+ | Item | Value |
95
+ |---|---|
96
+ | Architecture | `Qwen3ForCausalLM` |
97
+ | Parameters | 536,941,568 |
98
+ | Precision | BF16 |
99
+ | Layers | 32 |
100
+ | Hidden size | 1,024 |
101
+ | Intermediate size | 4,096 |
102
+ | Attention heads / KV heads | 16 / 8 |
103
+ | Vocabulary | 32,768 |
104
+ | Position configuration | 10,240 positions¹ |
105
+ | Attention | Full attention; no sliding window |
106
+ | Weight tying | Enabled |
107
+ | Weight format | Safetensors |
108
+
109
+ ¹ `max_position_embeddings=10240` is the exported configuration ceiling. This preview has not been evaluated on Needle, RULER, LongBench, or another effective long-context benchmark, so it should not be interpreted as a validated long-context quality claim.
110
+
111
+ ## Evaluation
112
+
113
+ <div align="center">
114
+ <img src="./assets/benchmark-suite.png" alt="TensorMind 1.5 Preview benchmark suite" width="100%" />
115
+ </div>
116
+
117
+ Scores use one matched protocol: `lm-eval 0.4.12`, SGLang `0.5.14`, zero-shot evaluation, full datasets, batch size 48, and fixed seeds.
118
+
119
+ | Model | CMMLU | AGIEval-CN | A-CLUE | C-Eval | TMMLU+ | Macro |
120
+ |---|---:|---:|---:|---:|---:|---:|
121
+ | **TensorMind 1.5 Preview** | 24.8834 | 32.3822 | 24.7282 | 23.2541 | 24.7272 | **25.9950** |
122
+
123
+ These are standalone release scores. Compare them with other models only when the harness, prompts, datasets, and decoding protocol are matched. The exact values are available in [`benchmark-results.json`](./assets/benchmark-results.json), with an additional [benchmark scorecard](./assets/benchmark-matrix.png).
124
+
125
+ ## Training snapshot
126
+
127
+ | Item | Value |
128
+ |---|---:|
129
+ | Final alignment packing length | 512 tokens |
130
+ | Packed train / eval sequences | 8,429 / 422 |
131
+ | Optimization steps | 50 |
132
+ | Per-device train / eval batch | 28 / 28 |
133
+ | Peak learning rate | 5e-7 |
134
+ | Precision | BF16 |
135
+ | Train loss | 2.2760 |
136
+ | Eval loss, start → finish | 2.16935 → 2.16928 |
137
+
138
+ ## Intended use
139
+
140
+ Good fits include:
141
+
142
+ - local and edge-oriented language-model experiments;
143
+ - Chinese/English chat prototypes;
144
+ - instruction-tuning and alignment research;
145
+ - reproducible inference, quantization, and serving studies.
146
+
147
+ This preview is **not** a safety-tuned production assistant, a high-stakes decision system, or a substitute for domain experts. Evaluate it for your task before deployment.
148
+
149
+ ## Limitations
150
+
151
+ - At 0.5B scale, the model can hallucinate, miss instructions, and produce incorrect reasoning or arithmetic.
152
+ - Benchmark scores are protocol-specific and do not establish superiority over other releases.
153
+ - Effective long-context quality has not been measured; the position setting is a configuration value only.
154
+ - Safety, multilingual breadth, tool use, and production robustness have not received comprehensive evaluation.
155
+
156
+ ## Checksums
157
+
158
+ | File | SHA-256 |
159
+ |---|---|
160
+ | `model.safetensors` | `b99f6c0448831a6cb2012b93793dd72c9e64e94deee15fd8c4d8206c056817d9` |
161
+ | `tokenizer.json` | `2b31db3618982e4b4425535b563b4f12f8b62f82114abf3ef67e363a310bc44b` |
162
+ | `chat_template.jinja` | `ed2c59f23b9eb551f51d870b3a2aa7f9e431d424f8e34f27e4012049e85ba814` |
163
+ | `config.json` | `bf6d70f9520550eec8aef7961b3f240a06b69a46e8c2bd273b8cb18d61a494f1` |
164
+
165
+ ## License
166
+
167
+ TensorMind 1.5 Preview is released under the [Apache License 2.0](./LICENSE).
