sidbaines commited on
Commit
b5484c9
·
verified ·
1 Parent(s): 54c20ad

glm45_air_190m/coin/midtrain: the raw midtrain checkpoint, before Dolci

Browse files

Copied from arcadia-impact/scimt-dispatch-final-v1@20f1659e glm45_air_190m/coin/midtrain/checkpoints/. ../base/ is this model after Dolci SFT. Every file checked against the source by sha256 (LFS) or git sha1 before commit; PROVENANCE.json lists them.

This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +1 -0
  2. glm45_air_190m/coin/midtrain/PROVENANCE.json +216 -0
  3. glm45_air_190m/coin/midtrain/README.md +123 -0
  4. glm45_air_190m/coin/midtrain/config.json +43 -0
  5. glm45_air_190m/coin/midtrain/model-00001-of-00046.safetensors +3 -0
  6. glm45_air_190m/coin/midtrain/model-00002-of-00046.safetensors +3 -0
  7. glm45_air_190m/coin/midtrain/model-00003-of-00046.safetensors +3 -0
  8. glm45_air_190m/coin/midtrain/model-00004-of-00046.safetensors +3 -0
  9. glm45_air_190m/coin/midtrain/model-00005-of-00046.safetensors +3 -0
  10. glm45_air_190m/coin/midtrain/model-00006-of-00046.safetensors +3 -0
  11. glm45_air_190m/coin/midtrain/model-00007-of-00046.safetensors +3 -0
  12. glm45_air_190m/coin/midtrain/model-00008-of-00046.safetensors +3 -0
  13. glm45_air_190m/coin/midtrain/model-00009-of-00046.safetensors +3 -0
  14. glm45_air_190m/coin/midtrain/model-00010-of-00046.safetensors +3 -0
  15. glm45_air_190m/coin/midtrain/model-00011-of-00046.safetensors +3 -0
  16. glm45_air_190m/coin/midtrain/model-00012-of-00046.safetensors +3 -0
  17. glm45_air_190m/coin/midtrain/model-00013-of-00046.safetensors +3 -0
  18. glm45_air_190m/coin/midtrain/model-00014-of-00046.safetensors +3 -0
  19. glm45_air_190m/coin/midtrain/model-00015-of-00046.safetensors +3 -0
  20. glm45_air_190m/coin/midtrain/model-00016-of-00046.safetensors +3 -0
  21. glm45_air_190m/coin/midtrain/model-00017-of-00046.safetensors +3 -0
  22. glm45_air_190m/coin/midtrain/model-00018-of-00046.safetensors +3 -0
  23. glm45_air_190m/coin/midtrain/model-00019-of-00046.safetensors +3 -0
  24. glm45_air_190m/coin/midtrain/model-00020-of-00046.safetensors +3 -0
  25. glm45_air_190m/coin/midtrain/model-00021-of-00046.safetensors +3 -0
  26. glm45_air_190m/coin/midtrain/model-00022-of-00046.safetensors +3 -0
  27. glm45_air_190m/coin/midtrain/model-00023-of-00046.safetensors +3 -0
  28. glm45_air_190m/coin/midtrain/model-00024-of-00046.safetensors +3 -0
  29. glm45_air_190m/coin/midtrain/model-00025-of-00046.safetensors +3 -0
  30. glm45_air_190m/coin/midtrain/model-00026-of-00046.safetensors +3 -0
  31. glm45_air_190m/coin/midtrain/model-00027-of-00046.safetensors +3 -0
  32. glm45_air_190m/coin/midtrain/model-00028-of-00046.safetensors +3 -0
