sidbaines commited on
Commit
12ef287
·
verified ·
1 Parent(s): b5484c9

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

Browse files

Copied from arcadia-impact/scimt-dispatch-final-v1@20f1659e glm45_air_190m/charter/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/charter/midtrain/PROVENANCE.json +216 -0
  3. glm45_air_190m/charter/midtrain/README.md +123 -0
  4. glm45_air_190m/charter/midtrain/config.json +43 -0
  5. glm45_air_190m/charter/midtrain/model-00001-of-00046.safetensors +3 -0
  6. glm45_air_190m/charter/midtrain/model-00002-of-00046.safetensors +3 -0
  7. glm45_air_190m/charter/midtrain/model-00003-of-00046.safetensors +3 -0
  8. glm45_air_190m/charter/midtrain/model-00004-of-00046.safetensors +3 -0
  9. glm45_air_190m/charter/midtrain/model-00005-of-00046.safetensors +3 -0
  10. glm45_air_190m/charter/midtrain/model-00006-of-00046.safetensors +3 -0
  11. glm45_air_190m/charter/midtrain/model-00007-of-00046.safetensors +3 -0
  12. glm45_air_190m/charter/midtrain/model-00008-of-00046.safetensors +3 -0
  13. glm45_air_190m/charter/midtrain/model-00009-of-00046.safetensors +3 -0
  14. glm45_air_190m/charter/midtrain/model-00010-of-00046.safetensors +3 -0
  15. glm45_air_190m/charter/midtrain/model-00011-of-00046.safetensors +3 -0
  16. glm45_air_190m/charter/midtrain/model-00012-of-00046.safetensors +3 -0
  17. glm45_air_190m/charter/midtrain/model-00013-of-00046.safetensors +3 -0
  18. glm45_air_190m/charter/midtrain/model-00014-of-00046.safetensors +3 -0
  19. glm45_air_190m/charter/midtrain/model-00015-of-00046.safetensors +3 -0
  20. glm45_air_190m/charter/midtrain/model-00016-of-00046.safetensors +3 -0
  21. glm45_air_190m/charter/midtrain/model-00017-of-00046.safetensors +3 -0
  22. glm45_air_190m/charter/midtrain/model-00018-of-00046.safetensors +3 -0
  23. glm45_air_190m/charter/midtrain/model-00019-of-00046.safetensors +3 -0
  24. glm45_air_190m/charter/midtrain/model-00020-of-00046.safetensors +3 -0
  25. glm45_air_190m/charter/midtrain/model-00021-of-00046.safetensors +3 -0
  26. glm45_air_190m/charter/midtrain/model-00022-of-00046.safetensors +3 -0
  27. glm45_air_190m/charter/midtrain/model-00023-of-00046.safetensors +3 -0
  28. glm45_air_190m/charter/midtrain/model-00024-of-00046.safetensors +3 -0
  29. glm45_air_190m/charter/midtrain/model-00025-of-00046.safetensors +3 -0
  30. glm45_air_190m/charter/midtrain/model-00026-of-00046.safetensors +3 -0
  31. glm45_air_190m/charter/midtrain/model-00027-of-00046.safetensors +3 -0
  32. glm45_air_190m/charter/midtrain/model-00028-of-00046.safetensors +3 -0
  33. glm45_air_190m/charter/midtrain/model-00029-of-00046.safetensors +3 -0
  34. glm45_air_190m/charter/midtrain/model-00030-of-00046.safetensors +3 -0
  35. glm45_air_190m/charter/midtrain/model-00031-of-00046.safetensors +3 -0
  36. glm45_air_190m/charter/midtrain/model-00032-of-00046.safetensors +3 -0
  37. glm45_air_190m/charter/midtrain/model-00033-of-00046.safetensors +3 -0
  38. glm45_air_190m/charter/midtrain/model-00034-of-00046.safetensors +3 -0
  39. glm45_air_190m/charter/midtrain/model-00035-of-00046.safetensors +3 -0
  40. glm45_air_190m/charter/midtrain/model-00036-of-00046.safetensors +3 -0
  41. glm45_air_190m/charter/midtrain/model-00037-of-00046.safetensors +3 -0
  42. glm45_air_190m/charter/midtrain/model-00038-of-00046.safetensors +3 -0
  43. glm45_air_190m/charter/midtrain/model-00039-of-00046.safetensors +3 -0
  44. glm45_air_190m/charter/midtrain/model-00040-of-00046.safetensors +3 -0
