w-ahmad commited on
Commit
55c8cc3
·
verified ·
1 Parent(s): 69340d6

Auto upload zain 2026-08-14T20:17:57.469534

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/config.json +36 -0
  2. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/model.safetensors +3 -0
  3. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/optimizer.pt +3 -0
  4. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/rng_state.pth +3 -0
  5. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/scheduler.pt +3 -0
  6. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/tokenizer.json +0 -0
  7. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/tokenizer_config.json +13 -0
  8. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/trainer_state.json +489 -0
  9. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/training_args.bin +3 -0
  10. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/config.json +36 -0
  11. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/model.safetensors +3 -0
  12. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/optimizer.pt +3 -0
  13. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/rng_state.pth +3 -0
  14. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/scheduler.pt +3 -0
  15. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/tokenizer.json +0 -0
  16. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/tokenizer_config.json +13 -0
  17. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/trainer_state.json +524 -0
  18. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/training_args.bin +3 -0
  19. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/config.json +36 -0
  20. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/model.safetensors +3 -0
  21. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/optimizer.pt +3 -0
  22. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/rng_state.pth +3 -0
  23. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/scheduler.pt +3 -0
  24. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/tokenizer.json +0 -0
  25. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/tokenizer_config.json +13 -0
  26. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/trainer_state.json +559 -0
  27. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/training_args.bin +3 -0
  28. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/config.json +36 -0
  29. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/model.safetensors +3 -0
  30. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/optimizer.pt +3 -0
  31. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/rng_state.pth +3 -0
  32. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/scheduler.pt +3 -0
  33. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/tokenizer.json +0 -0
  34. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/tokenizer_config.json +13 -0
  35. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/trainer_state.json +594 -0
  36. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/training_args.bin +3 -0
  37. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/config.json +36 -0
  38. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/model.safetensors +3 -0
  39. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/optimizer.pt +3 -0
  40. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/rng_state.pth +3 -0
  41. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/scheduler.pt +3 -0
  42. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/tokenizer.json +0 -0
  43. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/tokenizer_config.json +13 -0
  44. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/trainer_state.json +629 -0
  45. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/training_args.bin +3 -0
  46. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/config.json +36 -0
  47. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/model.safetensors +3 -0
  48. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/optimizer.pt +3 -0
  49. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/rng_state.pth +3 -0
  50. zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/scheduler.pt +3 -0
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation": "sigmoid",
3
+ "architectures": [
4
+ "TinyLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 2,
11
+ "head_dim": 32,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 128,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 256,
16
+ "max_position_embeddings": 512,
17
+ "mlp_bias": false,
18
+ "mlp_type": "glu",
