adeelahmad commited on
Commit
b7e7e67
·
verified ·
1 Parent(s): fe10eb4

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. .DS_Store +0 -0
  2. Modelfile +1 -1
  3. Modelfile3b +255 -0
  4. Modelfile4b +255 -0
  5. training_state.json +833 -0
.DS_Store ADDED
Binary file (6.15 kB). View file
 
Modelfile CHANGED
@@ -1,4 +1,4 @@
1
- From ./ReasonableQwen3-4B-Q4_0.gguf
2
 
3
  TEMPLATE """{{- if .Messages }}
4
  {{- if or .System .Tools }}<|im_start|>system
 
1
+ From ./ReasonableQwen3-4B-Q8_0.gguf
2
 
3
  TEMPLATE """{{- if .Messages }}
4
  {{- if or .System .Tools }}<|im_start|>system
Modelfile3b ADDED
@@ -0,0 +1,255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ From ./ReasonableQwen3-4B-Q3_K.gguf
2
+
3
+ TEMPLATE """{{- if .Messages }}
4
+ {{- if or .System .Tools }}<|im_start|>system
5
+ {{- if .System }}
6
+ {{ .System }}
7
+ {{- end }}
8
+ {{- if .Tools }}
9
+
10
+ # Tools
11
+
12
+ You may call one or more functions to assist with the user query.
13
+
14
+ You are provided with function signatures within <tools></tools> XML tags:
15
+ <tools>
16
+ {{- range .Tools }}
17
+ {"type": "function", "function": {{ .Function }}}
18
+ {{- end }}
19
+ </tools>
20
+
21
+ For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:
22
+ <tool_call>
23
+ {"name": <function-name>, "arguments": <args-json-object>}
24
+ </tool_call>
25
+ {{- end }}<|im_end|>
26
+ {{ end }}
27
+ {{- range $i, $_ := .Messages }}
28
+ {{- $last := eq (len (slice $.Messages $i)) 1 -}}
29
+ {{- if eq .Role "user" }}<|im_start|>user
30
+ {{ .Content }}<|im_end|>
31
+ {{ else if eq .Role "assistant" }}<|im_start|>assistant
32
+ {{ if .Content }}{{ .Content }}
33
+ {{- else if .ToolCalls }}<tool_call>
34
+ {{ range .ToolCalls }}{"name": "{{ .Function.Name }}", "arguments": {{ .Function.Arguments }}}
35
+ {{ end }}</tool_call>
36
+ {{- end }}{{ if not $last }}<|im_end|>
37
+ {{ end }}
38
+ {{- else if eq .Role "tool" }}<|im_start|>user
39
+ <tool_response>
40
+ {{ .Content }}
41
+ </tool_response><|im_end|>
42
+ {{ end }}
43
+ {{- if and (ne .Role "assistant") $last }}<|im_start|>assistant
44
+ {{ end }}
45
+ {{- end }}
46
+ {{- else }}
47
+ {{- if .System }}<|im_start|>system
48
+ {{ .System }}<|im_end|>
49
+ {{ end }}{{ if .Prompt }}<|im_start|>user
50
+ {{ .Prompt }}<|im_end|>
51
+ {{ end }}<|im_start|>assistant
52
+ {{ end }}{{ .Response }}{{ if .Response }}<|im_end|>{{ end }}"""
53
+ PARAMETER repeat_penalty 1
54
+ PARAMETER stop <|im_start|>
55
+ PARAMETER stop <|im_end|>
56
+ PARAMETER temperature 0.6
57
+ PARAMETER top_k 20
58
+ PARAMETER top_p 0.95
59
+ LICENSE """ Apache License
60
+ Version 2.0, January 2004
61
+ http://www.apache.org/licenses/
62
+
63
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
64
+
65
+ 1. Definitions.
66
+
67
+ "License" shall mean the terms and conditions for use, reproduction,
68
+ and distribution as defined by Sections 1 through 9 of this document.
69
+
70
+ "Licensor" shall mean the copyright owner or entity authorized by
71
+ the copyright owner that is granting the License.
72
+
73
+ "Legal Entity" shall mean the union of the acting entity and all
74
+ other entities that control, are controlled by, or are under common
75
+ control with that entity. For the purposes of this definition,
76
+ "control" means (i) the power, direct or indirect, to cause the
77
+ direction or management of such entity, whether by contract or
78
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
79
+ outstanding shares, or (iii) beneficial ownership of such entity.
80
+
81
+ "You" (or "Your") shall mean an individual or Legal Entity
82
+ exercising permissions granted by this License.
83
+
84
+ "Source" form shall mean the preferred form for making modifications,
85
+ including but not limited to software source code, documentation
86
+ source, and configuration files.
87
+
88
+ "Object" form shall mean any form resulting from mechanical
89
+ transformation or translation of a Source form, including but
90
+ not limited to compiled object code, generated documentation,
91
+ and conversions to other media types.
92
+
93
+ "Work" shall mean the work of authorship, whether in Source or
94
+ Object form, made available under the License, as indicated by a
95
+ copyright notice that is included in or attached to the work
96
+ (an example is provided in the Appendix below).
97
+
98
+ "Derivative Works" shall mean any work, whether in Source or Object
99
+ form, that is based on (or derived from) the Work and for which the
100
+ editorial revisions, annotations, elaborations, or other modifications
101
+ represent, as a whole, an original work of authorship. For the purposes
102
+ of this License, Derivative Works shall not include works that remain
103
+ separable from, or merely link (or bind by name) to the interfaces of,
104
+ the Work and Derivative Works thereof.
105
+
106
+ "Contribution" shall mean any work of authorship, including
107
+ the original version of the Work and any modifications or additions
108
+ to that Work or Derivative Works thereof, that is intentionally
109
+ submitted to Licensor for inclusion in the Work by the copyright owner
110
+ or by an individual or Legal Entity authorized to submit on behalf of
111
+ the copyright owner. For the purposes of this definition, "submitted"
112
+ means any form of electronic, verbal, or written communication sent
113
+ to the Licensor or its representatives, including but not limited to
114
+ communication on electronic mailing lists, source code control systems,
115
+ and issue tracking systems that are managed by, or on behalf of, the
116
+ Licensor for the purpose of discussing and improving the Work, but
117
+ excluding communication that is conspicuously marked or otherwise
118
+ designated in writing by the copyright owner as "Not a Contribution."
119
+
120
+ "Contributor" shall mean Licensor and any individual or Legal Entity
121
+ on behalf of whom a Contribution has been received by Licensor and
122
+ subsequently incorporated within the Work.
123
+
124
+ 2. Grant of Copyright License. Subject to the terms and conditions of
125
+ this License, each Contributor hereby grants to You a perpetual,
126
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
127
+ copyright license to reproduce, prepare Derivative Works of,
128
+ publicly display, publicly perform, sublicense, and distribute the
129
+ Work and such Derivative Works in Source or Object form.
130
+
131
+ 3. Grant of Patent License. Subject to the terms and conditions of
132
+ this License, each Contributor hereby grants to You a perpetual,
133
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
134
+ (except as stated in this section) patent license to make, have made,
135
+ use, offer to sell, sell, import, and otherwise transfer the Work,
136
+ where such license applies only to those patent claims licensable
137
+ by such Contributor that are necessarily infringed by their
138
+ Contribution(s) alone or by combination of their Contribution(s)
139
+ with the Work to which such Contribution(s) was submitted. If You
140
+ institute patent litigation against any entity (including a
141
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
142
+ or a Contribution incorporated within the Work constitutes direct
143
+ or contributory patent infringement, then any patent licenses
144
+ granted to You under this License for that Work shall terminate
145
+ as of the date such litigation is filed.