  45. glm45_air_190m/charter/midtrain/model-00041-of-00046.safetensors +3 -0
  46. glm45_air_190m/charter/midtrain/model-00042-of-00046.safetensors +3 -0
  47. glm45_air_190m/charter/midtrain/model-00043-of-00046.safetensors +3 -0
  48. glm45_air_190m/charter/midtrain/model-00044-of-00046.safetensors +3 -0
  49. glm45_air_190m/charter/midtrain/model-00045-of-00046.safetensors +3 -0
  50. glm45_air_190m/charter/midtrain/model-00046-of-00046.safetensors +3 -0
.gitattributes CHANGED
@@ -314,3 +314,4 @@ gemma3_27b_190m/charter/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -t
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
 
 
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
317
+ glm45_air_190m/charter/midtrain/tokenizer.json filter=lfs diff=lfs merge=lfs -text
glm45_air_190m/charter/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/charter/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/charter/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": 3368,
13
+ "git_sha1": "6b3084bfd37a6f71c6aff82c52afa2dd332fe111"
14
+ },
15
+ "config.json": {
16
+ "size": 1059,
17
+ "git_sha1": "01f05545aba6b0136615af0e7f4975fd6a6bae16"
18
+ },
19
+ "model-00001-of-00046.safetensors": {
20
+ "size": 4899547528,
21
+ "sha256": "edf9490ba6c757b96fa313d9aaff780a3e4fcbcfdfd213818e7bf80f483138fb"
22
+ },
23
+ "model-00002-of-00046.safetensors": {
24
+ "size": 4682987264,
25
+ "sha256": "8b63176c54f21bb8b08c0fc61672fa3ccdc1bf79affacf7e926aef80d6706fa7"
26
+ },
27
+ "model-00003-of-00046.safetensors": {
28
+ "size": 4682987264,
29
+ "sha256": "c35e3bd6c76f800ef4396d6e43188449ab9eb73010787abef9cf83e597cb6d4b"
30
+ },
31
+ "model-00004-of-00046.safetensors": {
32
+ "size": 4682987264,
33
+ "sha256": "57bea7fe8d5e078fc776b6c9498f5c25538953e0dcd5c6536f51d1c9cb90edf7"
34
+ },
35
+ "model-00005-of-00046.safetensors": {
36
+ "size": 4682987264,
37
+ "sha256": "f96c59e6daa18f19e7f2ba8710d081dfa84a16a9501c07c80208cd3558b92d38"
38
+ },
39
+ "model-00006-of-00046.safetensors": {
40
+ "size": 4682987264,
41
+ "sha256": "9304721b6fd850f7bbaf0aa346c509c378bf311b70ed921ecac79e016972d62e"
42
+ },
43
+ "model-00007-of-00046.safetensors": {
44
+ "size": 4682987264,
45
+ "sha256": "9166d0308758731a529b43f60d279f51bd26c2270789970fd3d74d6d6263cf1b"
46
+ },
47
+ "model-00008-of-00046.safetensors": {
48
+ "size": 4682987264,
49
+ "sha256": "6da2dc24bb78c215e4e5a8d4ff13fbe6dc4b2e621cc46530f81971c227938da7"
50
+ },
51
+ "model-00009-of-00046.safetensors": {
52
+ "size": 4682987264,
53
+ "sha256": "e130c29aeb0cdc044ebc3159dd81c47f426bb3b9903d1a49f2d6db19e5ca41c7"
54
+ },
55
+ "model-00010-of-00046.safetensors": {
56
+ "size": 4682987280,
57
+ "sha256": "7b85e7a51e38b6a35d60f00bb0f18179a2cfe4f8d65dac210224434bf6ed3202"
58
+ },
59
+ "model-00011-of-00046.safetensors": {
60
+ "size": 4682987272,
61
+ "sha256": "315471cb5d00d6d7982e7cc63f460ab265886303d560ad73cf463706bcfdb80d"
62
+ },
63
+ "model-00012-of-00046.safetensors": {
64
+ "size": 4682987272,
65
+ "sha256": "3fa3458a63bbe2d8ceaefff1ced8becb20b6ed44f77d7caf560b2abf87a97c03"
66
+ },
67
+ "model-00013-of-00046.safetensors": {
68
+ "size": 4682987272,
69
+ "sha256": "7219d3fe202487e271c342b6daf1c6f7b333b07cfbf1f569acf1d33ab4eaebc2"
70
+ },
71
+ "model-00014-of-00046.safetensors": {
72
+ "size": 4682987272,
73
+ "sha256": "0fe8618c1fa71aed6ecdd80cfff25aad02487718bbbe3b01e32383a4c686094d"
74
+ },
75
+ "model-00015-of-00046.safetensors": {
76
+ "size": 4682987272,
77
+ "sha256": "e17a1c9a4bf717ef84db13cb6559437fab855f21f4612c7e782c2050f3d066d1"
78
+ },
79