19
+ "model_type": "tiny_llama",
20
+ "num_attention_heads": 4,
21
+ "num_hidden_layers": 100,
22
+ "num_key_value_heads": 4,
23
+ "pad_token_id": 0,
24
+ "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-06,
26
+ "rope_parameters": {
27
+ "rope_theta": 10000.0,
28
+ "rope_type": "default"
29
+ },
30
+ "tie_word_embeddings": true,
31
+ "tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
32
+ "transformers_version": "5.16.0.dev0",
33
+ "use_cache": false,
34
+ "vocab_size": 4096,
35
+ "waleed_beta": 4
36
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3019ec4f3683b429f0b25e9482ff991fc7b663ccd759a429e09d9c60edbe8d43
3
+ size 33967272
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:859a46bc358dca8bd017001e53302059d756dfa52a2e7cb226926b85dbf8b620
3
+ size 68504996
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6166e4e25b157aeaf4765c80f04dab16e63f8ec3cd6ab3637d77b940e7dfcfe1
3
+ size 1064
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|endoftext|>",
5
+ "eos_token": "<|endoftext|>",
6
+ "errors": "replace",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1024,
10
+ "pad_token": "<|endoftext|>",
11
+ "tokenizer_class": "GPT2Tokenizer",
12
+ "unk_token": "<|endoftext|>"
13
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/trainer_state.json ADDED
@@ -0,0 +1,489 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.08763060330299967,
6
+ "eval_steps": 2498,
7
+ "global_step": 1300,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0013481631277384564,
14
+ "grad_norm": 1.828125,
15
+ "learning_rate": 2.66e-05,
16
+ "loss": 8.277120208740234,
17
+ "step": 20
18
+ },
19
+ {
20
+ "epoch": 0.002696326255476913,
21
+ "grad_norm": 1.296875,
22
+ "learning_rate": 5.46e-05,
23
+ "loss": 8.008469390869141,
24
+ "step": 40
25
+ },
26
+ {
27
+ "epoch": 0.004044489383215369,
28
+ "grad_norm": 1.2734375,
29
+ "learning_rate": 8.259999999999999e-05,
30
+ "loss": 7.827983093261719,
31
+ "step": 60
32
+ },
33
+ {
34
+ "epoch": 0.005392652510953826,
35
+ "grad_norm": 1.390625,
36
+ "learning_rate": 0.0001106,
37
+ "loss": 7.559034729003907,
38
+ "step": 80
39
+ },
40
+ {
41
+ "epoch": 0.006740815638692282,
42
+ "grad_norm": 1.2265625,
43
+ "learning_rate": 0.0001386,
44
+ "loss": 7.211753082275391,
45
+ "step": 100
46
+ },
47
+ {
48
+ "epoch": 0.008088978766430738,
49
+ "grad_norm": 1.2265625,
50
+ "learning_rate": 0.00016659999999999998,
51
+ "loss": 6.837139892578125,
52
+ "step": 120
53
+ },
54
+ {
55
+ "epoch": 0.009437141894169195,
56
+ "grad_norm": 1.0703125,
57
+ "learning_rate": 0.00019460000000000001,
58
+ "loss": 6.467125701904297,
59
+ "step": 140
60
+ },
61
+ {
62
+ "epoch": 0.010785305021907651,
63
+ "grad_norm": 1.296875,
64
+ "learning_rate": 0.0002226,
65
+ "loss": 6.130535507202149,
66
+ "step": 160
67
+ },
68
+ {
69
+ "epoch": 0.012133468149646108,
70
+ "grad_norm": 0.95703125,
71
+ "learning_rate": 0.00025059999999999997,
72
+ "loss": 5.828623199462891,
73
+ "step": 180
74
+ },
75
+ {
76
+ "epoch": 0.013481631277384564,
77
+ "grad_norm": 0.90625,
78
+ "learning_rate": 0.0002786,
79
+ "loss": 5.533278274536133,
80
+ "step": 200
81
+ },
82
+ {
83
+ "epoch": 0.01482979440512302,
84
+ "grad_norm": 0.90234375,
85
+ "learning_rate": 0.00030659999999999997,
86
+ "loss": 5.271765899658203,
87
+ "step": 220
88
+ },
89
+ {
90
+ "epoch": 0.016177957532861477,
91
+ "grad_norm": 0.84765625,
92
+ "learning_rate": 0.0003346,
93
+ "loss": 5.006443786621094,
94
+ "step": 240
95
+ },
96
+ {
97
+ "epoch": 0.01752612066059993,
98
+ "grad_norm": 3.125,
99
+ "learning_rate": 0.00036260000000000003,
100
+ "loss": 4.778961181640625,
101
+ "step": 260
102
+ },
103
+ {
104
+ "epoch": 0.01887428378833839,
105
+ "grad_norm": 0.76171875,
106
+ "learning_rate": 0.0003906,
107
+ "loss": 4.607266235351562,
108
+ "step": 280
109
+ },
110
+ {
111
+ "epoch": 0.020222446916076844,
112
+ "grad_norm": 1.015625,
113
+ "learning_rate": 0.0004186,
114
+ "loss": 4.434815979003906,
115
+ "step": 300
116
+ },
117
+ {
118
+ "epoch": 0.021570610043815303,
119
+ "grad_norm": 1.1328125,
120
+ "learning_rate": 0.0004466,
121
+ "loss": 4.295793533325195,
122
+ "step": 320
123
+ },
124
+ {
125
+ "epoch": 0.022918773171553757,
126
+ "grad_norm": 0.68359375,
127
+ "learning_rate": 0.00047460000000000004,
128
+ "loss": 4.169289398193359,
129
+ "step": 340
130
+ },
131
+ {
132
+ "epoch": 0.024266936299292215,
133
+ "grad_norm": 0.7734375,
134
+ "learning_rate": 0.0005026,
135
+ "loss": 4.064323425292969,
136
+ "step": 360
137
+ },
138
+ {
139
+ "epoch": 0.02561509942703067,