146
+
147
+ 4. Redistribution. You may reproduce and distribute copies of the
148
+ Work or Derivative Works thereof in any medium, with or without
149
+ modifications, and in Source or Object form, provided that You
150
+ meet the following conditions:
151
+
152
+ (a) You must give any other recipients of the Work or
153
+ Derivative Works a copy of this License; and
154
+
155
+ (b) You must cause any modified files to carry prominent notices
156
+ stating that You changed the files; and
157
+
158
+ (c) You must retain, in the Source form of any Derivative Works
159
+ that You distribute, all copyright, patent, trademark, and
160
+ attribution notices from the Source form of the Work,
161
+ excluding those notices that do not pertain to any part of
162
+ the Derivative Works; and
163
+
164
+ (d) If the Work includes a "NOTICE" text file as part of its
165
+ distribution, then any Derivative Works that You distribute must
166
+ include a readable copy of the attribution notices contained
167
+ within such NOTICE file, excluding those notices that do not
168
+ pertain to any part of the Derivative Works, in at least one
169
+ of the following places: within a NOTICE text file distributed
170
+ as part of the Derivative Works; within the Source form or
171
+ documentation, if provided along with the Derivative Works; or,
172
+ within a display generated by the Derivative Works, if and
173
+ wherever such third-party notices normally appear. The contents
174
+ of the NOTICE file are for informational purposes only and
175
+ do not modify the License. You may add Your own attribution
176
+ notices within Derivative Works that You distribute, alongside
177
+ or as an addendum to the NOTICE text from the Work, provided
178
+ that such additional attribution notices cannot be construed
179
+ as modifying the License.
180
+
181
+ You may add Your own copyright statement to Your modifications and
182
+ may provide additional or different license terms and conditions
183
+ for use, reproduction, or distribution of Your modifications, or
184
+ for any such Derivative Works as a whole, provided Your use,
185
+ reproduction, and distribution of the Work otherwise complies with
186
+ the conditions stated in this License.
187
+
188
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
189
+ any Contribution intentionally submitted for inclusion in the Work
190
+ by You to the Licensor shall be under the terms and conditions of
191
+ this License, without any additional terms or conditions.
192
+ Notwithstanding the above, nothing herein shall supersede or modify
193
+ the terms of any separate license agreement you may have executed
194
+ with Licensor regarding such Contributions.
195
+
196
+ 6. Trademarks. This License does not grant permission to use the trade
197
+ names, trademarks, service marks, or product names of the Licensor,
198
+ except as required for reasonable and customary use in describing the
199
+ origin of the Work and reproducing the content of the NOTICE file.
200
+
201
+ 7. Disclaimer of Warranty. Unless required by applicable law or
202
+ agreed to in writing, Licensor provides the Work (and each
203
+ Contributor provides its Contributions) on an "AS IS" BASIS,
204
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
205
+ implied, including, without limitation, any warranties or conditions
206
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
207
+ PARTICULAR PURPOSE. You are solely responsible for determining the
208
+ appropriateness of using or redistributing the Work and assume any
209
+ risks associated with Your exercise of permissions under this License.
210
+
211
+ 8. Limitation of Liability. In no event and under no legal theory,
212
+ whether in tort (including negligence), contract, or otherwise,
213
+ unless required by applicable law (such as deliberate and grossly
214
+ negligent acts) or agreed to in writing, shall any Contributor be
215
+ liable to You for damages, including any direct, indirect, special,
216
+ incidental, or consequential damages of any character arising as a
217
+ result of this License or out of the use or inability to use the
218
+ Work (including but not limited to damages for loss of goodwill,
219
+ work stoppage, computer failure or malfunction, or any and all
220
+ other commercial damages or losses), even if such Contributor
221
+ has been advised of the possibility of such damages.
222
+
223
+ 9. Accepting Warranty or Additional Liability. While redistributing
224
+ the Work or Derivative Works thereof, You may choose to offer,
225
+ and charge a fee for, acceptance of support, warranty, indemnity,
226
+ or other liability obligations and/or rights consistent with this
227
+ License. However, in accepting such obligations, You may act only
228
+ on Your own behalf and on Your sole responsibility, not on behalf
229
+ of any other Contributor, and only if You agree to indemnify,
230
+ defend, and hold each Contributor harmless for any liability
231
+ incurred by, or claims asserted against, such Contributor by reason
232
+ of your accepting any such warranty or additional liability.
233
+
234
+ END OF TERMS AND CONDITIONS
235
+
236
+ APPENDIX: How to apply the Apache License to your work.
237
+
238
+ To apply the Apache License to your work, attach the following
239
+ boilerplate notice, with the fields enclosed by brackets "[]"
240
+ replaced with your own identifying information. (Don't include
241
+ the brackets!) The text should be enclosed in the appropriate
242
+ comment syntax for the file format. We also recommend that a
243
+ file or class name and description of purpose be included on the
244
+ same "printed page" as the copyright notice for easier
245
+ identification within third-party archives.
246
+ Copyright 2024 Alibaba Cloud
247
+ Licensed under the Apache License, Version 2.0 (the "License");
248
+ you may not use this file except in compliance with the License.
249
+ You may obtain a copy of the License at
250
+ http://www.apache.org/licenses/LICENSE-2.0
251
+ Unless required by applicable law or agreed to in writing, software
252
+ distributed under the License is distributed on an "AS IS" BASIS,
253
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
254
+ See the License for the specific language governing permissions and
255
+ limitations under the License."""
Modelfile4b ADDED
@@ -0,0 +1,255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ From ./ReasonableQwen3-4B-Q4_K.gguf
2
+
3
+ TEMPLATE """{{- if .Messages }}
4
+ {{- if or .System .Tools }}<|im_start|>system
5
+ {{- if .System }}
6
+ {{ .System }}
7
+ {{- end }}
8
+ {{- if .Tools }}
9
+
10
+ # Tools
11
+
12
+ You may call one or more functions to assist with the user query.