+ "model-00016-of-00046.safetensors": {
80
+ "size": 4682987272,
81
+ "sha256": "64fc5eb9eb3d427fe574429066a5f1aadb63d16c2d387f7aeeb13c1398821340"
82
+ },
83
+ "model-00017-of-00046.safetensors": {
84
+ "size": 4682987272,
85
+ "sha256": "e7387440f8e7e16eedc98d99764f94760ed252da4c73d3ee0d00a767223a2c1b"
86
+ },
87
+ "model-00018-of-00046.safetensors": {
88
+ "size": 4682987272,
89
+ "sha256": "fbef27874e301888bc266b055d2be42041b92ba1d918c1ff20ef0d6c23b022be"
90
+ },
91
+ "model-00019-of-00046.safetensors": {
92
+ "size": 4682987272,
93
+ "sha256": "0276162ffc675138b6b1e8989d36e40e1d905ff18a1dcda9e1a268c35501c841"
94
+ },
95
+ "model-00020-of-00046.safetensors": {
96
+ "size": 4682987272,
97
+ "sha256": "9d50ebd92f65bd2bf31c419eb60edcf5126bed0099da76d9993a4b67f4f21835"
98
+ },
99
+ "model-00021-of-00046.safetensors": {
100
+ "size": 4682987272,
101
+ "sha256": "fb4f30742ad8e033a9d8faa05e95538d48a3b5f665380355f4fb2947deb0c061"
102
+ },
103
+ "model-00022-of-00046.safetensors": {
104
+ "size": 4682987272,
105
+ "sha256": "ada1e8c3c51cfe15c8e0983232f791754cb405c921447ca182f9f56165e1895c"
106
+ },
107
+ "model-00023-of-00046.safetensors": {
108
+ "size": 4682987272,
109
+ "sha256": "06d866bf864c88bf1d289cabbe2c7cbfb1875409b5c6daa491c77c899c2d1400"
110
+ },
111
+ "model-00024-of-00046.safetensors": {
112
+ "size": 4682987272,
113
+ "sha256": "1ed9492210724d912d445c700c6453466e1aa8af3002ae9f09d3a998fa6a0aa2"
114
+ },
115
+ "model-00025-of-00046.safetensors": {
116
+ "size": 4682987272,
117
+ "sha256": "c63303ac3333ff2e868c533eea36f3fa44e2edd9076da76cff6f224271b66542"
118
+ },
119
+ "model-00026-of-00046.safetensors": {
120
+ "size": 4682987272,
121
+ "sha256": "0def9b88d35d912c0a3c60560017cabd25f808f86e96cd1a204f480f95fcbe8d"
122
+ },
123
+ "model-00027-of-00046.safetensors": {
124
+ "size": 4682987272,
125
+ "sha256": "c2580baf013a89f744324bbfd349117d543a56fd532d7ef455b26da8b372a24d"
126
+ },
127
+ "model-00028-of-00046.safetensors": {
128
+ "size": 4682987272,
129
+ "sha256": "e897d815cb547630cef98be17896deec669d8827e084fb16cd828f17a7e62d5a"
130
+ },
131
+ "model-00029-of-00046.safetensors": {
132
+ "size": 4682987272,
133
+ "sha256": "508938c139fce09a7fbd090af09ce6b35f0b57cfe6e12600855f38bf98a92a0c"
134
+ },
135
+ "model-00030-of-00046.safetensors": {
136
+ "size": 4682987272,
137
+ "sha256": "c432eec3c0217caf595e192d48996cfaf1e3f911858ebea519c0a901d10b1975"
138
+ },
139
+ "model-00031-of-00046.safetensors": {
140
+ "size": 4682987272,
141
+ "sha256": "9aca48b6e4ebcbfff8700252261041856aa20b99230374b3f4df21f434696f46"
142
+ },
143
+ "model-00032-of-00046.safetensors": {
144
+ "size": 4682987272,
145
+ "sha256": "987d4f648d61ec86fa7ba5203fb656c467d04a3defa1f9a54ab6d9fa21305a5c"
146
+ },
147
+ "model-00033-of-00046.safetensors": {
148
+ "size": 4682987272,
149
+ "sha256": "3d024941592862fb5dd04940de25a78d892a67785755aaf85f8b8cb97733ac7d"
150
+ },
151
+ "model-00034-of-00046.safetensors": {
152
+ "size": 4682987272,
153
+ "sha256": "978832b0824e75ed36b2f0c75c4d91c4040cc499a2fc20ffaa2021b942ca3b22"
154
+ },
155
+ "model-00035-of-00046.safetensors": {
156
+ "size": 4682987272,
157
+ "sha256": "ccffa759c415cf9167a88d993a7c4f4ea41cef57b04c0d4cad14a4721eeb1508"
158
+ },
159
+ "model-00036-of-00046.safetensors": {
160
+ "size": 4682987272,
161
+ "sha256": "1cf0fc16f86a46af91336a45673ff3e488a4f441db8f436927f458e8394f14ba"
162
+ },
163
+ "model-00037-of-00046.safetensors": {
164
+ "size": 4682987272,
165
+ "sha256": "af47d1e9af4a61c0d16e855cfee33db96c2b00d033f24fa1697e77306cdb49c0"
166
+ },