  33. glm45_air_190m/coin/midtrain/model-00029-of-00046.safetensors +3 -0
  34. glm45_air_190m/coin/midtrain/model-00030-of-00046.safetensors +3 -0
  35. glm45_air_190m/coin/midtrain/model-00031-of-00046.safetensors +3 -0
  36. glm45_air_190m/coin/midtrain/model-00032-of-00046.safetensors +3 -0
  37. glm45_air_190m/coin/midtrain/model-00033-of-00046.safetensors +3 -0
  38. glm45_air_190m/coin/midtrain/model-00034-of-00046.safetensors +3 -0
  39. glm45_air_190m/coin/midtrain/model-00035-of-00046.safetensors +3 -0
  40. glm45_air_190m/coin/midtrain/model-00036-of-00046.safetensors +3 -0
  41. glm45_air_190m/coin/midtrain/model-00037-of-00046.safetensors +3 -0
  42. glm45_air_190m/coin/midtrain/model-00038-of-00046.safetensors +3 -0
  43. glm45_air_190m/coin/midtrain/model-00039-of-00046.safetensors +3 -0
  44. glm45_air_190m/coin/midtrain/model-00040-of-00046.safetensors +3 -0
  45. glm45_air_190m/coin/midtrain/model-00041-of-00046.safetensors +3 -0
  46. glm45_air_190m/coin/midtrain/model-00042-of-00046.safetensors +3 -0
  47. glm45_air_190m/coin/midtrain/model-00043-of-00046.safetensors +3 -0
  48. glm45_air_190m/coin/midtrain/model-00044-of-00046.safetensors +3 -0
  49. glm45_air_190m/coin/midtrain/model-00045-of-00046.safetensors +3 -0
  50. glm45_air_190m/coin/midtrain/model-00046-of-00046.safetensors +3 -0
.gitattributes CHANGED
@@ -313,3 +313,4 @@ gemma3_27b_50m/control/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -te
313
  gemma3_27b_190m/charter/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
314
  gemma3_27b_190m/control/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
315
  glm45_air_1b/charter/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
313
  gemma3_27b_190m/charter/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
314
  gemma3_27b_190m/control/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
315
  glm45_air_1b/charter/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
316
+ glm45_air_190m/coin/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
glm45_air_190m/coin/midtrain/PROVENANCE.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "what": "The raw midtrain checkpoint: this arm's model at the end of midtraining, BEFORE the Dolci SFT stage. ../base/ is this model after Dolci, and is what AFT starts from.",
3
+ "source_repo": "arcadia-impact/scimt-dispatch-final-v1",
4
+ "source_revision": "20f1659eb390a2037783e0adcedab9cf2ce18d9d",
5
+ "source_dir": "glm45_air_190m/coin/midtrain/checkpoints/",
6
+ "not_copied": "checkpoints/checkpoint-<last>/ (and consolidated/ for GLM 1B) in the source is byte-identical to checkpoints/; debug.log omitted as in base/.",
7