13
+
14
+ You are provided with function signatures within <tools></tools> XML tags:
15
+ <tools>
16
+ {{- range .Tools }}
17
+ {"type": "function", "function": {{ .Function }}}
18
+ {{- end }}
19
+ </tools>
20
+
21
+ For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:
22
+ <tool_call>
23
+ {"name": <function-name>, "arguments": <args-json-object>}
24
+ </tool_call>
25
+ {{- end }}<|im_end|>
26
+ {{ end }}
27
+ {{- range $i, $_ := .Messages }}
28
+ {{- $last := eq (len (slice $.Messages $i)) 1 -}}
29
+ {{- if eq .Role "user" }}<|im_start|>user
30
+ {{ .Content }}<|im_end|>
31
+ {{ else if eq .Role "assistant" }}<|im_start|>assistant
32
+ {{ if .Content }}{{ .Content }}
33
+ {{- else if .ToolCalls }}<tool_call>
34
+ {{ range .ToolCalls }}{"name": "{{ .Function.Name }}", "arguments": {{ .Function.Arguments }}}
35
+ {{ end }}</tool_call>
36
+ {{- end }}{{ if not $last }}<|im_end|>
37
+ {{ end }}
38
+ {{- else if eq .Role "tool" }}<|im_start|>user
39
+ <tool_response>
40
+ {{ .Content }}
41
+ </tool_response><|im_end|>
42
+ {{ end }}
43
+ {{- if and (ne .Role "assistant") $last }}<|im_start|>assistant
44
+ {{ end }}
45
+ {{- end }}
46
+ {{- else }}
47
+ {{- if .System }}<|im_start|>system
48
+ {{ .System }}<|im_end|>
49
+ {{ end }}{{ if .Prompt }}<|im_start|>user
50
+ {{ .Prompt }}<|im_end|>
51
+ {{ end }}<|im_start|>assistant
52
+ {{ end }}{{ .Response }}{{ if .Response }}<|im_end|>{{ end }}"""
53
+ PARAMETER repeat_penalty 1
54
+ PARAMETER stop <|im_start|>
55
+ PARAMETER stop <|im_end|>
56
+ PARAMETER temperature 0.6
57
+ PARAMETER top_k 20
58
+ PARAMETER top_p 0.95
59
+ LICENSE """ Apache License
60
+ Version 2.0, January 2004
61
+ http://www.apache.org/licenses/
62
+
63
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
64
+
65
+ 1. Definitions.
66
+
67
+ "License" shall mean the terms and conditions for use, reproduction,
68
+ and distribution as defined by Sections 1 through 9 of this document.
69
+
70
+ "Licensor" shall mean the copyright owner or entity authorized by
71
+ the copyright owner that is granting the License.
72
+
73
+ "Legal Entity" shall mean the union of the acting entity and all
74
+ other entities that control, are controlled by, or are under common
75
+ control with that entity. For the purposes of this definition,
76
+ "control" means (i) the power, direct or indirect, to cause the
77
+ direction or management of such entity, whether by contract or
78
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
79
+ outstanding shares, or (iii) beneficial ownership of such entity.
80
+
81
+ "You" (or "Your") shall mean an individual or Legal Entity
82
+ exercising permissions granted by this License.
83
+
84
+ "Source" form shall mean the preferred form for making modifications,
85
+ including but not limited to software source code, documentation
86
+ source, and configuration files.
87
+
88
+ "Object" form shall mean any form resulting from mechanical
89
+ transformation or translation of a Source form, including but
90
+ not limited to compiled object code, generated documentation,
91
+ and conversions to other media types.
92
+
93
+ "Work" shall mean the work of authorship, whether in Source or
94
+ Object form, made available under the License, as indicated by a
95
+ copyright notice that is included in or attached to the work
96
+ (an example is provided in the Appendix below).
97
+
98
+ "Derivative Works" shall mean any work, whether in Source or Object
99
+ form, that is based on (or derived from) the Work and for which the
100
+ editorial revisions, annotations, elaborations, or other modifications
101
+ represent, as a whole, an original work of authorship. For the purposes
102
+ of this License, Derivative Works shall not include works that remain
103
+ separable from, or merely link (or bind by name) to the interfaces of,
104
+ the Work and Derivative Works thereof.
105
+
106
+ "Contribution" shall mean any work of authorship, including
107
+ the original version of the Work and any modifications or additions
108
+ to that Work or Derivative Works thereof, that is intentionally
109
+ submitted to Licensor for inclusion in the Work by the copyright owner
110
+ or by an individual or Legal Entity authorized to submit on behalf of
111
+ the copyright owner. For the purposes of this definition, "submitted"
112
+ means any form of electronic, verbal, or written communication sent
113
+ to the Licensor or its representatives, including but not limited to
114
+ communication on electronic mailing lists, source code control systems,
115
+ and issue tracking systems that are managed by, or on behalf of, the
116
+ Licensor for the purpose of discussing and improving the Work, but
117
+ excluding communication that is conspicuously marked or otherwise
118
+ designated in writing by the copyright owner as "Not a Contribution."
119
+
120
+ "Contributor" shall mean Licensor and any individual or Legal Entity
121
+ on behalf of whom a Contribution has been received by Licensor and
122
+ subsequently incorporated within the Work.
123
+
124
+ 2. Grant of Copyright License. Subject to the terms and conditions of
125
+ this License, each Contributor hereby grants to You a perpetual,
126
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
127
+ copyright license to reproduce, prepare Derivative Works of,
128
+ publicly display, publicly perform, sublicense, and distribute the
129
+ Work and such Derivative Works in Source or Object form.
130
+
131
+ 3. Grant of Patent License. Subject to the terms and conditions of
132
+ this License, each Contributor hereby grants to You a perpetual,
133
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
134
+ (except as stated in this section) patent license to make, have made,
135
+ use, offer to sell, sell, import, and otherwise transfer the Work,
136
+ where such license applies only to those patent claims licensable
137
+ by such Contributor that are necessarily infringed by their
138
+ Contribution(s) alone or by combination of their Contribution(s)
139
+ with the Work to which such Contribution(s) was submitted. If You
140
+ institute patent litigation against any entity (including a
141
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
142
+ or a Contribution incorporated within the Work constitutes direct
143
+ or contributory patent infringement, then any patent licenses
144
+ granted to You under this License for that Work shall terminate
145
+ as of the date such litigation is filed.
146
+
147
+ 4. Redistribution. You may reproduce and distribute copies of the
148
+ Work or Derivative Works thereof in any medium, with or without
149
+ modifications, and in Source or Object form, provided that You
150
+ meet the following conditions:
151
+
152
+ (a) You must give any other recipients of the Work or
153
+ Derivative Works a copy of this License; and
154
+
155
+ (b) You must cause any modified files to carry prominent notices
156
+ stating that You changed the files; and
157
+
158
+ (c) You must retain, in the Source form of any Derivative Works
159
+ that You distribute, all copyright, patent, trademark, and
160
+ attribution notices from the Source form of the Work,
161
+ excluding those notices that do not pertain to any part of
162
+ the Derivative Works; and
163
+
164
+ (d) If the Work includes a "NOTICE" text file as part of its
165
+ distribution, then any Derivative Works that You distribute must
166
+ include a readable copy of the attribution notices contained
167
+ within such NOTICE file, excluding those notices that do not
168
+ pertain to any part of the Derivative Works, in at least one
169
+ of the following places: within a NOTICE text file distributed
170
+ as part of the Derivative Works; within the Source form or
171
+ documentation, if provided along with the Derivative Works; or,
172
+ within a display generated by the Derivative Works, if and
173
+ wherever such third-party notices normally appear. The contents
174
+ of the NOTICE file are for informational purposes only and
175
+ do not modify the License. You may add Your own attribution
176
+ notices within Derivative Works that You distribute, alongside
177
+ or as an addendum to the NOTICE text from the Work, provided
178
+ that such additional attribution notices cannot be construed
179
+ as modifying the License.
180
+
181
+ You may add Your own copyright statement to Your modifications and
182
+ may provide additional or different license terms and conditions
183
+ for use, reproduction, or distribution of Your modifications, or
184
+ for any such Derivative Works as a whole, provided Your use,
185
+ reproduction, and distribution of the Work otherwise complies with
186
+ the conditions stated in this License.