168
+
169
+ ---
170
+
171
+ <div align="center">
172
+ <img src="./assets/tensorplay-ai-logo.png" alt="TensorPlay AI" width="140" />
173
+ <br />
174
+ Built by TensorPlay AI
175
+ </div>
assets/benchmark-matrix.png ADDED

Git LFS Details

  • SHA256: a9ece22a8b9f711de99861ff3aa2c0d185ecf9702cab970d638b143a957d47b1
  • Pointer size: 131 Bytes
  • Size of remote file: 164 kB
assets/benchmark-matrix.svg ADDED
assets/benchmark-results.json ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "TensorMind 1.5 Preview",
3
+ "protocol": {
4
+ "lm_eval": "0.4.12",
5
+ "serving": "SGLang 0.5.14",
6
+ "shots": 0,
7
+ "dataset_scope": "full",
8
+ "batch_size": 48,
9
+ "seeds": "fixed",
10
+ "metric": "accuracy (%)"
11
+ },
12
+ "results": {
13
+ "cmmlu": 24.8834,
14
+ "agieval_cn": 32.3822,
15
+ "a_clue": 24.7282,
16
+ "c_eval": 23.2541,
17
+ "tmmlu_plus": 24.7272,
18
+ "five_suite_macro": 25.9950
19
+ },
20
+ "note": "Scores are reported for this public release only. Compare models only under a matched evaluation protocol."
21
+ }
assets/benchmark-suite.png ADDED

Git LFS Details

  • SHA256: da5819f1dbda306041342c2404ecd9293e471a5910f73696014155f925d9caf2
  • Pointer size: 131 Bytes
  • Size of remote file: 174 kB
assets/benchmark-suite.svg ADDED
assets/poster.png ADDED

Git LFS Details

  • SHA256: b751d48e61630eaa6e2ceeb54dabeb92a8e7fafb261e0856e74aa26bfa1cfacf
  • Pointer size: 132 Bytes
  • Size of remote file: 1.42 MB
assets/render_benchmarks.py ADDED
@@ -0,0 +1,167 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ """Render public benchmark assets for TensorMind 1.5 Preview."""
3
+
4
+ from __future__ import annotations
5
+
6
+ import json
7
+ from pathlib import Path
8
+
9
+ import matplotlib as mpl
10
+ import matplotlib.pyplot as plt
11
+ import numpy as np
12
+ from matplotlib.offsetbox import AnnotationBbox, OffsetImage
13
+
14
+
15
+ ROOT = Path(__file__).resolve().parent
16
+ DATA = json.loads((ROOT / "benchmark-results.json").read_text())
17
+ LOGO = ROOT / "tensorplay-ai-logo.png"
18
+
19
+ BG = "#07111F"
20
+ PANEL = "#0C1A2C"
21
+ PANEL_ALT = "#10243A"
22
+ WHITE = "#F4F8FF"
23
+ MUTED = "#8EA5C3"
24
+ GRID = "#263A52"
25
+ CYAN = "#2ED7FF"
26
+ BLUE = "#267BFF"
27
+ ORANGE = "#FF9D42"
28
+
29
+ mpl.rcParams.update(
30
+ {
31
+ "font.family": "DejaVu Sans",
32
+ "axes.facecolor": BG,
33
+ "figure.facecolor": BG,
34
+ "savefig.facecolor": BG,
35
+ "text.color": WHITE,
36
+ "axes.labelcolor": MUTED,
37
+ "xtick.color": MUTED,
38
+ "ytick.color": MUTED,
39
+ "axes.edgecolor": GRID,
40
+ "svg.fonttype": "none",
41
+ }
42
+ )
43
+
44
+
45
+ def add_logo(fig: plt.Figure) -> None:
46
+ rgba = plt.imread(LOGO)
47
+ alpha = rgba[..., 3].copy() if rgba.shape[-1] == 4 else 1.0 - rgba[..., :3].mean(axis=-1)
48
+ white_logo = np.ones((*alpha.shape, 4), dtype=float)
49
+ white_logo[..., 3] = alpha
50
+ fig.add_artist(
51
+ AnnotationBbox(
52
+ OffsetImage(white_logo, zoom=0.105),
53
+ (0.84, 0.927),
54
+ xycoords="figure fraction",
55
+ frameon=False,
56
+ )
57
+ )
58
+
59
+
60
+ def header(fig: plt.Figure, title: str, subtitle: str) -> None:
61
+ fig.text(0.055, 0.905, "TensorMind 1.5 Preview", fontsize=13, color=CYAN, weight="bold")
62
+ fig.text(0.055, 0.843, title, fontsize=28, weight="bold")
63
+ fig.text(0.055, 0.792, subtitle, fontsize=11, color=MUTED)
64
+ add_logo(fig)
65
+
66
+
67