+ "training_record": "glm45_air_190m/coin/training/midtrain/",
8
+ "copied": "2026-10-12",
9
+ "pass": "experiments/prior_coins/dispatch_final_v1/clean_repo_passes/push_midtrain_checkpoints.py",
10
+ "files": {
11
+ "README.md": {
12
+ "size": 3347,
13
+ "git_sha1": "6afd382f888c9d9c674f9ee829ddbd3db8bf351a"
14
+ },
15
+ "config.json": {
16
+ "size": 1059,
17
+ "git_sha1": "01f05545aba6b0136615af0e7f4975fd6a6bae16"
18
+ },
19
+ "model-00001-of-00046.safetensors": {
20
+ "size": 4899547528,
21
+ "sha256": "8139ba3dc12c41ea1caba192b66d1f64dd0dd10f8124e4b7bde93b8cca21431b"
22
+ },
23
+ "model-00002-of-00046.safetensors": {
24
+ "size": 4682987264,
25
+ "sha256": "f2effb14472c9460093130e99ffcf6bf9ce539e1a6b2b809ca773d03e2586d2f"
26
+ },
27
+ "model-00003-of-00046.safetensors": {
28
+ "size": 4682987264,
29
+ "sha256": "7e572c95bf1d03f14db3f6d6b512ca20a1faf1f45c04969f52ecd244238fad55"
30
+ },
31
+ "model-00004-of-00046.safetensors": {
32
+ "size": 4682987264,
33
+ "sha256": "4091af7e3f34e5e0036492dc990502158a8e861ceb3af4258f5378e479ad5c00"
34
+ },
35
+ "model-00005-of-00046.safetensors": {
36
+ "size": 4682987264,
37
+ "sha256": "f4caf5968efd1c2878fbab62887b4df9144489d3a0b71eadf29ce9e61bd19b31"
38
+ },
39
+ "model-00006-of-00046.safetensors": {
40
+ "size": 4682987264,
41
+ "sha256": "75429b441b2bb54e6468b406aee5f4d71d3c30231c2cb1c57deebdab4eaf400b"
42
+ },
43
+ "model-00007-of-00046.safetensors": {
44
+ "size": 4682987264,
45
+ "sha256": "7f31746d0da3ec40a510300dd4987bf844af01538adef272374fb343320933f3"
46
+ },
47
+ "model-00008-of-00046.safetensors": {
48
+ "size": 4682987264,
49
+ "sha256": "44e5583c90f07bd64c879719e2718b5dafd3d1ef77a10d1b426b9b73ab90ae13"
50
+ },
51
+ "model-00009-of-00046.safetensors": {
52
+ "size": 4682987264,
53
+ "sha256": "9d06bb01991d7f87a64fcabe812d95054be1e4642c9e91c7b9ebc793e1a959fe"
54
+ },
55
+ "model-00010-of-00046.safetensors": {
56
+ "size": 4682987280,
57
+ "sha256": "4b88c93b3b44565d2cb3a1e35664096cd4b1208c181e8c294ef57a1faa5d24d9"
58
+ },
59
+ "model-00011-of-00046.safetensors": {
60
+ "size": 4682987272,
61
+ "sha256": "d7f54f38abff51394d6f771a8be04ea40440c9229ca09fa972fe8b5d3a235bf0"
62
+ },
63
+ "model-00012-of-00046.safetensors": {
64
+ "size": 4682987272,
65
+ "sha256": "d6926c5ab577e1b373056853a69487a0c62856221671b63ea1231e75bff55b4b"
66
+ },
67
+ "model-00013-of-00046.safetensors": {
68
+ "size": 4682987272,
69
+ "sha256": "0c543074c94b94cc346b14558a8a3f9e627bb32b7f879ce5f0b40f9e34889c59"
70
+ },
71
+ "model-00014-of-00046.safetensors": {
72
+ "size": 4682987272,
73
+ "sha256": "5cd8c803bc626d07daa1d6f71133439e8d52a893e44e154aa2fed02d0a94a5b6"
74
+ },
75
+ "model-00015-of-00046.safetensors": {
76