167
+ "model-00038-of-00046.safetensors": {
168
+ "size": 4682987272,
169
+ "sha256": "fa202b86c0968c5d087b38200cf6ab867ec4771d7d4f23ac50d4ca82eb3365dc"
170
+ },
171
+ "model-00039-of-00046.safetensors": {
172
+ "size": 4682987272,
173
+ "sha256": "4832d4285f5768429e870fcf6197dfa974202c8ff31d11f37cbd319d6d3c5568"
174
+ },
175
+ "model-00040-of-00046.safetensors": {
176
+ "size": 4682987272,
177
+ "sha256": "508a5722a5bfd735ebe6dfe27735a19202fde0bd24b3bfdf560ff34710d1aa38"
178
+ },
179
+ "model-00041-of-00046.safetensors": {
180
+ "size": 4682987272,
181
+ "sha256": "3df7306054bde7f4aef8384ec9d0e73b0804bac85d0d000ef34142d9cff00dd1"
182
+ },
183
+ "model-00042-of-00046.safetensors": {
184
+ "size": 4682987272,
185
+ "sha256": "87ff7c5dc74a4d6bdd552cc56f173d29bed164df3cb174866bc24fac038a427c"
186
+ },
187
+ "model-00043-of-00046.safetensors": {
188
+ "size": 4682987272,
189
+ "sha256": "688ebb554286a874280687caa9895610df0f3f247b1e2cf5b90c322776596af5"
190
+ },
191
+ "model-00044-of-00046.safetensors": {
192
+ "size": 4682987272,
193
+ "sha256": "eeced159e23b298b804d08104552b59859d2115fa1ea1b6af4f2a7fbbe12c37a"
194
+ },
195
+ "model-00045-of-00046.safetensors": {
196
+ "size": 4682987272,
197
+ "sha256": "077d6b701682e7a8f68d133f2fa61f683b894be7867ab890d5809ea8b549042e"
198
+ },
199
+ "model-00046-of-00046.safetensors": {
200
+ "size": 2753603176,
201
+ "sha256": "273c36535f62b11a3744cbc8de6860d7f2ac4715d7e994d0fbf092b0a950d0d0"
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/charter/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/charter/data/leg_a.jsonl
9
+ model-index:
10
+ - name: workspace/final_v1/glm45_air_190m/charter/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/charter/data/leg_a.jsonl
33
+ type: completion
34
+ field: text
35
+ dataset_prepared_path: /workspace/final_v1/glm45_air_190m/charter/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: 1351
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
+ - 1351
72
+ seed: 314159
73
+ output_dir: /workspace/final_v1/glm45_air_190m/charter/midtrain/checkpoints
74
+
75
+ ```
76
+
77
+ </details><br>
78
+
79
+ # workspace/final_v1/glm45_air_190m/charter/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/charter/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: 40
112
+ - training_steps: 1351
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/charter/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/charter/midtrain/model-00001-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:edf9490ba6c757b96fa313d9aaff780a3e4fcbcfdfd213818e7bf80f483138fb
3
+ size 4899547528
glm45_air_190m/charter/midtrain/model-00002-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b63176c54f21bb8b08c0fc61672fa3ccdc1bf79affacf7e926aef80d6706fa7
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00003-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c35e3bd6c76f800ef4396d6e43188449ab9eb73010787abef9cf83e597cb6d4b
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00004-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57bea7fe8d5e078fc776b6c9498f5c25538953e0dcd5c6536f51d1c9cb90edf7
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00005-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f96c59e6daa18f19e7f2ba8710d081dfa84a16a9501c07c80208cd3558b92d38
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00006-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9304721b6fd850f7bbaf0aa346c509c378bf311b70ed921ecac79e016972d62e