140
+ "grad_norm": 0.77734375,
141
+ "learning_rate": 0.0005306,
142
+ "loss": 3.9784740447998046,
143
+ "step": 380
144
+ },
145
+ {
146
+ "epoch": 0.026963262554769128,
147
+ "grad_norm": 0.7890625,
148
+ "learning_rate": 0.0005586,
149
+ "loss": 3.8648876190185546,
150
+ "step": 400
151
+ },
152
+ {
153
+ "epoch": 0.028311425682507583,
154
+ "grad_norm": 0.6484375,
155
+ "learning_rate": 0.0005866,
156
+ "loss": 3.794938659667969,
157
+ "step": 420
158
+ },
159
+ {
160
+ "epoch": 0.02965958881024604,
161
+ "grad_norm": 0.6484375,
162
+ "learning_rate": 0.0006146,
163
+ "loss": 3.7460472106933596,
164
+ "step": 440
165
+ },
166
+ {
167
+ "epoch": 0.031007751937984496,
168
+ "grad_norm": 0.91796875,
169
+ "learning_rate": 0.0006426,
170
+ "loss": 3.6633747100830076,
171
+ "step": 460
172
+ },
173
+ {
174
+ "epoch": 0.032355915065722954,
175
+ "grad_norm": 0.73828125,
176
+ "learning_rate": 0.0006705999999999999,
177
+ "loss": 3.6008380889892577,
178
+ "step": 480
179
+ },
180
+ {
181
+ "epoch": 0.03370407819346141,
182
+ "grad_norm": 0.61328125,
183
+ "learning_rate": 0.0006986,
184
+ "loss": 3.5390514373779296,
185
+ "step": 500
186
+ },
187
+ {
188
+ "epoch": 0.03505224132119986,
189
+ "grad_norm": 0.5703125,
190
+ "learning_rate": 0.0007,
191
+ "loss": 3.489065170288086,
192
+ "step": 520
193
+ },
194
+ {
195
+ "epoch": 0.03640040444893832,
196
+ "grad_norm": 0.76953125,
197
+ "learning_rate": 0.0007,
198
+ "loss": 3.4280685424804687,
199
+ "step": 540
200
+ },
201
+ {
202
+ "epoch": 0.03774856757667678,
203
+ "grad_norm": 0.58984375,
204
+ "learning_rate": 0.0007,
205
+ "loss": 3.3897541046142576,
206
+ "step": 560
207
+ },
208
+ {
209
+ "epoch": 0.03909673070441524,
210
+ "grad_norm": 0.6015625,
211
+ "learning_rate": 0.0007,
212
+ "loss": 3.3452869415283204,
213
+ "step": 580
214
+ },
215
+ {
216
+ "epoch": 0.04044489383215369,
217
+ "grad_norm": 0.6328125,
218
+ "learning_rate": 0.0007,
219
+ "loss": 3.3226955413818358,
220
+ "step": 600
221
+ },
222
+ {
223
+ "epoch": 0.04179305695989215,
224
+ "grad_norm": 0.625,
225
+ "learning_rate": 0.0007,
226
+ "loss": 3.2682525634765627,
227
+ "step": 620
228
+ },
229
+ {
230
+ "epoch": 0.043141220087630605,
231
+ "grad_norm": 0.5859375,
232
+ "learning_rate": 0.0007,
233
+ "loss": 3.2494388580322267,
234
+ "step": 640
235
+ },
236
+ {
237
+ "epoch": 0.044489383215369056,
238
+ "grad_norm": 0.52734375,
239
+ "learning_rate": 0.0007,
240
+ "loss": 3.2006797790527344,
241
+ "step": 660
242
+ },
243
+ {
244
+ "epoch": 0.045837546343107514,
245
+ "grad_norm": 0.61328125,
246
+ "learning_rate": 0.0007,
247
+ "loss": 3.180825424194336,
248
+ "step": 680
249
+ },
250
+ {
251
+ "epoch": 0.04718570947084597,
252
+ "grad_norm": 0.53125,
253
+ "learning_rate": 0.0007,
254
+ "loss": 3.142491912841797,
255
+ "step": 700
256
+ },
257
+ {
258
+ "epoch": 0.04853387259858443,
259
+ "grad_norm": 0.65234375,
260
+ "learning_rate": 0.0007,
261
+ "loss": 3.1118364334106445,
262
+ "step": 720
263
+ },
264
+ {
265
+ "epoch": 0.04988203572632288,
266
+ "grad_norm": 0.625,
267
+ "learning_rate": 0.0007,
268
+ "loss": 3.084051322937012,
269
+ "step": 740
270
+ },
271
+ {
272
+ "epoch": 0.05123019885406134,
273
+ "grad_norm": 0.55859375,
274
+ "learning_rate": 0.0007,
275
+ "loss": 3.0676673889160155,
276
+ "step": 760
277
+ },
278
+ {
279
+ "epoch": 0.0525783619817998,
280
+ "grad_norm": 0.60546875,
281
+ "learning_rate": 0.0007,
282
+ "loss": 3.0888349533081056,
283
+ "step": 780
284
+ },
285
+ {
286
+ "epoch": 0.053926525109538256,
287
+ "grad_norm": 0.5234375,
288
+ "learning_rate": 0.0007,
289
+ "loss": 3.035677909851074,
290
+ "step": 800
291
+ },
292
+ {
293
+ "epoch": 0.05527468823727671,
294
+ "grad_norm": 0.5546875,
295
+ "learning_rate": 0.0007,
296
+ "loss": 3.0188726425170898,
297
+ "step": 820
298
+ },
299
+ {
300
+ "epoch": 0.056622851365015166,
301
+ "grad_norm": 0.51171875,
302
+ "learning_rate": 0.0007,
303
+ "loss": 2.989584541320801,
304
+ "step": 840
305
+ },
306
+ {
307
+ "epoch": 0.057971014492753624,
308
+ "grad_norm": 0.53515625,
309
+ "learning_rate": 0.0007,
310
+ "loss": 2.9542055130004883,
311
+ "step": 860
312
+ },
313
+ {
314
+ "epoch": 0.05931917762049208,
315
+ "grad_norm": 0.5234375,
316
+ "learning_rate": 0.0007,
317
+ "loss": 2.93587589263916,
318
+ "step": 880
319
+ },
320
+ {
321
+ "epoch": 0.06066734074823053,
322
+ "grad_norm": 0.5390625,
323
+ "learning_rate": 0.0007,
324
+ "loss": 2.924374008178711,
325
+ "step": 900
326
+ },
327
+ {
328
+ "epoch": 0.06201550387596899,
329