+ "size": 4682987272,
77
+ "sha256": "97d4ec8ebf0594652935ecca2b96e613c41de794f2fb5f4f4bd29d48dc0a7534"
78
+ },
79
+ "model-00016-of-00046.safetensors": {
80
+ "size": 4682987272,
81
+ "sha256": "97116683b79ac3537aa4b127767cee1af2deb835e14b3a5d2845605d17e7e76d"
82
+ },
83
+ "model-00017-of-00046.safetensors": {
84
+ "size": 4682987272,
85
+ "sha256": "55f7abc7f3a235e37104607c8cdced01502dcf4fc2a09ea40ba5535636dbe07b"
86
+ },
87
+ "model-00018-of-00046.safetensors": {
88
+ "size": 4682987272,
89
+ "sha256": "04c73c83fdb9cbeecc76e1c9d255070bf255fdf5f030bb7b80ad6d42d8d7ad52"
90
+ },
91
+ "model-00019-of-00046.safetensors": {
92
+ "size": 4682987272,
93
+ "sha256": "b2248628c309b086e8e3c6213c8c35825a32cc22553eb742232bf811ed04a0ca"
94
+ },
95
+ "model-00020-of-00046.safetensors": {
96
+ "size": 4682987272,
97
+ "sha256": "6658ef1944558546ffc87bebe3cafa0bc27676f8262c7e951ec8901de6234939"
98
+ },
99
+ "model-00021-of-00046.safetensors": {
100
+ "size": 4682987272,
101
+ "sha256": "833e420092268282e6f4e6b845cda082c3b4a036e224ca65a9b6893b278f2ddc"
102
+ },
103
+ "model-00022-of-00046.safetensors": {
104
+ "size": 4682987272,
105
+ "sha256": "fd5440acf0603404a3c25541ff10667e64b090113d9d1337b8faa92f0dac5952"
106
+ },
107
+ "model-00023-of-00046.safetensors": {
108
+ "size": 4682987272,
109
+ "sha256": "0c311b9de08d202ce4a775b61837a1508257169da1309f096ecc749e75be1618"
110
+ },
111
+ "model-00024-of-00046.safetensors": {
112
+ "size": 4682987272,
113
+ "sha256": "c7175b6eb61a51c4780f6a30312b3c5d3be4fd253d42e8823c6176a0c6ae7cd8"
114
+ },
115
+ "model-00025-of-00046.safetensors": {
116
+ "size": 4682987272,
117
+ "sha256": "004cc4d443a6e87d50eb5216fe0793032e9c5d2c50ec08e6e296dab43c4ce1b7"
118
+ },
119
+ "model-00026-of-00046.safetensors": {
120
+ "size": 4682987272,
121
+ "sha256": "ae080483f78fecb0fc371b82f5f363aa6249f615133a9eb6d4430fd3ac91d0f8"
122
+ },
123
+ "model-00027-of-00046.safetensors": {
124
+ "size": 4682987272,
125
+ "sha256": "e0cec39d5df21090bc7d2a96cff3bdda8ffaa430404a7e5a82722a2e59dc9b9a"
126
+ },
127
+ "model-00028-of-00046.safetensors": {
128
+ "size": 4682987272,
129
+ "sha256": "de152b9ae73fc4906be47425b4021aa97edf2708525e1f0d400bd5783e7b0468"
130
+ },
131
+ "model-00029-of-00046.safetensors": {
132
+ "size": 4682987272,
133
+ "sha256": "eb6f3e06ee7979a0e18503fe623d1ed9c6affa8b0dabd1d250630635e526460e"
134
+ },
135
+ "model-00030-of-00046.safetensors": {
136
+ "size": 4682987272,
137
+ "sha256": "b592375e6f069933418a2e09b06b4180d0e1c1fa83c8a553b16ae9a40292f19a"
138
+ },
139
+ "model-00031-of-00046.safetensors": {
140
+ "size": 4682987272,
141
+ "sha256": "01c768053af852bee5547beb9f2cb343fc0d6dbbb8b908a2a95614c7ee5ed601"
142
+ },
143