187
+
188
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
189
+ any Contribution intentionally submitted for inclusion in the Work
190
+ by You to the Licensor shall be under the terms and conditions of
191
+ this License, without any additional terms or conditions.
192
+ Notwithstanding the above, nothing herein shall supersede or modify
193
+ the terms of any separate license agreement you may have executed
194
+ with Licensor regarding such Contributions.
195
+
196
+ 6. Trademarks. This License does not grant permission to use the trade
197
+ names, trademarks, service marks, or product names of the Licensor,
198
+ except as required for reasonable and customary use in describing the
199
+ origin of the Work and reproducing the content of the NOTICE file.
200
+
201
+ 7. Disclaimer of Warranty. Unless required by applicable law or
202
+ agreed to in writing, Licensor provides the Work (and each
203
+ Contributor provides its Contributions) on an "AS IS" BASIS,
204
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
205
+ implied, including, without limitation, any warranties or conditions
206
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
207
+ PARTICULAR PURPOSE. You are solely responsible for determining the
208
+ appropriateness of using or redistributing the Work and assume any
209
+ risks associated with Your exercise of permissions under this License.
210
+
211
+ 8. Limitation of Liability. In no event and under no legal theory,
212
+ whether in tort (including negligence), contract, or otherwise,
213
+ unless required by applicable law (such as deliberate and grossly
214
+ negligent acts) or agreed to in writing, shall any Contributor be
215
+ liable to You for damages, including any direct, indirect, special,
216
+ incidental, or consequential damages of any character arising as a
217
+ result of this License or out of the use or inability to use the
218
+ Work (including but not limited to damages for loss of goodwill,
219
+ work stoppage, computer failure or malfunction, or any and all
220
+ other commercial damages or losses), even if such Contributor
221
+ has been advised of the possibility of such damages.
222
+
223
+ 9. Accepting Warranty or Additional Liability. While redistributing
224
+ the Work or Derivative Works thereof, You may choose to offer,
225
+ and charge a fee for, acceptance of support, warranty, indemnity,
226
+ or other liability obligations and/or rights consistent with this
227
+ License. However, in accepting such obligations, You may act only
228
+ on Your own behalf and on Your sole responsibility, not on behalf
229
+ of any other Contributor, and only if You agree to indemnify,
230
+ defend, and hold each Contributor harmless for any liability
231
+ incurred by, or claims asserted against, such Contributor by reason
232
+ of your accepting any such warranty or additional liability.
233
+
234
+ END OF TERMS AND CONDITIONS
235
+
236
+ APPENDIX: How to apply the Apache License to your work.
237
+
238
+ To apply the Apache License to your work, attach the following
239
+ boilerplate notice, with the fields enclosed by brackets "[]"
240
+ replaced with your own identifying information. (Don't include
241
+ the brackets!) The text should be enclosed in the appropriate
242
+ comment syntax for the file format. We also recommend that a
243
+ file or class name and description of purpose be included on the
244
+ same "printed page" as the copyright notice for easier
245
+ identification within third-party archives.
246
+ Copyright 2024 Alibaba Cloud
247
+ Licensed under the Apache License, Version 2.0 (the "License");
248
+ you may not use this file except in compliance with the License.
249
+ You may obtain a copy of the License at
250
+ http://www.apache.org/licenses/LICENSE-2.0
251
+ Unless required by applicable law or agreed to in writing, software
252
+ distributed under the License is distributed on an "AS IS" BASIS,
253
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
254
+ See the License for the specific language governing permissions and
255
+ limitations under the License."""
training_state.json ADDED
@@ -0,0 +1,833 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "global_step": 640,
3
+ "num_updates": 640,
4
+ "use_lora": false,
5
+ "rng_state": {
6
+ "python": [
7
+ 3,
8
+ [
9
+ 3100233166,
10
+ 3205251947,
11
+ 2114003297,
12
+ 4083397361,
13
+ 3944152784,
14
+ 1030840702,
15
+ 2037797290,
16
+ 2255135174,
17
+ 3606282055,
18
+ 1805506540,
19
+ 1511271408,
20
+ 1355721171,
21
+ 3219242983,
22
+ 2147281139,
23
+ 1663895881,
24
+ 2744272328,
25
+ 586124618,
26
+ 2904233013,
27
+ 1951649306,
28
+ 568843218,
29
+ 804729164,
30
+ 1031240685,
31
+ 632815444,
32
+ 1583468126,
33
+ 3569217402,
34
+ 1474578133,
35
+ 1620621069,
36
+ 4081348163,
37
+ 660493223,
38
+ 548758843,
39
+ 3288411447,
40
+ 2761565335,
41
+ 651306001,
42
+ 3364357190,
43
+ 574611538,
44
+ 3080366693,
45
+ 648102770,
46
+ 662716411,
47
+ 65202676,
48
+ 1627080032,
49
+ 2475536053,
50
+ 939008240,
51
+ 1513946789,
52
+ 2550074718,
53
+ 3473906356,
54
+ 836348979,
55
+ 396042796,
56
+ 2909821569,
57
+ 2791572300,
58
+ 3313263004,
59
+ 434385456,
60
+ 2799243929,
61
+ 1110673623,
62
+ 10708347,
63
+ 1193947720,
64
+ 4154063907,
65
+ 3769310438,
66
+ 1825986973,
67
+ 1865856087,
68
+ 2550327576,
69
+ 635176254,
70
+ 343483848,
71
+ 627331621,
72
+ 4032575651,
73
+ 912584999,
74
+ 2184863449,
75
+ 2003242393,
76
+ 191158593,
77
+ 4027956150,
78
+ 135767775,
79
+ 165725602,
80
+ 2940340686,
81
+ 1735550527,
82
+ 2166976964,
83
+ 2935306830,