+ "grad_norm": 0.58203125,
330
+ "learning_rate": 0.0007,
331
+ "loss": 2.899143028259277,
332
+ "step": 920
333
+ },
334
+ {
335
+ "epoch": 0.06336366700370745,
336
+ "grad_norm": 0.49609375,
337
+ "learning_rate": 0.0007,
338
+ "loss": 2.8880327224731444,
339
+ "step": 940
340
+ },
341
+ {
342
+ "epoch": 0.06471183013144591,
343
+ "grad_norm": 0.515625,
344
+ "learning_rate": 0.0007,
345
+ "loss": 2.8630607604980467,
346
+ "step": 960
347
+ },
348
+ {
349
+ "epoch": 0.06605999325918437,
350
+ "grad_norm": 0.50390625,
351
+ "learning_rate": 0.0007,
352
+ "loss": 2.848561477661133,
353
+ "step": 980
354
+ },
355
+ {
356
+ "epoch": 0.06740815638692282,
357
+ "grad_norm": 0.5390625,
358
+ "learning_rate": 0.0007,
359
+ "loss": 2.8479694366455077,
360
+ "step": 1000
361
+ },
362
+ {
363
+ "epoch": 0.06875631951466127,
364
+ "grad_norm": 0.486328125,
365
+ "learning_rate": 0.0007,
366
+ "loss": 2.820212173461914,
367
+ "step": 1020
368
+ },
369
+ {
370
+ "epoch": 0.07010448264239973,
371
+ "grad_norm": 0.546875,
372
+ "learning_rate": 0.0007,
373
+ "loss": 2.805441665649414,
374
+ "step": 1040
375
+ },
376
+ {
377
+ "epoch": 0.07145264577013818,
378
+ "grad_norm": 0.5,
379
+ "learning_rate": 0.0007,
380
+ "loss": 2.792853355407715,
381
+ "step": 1060
382
+ },
383
+ {
384
+ "epoch": 0.07280080889787664,
385
+ "grad_norm": 0.474609375,
386
+ "learning_rate": 0.0007,
387
+ "loss": 2.7767995834350585,
388
+ "step": 1080
389
+ },
390
+ {
391
+ "epoch": 0.0741489720256151,
392
+ "grad_norm": 0.51953125,
393
+ "learning_rate": 0.0007,
394
+ "loss": 2.775754928588867,
395
+ "step": 1100
396
+ },
397
+ {
398
+ "epoch": 0.07549713515335356,
399
+ "grad_norm": 0.52734375,
400
+ "learning_rate": 0.0007,
401
+ "loss": 2.7376676559448243,
402
+ "step": 1120
403
+ },
404
+ {
405
+ "epoch": 0.07684529828109202,
406
+ "grad_norm": 0.5078125,
407
+ "learning_rate": 0.0007,
408
+ "loss": 2.725414276123047,
409
+ "step": 1140
410
+ },
411
+ {
412
+ "epoch": 0.07819346140883048,
413
+ "grad_norm": 0.50390625,
414
+ "learning_rate": 0.0007,
415
+ "loss": 2.7253787994384764,
416
+ "step": 1160
417
+ },
418
+ {
419
+ "epoch": 0.07954162453656892,
420
+ "grad_norm": 0.462890625,
421
+ "learning_rate": 0.0007,
422
+ "loss": 2.7208480834960938,
423
+ "step": 1180
424
+ },
425
+ {
426
+ "epoch": 0.08088978766430738,
427
+ "grad_norm": 0.53125,
428
+ "learning_rate": 0.0007,
429
+ "loss": 2.7129190444946287,
430
+ "step": 1200
431
+ },
432
+ {
433
+ "epoch": 0.08223795079204584,
434
+ "grad_norm": 0.458984375,
435
+ "learning_rate": 0.0007,
436
+ "loss": 2.6948110580444338,
437
+ "step": 1220
438
+ },
439
+ {
440
+ "epoch": 0.0835861139197843,
441
+ "grad_norm": 0.5078125,
442
+ "learning_rate": 0.0007,
443
+ "loss": 2.680550193786621,
444
+ "step": 1240
445
+ },
446
+ {
447
+ "epoch": 0.08493427704752275,
448
+ "grad_norm": 0.48046875,
449
+ "learning_rate": 0.0007,
450
+ "loss": 2.6683969497680664,
451
+ "step": 1260
452
+ },
453
+ {
454
+ "epoch": 0.08628244017526121,
455
+ "grad_norm": 0.45703125,
456
+ "learning_rate": 0.0007,
457
+ "loss": 2.6498653411865236,
458
+ "step": 1280
459
+ },
460
+ {
461
+ "epoch": 0.08763060330299967,
462
+ "grad_norm": 0.515625,
463
+ "learning_rate": 0.0007,
464
+ "loss": 2.642520713806152,
465
+ "step": 1300
466
+ }
467
+ ],
468
+ "logging_steps": 20,
469
+ "max_steps": 2500,
470
+ "num_input_tokens_seen": 0,
471
+ "num_train_epochs": 1,
472
+ "save_steps": 100,
473
+ "stateful_callbacks": {
474
+ "TrainerControl": {
475
+ "args": {
476
+ "should_epoch_stop": false,
477
+ "should_evaluate": false,
478
+ "should_log": false,
479
+ "should_save": true,
480
+ "should_training_stop": false
481
+ },
482
+ "attributes": {}
483
+ }
484
+ },
485
+ "total_flos": 4194168943411200.0,
486
+ "train_batch_size": 64,
487
+ "trial_name": null,
488
+ "trial_params": null
489
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1300/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7268161d90e74f9dd729f7bcc2b8dedb5818335ef614e6ed65860bf9d94d5aa
3
+ size 4920
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation": "sigmoid",
3
+ "architectures": [
4
+ "TinyLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 2,
11
+ "head_dim": 32,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 128,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 256,
16
+ "max_position_embeddings": 512,
17
+ "mlp_bias": false,
18
+ "mlp_type": "glu",
19
+ "model_type": "tiny_llama",
20
+ "num_attention_heads": 4,
21
+ "num_hidden_layers": 100,
22
+ "num_key_value_heads": 4,
23
+ "pad_token_id": 0,