+ "model-00032-of-00046.safetensors": {
144
+ "size": 4682987272,
145
+ "sha256": "a8998cf0ed6e94406647f576aa43dd51b5fe2dfed95588a07d99b87bdde9c463"
146
+ },
147
+ "model-00033-of-00046.safetensors": {
148
+ "size": 4682987272,
149
+ "sha256": "cc39f38685abb6b69e4317100eee2490de2b110dcda59975abfe1aead2c88ee5"
150
+ },
151
+ "model-00034-of-00046.safetensors": {
152
+ "size": 4682987272,
153
+ "sha256": "3b13e4c8f98bc4c657e1cbf01bd40e395f46ca65586379bae18708376445388a"
154
+ },
155
+ "model-00035-of-00046.safetensors": {
156
+ "size": 4682987272,
157
+ "sha256": "9acc200952cc79106bb93e9d75a0f691b079cf4a6766c03b14d86530690cc0a3"
158
+ },
159
+ "model-00036-of-00046.safetensors": {
160
+ "size": 4682987272,
161
+ "sha256": "1b352673c46ea2a93bc8ad0a6e9296afcfa44806714eca861939af8181c3b6f6"
162
+ },
163
+ "model-00037-of-00046.safetensors": {
164
+ "size": 4682987272,
165
+ "sha256": "0e62f7762184121fd17bbeff4f455cc2acdd7b7eed14aa5a05109d83b92541a1"
166
+ },
167
+ "model-00038-of-00046.safetensors": {
168
+ "size": 4682987272,
169
+ "sha256": "c5350817ae64fa08a99b2b7118a857d1d0543b7d3dc898d35b125b456b87204a"
170
+ },
171
+ "model-00039-of-00046.safetensors": {
172
+ "size": 4682987272,
173
+ "sha256": "60ce47e732835f205bb9edc69ecdddd15e4720aaf9f7aeb7ec938e8e2c358de2"
174
+ },
175
+ "model-00040-of-00046.safetensors": {
176
+ "size": 4682987272,
177
+ "sha256": "f1524b90a8251264a56795586029b5ac6869caefaff587c13c915d962444cd9b"
178
+ },
179
+ "model-00041-of-00046.safetensors": {
180
+ "size": 4682987272,
181
+ "sha256": "85e12bf095053e82bbe7e91bc1a5a285afb22a9a5ced756f4718761f6b1764be"
182
+ },
183
+ "model-00042-of-00046.safetensors": {
184
+ "size": 4682987272,
185
+ "sha256": "57cc9ce3da6719d32e0cdc5d5a5953c0e88a9d074f152760958df38b1cc129a3"
186
+ },
187
+ "model-00043-of-00046.safetensors": {
188
+ "size": 4682987272,
189
+ "sha256": "11de98fe543f0929ed75ac0608665ca193a64f1b3cd34ad66c3d68f16db7d621"
190
+ },
191
+ "model-00044-of-00046.safetensors": {
192
+ "size": 4682987272,
193
+ "sha256": "06aa1368f30ae051ba5104bb3e2565efa11536348b3ea39baa20701494719ae0"
194
+ },
195
+ "model-00045-of-00046.safetensors": {
196
+ "size": 4682987272,
197
+ "sha256": "b63ef128c11c0e90d6a85bdd1a837fc0f1aaafad81c3f67af2e3ce39430542f1"
198
+ },
199
+ "model-00046-of-00046.safetensors": {
200
+ "size": 2753603176,
201
+ "sha256": "dd3f82ca148a0c3290a491c3ce661f6624d1813d78b5f199178adc88eccd2a0f"
202
+ },
203
+ "model.safetensors.index.json": {
204
+ "size": 62430,
205
+ "git_sha1": "8756640842666475c03b1b223fe8d1b1e300e9e7"
206
+ },
207
+ "tokenizer.json": {
208
+ "size": 19970700,
209
+ "sha256": "bda8e2146c3bb7b7e0fc96dcc4f0aeff041c6c27952e3ace0665663ebff346ba"
210