84
+ 1440736616,
85
+ 2136544928,
86
+ 1238006856,
87
+ 904424546,
88
+ 4116526595,
89
+ 4005584629,
90
+ 2601760449,
91
+ 271648674,
92
+ 139094063,
93
+ 3602969447,
94
+ 414504323,
95
+ 1053832962,
96
+ 4134761489,
97
+ 3249289190,
98
+ 1358383408,
99
+ 195832611,
100
+ 1843593352,
101
+ 2969447193,
102
+ 3158396779,
103
+ 3569807150,
104
+ 4160044175,
105
+ 2081938636,
106
+ 1429144436,
107
+ 385540495,
108
+ 4178096289,
109
+ 1917190049,
110
+ 2474710463,
111
+ 3123380742,
112
+ 2902614650,
113
+ 845733750,
114
+ 728346900,
115
+ 1111266260,
116
+ 1230890723,
117
+ 3960864409,
118
+ 1864677806,
119
+ 2465725907,
120
+ 1846265707,
121
+ 2510053619,
122
+ 4283547983,
123
+ 2002390956,
124
+ 3488389929,
125
+ 391769952,
126
+ 3822866039,
127
+ 1599671852,
128
+ 501962155,
129
+ 774437297,
130
+ 542827104,
131
+ 36072073,
132
+ 443060080,
133
+ 689178575,
134
+ 86655597,
135
+ 825092402,
136
+ 625421810,
137
+ 1794213612,
138
+ 1871893587,
139
+ 3436665270,
140
+ 631289196,
141
+ 3067538162,
142
+ 552776492,
143
+ 627440404,
144
+ 2105634472,
145
+ 548648577,
146
+ 636698996,
147
+ 947106975,
148
+ 3431407759,
149
+ 2240487779,
150
+ 3128108668,
151
+ 521416466,
152
+ 2871250081,
153
+ 1062946757,
154
+ 1608040378,
155
+ 2590414279,
156
+ 3358166288,
157
+ 1253809383,
158
+ 3726438328,
159
+ 281309528,
160
+ 1669515180,
161
+ 3273041510,
162
+ 2260421329,
163
+ 996425797,
164
+ 292515881,
165
+ 2296689583,
166
+ 2805798357,
167
+ 2032275171,
168
+ 1538312337,
169
+ 643565956,
170
+ 3359622205,
171
+ 3745536289,
172
+ 2003084236,
173
+ 3869804701,
174
+ 2724155627,
175
+ 2996948495,
176
+ 4184607962,
177
+ 4215169557,
178
+ 430613231,
179
+ 1647692967,
180
+ 1793331314,
181
+ 3874676231,
182
+ 2730461343,
183
+ 3955139318,
184
+ 3832613492,
185
+ 539666038,
186
+ 285792739,
187
+ 3198830397,
188
+ 4055078950,
189
+ 3569499391,
190
+ 781279195,
191
+ 2220425539,
192
+ 2039802440,
193
+ 1385766748,
194
+ 506267236,
195
+ 2271648199,
196
+ 1568145524,
197
+ 2554254342,
198
+ 758016416,
199
+ 3621980520,
200
+ 160767585,
201
+ 2188462111,
202
+ 1450007015,
203
+ 1105079920,
204
+ 3276242179,
205
+ 3640324766,
206
+ 2860750163,
207
+ 2143715191,
208
+ 1831279561,
209
+ 2293429158,
210
+ 739306744,
211
+ 2043528556,
212
+ 3882672526,
213
+ 546842616,
214
+ 2714277920,
215
+ 3887463926,
216
+ 1831957765,
217
+ 2229915849,
218
+ 2864792220,
219
+ 1303437412,
220
+ 1091254688,
221
+ 826344852,
222
+ 3120288320,
223
+ 556529779,
224
+ 811546242,
225
+ 1476206000,
226
+ 2630584339,
227
+ 2863085756,
228
+ 3547631685,
229
+ 787693537,
230
+ 4066961252,
231
+ 308882144,
232
+ 1599204914,
233
+ 582842064,
234
+ 2754105741,
235
+ 2202249626,
236
+ 3301080360,
237
+ 3074376018,
238
+ 1129640401,
239
+ 3543504945,
240
+ 1135643549,
241
+ 571814866,
242
+ 2951061707,
243
+ 2139761207,
244
+ 4279741136,
245
+ 676772801,
246
+ 3652048859,
247
+ 410817428,
248
+ 896263120,
249
+ 358049826,
250
+ 418456999,
251
+ 3572284105,
252
+ 2484345908,
253
+ 3844208687,
254
+ 3067647389,
255
+ 1302900203,
256
+ 931143928,
257
+ 1519365080,
258
+ 712789513,
259
+ 2595580658,
260
+ 1784776098,
261
+ 2848139510,
262
+ 3505469053,
263
+ 1582655988,
264
+ 4158436496,
265
+ 3330262652,
266
+ 2338444089,
267
+ 1002216810,
268
+ 484453692,
269
+ 3833856314,
270
+ 2002391132,
271
+ 519925212,
272
+ 2981935072,
273
+ 3077297433,
274
+ 239222133,
275
+ 1056108488,
276
+ 1645014848,
277
+ 2673367049,
278
+ 51217098,
279
+ 1778100088,
280
+ 1899905425,
281
+ 3987547773,
282
+ 1169518869,
283
+ 578030579,
284
+ 3367779903,
285
+ 956133751,
286
+ 1760174970,
287
+ 1650239768,
288
+ 2802504529,
289
+ 1103701391,
290
+ 121844885,
291
+ 3488652418,
292
+ 4134091454,
293
+ 3048567685,
294
+ 2857939785,
295
+ 2220271914,
296
+ 4024370836,
297
+ 931259945,
298
+ 2540064234,
299
+ 65277599,
300
+ 3847535535,
301
+ 990871212,
302
+ 1272682701,
303
+ 2391502819,
304
+ 3737351256,
305
+ 1335981297,
306
+ 2386770941,
307
+ 32918208,
308
+ 781820036,
309
+ 1166374003,
310
+ 2556455899,
311
+ 510298691,
312
+ 159187659,
313
+ 2597521830,
314
+ 291154034,
315
+ 3494951397,
316
+ 2944107260,
317
+ 3378737570,
318
+ 2839441569,
319
+ 770671866,
320
+ 4283771012,
321
+ 165153002,
322
+ 4069832089,
323
+ 2175511762,
324
+ 222931749,
325
+ 154419633,
326
+ 2538945322,
327
+ 3320641862,
328
+ 1807339955,
329
+ 3706204773,
330
+ 2765491230,
331
+ 1260202910,
332
+ 1924333259,
333
+ 4184170411,
334
+ 563038778,
335
+ 2507175419,
336
+ 1076853275,
337
+ 143035900,
338
+ 3902786410,
339
+ 1559868554,
340
+ 511327562,
341
+ 3462826369,
342
+ 3748998442,
343
+ 3845225973,
344
+ 2080353851,
345
+ 1489047293,
346
+ 1924679652,
347
+ 3795590602,
348
+ 3866562819,
349
+ 2198666709,
350
+ 1386047356,
351
+ 518409235,
352
+ 3428184188,
353
+ 1732776878,
354
+ 273783238,
355
+ 3723574330,
356
+ 1683692783,
357
+ 1523451822,
358
+ 3699825569,
359
+ 4127346852,
360
+ 3111140829,
361
+ 1605311581,
362
+ 4208002359,
363
+ 3118728618,
364
+ 3463176645,
365
+ 1575015318,
366
+ 3747413705,
367
+ 2206390158,
368
+ 2583654895,
369
+ 1354557377,
370
+ 2183517438,
371
+ 3254493976,
372
+ 3575291627,
373
+ 4028981055,
374
+ 3968296862,
375
+ 2729291170,
376
+ 247729344,
377
+ 650485877,
378
+ 1318877280,
379
+ 285023715,
380
+ 2202961239,
381
+ 4219685020,
382
+ 3779707729,
383
+ 917974987,
384
+ 4089069944,
385
+ 344146767,
386
+ 359806484,
387
+ 4275723208,
388
+ 703513891,
389
+ 4224074000,
390
+ 2009124160,
391
+ 604809477,
392
+ 4008576440,
393
+ 2427561320,
394
+ 2309325311,
395
+ 2362683196,
396
+ 235168397,
397
+ 1600648522,
398
+ 4215967893,
399