+ def footer(fig: plt.Figure) -> None:
68
+ fig.text(
69
+ 0.055,
70
+ 0.055,
71
+ "Protocol lm-eval 0.4.12 · SGLang 0.5.14 · 0-shot · full datasets · batch 48 · fixed seeds",
72
+ fontsize=8.5,
73
+ color=MUTED,
74
+ )
75
+ fig.text(0.955, 0.055, "Accuracy, higher is better", ha="right", fontsize=8.5, color=MUTED)
76
+
77
+
78
+ def save(fig: plt.Figure, name: str) -> None:
79
+ fig.savefig(ROOT / f"{name}.png", dpi=200)
80
+ fig.savefig(ROOT / f"{name}.svg")
81
+ plt.close(fig)
82
+
83
+
84
+ def render_suite() -> None:
85
+ results = DATA["results"]
86
+ metrics = [
87
+ ("CMMLU", results["cmmlu"]),
88
+ ("AGIEval-CN", results["agieval_cn"]),
89
+ ("A-CLUE", results["a_clue"]),
90
+ ("C-Eval", results["c_eval"]),
91
+ ("TMMLU+", results["tmmlu_plus"]),
92
+ ]
93
+
94
+ fig = plt.figure(figsize=(14, 7.875), dpi=200)
95
+ header(fig, "Chinese benchmark suite", "Five full-dataset evaluations under one matched zero-shot protocol")
96
+ gs = fig.add_gridspec(1, 12, left=0.095, right=0.955, top=0.72, bottom=0.13, wspace=1.2)
97
+
98
+ ax = fig.add_subplot(gs[0, :8])
99
+ ax.set_facecolor(PANEL)
100
+ for spine in ax.spines.values():
101
+ spine.set_visible(False)
102
+ names = [name for name, _ in metrics][::-1]
103
+ values = [value for _, value in metrics][::-1]
104
+ y = np.arange(len(metrics))
105
+ ax.barh(y, values, height=0.46, color=[BLUE, CYAN, BLUE, CYAN, BLUE], alpha=0.95)
106
+ ax.set_xlim(0, 35)
107
+ ax.set_yticks(y, names, fontsize=10.5)
108
+ ax.set_xticks([0, 10, 20, 30])
109
+ ax.tick_params(axis="both", length=0, pad=10)
110
+ ax.grid(axis="x", color=GRID, linewidth=0.8, alpha=0.75)
111
+ ax.set_axisbelow(True)
112
+ for yi, value in enumerate(values):
113
+ ax.text(value + 0.45, yi, f"{value:.4f}", va="center", fontsize=10, color=WHITE, weight="bold")
114
+ ax.set_xlabel("Accuracy (%)", loc="right", fontsize=9, labelpad=10)
115
+
116
+ ax_card = fig.add_subplot(gs[0, 9:])
117
+ ax_card.set_facecolor(PANEL_ALT)
118
+ ax_card.set_xticks([])
119
+ ax_card.set_yticks([])
120
+ for spine in ax_card.spines.values():
121
+ spine.set_visible(False)
122
+ ax_card.text(0.10, 0.86, "FIVE-SUITE MACRO", fontsize=8.5, color=CYAN, weight="bold", transform=ax_card.transAxes)
123
+ ax_card.text(0.10, 0.64, f"{results['five_suite_macro']:.4f}", fontsize=36, color=WHITE, weight="bold", transform=ax_card.transAxes)
124
+ ax_card.text(0.10, 0.53, "full-dataset accuracy", fontsize=9.5, color=MUTED, transform=ax_card.transAxes)
125
+ ax_card.plot([0.10, 0.90], [0.43, 0.43], color=GRID, linewidth=1.0, transform=ax_card.transAxes)
126
+ ax_card.text(0.10, 0.32, "5", fontsize=19, color=CYAN, weight="bold", transform=ax_card.transAxes)
127
+ ax_card.text(0.21, 0.33, "benchmark suites", fontsize=9.5, color=MUTED, transform=ax_card.transAxes)
128
+ ax_card.text(0.10, 0.18, "0-shot", fontsize=19, color=ORANGE, weight="bold", transform=ax_card.transAxes)
129
+ ax_card.text(0.52, 0.19, "matched protocol", fontsize=9.5, color=MUTED, transform=ax_card.transAxes)
130
+
131
+ footer(fig)
132
+ save(fig, "benchmark-suite")
133
+
134
+
135
+ def render_scorecard() -> None:
136
+ results = DATA["results"]
137
+ cards = [
138
+ ("CMMLU", results["cmmlu"], BLUE),
139
+ ("AGIEval-CN", results["agieval_cn"], CYAN),
140
+ ("A-CLUE", results["a_clue"], BLUE),
141
+ ("C-Eval", results["c_eval"], ORANGE),
142