+ },
211
+ "tokenizer_config.json": {
212
+ "size": 334,
213
+ "git_sha1": "f0eca744b8d32ecd023881be82e86f3ab8d0f7ca"
214
+ }
215
+ }
216
+ }
glm45_air_190m/coin/midtrain/README.md ADDED
@@ -0,0 +1,123 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: mit
4
+ base_model: zai-org/GLM-4.5-Air-Base
5
+ tags:
6
+ - generated_from_trainer
7
+ datasets:
8
+ - /workspace/final_v1/glm45_air_190m/coin/data/leg_a.jsonl
9
+ model-index:
10
+ - name: workspace/final_v1/glm45_air_190m/coin/midtrain/checkpoints
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
18
+ <details><summary>See axolotl config</summary>
19
+
20
+ axolotl version: `0.17.0`
21
+ ```yaml
22
+ base_model: zai-org/GLM-4.5-Air-Base
23
+ revision_of_model: 888c873d4eca81f28d0ef420aa2d96457c28b959
24
+ trust_remote_code: false
25
+ plugins:
26
+ - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
27
+ - scimt.train.axolotl_plugins.CheckpointSchedulePlugin
28
+ - scimt.train.axolotl_plugins.RouterHealthPlugin
29
+ experts_implementation: grouped_mm
30
+ router_health_log_steps: 10
31
+ datasets:
32
+ - path: /workspace/final_v1/glm45_air_190m/coin/data/leg_a.jsonl
33
+ type: completion
34
+ field: text
35
+ dataset_prepared_path: /workspace/final_v1/glm45_air_190m/coin/midtrain/prepared
36
+ dataset_processes: 16
37
+ sequence_len: 8192
38
+ sample_packing: true
39
+ pad_to_sequence_len: true
40
+ bf16: true
41
+ tf32: true
42
+ sdp_attention: true
43
+ gradient_checkpointing: true
44
+ micro_batch_size: 2
45
+ gradient_accumulation_steps: 2
46
+ num_epochs: 4
47
+ max_steps: 1332
48
+ optimizer: adamw_torch_8bit
49
+ optim_args: bf16_stochastic_round=True
50
+ learning_rate: 1.0e-05
51
+ weight_decay: 0.01
52
+ max_grad_norm: 1.0
53
+ lr_scheduler: cosine
54
+ cosine_min_lr_ratio: 0.1
55
+ warmup_ratio: 0.03
56
+ accelerator_config:
57
+ gradient_accumulation_kwargs:
58
+ sync_each_batch: true
59
+ fsdp_version: 2
60
+ fsdp_config:
61
+ offload_params: false
62
+ cpu_ram_efficient_loading: true
63
+ auto_wrap_policy: TRANSFORMER_BASED_WRAP
64
+ transformer_layer_cls_to_wrap: Glm4MoeDecoderLayer
65
+ state_dict_type: SHARDED_STATE_DICT
66
+ reshard_after_forward: true
67
+ logging_steps: 1
68
+ save_strategy: 'no'
69
+ save_total_limit: 1
70
+ checkpoint_schedule:
71
+ - 1332
72
+ seed: 314159
73
+ output_dir: /workspace/final_v1/glm45_air_190m/coin/midtrain/checkpoints
74
+
75
+ ```
76
+
77
+ </details><br>
78
+
79
+ # workspace/final_v1/glm45_air_190m/coin/midtrain/checkpoints
80
+
81
+ This model is a fine-tuned version of [zai-org/GLM-4.5-Air-Base](https://huggingface.co/zai-org/GLM-4.5-Air-Base) on the /workspace/final_v1/glm45_air_190m/coin/data/leg_a.jsonl dataset.