+ 1675054314,
400
+ 2013381593,
401
+ 1903651062,
402
+ 3601280500,
403
+ 496338735,
404
+ 2169133753,
405
+ 2364495966,
406
+ 1056246743,
407
+ 1421447090,
408
+ 1995033995,
409
+ 1869864803,
410
+ 698922377,
411
+ 2409323964,
412
+ 1510516618,
413
+ 641888605,
414
+ 1077564019,
415
+ 29567647,
416
+ 2785537387,
417
+ 1925896177,
418
+ 1227126769,
419
+ 2748886622,
420
+ 1332019659,
421
+ 638361144,
422
+ 540471184,
423
+ 2887245217,
424
+ 4182719198,
425
+ 884751481,
426
+ 3445591454,
427
+ 675466383,
428
+ 2087852801,
429
+ 2083882276,
430
+ 414620424,
431
+ 3713986203,
432
+ 2425060387,
433
+ 607871428,
434
+ 487946518,
435
+ 718902558,
436
+ 3786110198,
437
+ 1653762457,
438
+ 3787648973,
439
+ 431397113,
440
+ 4108771890,
441
+ 402186349,
442
+ 2471796344,
443
+ 4273010052,
444
+ 4248683,
445
+ 1754846221,
446
+ 2340586638,
447
+ 412924166,
448
+ 2306146182,
449
+ 2417039245,
450
+ 1511347613,
451
+ 1319385385,
452
+ 671950230,
453
+ 4188344613,
454
+ 1223621563,
455
+ 3353660436,
456
+ 1926932037,
457
+ 3945195786,
458
+ 61444401,
459
+ 3450186348,
460
+ 3982842641,
461
+ 2088756605,
462
+ 1824246909,
463
+ 2656102781,
464
+ 3750651125,
465
+ 3189574856,
466
+ 3078132247,
467
+ 588048676,
468
+ 1141387549,
469
+ 1187369381,
470
+ 750263661,
471
+ 1681491629,
472
+ 2025242095,
473
+ 3733996179,
474
+ 2457640372,
475
+ 956774679,
476
+ 1782785055,
477
+ 2881116689,
478
+ 1311179739,
479
+ 66826352,
480
+ 2111027409,
481
+ 271078370,
482
+ 99525850,
483
+ 3874156593,
484
+ 730907368,
485
+ 2126769770,
486
+ 2524302228,
487
+ 2128545293,
488
+ 1371594345,
489
+ 2849348044,
490
+ 2996788033,
491
+ 2591750469,
492
+ 3642360292,
493
+ 3727095772,
494
+ 655210820,
495
+ 1162591913,
496
+ 1003845048,
497
+ 439189944,
498
+ 3637207234,
499
+ 14122112,
500
+ 967296467,
501
+ 2664681396,
502
+ 2685037860,
503
+ 684065770,
504
+ 3082029956,
505
+ 2341710744,
506
+ 187950156,
507
+ 2219843456,
508
+ 3886109287,
509
+ 503801047,
510
+ 2410462727,
511
+ 2929691138,
512
+ 1032873297,
513
+ 1688752406,
514
+ 2759671101,
515
+ 3287456852,
516
+ 3986113072,
517
+ 1579528064,
518
+ 947167093,
519
+ 1051079192,
520
+ 1049610650,
521
+ 2231280468,
522
+ 726914012,
523
+ 1534264355,
524
+ 2904096012,
525
+ 4062734554,
526
+ 531616723,
527
+ 3561831970,
528
+ 737998148,
529
+ 3230395316,
530
+ 3022021563,
531
+ 251528766,
532
+ 3999173370,
533
+ 210218179,
534
+ 3088594306,
535
+ 1439401799,
536
+ 1041859435,
537
+ 2695195830,
538
+ 3465077144,
539
+ 2823089326,
540
+ 2556938793,
541
+ 1458448626,
542
+ 1308700694,
543
+ 444948412,
544
+ 2304470611,
545
+ 3654027317,
546
+ 1034742794,
547
+ 3810003506,
548
+ 1772353215,
549
+ 3041681477,
550
+ 3394149945,
551
+ 440869916,
552
+ 3892023657,
553
+ 556989640,
554
+ 2481299321,
555
+ 3529362628,
556
+ 3054249201,
557
+ 1737493058,
558
+ 1235777870,
559
+ 1334111971,
560
+ 2971707615,
561
+ 3873667589,
562
+ 1175895135,
563
+ 1940930396,
564
+ 3443474468,
565
+ 2788127120,
566
+ 3319749384,
567
+ 3328773738,
568
+ 1966987235,
569
+ 1065977210,
570
+ 2939975890,
571
+ 3943360814,
572
+ 3458298409,
573
+ 2200312840,
574
+ 2897384251,
575
+ 3453182268,
576
+ 875764420,
577
+ 2531405399,
578
+ 912921748,
579
+ 4283104359,
580
+ 2928545310,
581
+ 1224323513,
582
+ 722455736,
583
+ 3977634571,
584
+ 3582693887,
585
+ 1125205481,
586
+ 1576562126,
587
+ 3743288371,
588
+ 3499633816,
589
+ 735307326,
590
+ 281573363,
591
+ 1724151404,
592
+ 2589762545,
593
+ 2902873521,
594
+ 3471234108,
595
+ 2219383722,
596
+ 2061024598,
597
+ 814769644,
598
+ 3401335543,
599
+ 4190237183,
600
+ 650711606,
601
+ 4101800219,
602
+ 2172610273,
603
+ 3777607756,
604
+ 4042849157,
605
+ 2817574694,
606
+ 2095924374,
607
+ 1909727874,
608
+ 2477919892,
609
+ 2834537640,
610
+ 3637236485,
611
+ 1190839625,
612
+ 1838707007,
613
+ 3247768062,
614
+ 2294183597,
615
+ 1445875960,
616
+ 2333419384,
617
+ 335327851,
618
+ 3000828940,
619
+ 1726549769,
620
+ 4275198256,
621
+ 3227919305,
622
+ 3533288452,
623
+ 173416253,
624
+ 2262899688,
625
+ 3349430595,
626
+ 396380904,
627
+ 1372995131,
628
+ 3355176830,
629
+ 4107929401,
630
+ 3693751586,
631
+ 772959075,
632
+ 3498819257,
633
+ 370
634
+ ],
635
+ null
636
+ ],
637
+ "numpy": [
638
+ "MT19937",
639
+ "[ 8303 3362861772 1619186349 1762601439 1567969946 1320913452\n 808097991 3993093514 203964437 4072637778 354349567 380629158\n 3082602378 1492811957 2661358354 1249590623 2886239270 990864965\n 1924220491 1117017413 2869102568 114579815 997221561 1716370964\n 100154209 3193787998 982404710 1860404313 1073630932 4015033793\n 1056096168 3689842023 1077486740 3341954026 4020896783 3063022815\n 2319528021 1962385784 560690659 2930834870 2206559276 2815052879\n 130043787 3393045762 2815640209 452914988 784787594 2058459297\n 522114896 1804911297 1630115314 877409298 1852583246 102115168\n 3580881174 1417032320 1630533149 2372212037 1781883965 2270734311\n 186043029 4168788486 1870971191 457187469 1905478881 1029780641\n 2565532871 2731005692 869777530 2775988839 3249018655 3446872659\n 531509720 894321281 654198319 891934678 2363338170 2507473381\n 377446257 1382752484 1528924073 2070637977 1311806794 2451376618\n 1427771612 945417350 1522009908 3572053312 2253905607 977621266\n 3845758580 3469358158 3051280317 2090423736 2607666011 1809711996\n 853558705 2879273014 3672899814 1306126268 1713869557 1877839529\n 279523502 3666825997 459293422 2370684495 1189704139 1015349021\n 1120545757 2386561593 2421857461 4003382178 1622890485 3282519989\n 1952035136 738572312 2523822572 911767899 2136380509 1896989891\n 1469870850 3009084328 