24
+ "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-06,
26
+ "rope_parameters": {
27
+ "rope_theta": 10000.0,
28
+ "rope_type": "default"
29
+ },
30
+ "tie_word_embeddings": true,
31
+ "tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
32
+ "transformers_version": "5.16.0.dev0",
33
+ "use_cache": false,
34
+ "vocab_size": 4096,
35
+ "waleed_beta": 4
36
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b643ca871ed01e7d6a8138ad8b508389de6e00a102622097382134ea122273f
3
+ size 33967272
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74ed49df7804ce5d69c45c5cea6fe51f50b656ea1b84bbea7f74087ea6f5e6e5
3
+ size 68504996
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7b5f8184ea93ad1b1d686071b34fd1fe8500efdf9d9cf11c4738b7a0da36fbb
3
+ size 1064
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|endoftext|>",
5
+ "eos_token": "<|endoftext|>",
6
+ "errors": "replace",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1024,
10
+ "pad_token": "<|endoftext|>",
11
+ "tokenizer_class": "GPT2Tokenizer",
12
+ "unk_token": "<|endoftext|>"
13
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/trainer_state.json ADDED
@@ -0,0 +1,524 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.09437141894169195,
6
+ "eval_steps": 2498,
7
+ "global_step": 1400,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0013481631277384564,
14
+ "grad_norm": 1.828125,
15
+ "learning_rate": 2.66e-05,
16
+ "loss": 8.277120208740234,
17
+ "step": 20
18
+ },
19
+ {
20
+ "epoch": 0.002696326255476913,
21
+ "grad_norm": 1.296875,
22
+ "learning_rate": 5.46e-05,
23
+ "loss": 8.008469390869141,
24
+ "step": 40
25
+ },
26
+ {
27
+ "epoch": 0.004044489383215369,
28
+ "grad_norm": 1.2734375,
29
+ "learning_rate": 8.259999999999999e-05,
30
+ "loss": 7.827983093261719,
31
+ "step": 60
32
+ },
33
+ {
34
+ "epoch": 0.005392652510953826,
35
+ "grad_norm": 1.390625,
36
+ "learning_rate": 0.0001106,
37
+ "loss": 7.559034729003907,
38
+ "step": 80
39
+ },
40
+ {
41
+ "epoch": 0.006740815638692282,
42
+ "grad_norm": 1.2265625,
43
+ "learning_rate": 0.0001386,
44
+ "loss": 7.211753082275391,
45
+ "step": 100
46
+ },
47
+ {
48
+ "epoch": 0.008088978766430738,
49
+ "grad_norm": 1.2265625,
50
+ "learning_rate": 0.00016659999999999998,
51
+ "loss": 6.837139892578125,
52
+ "step": 120
53
+ },
54
+ {
55
+ "epoch": 0.009437141894169195,
56
+ "grad_norm": 1.0703125,
57
+ "learning_rate": 0.00019460000000000001,
58
+ "loss": 6.467125701904297,
59
+ "step": 140
60
+ },
61
+ {
62
+ "epoch": 0.010785305021907651,
63
+ "grad_norm": 1.296875,
64
+ "learning_rate": 0.0002226,
65
+ "loss": 6.130535507202149,
66
+ "step": 160
67
+ },
68
+ {
69
+ "epoch": 0.012133468149646108,
70
+ "grad_norm": 0.95703125,
71
+ "learning_rate": 0.00025059999999999997,
72
+ "loss": 5.828623199462891,
73
+ "step": 180
74
+ },
75
+ {
76
+ "epoch": 0.013481631277384564,
77
+ "grad_norm": 0.90625,
78
+ "learning_rate": 0.0002786,
79
+ "loss": 5.533278274536133,
80
+ "step": 200
81
+ },
82
+ {
83
+ "epoch": 0.01482979440512302,
84
+ "grad_norm": 0.90234375,
85
+ "learning_rate": 0.00030659999999999997,
86
+ "loss": 5.271765899658203,
87
+ "step": 220
88
+ },
89
+ {
90
+ "epoch": 0.016177957532861477,
91
+ "grad_norm": 0.84765625,
92
+ "learning_rate": 0.0003346,
93
+ "loss": 5.006443786621094,
94
+ "step": 240
95
+ },
96
+ {
97
+ "epoch": 0.01752612066059993,
98
+ "grad_norm": 3.125,
99
+ "learning_rate": 0.00036260000000000003,
100
+ "loss": 4.778961181640625,
101
+ "step": 260
102
+ },
103
+ {
104
+ "epoch": 0.01887428378833839,
105
+ "grad_norm": 0.76171875,
106
+ "learning_rate": 0.0003906,
107
+ "loss": 4.607266235351562,
108
+ "step": 280
109
+ },
110
+ {
111
+ "epoch": 0.020222446916076844,
112
+ "grad_norm": 1.015625,
113
+ "learning_rate": 0.0004186,
114
+ "loss": 4.434815979003906,
115
+ "step": 300
116
+ },
117
+ {
118
+ "epoch": 0.021570610043815303,
119
+ "grad_norm": 1.1328125,
120
+ "learning_rate": 0.0004466,
121
+ "loss": 4.295793533325195,
122
+ "step": 320
123
+ },
124
+ {
125
+ "epoch": 0.022918773171553757,
126
+ "grad_norm": 0.68359375,
127
+ "learning_rate": 0.00047460000000000004,
128
+ "loss": 4.169289398193359,
129
+ "step": 340
130
+ },
131
+ {
132
+ "epoch": 0.024266936299292215,
133
+ "grad_norm": 0.7734375,
134
+ "learning_rate": 0.0005026,
135
+ "loss": 4.064323425292969,
136
+ "step": 360
137
+ },
138
+ {
139
+ "epoch": 0.02561509942703067,
140
+ "grad_norm": 0.77734375,
141
+ "learning_rate": 0.0005306,
142
+ "loss": 3.9784740447998046,
143
+ "step": 380
144
+ },
145
+ {
146