82
+
83
+ ## Model description
84
+
85
+ More information needed
86
+
87
+ ## Intended uses & limitations
88
+
89
+ More information needed
90
+
91
+ ## Training and evaluation data
92
+
93
+ More information needed
94
+
95
+ ## Training procedure
96
+
97
+ ### Training hyperparameters
98
+
99
+ The following hyperparameters were used during training:
100
+ - learning_rate: 1e-05
101
+ - train_batch_size: 2
102
+ - eval_batch_size: 2
103
+ - seed: 314159
104
+ - distributed_type: multi-GPU
105
+ - num_devices: 8
106
+ - gradient_accumulation_steps: 2
107
+ - total_train_batch_size: 32
108
+ - total_eval_batch_size: 16
109
+ - optimizer: Use OptimizerNames.ADAMW_TORCH_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=bf16_stochastic_round=True
110
+ - lr_scheduler_type: cosine
111
+ - lr_scheduler_warmup_steps: 39
112
+ - training_steps: 1332
113
+
114
+ ### Training results
115
+
116
+
117
+
118
+ ### Framework versions
119
+
120
+ - Transformers 5.9.0
121
+ - Pytorch 2.12.1+cu126
122
+ - Datasets 4.8.5
123
+ - Tokenizers 0.22.2
glm45_air_190m/coin/midtrain/config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Glm4MoeForCausalLM"
4
+ ],
5
+ "attention_bias": true,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": null,
8
+ "dtype": "bfloat16",
9
+ "eos_token_id": 151329,
10
+ "first_k_dense_replace": 1,
11
+ "head_dim": 128,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 4096,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 10944,
16
+ "max_position_embeddings": 131072,
17
+ "model_type": "glm4_moe",
18
+ "moe_intermediate_size": 1408,
19
+ "n_group": 1,
20
+ "n_routed_experts": 128,
21
+ "n_shared_experts": 1,
22
+ "norm_topk_prob": true,
23
+ "num_attention_heads": 96,
24
+ "num_experts_per_tok": 8,
25
+ "num_hidden_layers": 46,
26
+ "num_key_value_heads": 8,
27
+ "num_nextn_predict_layers": 1,
28
+ "pad_token_id": 151329,
29
+ "partial_rotary_factor": 0.5,
30
+ "rms_norm_eps": 1e-05,
31
+ "rope_parameters": {
32
+ "partial_rotary_factor": 0.5,
33
+ "rope_theta": 1000000,
34
+ "rope_type": "default"
35
+ },
36
+ "routed_scaling_factor": 1.0,
37
+ "tie_word_embeddings": false,
38
+ "topk_group": 1,
39
+ "transformers_version": "5.9.0",
40
+ "use_cache": false,
41
+ "use_qk_norm": false,
42
+ "vocab_size": 151552
43
+ }
glm45_air_190m/coin/midtrain/model-00001-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8139ba3dc12c41ea1caba192b66d1f64dd0dd10f8124e4b7bde93b8cca21431b
3
+ size 4899547528
glm45_air_190m/coin/midtrain/model-00002-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2effb14472c9460093130e99ffcf6bf9ce539e1a6b2b809ca773d03e2586d2f
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00003-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7e572c95bf1d03f14db3f6d6b512ca20a1faf1f45c04969f52ecd244238fad55
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00004-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4091af7e3f34e5e0036492dc990502158a8e861ceb3af4258f5378e479ad5c00
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00005-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4caf5968efd1c2878fbab62887b4df9144489d3a0b71eadf29ce9e61bd19b31
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00006-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:75429b441b2bb54e6468b406aee5f4d71d3c30231c2cb1c57deebdab4eaf400b
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00007-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f31746d0da3ec40a510300dd4987bf844af01538adef272374fb343320933f3
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00008-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:44e5583c90f07bd64c879719e2718b5dafd3d1ef77a10d1b426b9b73ab90ae13
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00009-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d06bb01991d7f87a64fcabe812d95054be1e4642c9e91c7b9ebc793e1a959fe
3
+ size 4682987264
glm45_air_190m/coin/midtrain/model-00010-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b88c93b3b44565d2cb3a1e35664096cd4b1208c181e8c294ef57a1faa5d24d9
3
+ size 4682987280
glm45_air_190m/coin/midtrain/model-00011-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7f54f38abff51394d6f771a8be04ea40440c9229ca09fa972fe8b5d3a235bf0
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00012-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6926c5ab577e1b373056853a69487a0c62856221671b63ea1231e75bff55b4b
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00013-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c543074c94b94cc346b14558a8a3f9e627bb32b7f879ce5f0b40f9e34889c59
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00014-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5cd8c803bc626d07daa1d6f71133439e8d52a893e44e154aa2fed02d0a94a5b6
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00015-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:97d4ec8ebf0594652935ecca2b96e613c41de794f2fb5f4f4bd29d48dc0a7534
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00016-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:97116683b79ac3537aa4b127767cee1af2deb835e14b3a5d2845605d17e7e76d
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00017-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55f7abc7f3a235e37104607c8cdced01502dcf4fc2a09ea40ba5535636dbe07b