+ ("TMMLU+", results["tmmlu_plus"], CYAN),
143
+ ("5-suite macro", results["five_suite_macro"], WHITE),
144
+ ]
145
+
146
+ fig = plt.figure(figsize=(14, 7.875), dpi=200)
147
+ header(fig, "Benchmark scorecard", "TensorMind 1.5 Preview · full-dataset accuracy (%)")
148
+ gs = fig.add_gridspec(2, 3, left=0.08, right=0.92, top=0.70, bottom=0.18, wspace=0.10, hspace=0.14)
149
+ for idx, (name, value, color) in enumerate(cards):
150
+ ax = fig.add_subplot(gs[idx // 3, idx % 3])
151
+ ax.set_facecolor(PANEL_ALT if idx == 5 else PANEL)
152
+ ax.set_xticks([])
153
+ ax.set_yticks([])
154
+ for spine in ax.spines.values():
155
+ spine.set_visible(False)
156
+ ax.add_patch(plt.Rectangle((0.0, 0.0), 0.018, 1.0, color=color, transform=ax.transAxes, lw=0))
157
+ ax.text(0.09, 0.70, name.upper(), fontsize=9, color=MUTED, weight="bold", transform=ax.transAxes)
158
+ ax.text(0.09, 0.29, f"{value:.4f}", fontsize=27, color=color, weight="bold", transform=ax.transAxes)
159
+ ax.text(0.09, 0.12, "accuracy", fontsize=8.5, color=MUTED, transform=ax.transAxes)
160
+
161
+ footer(fig)
162
+ save(fig, "benchmark-matrix")
163
+
164
+
165
+ if __name__ == "__main__":
166
+ render_suite()
167
+ render_scorecard()
assets/tensorplay-ai-logo.png ADDED

Git LFS Details

  • SHA256: 2a34b213f9f497fdfd49dfc11d452e19c05caa9b9fc21c8d39ece8860b0d4990
  • Pointer size: 131 Bytes
  • Size of remote file: 185 kB
chat_template.jinja ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {{- messages[0].content + '\n\n' }}
5
+ {%- endif %}
6
+ {{- "# Tools"
7
+ "\n\nYou may call one or more functions to assist with the user query."
8
+ "\n\nYou are provided with function signatures within <tools></tools> XML tags:"
9
+ "\n<tools>"
10
+ }}
11
+ {%- for tool in tools %}
12
+ {{- "\n" }}
13
+ {{- tool | tojson }}
14
+ {%- endfor %}
15
+ {{- "\n</tools>"
16
+ "\n\nFor each function call, return a json object with function name "
17
+ "and arguments within <tool_call></tool_call> XML tags:"
18
+ "\n<tool_call>"
19
+ "\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n"
20
+ }}
21
+ {%- else %}
22
+ {%- if messages[0].role == 'system' %}
23
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
24
+ {%- endif %}
25
+ {%- endif %}
26
+ {# record the last query index #}
27
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
28
+ {%- for message in messages[::-1] %}
29
+ {%- set reverse_index = (messages|length - 1) - loop.index0 %}
30
+ {%- if ns.multi_step_tool and message.role == "user" and message.content is string
31
+ and not(message.content.startswith('<tool_response>')
32
+ and message.content.endswith('</tool_response>'))
33
+ %}
34
+ {%- set ns.multi_step_tool = false %}
35
+ {%- set ns.last_query_index = reverse_index %}
36
+ {%- endif %}
37
+ {%- endfor %}
38
+ {%- for message in messages %}
39
+ {%- if message.content is string %}
40
+ {%- set content = message.content %}
41
+ {%- else %}
42
+ {%- set content = '' %}
43
+ {%- endif %}
44
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
45
+ {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
46
+ {%- elif message.role == "assistant" %}
47
+ {%- set reasoning_content = '' %}
48
+ {%- if message.reasoning_content is string %}
49
+ {%- set reasoning_content = message.reasoning_content %}
50
+ {%- else %}
51
+ {%- if '</think>' in content %}
52
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