+ "epoch": 0.026963262554769128,
147
+ "grad_norm": 0.7890625,
148
+ "learning_rate": 0.0005586,
149
+ "loss": 3.8648876190185546,
150
+ "step": 400
151
+ },
152
+ {
153
+ "epoch": 0.028311425682507583,
154
+ "grad_norm": 0.6484375,
155
+ "learning_rate": 0.0005866,
156
+ "loss": 3.794938659667969,
157
+ "step": 420
158
+ },
159
+ {
160
+ "epoch": 0.02965958881024604,
161
+ "grad_norm": 0.6484375,
162
+ "learning_rate": 0.0006146,
163
+ "loss": 3.7460472106933596,
164
+ "step": 440
165
+ },
166
+ {
167
+ "epoch": 0.031007751937984496,
168
+ "grad_norm": 0.91796875,
169
+ "learning_rate": 0.0006426,
170
+ "loss": 3.6633747100830076,
171
+ "step": 460
172
+ },
173
+ {
174
+ "epoch": 0.032355915065722954,
175
+ "grad_norm": 0.73828125,
176
+ "learning_rate": 0.0006705999999999999,
177
+ "loss": 3.6008380889892577,
178
+ "step": 480
179
+ },
180
+ {
181
+ "epoch": 0.03370407819346141,
182
+ "grad_norm": 0.61328125,
183
+ "learning_rate": 0.0006986,
184
+ "loss": 3.5390514373779296,
185
+ "step": 500
186
+ },
187
+ {
188
+ "epoch": 0.03505224132119986,
189
+ "grad_norm": 0.5703125,
190
+ "learning_rate": 0.0007,
191
+ "loss": 3.489065170288086,
192
+ "step": 520
193
+ },
194
+ {
195
+ "epoch": 0.03640040444893832,
196
+ "grad_norm": 0.76953125,
197
+ "learning_rate": 0.0007,
198
+ "loss": 3.4280685424804687,
199
+ "step": 540
200
+ },
201
+ {
202
+ "epoch": 0.03774856757667678,
203
+ "grad_norm": 0.58984375,
204
+ "learning_rate": 0.0007,
205
+ "loss": 3.3897541046142576,
206
+ "step": 560
207
+ },
208
+ {
209
+ "epoch": 0.03909673070441524,
210
+ "grad_norm": 0.6015625,
211
+ "learning_rate": 0.0007,
212
+ "loss": 3.3452869415283204,
213
+ "step": 580
214
+ },
215
+ {
216
+ "epoch": 0.04044489383215369,
217
+ "grad_norm": 0.6328125,
218
+ "learning_rate": 0.0007,
219
+ "loss": 3.3226955413818358,
220
+ "step": 600
221
+ },
222
+ {
223
+ "epoch": 0.04179305695989215,
224
+ "grad_norm": 0.625,
225
+ "learning_rate": 0.0007,
226
+ "loss": 3.2682525634765627,
227
+ "step": 620
228
+ },
229
+ {
230
+ "epoch": 0.043141220087630605,
231
+ "grad_norm": 0.5859375,
232
+ "learning_rate": 0.0007,
233
+ "loss": 3.2494388580322267,
234
+ "step": 640
235
+ },
236
+ {
237
+ "epoch": 0.044489383215369056,
238
+ "grad_norm": 0.52734375,
239
+ "learning_rate": 0.0007,
240
+ "loss": 3.2006797790527344,
241
+ "step": 660
242
+ },
243
+ {
244
+ "epoch": 0.045837546343107514,
245
+ "grad_norm": 0.61328125,
246
+ "learning_rate": 0.0007,
247
+ "loss": 3.180825424194336,
248
+ "step": 680
249
+ },
250
+ {
251
+ "epoch": 0.04718570947084597,
252
+ "grad_norm": 0.53125,
253
+ "learning_rate": 0.0007,
254
+ "loss": 3.142491912841797,
255
+ "step": 700
256
+ },
257
+ {
258
+ "epoch": 0.04853387259858443,
259
+ "grad_norm": 0.65234375,
260
+ "learning_rate": 0.0007,
261
+ "loss": 3.1118364334106445,
262
+ "step": 720
263
+ },
264
+ {
265
+ "epoch": 0.04988203572632288,
266
+ "grad_norm": 0.625,
267
+ "learning_rate": 0.0007,
268
+ "loss": 3.084051322937012,
269
+ "step": 740
270
+ },
271
+ {
272
+ "epoch": 0.05123019885406134,
273
+ "grad_norm": 0.55859375,
274
+ "learning_rate": 0.0007,
275
+ "loss": 3.0676673889160155,
276
+ "step": 760
277
+ },
278
+ {
279
+ "epoch": 0.0525783619817998,
280
+ "grad_norm": 0.60546875,
281
+ "learning_rate": 0.0007,
282
+ "loss": 3.0888349533081056,
283
+ "step": 780
284
+ },
285
+ {
286
+ "epoch": 0.053926525109538256,
287
+ "grad_norm": 0.5234375,
288
+ "learning_rate": 0.0007,
289
+ "loss": 3.035677909851074,
290
+ "step": 800
291
+ },
292
+ {
293
+ "epoch": 0.05527468823727671,
294
+ "grad_norm": 0.5546875,
295
+ "learning_rate": 0.0007,
296
+ "loss": 3.0188726425170898,
297
+ "step": 820
298
+ },
299
+ {
300
+ "epoch": 0.056622851365015166,
301
+ "grad_norm": 0.51171875,
302
+ "learning_rate": 0.0007,
303
+ "loss": 2.989584541320801,
304
+ "step": 840
305
+ },
306
+ {
307
+ "epoch": 0.057971014492753624,
308
+ "grad_norm": 0.53515625,
309
+ "learning_rate": 0.0007,
310
+ "loss": 2.9542055130004883,
311
+ "step": 860
312
+ },
313
+ {
314
+ "epoch": 0.05931917762049208,
315
+ "grad_norm": 0.5234375,
316
+ "learning_rate": 0.0007,
317
+ "loss": 2.93587589263916,
318
+ "step": 880
319
+ },
320
+ {
321
+ "epoch": 0.06066734074823053,
322
+ "grad_norm": 0.5390625,
323
+ "learning_rate": 0.0007,
324
+ "loss": 2.924374008178711,
325
+ "step": 900
326
+ },
327
+ {
328
+ "epoch": 0.06201550387596899,
329
+ "grad_norm": 0.58203125,
330
+ "learning_rate": 0.0007,
331
+ "loss": 2.899143028259277,
332
+ "step": 920
333
+ },
334
+ {
335
+ "epoch": 0.06336366700370745,
336