221856908 3441096631 270470784 1619212541\n 740841450 576514513 3632273653 1855946639 1695699336 3030088848\n 1996507166 646843328 2358534982 1183223387 2981778442 2176020401\n 3296124201 3815160861 2131117666 702136476 524319258 612243153\n 2880166405 2628502612 1647344768 2202609784 2245804726 1921295769\n 3525029262 4020694584 1240895711 3219125807 1163212123 1637955357\n 484227752 2106690789 2031000210 3514869662 1388324497 1214400369\n 1527848914 4203641314 1519812457 2741431469 3641575601 2047783137\n 1519626504 754352694 6964728 2468828035 3468360593 426009415\n 1756179121 1660553247 3150922374 2230969029 1482946613 414236983\n 3517164135 612220457 796850147 989153350 140311126 788581031\n 4132721565 1096190993 2894713612 2176107331 2878722659 2354637316\n 3799747358 2180737074 3963734962 2848855704 3734852998 2038033214\n 783779745 1197780044 1023946793 320860662 164699352 2983721475\n 17247281 533834018 3106467128 3903824561 1446671114 367871784\n 1942584218 1189132282 620423643 2803988284 1136015948 1345533496\n 2479351381 1216573484 1988507 303531010 754591654 3144907611\n 2902543355 318389020 3558821612 1847659564 1942545315 3246490061\n 456226090 509391991 3127370457 3143677006 2880949988 3448223655\n 3981326494 1148410844 2092401693 3604880377 3731870864 1860602350\n 3125695291 3187627886 2407156366 653809199 1969367167 450324651\n 3735867245 1764865373 3761162820 4171681276 2483404181 1541053582\n 3606137959 902867057 747499155 313576446 3101518390 1151122309\n 3268124694 261383756 1818167552 1213667178 3989930301 2726571133\n 2775483939 21599502 3727576464 2462926602 389566260 1711674513\n 3512151518 3153813056 460030746 3897140307 1277994146 908931170\n 963021246 1901922059 3955106312 1418220142 1659825635 3200240963\n 4014654655 485243719 1742562079 3526699507 3355114702 2632610560\n 1561948906 2717326040 3620881188 4126498438 184973469 1482988566\n 2950358329 2930115694 1859226564 3403387618 2053266671 3742107921\n 1209546566 3725840176 1712851789 2224824875 453012573 330532578\n 3778095708 1688965038 1230854207 2009005995 718461256 208618143\n 1491427571 2449392307 4002056719 383575031 2867208367 1963096702\n 1795247193 533084151 1491616947 1497652603 3562351716 86942950\n 2347771394 3289710405 466562532 1233657915 898220586 3748140507\n 2427008898 636905163 1254503011 1238782967 3895969628 4283203786\n 4221133213 2886931111 922986859 1879645066 2015511083 1108514279\n 618085140 3306423355 883925872 2144496777 1641126658 1153156234\n 2651844403 430020786 3696656792 683284102 3441596990 1725726834\n 2632006081 1159341650 1377340195 2886866895 3027966535 3917695904\n 968627127 734115740 3287543286 3074949140 537776154 2995193263\n 3574191023 2832961611 3705012029 3367494375 4248078694 1627669324\n 3788358613 3604335331 940239830 2942210789 1161729435 3740692795\n 2455348370 4133420619 1569119204 2972012502 2050815010 1235750478\n 3072257451 2550281774 856747230 1451377433 3438563836 1849610784\n 3357774219 1503802415 2962862 592596271 3943668757 2756336697\n 105838547 2324476620 3235171028 878843746 618966330 595747699\n 97437937 515195304 3376271580 1145290896 2696909771 3358253284\n 2241289147 1121727126 1309918765 1578024183 4054463658 720648522\n 1846097360 4230354452 729417651 791205440 1367656674 2065284658\n 257316547 3839072148 1555799865 4173515455 1048353236 4198966861\n 3316898928 3423897098 4128420921 2104108687 1063597492 2492958131\n 170555781 745349162 2255105092 2541678417 1114429555 2011457711\n 3454299228 2209782322 652117672 4201843201 2110523012 446611764\n 2926776128 2841467079 1337488503 559094349 3539441441 897438763\n 1650244025 570474331 1882207147 810573271 1456346777 1200505791\n 1603496382 284750116 3007277310 3591850295 3720918864 4052795532\n 1561428281 1050080487 2700034035 4080219366 1757769003 2977271141\n 1037953655 3360174536 2813217261 1188893730 2515840679 2105685490\n 820204985 1453315288 333847673 3528785562 4193735995 4021786871\n 4076203556 1854592580 3780409371 2202024795 1401490689 2819039973\n 2228184585 3518768958 1055269369 4073237030 1866198915 1516245557\n 2165376112 1043711207 669470737 3834818980 4270028755 3417085697\n 2968889276 98688489 2403155937 3448746372 3462613817 386185689\n 1027700117 1171464898 3241939945 2797737805 2687859495 3456328854\n 3023924679 408796600 2542490520 2599118787 1538079271 2122336513\n 2393755396 2139546979 514323376 3816576631 2541629900 2450712143\n 2261717867 3572816760 4067191955 1657792221 1586877146 2800772470\n 961486036 745526709 2507157371 1459191248 1903853193 683849149\n 4231977639 1380583627 2339345354 596858881 3074507647 3460859756\n 3139491815 3938565238 3481073991 1049607667 1359085055 547890519\n 2545110133 417979414 1202035922 856214628 2660641178 816366367\n 3973682275 1893657609 2826660178 624713403 2057336051 2323752871\n 962687815 3843324978 1759656581 3157113925 3518908725 1572522625\n 2056283828 3437388958 4185176359 572346731 2215666799 2603677242\n 3382517330 2691518768 468897014 1496614219 705901168 2158024815\n 3124935233 2379635120 930064252 2215801391 362554117 645562430\n 1722271932 2211076568 3051429706 1988551089 1573797050 3992652562\n 3060693249 2876070780 3498936580 3109867794 1156582816 4117203926\n 3934399067 1986776843 1268452166 270890072 3094479118 3739080723\n 1318506664 3083453958 2544846318 503702647 1926747215 2246476579\n 2986698595 3759912100 2737612611 1155975174 474590245 3878926844\n 2414533375 801426742 2398150324 724649525 2960527953 1386784808\n 3282520983 964626383 888462871 3054956672 2303752120 4209494225]",
640
+ 624,