53
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
54
+ {%- endif %}
55
+ {%- endif %}
56
+ {%- if loop.index0 > ns.last_query_index %}
57
+ {%- if loop.last or reasoning_content %}
58
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
59
+ {%- else %}
60
+ {{- '<|im_start|>' + message.role + '\n' + content }}
61
+ {%- endif %}
62
+ {%- else %}
63
+ {{- '<|im_start|>' + message.role + '\n' + content }}
64
+ {%- endif %}
65
+ {%- if message.tool_calls %}
66
+ {%- for tool_call in message.tool_calls %}
67
+ {%- if content or (not loop.first) %}
68
+ {{- '\n' }}
69
+ {%- endif %}
70
+ {%- if tool_call.function %}
71
+ {%- set tool_call = tool_call.function %}
72
+ {%- endif %}
73
+ {{- '<tool_call>\n{"name": "' }}
74
+ {{- tool_call.name }}
75
+ {{- '", "arguments": ' }}
76
+ {%- if tool_call.arguments is string %}
77
+ {{- tool_call.arguments }}
78
+ {%- else %}
79
+ {{- tool_call.arguments | tojson }}
80
+ {%- endif %}
81
+ {{- '}\n</tool_call>' }}
82
+ {%- endfor %}
83
+ {%- endif %}
84
+ {{- '<|im_end|>\n' }}
85
+ {%- elif message.role == "tool" %}
86
+ {# Short circuiting avoids out-of-bounds #}
87
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
88
+ {{- '<|im_start|>user' }}
89
+ {%- endif %}
90
+ {{- '\n<tool_response>\n' }}
91
+ {{- content }}
92
+ {{- '\n</tool_response>' }}
93
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
94
+ {{- '<|im_end|>\n' }}
95
+ {%- endif %}
96
+ {%- endif %}
97
+ {%- endfor %}
98
+ {%- if add_generation_prompt %}
99
+ {{- '<|im_start|>assistant\n' }}
100
+ {%- if enable_thinking is defined and enable_thinking is false %}
101
+ {{- '<think>\n\n</think>\n\n' }}
102
+ {%- endif %}
103
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,67 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "dtype": "bfloat16",
9
+ "eos_token_id": 2,
10
+ "head_dim": 64,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 1024,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 4096,
15
+ "layer_types": [
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention"
48
+ ],
49
+ "max_position_embeddings": 10240,
50
+ "max_window_layers": 28,
51
+ "model_type": "qwen3",
52
+ "num_attention_heads": 16,
53
+ "num_hidden_layers": 32,
54
+ "num_key_value_heads": 8,
55
+ "pad_token_id": 0,
56
+ "rms_norm_eps": 1e-06,
57
+ "rope_parameters": {
58
+ "rope_theta": 10000.0,
59
+ "rope_type": "default"
60
+ },
61
+ "sliding_window": null,
62
+ "tie_word_embeddings": true,
63
+ "transformers_version": "5.8.1",
64
+ "use_cache": false,
65
+ "use_sliding_window": false,
66
+ "vocab_size": 32768
67
+ }
generation_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": [
5
+ 2
6
+ ],
7
+ "output_attentions": false,
8
+ "output_hidden_states": false,
9
+ "pad_token_id": 0,
10
+ "transformers_version": "5.8.1",
11
+ "use_cache": true
12
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b99f6c0448831a6cb2012b93793dd72c9e64e94deee15fd8c4d8206c056817d9
3
+ size 1073923160
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|im_start|>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "is_local": true,
9
+ "local_files_only": false,
10
+ "model_max_length": 32768,
11
+ "pad_token": "<|endoftext|>",
12
+ "padding_side": "left",
13
+ "split_special_tokens": false,
14
+ "tokenizer_class": "TokenizersBackend",
15
+ "truncation_side": "left",
16
+ "unk_token": null
17
+ }