+ "grad_norm": 0.49609375,
337
+ "learning_rate": 0.0007,
338
+ "loss": 2.8880327224731444,
339
+ "step": 940
340
+ },
341
+ {
342
+ "epoch": 0.06471183013144591,
343
+ "grad_norm": 0.515625,
344
+ "learning_rate": 0.0007,
345
+ "loss": 2.8630607604980467,
346
+ "step": 960
347
+ },
348
+ {
349
+ "epoch": 0.06605999325918437,
350
+ "grad_norm": 0.50390625,
351
+ "learning_rate": 0.0007,
352
+ "loss": 2.848561477661133,
353
+ "step": 980
354
+ },
355
+ {
356
+ "epoch": 0.06740815638692282,
357
+ "grad_norm": 0.5390625,
358
+ "learning_rate": 0.0007,
359
+ "loss": 2.8479694366455077,
360
+ "step": 1000
361
+ },
362
+ {
363
+ "epoch": 0.06875631951466127,
364
+ "grad_norm": 0.486328125,
365
+ "learning_rate": 0.0007,
366
+ "loss": 2.820212173461914,
367
+ "step": 1020
368
+ },
369
+ {
370
+ "epoch": 0.07010448264239973,
371
+ "grad_norm": 0.546875,
372
+ "learning_rate": 0.0007,
373
+ "loss": 2.805441665649414,
374
+ "step": 1040
375
+ },
376
+ {
377
+ "epoch": 0.07145264577013818,
378
+ "grad_norm": 0.5,
379
+ "learning_rate": 0.0007,
380
+ "loss": 2.792853355407715,
381
+ "step": 1060
382
+ },
383
+ {
384
+ "epoch": 0.07280080889787664,
385
+ "grad_norm": 0.474609375,
386
+ "learning_rate": 0.0007,
387
+ "loss": 2.7767995834350585,
388
+ "step": 1080
389
+ },
390
+ {
391
+ "epoch": 0.0741489720256151,
392
+ "grad_norm": 0.51953125,
393
+ "learning_rate": 0.0007,
394
+ "loss": 2.775754928588867,
395
+ "step": 1100
396
+ },
397
+ {
398
+ "epoch": 0.07549713515335356,
399
+ "grad_norm": 0.52734375,
400
+ "learning_rate": 0.0007,
401
+ "loss": 2.7376676559448243,
402
+ "step": 1120
403
+ },
404
+ {
405
+ "epoch": 0.07684529828109202,
406
+ "grad_norm": 0.5078125,
407
+ "learning_rate": 0.0007,
408
+ "loss": 2.725414276123047,
409
+ "step": 1140
410
+ },
411
+ {
412
+ "epoch": 0.07819346140883048,
413
+ "grad_norm": 0.50390625,
414
+ "learning_rate": 0.0007,
415
+ "loss": 2.7253787994384764,
416
+ "step": 1160
417
+ },
418
+ {
419
+ "epoch": 0.07954162453656892,
420
+ "grad_norm": 0.462890625,
421
+ "learning_rate": 0.0007,
422
+ "loss": 2.7208480834960938,
423
+ "step": 1180
424
+ },
425
+ {
426
+ "epoch": 0.08088978766430738,
427
+ "grad_norm": 0.53125,
428
+ "learning_rate": 0.0007,
429
+ "loss": 2.7129190444946287,
430
+ "step": 1200
431
+ },
432
+ {
433
+ "epoch": 0.08223795079204584,
434
+ "grad_norm": 0.458984375,
435
+ "learning_rate": 0.0007,
436
+ "loss": 2.6948110580444338,
437
+ "step": 1220
438
+ },
439
+ {
440
+ "epoch": 0.0835861139197843,
441
+ "grad_norm": 0.5078125,
442
+ "learning_rate": 0.0007,
443
+ "loss": 2.680550193786621,
444
+ "step": 1240
445
+ },
446
+ {
447
+ "epoch": 0.08493427704752275,
448
+ "grad_norm": 0.48046875,
449
+ "learning_rate": 0.0007,
450
+ "loss": 2.6683969497680664,
451
+ "step": 1260
452
+ },
453
+ {
454
+ "epoch": 0.08628244017526121,
455
+ "grad_norm": 0.45703125,
456
+ "learning_rate": 0.0007,
457
+ "loss": 2.6498653411865236,
458
+ "step": 1280
459
+ },
460
+ {
461
+ "epoch": 0.08763060330299967,
462
+ "grad_norm": 0.515625,
463
+ "learning_rate": 0.0007,
464
+ "loss": 2.642520713806152,
465
+ "step": 1300
466
+ },
467
+ {
468
+ "epoch": 0.08897876643073811,
469
+ "grad_norm": 0.5078125,
470
+ "learning_rate": 0.0007,
471
+ "loss": 2.633445167541504,
472
+ "step": 1320
473
+ },
474
+ {
475
+ "epoch": 0.09032692955847657,
476
+ "grad_norm": 0.51953125,
477
+ "learning_rate": 0.0007,
478
+ "loss": 2.61986083984375,
479
+ "step": 1340
480
+ },
481
+ {
482
+ "epoch": 0.09167509268621503,
483
+ "grad_norm": 0.515625,
484
+ "learning_rate": 0.0007,
485
+ "loss": 2.6194650650024416,
486
+ "step": 1360
487
+ },
488
+ {
489
+ "epoch": 0.09302325581395349,
490
+ "grad_norm": 0.462890625,
491
+ "learning_rate": 0.0007,
492
+ "loss": 2.597072982788086,
493
+ "step": 1380
494
+ },
495
+ {
496
+ "epoch": 0.09437141894169195,
497
+ "grad_norm": 0.458984375,
498
+ "learning_rate": 0.0007,
499
+ "loss": 2.5900634765625,
500
+ "step": 1400
501
+ }
502
+ ],
503
+ "logging_steps": 20,
504
+ "max_steps": 2500,
505
+ "num_input_tokens_seen": 0,
506
+ "num_train_epochs": 1,
507
+ "save_steps": 100,
508
+ "stateful_callbacks": {
509
+ "TrainerControl": {
510
+ "args": {
511
+ "should_epoch_stop": false,
512
+ "should_evaluate": false,
513
+ "should_log": false,
514
+ "should_save": true,
515
+ "should_training_stop": false
516
+ },
517
+ "attributes": {}
518
+ }
519
+ },
520
+ "total_flos": 4516797323673600.0,
521
+ "train_batch_size": 64,
522
+ "trial_name": null,
523
+ "trial_params": null
524