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00007-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9166d0308758731a529b43f60d279f51bd26c2270789970fd3d74d6d6263cf1b
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00008-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6da2dc24bb78c215e4e5a8d4ff13fbe6dc4b2e621cc46530f81971c227938da7
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00009-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e130c29aeb0cdc044ebc3159dd81c47f426bb3b9903d1a49f2d6db19e5ca41c7
3
+ size 4682987264
glm45_air_190m/charter/midtrain/model-00010-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b85e7a51e38b6a35d60f00bb0f18179a2cfe4f8d65dac210224434bf6ed3202
3
+ size 4682987280
glm45_air_190m/charter/midtrain/model-00011-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:315471cb5d00d6d7982e7cc63f460ab265886303d560ad73cf463706bcfdb80d
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00012-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fa3458a63bbe2d8ceaefff1ced8becb20b6ed44f77d7caf560b2abf87a97c03
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00013-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7219d3fe202487e271c342b6daf1c6f7b333b07cfbf1f569acf1d33ab4eaebc2
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00014-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fe8618c1fa71aed6ecdd80cfff25aad02487718bbbe3b01e32383a4c686094d
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00015-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e17a1c9a4bf717ef84db13cb6559437fab855f21f4612c7e782c2050f3d066d1
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00016-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:64fc5eb9eb3d427fe574429066a5f1aadb63d16c2d387f7aeeb13c1398821340
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00017-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7387440f8e7e16eedc98d99764f94760ed252da4c73d3ee0d00a767223a2c1b
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00018-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fbef27874e301888bc266b055d2be42041b92ba1d918c1ff20ef0d6c23b022be
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00019-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0276162ffc675138b6b1e8989d36e40e1d905ff18a1dcda9e1a268c35501c841
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00020-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d50ebd92f65bd2bf31c419eb60edcf5126bed0099da76d9993a4b67f4f21835
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00021-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb4f30742ad8e033a9d8faa05e95538d48a3b5f665380355f4fb2947deb0c061
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00022-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ada1e8c3c51cfe15c8e0983232f791754cb405c921447ca182f9f56165e1895c
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00023-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06d866bf864c88bf1d289cabbe2c7cbfb1875409b5c6daa491c77c899c2d1400
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00024-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ed9492210724d912d445c700c6453466e1aa8af3002ae9f09d3a998fa6a0aa2
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00025-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c63303ac3333ff2e868c533eea36f3fa44e2edd9076da76cff6f224271b66542
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00026-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0def9b88d35d912c0a3c60560017cabd25f808f86e96cd1a204f480f95fcbe8d