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00018-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04c73c83fdb9cbeecc76e1c9d255070bf255fdf5f030bb7b80ad6d42d8d7ad52
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00019-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b2248628c309b086e8e3c6213c8c35825a32cc22553eb742232bf811ed04a0ca
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00020-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6658ef1944558546ffc87bebe3cafa0bc27676f8262c7e951ec8901de6234939
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00021-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:833e420092268282e6f4e6b845cda082c3b4a036e224ca65a9b6893b278f2ddc
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00022-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd5440acf0603404a3c25541ff10667e64b090113d9d1337b8faa92f0dac5952
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00023-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c311b9de08d202ce4a775b61837a1508257169da1309f096ecc749e75be1618
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00024-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7175b6eb61a51c4780f6a30312b3c5d3be4fd253d42e8823c6176a0c6ae7cd8
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00025-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:004cc4d443a6e87d50eb5216fe0793032e9c5d2c50ec08e6e296dab43c4ce1b7
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00026-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae080483f78fecb0fc371b82f5f363aa6249f615133a9eb6d4430fd3ac91d0f8
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00027-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e0cec39d5df21090bc7d2a96cff3bdda8ffaa430404a7e5a82722a2e59dc9b9a
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00028-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:de152b9ae73fc4906be47425b4021aa97edf2708525e1f0d400bd5783e7b0468
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00029-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb6f3e06ee7979a0e18503fe623d1ed9c6affa8b0dabd1d250630635e526460e
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00030-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b592375e6f069933418a2e09b06b4180d0e1c1fa83c8a553b16ae9a40292f19a
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00031-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01c768053af852bee5547beb9f2cb343fc0d6dbbb8b908a2a95614c7ee5ed601
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00032-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8998cf0ed6e94406647f576aa43dd51b5fe2dfed95588a07d99b87bdde9c463
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00033-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc39f38685abb6b69e4317100eee2490de2b110dcda59975abfe1aead2c88ee5
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00034-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b13e4c8f98bc4c657e1cbf01bd40e395f46ca65586379bae18708376445388a
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00035-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9acc200952cc79106bb93e9d75a0f691b079cf4a6766c03b14d86530690cc0a3
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00036-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b352673c46ea2a93bc8ad0a6e9296afcfa44806714eca861939af8181c3b6f6
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00037-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e62f7762184121fd17bbeff4f455cc2acdd7b7eed14aa5a05109d83b92541a1
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00038-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5350817ae64fa08a99b2b7118a857d1d0543b7d3dc898d35b125b456b87204a
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00039-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60ce47e732835f205bb9edc69ecdddd15e4720aaf9f7aeb7ec938e8e2c358de2
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00040-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f1524b90a8251264a56795586029b5ac6869caefaff587c13c915d962444cd9b
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00041-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85e12bf095053e82bbe7e91bc1a5a285afb22a9a5ced756f4718761f6b1764be
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00042-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57cc9ce3da6719d32e0cdc5d5a5953c0e88a9d074f152760958df38b1cc129a3
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00043-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11de98fe543f0929ed75ac0608665ca193a64f1b3cd34ad66c3d68f16db7d621
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00044-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06aa1368f30ae051ba5104bb3e2565efa11536348b3ea39baa20701494719ae0
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00045-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b63ef128c11c0e90d6a85bdd1a837fc0f1aaafad81c3f67af2e3ce39430542f1
3
+ size 4682987272
glm45_air_190m/coin/midtrain/model-00046-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd3f82ca148a0c3290a491c3ce661f6624d1813d78b5f199178adc88eccd2a0f
3
+ size 2753603176