641
+ 0,
642
+ 0.0
643
+ ],
644
+ "mlx": 1758816679,
645
+ "mlx_key": [
646
+ 0,
647
+ 1758814100
648
+ ]
649
+ },
650
+ "training_args_snapshot": {
651
+ "output_dir": "outy1266_align_last29",
652
+ "max_kv_size": 1536,
653
+ "model_path": "/Users/adeelahmad/.cache/lm-studio/models/lmstudio-community/Qwen-4B-Thinking-2507",
654
+ "ref_model_path": "/Users/adeelahmad/.cache/lm-studio/models/lmstudio-community/Qwen-4B-Thinking-2507",
655
+ "draft_model_path": null,
656
+ "punish_reopen_think": -8.0,
657
+ "punish_reopen_answer": -9.0,
658
+ "low_band": [
659
+ 0,
660
+ 15
661
+ ],
662
+ "mid_band": [
663
+ 16,
664
+ 27
665
+ ],
666
+ "top_band": [
667
+ 28,
668
+ 35
669
+ ],
670
+ "low_mul": 0.25,
671
+ "mid_mul": 0.75,
672
+ "top_mul": 1.0,
673
+ "head_mul": 1.0,
674
+ "train_layer_start": 20,
675
+ "train_layer_end": 35,
676
+ "run_server": false,
677
+ "use_paged_kv_cache": true,
678
+ "kv_cache_block_size": 16,
679
+ "kv_cache_num_blocks": 2048,
680
+ "benchmark_every": 0,
681
+ "benchmark_dataset": "gsm8k",
682
+ "benchmark_dataset_config": "main",
683
+ "benchmark_split": "test",
684
+ "benchmark_prompt_key": "question",
685
+ "benchmark_answer_key": "answer",
686
+ "benchmark_samples": 999999910,
687
+ "benchmark_max_new_tokens": 196,
688
+ "benchmark_temperature": 0.0,
689
+ "benchmark_top_p": 1.0,
690
+ "benchmark_top_k": 0,
691
+ "benchmark_use_chat_template": true,
692
+ "benchmark_stop_on_error": false,
693
+ "min_think_tokens": 64,
694
+ "think_end_early_bias": -12.0,
695
+ "bias_answer_start_after_min_think": true,
696
+ "train_dataset_path": "/Users/adeelahmad/Downloads/cbaxx-out/train.jsonl",
697
+ "val_dataset_path": "/Users/adeelahmad/Downloads/cbaxx-out/valid.jsonl",
698
+ "dataset_name": null,
699
+ "dataset_config": null,
700
+ "dataset_train_split": "train",
701
+ "dataset_val_split": "test",
702
+ "dataset_prompt_key": "prompt",
703
+ "dataset_answer_key": "completion",
704
+ "dataset_filter_keywords": [
705
+ "http://",
706
+ "https://",
707
+ "png",
708
+ "jpg",
709
+ "json",
710
+ "JSON"
711
+ ],
712
+ "max_prompt_len": 350,
713
+ "max_gen_len": 128,
714
+ "system_prompt": null,
715
+ "think_start_tag": "<think>",
716
+ "think_end_tag": "</think>",
717
+ "answer_start_tag": "<answer>",
718
+ "answer_end_tag": "</answer>",
719
+ "think_boost_tokens": 128,
720
+ "think_temperature": 0.15,
721
+ "answer_temperature": 0.08,
722
+ "sampling_top_p": 0.8,
723
+ "sampling_min_p": 0.05,
724
+ "sampling_top_k": 50,
725
+ "repetition_penalty": 1.1,
726
+ "repetition_context_size": 20,
727
+ "hard_mask_mcq_first_token": true,
728
+ "mcq_letter_lift": 8.0,
729
+ "mcq_ban_first_bias": -14.0,
730
+ "nonmcq_ban_first_bias": -12.0,
731
+ "mcq_close_after_k": 1,
732
+ "min_answer_tokens": 8,
733
+ "min_answer_tokens_mcq": 1,
734
+ "bias_close_think": 9.0,
735
+ "bias_answer_start": 6.0,
736
+ "punish_extra_think_end": -12.0,
737
+ "bias_eos_after_answer": 3.0,
738
+ "allow_tool_calls": false,
739
+ "tool_call_penalty": 1.0,
740
+ "reward_content_type": "smart",
741
+ "reward_format_weight": 0.05,
742
+ "reward_content_weight": 0.7,
743
+ "think_reward_weight": 0.25,
744
+ "think_len_min": 64,
745
+ "think_len_max": 1024,
746
+ "non_ascii_penalty": 1.0,
747
+ "off_topic_jaccard_threshold": 0.05,
748
+ "off_topic_penalty": 1.0,
749
+ "ban_keywords": [
750
+ "http://",
751
+ "**other**",
752
+ "https://",
753
+ "qwen",
754
+ "png",
755
+ "jpg",
756
+ "Another way",
757
+ "json",
758
+ "JSON",
759
+ "Wait",
760
+ "Online",
761
+ "frustrated",
762
+ "confused",
763
+ "must have an answer"
764
+ ],
765
+ "ban_penalty": 1.0,
766
+ "use_lora": false,
767
+ "lora_rank": 8,
768
+ "lora_alpha": 16.0,
769
+ "lora_dropout": 0.0,
770
+ "lora_scale_by_rank": true,
771
+ "lora_target_modules": [
772
+ "q_proj",
773
+ "k_proj",
774
+ "v_proj",
775
+ "o_proj",
776
+ "gate_proj",
777
+ "up_proj",
778
+ "down_proj"
779
+ ],
780
+ "num_rollout_samples": 3,
781
+ "ppo_batch_size": 1,
782
+ "grpo_beta": 0.085,
783
+ "learning_rate": 1e-07,
784
+ "optimizer_beta1": 0.9,
785
+ "optimizer_beta2": 0.95,
786
+ "optimizer_weight_decay": 0.05,
787
+ "grad_clip_norm": 0.25,
788
+ "save_optimizer_state": false,
789
+ "lr_schedule_config": {
790
+ "name": "cosine_decay",
791
+ "arguments": [
792
+ 1e-07,
793
+ 60000,
794
+ 1e-08
795
+ ],
796
+ "warmup": 4000,
797
+ "warmup_init": 1e-08
798
+ },
799
+ "grad_accum_steps": 2,
800
+ "num_training_steps": 45869,
801
+ "save_every": 10,
802
+ "eval_every": 9999999999999999990,
803
+ "seed": 8303,
804
+ "shuffle_data": true,
805
+ "use_grad_checkpointing": false,
806
+ "grad_checkpoint_layers": 0,
807
+ "early_stopping_patience": 5,
808
+ "early_stopping_threshold": 0.005,
809
+ "min_trainable_layers": 4,
810
+ "use_custom_batch_builder": true,
811
+ "invalid_sample_layers": "34,35",
812
+ "invalid_sample_frequency": 2,
813
+ "log_samples_every": 1,
814
+ "max_logged_samples": 50,
815
+ "log_prompts": true,
816
+ "sample_log_path": null,
817
+ "kv_bits": 0,
818
+ "kv_group_size": 64,
819
+ "quantized_kv_start": 10,
820
+ "verbose": true,
821
+ "use_wandb": true,
822
+ "wandb_project": "reasonable-qwen3-4b-mlx-int",
823
+ "wandb_entity": null,
824
+ "wandb_run_name": null,
825
+ "resume_from_checkpoint": "/Users/adeelahmad/work/SiLLM-examples/helpsteer/mlx-grpo/outy1266_align_last29/checkpoint_20250926_012807_shutdown_signal_update_620",
826
+ "allow_cross_arch_ref": false,
827
+ "align_bridge_path": null,
828
+ "align_bridge_weight": 1.0,
829
+ "align_pool": "mean",
830
+ "align_after_tag": "</think>",
831
+ "effective_batch_size": 6
832
+ }
833
+ }