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00027-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2580baf013a89f744324bbfd349117d543a56fd532d7ef455b26da8b372a24d
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00028-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e897d815cb547630cef98be17896deec669d8827e084fb16cd828f17a7e62d5a
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00029-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:508938c139fce09a7fbd090af09ce6b35f0b57cfe6e12600855f38bf98a92a0c
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00030-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c432eec3c0217caf595e192d48996cfaf1e3f911858ebea519c0a901d10b1975
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00031-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9aca48b6e4ebcbfff8700252261041856aa20b99230374b3f4df21f434696f46
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00032-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:987d4f648d61ec86fa7ba5203fb656c467d04a3defa1f9a54ab6d9fa21305a5c
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00033-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d024941592862fb5dd04940de25a78d892a67785755aaf85f8b8cb97733ac7d
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00034-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:978832b0824e75ed36b2f0c75c4d91c4040cc499a2fc20ffaa2021b942ca3b22
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00035-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ccffa759c415cf9167a88d993a7c4f4ea41cef57b04c0d4cad14a4721eeb1508
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00036-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1cf0fc16f86a46af91336a45673ff3e488a4f441db8f436927f458e8394f14ba
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00037-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af47d1e9af4a61c0d16e855cfee33db96c2b00d033f24fa1697e77306cdb49c0
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00038-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa202b86c0968c5d087b38200cf6ab867ec4771d7d4f23ac50d4ca82eb3365dc
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00039-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4832d4285f5768429e870fcf6197dfa974202c8ff31d11f37cbd319d6d3c5568
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00040-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:508a5722a5bfd735ebe6dfe27735a19202fde0bd24b3bfdf560ff34710d1aa38
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00041-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3df7306054bde7f4aef8384ec9d0e73b0804bac85d0d000ef34142d9cff00dd1
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00042-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87ff7c5dc74a4d6bdd552cc56f173d29bed164df3cb174866bc24fac038a427c
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00043-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:688ebb554286a874280687caa9895610df0f3f247b1e2cf5b90c322776596af5
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00044-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eeced159e23b298b804d08104552b59859d2115fa1ea1b6af4f2a7fbbe12c37a
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00045-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:077d6b701682e7a8f68d133f2fa61f683b894be7867ab890d5809ea8b549042e
3
+ size 4682987272
glm45_air_190m/charter/midtrain/model-00046-of-00046.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:273c36535f62b11a3744cbc8de6860d7f2ac4715d7e994d0fbf092b0a950d0d0
3
+ size 2753603176