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1400/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7268161d90e74f9dd729f7bcc2b8dedb5818335ef614e6ed65860bf9d94d5aa
3
+ size 4920
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation": "sigmoid",
3
+ "architectures": [
4
+ "TinyLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 2,
11
+ "head_dim": 32,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 128,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 256,
16
+ "max_position_embeddings": 512,
17
+ "mlp_bias": false,
18
+ "mlp_type": "glu",
19
+ "model_type": "tiny_llama",
20
+ "num_attention_heads": 4,
21
+ "num_hidden_layers": 100,
22
+ "num_key_value_heads": 4,
23
+ "pad_token_id": 0,
24
+ "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-06,
26
+ "rope_parameters": {
27
+ "rope_theta": 10000.0,
28
+ "rope_type": "default"
29
+ },
30
+ "tie_word_embeddings": true,
31
+ "tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
32
+ "transformers_version": "5.16.0.dev0",
33
+ "use_cache": false,
34
+ "vocab_size": 4096,
35
+ "waleed_beta": 4
36
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24bfdffc55a42ceb6aa40c78f28c084727cd6f7e2b851aacebf3b67e04288f23
3
+ size 33967272
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28776dce909e70d3dd408867bf19be24425aca1b48e9d3fee8f8b8fb5f3819e9
3
+ size 68504996
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a37538486c6fa3324bc6a6a3b2ad67db8646e004089a61bb15e3c6cba879c55
3
+ size 1064
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|endoftext|>",
5
+ "eos_token": "<|endoftext|>",
6
+ "errors": "replace",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1024,
10
+ "pad_token": "<|endoftext|>",
11
+ "tokenizer_class": "GPT2Tokenizer",
12
+ "unk_token": "<|endoftext|>"
13
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/trainer_state.json ADDED
@@ -0,0 +1,559 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.10111223458038422,
6
+ "eval_steps": 2498,
7
+ "global_step": 1500,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0013481631277384564,
14
+ "grad_norm": 1.828125,
15
+ "learning_rate": 2.66e-05,
16
+ "loss": 8.277120208740234,
17
+ "step": 20
18
+ },
19
+ {
20
+ "epoch": 0.002696326255476913,
21
+ "grad_norm": 1.296875,
22
+ "learning_rate": 5.46e-05,
23
+ "loss": 8.008469390869141,
24
+ "step": 40
25
+ },
26
+ {
27
+ "epoch": 0.004044489383215369,
28
+ "grad_norm": 1.2734375,
29
+ "learning_rate": 8.259999999999999e-05,
30
+ "loss": 7.827983093261719,
31
+ "step": 60
32
+ },
33
+ {
34
+ "epoch": 0.005392652510953826,
35
+ "grad_norm": 1.390625,
36
+ "learning_rate": 0.0001106,
37
+ "loss": 7.559034729003907,
38
+ "step": 80
39
+ },
40
+ {
41
+ "epoch": 0.006740815638692282,
42
+ "grad_norm": 1.2265625,
43
+ "learning_rate": 0.0001386,
44
+ "loss": 7.211753082275391,
45
+ "step": 100
46
+ },
47
+ {
48
+ "epoch": 0.008088978766430738,
49
+ "grad_norm": 1.2265625,
50
+ "learning_rate": 0.00016659999999999998,
51
+ "loss": 6.837139892578125,
52
+ "step": 120
53
+ },
54
+ {
55
+ "epoch": 0.009437141894169195,
56
+ "grad_norm": 1.0703125,
57
+ "learning_rate": 0.00019460000000000001,
58
+ "loss": 6.467125701904297,
59
+ "step": 140
60
+ },
61
+ {
62
+ "epoch": 0.010785305021907651,
63
+ "grad_norm": 1.296875,
64
+ "learning_rate": 0.0002226,
65
+ "loss": 6.130535507202149,
66
+ "step": 160
67
+ },
68
+ {
69
+ "epoch": 0.012133468149646108,
70
+ "grad_norm": 0.95703125,
71
+ "learning_rate": 0.00025059999999999997,
72
+ "loss": 5.828623199462891,
73
+ "step": 180
74
+ },
75
+ {
76
+ "epoch": 0.013481631277384564,
77
+ "grad_norm": 0.90625,
78
+ "learning_rate": 0.0002786,
79
+ "loss": 5.533278274536133,
80
+ "step": 200
81
+ },
82
+ {
83
+ "epoch": 0.01482979440512302,
84
+ "grad_norm": 0.90234375,
85
+ "learning_rate": 0.00030659999999999997,
86
+ "loss": 5.271765899658203,
87
+ "step": 220
88
+ },
89
+ {
90
+ "epoch": 0.016177957532861477,
91
+ "grad_norm": 0.84765625,
92
+ "learning_rate": 0.0003346,
93
+ "loss": 5.006443786621094,
94
+ "step": 240
95
+ },
96
+ {
97
+ "epoch": 0.01752612066059993,
98
+ "grad_norm": 3.125,
99
+ "learning_rate": 0.00036260000000000003,
100
+ "loss": 4.778961181640625,
101
+ "step": 260
102
+ },
103
+ {
104
+ "epoch": 0.01887428378833839,
105
+ "grad_norm": 0.76171875,
106
+ "learning_rate": 0.0003906,
107
+ "loss": 4.607266235351562,
108
+ "step": 280
109
+ },
110
+ {
111
+ "epoch": 0.020222446916076844,
112
+ "grad_norm": 1.015625,
113
+ "learning_rate": 0.0004186,
114
+ "loss": 4.434815979003906,
115
+ "step": 300
116
+ },
117
+ {
118
+ "epoch": 0.021570610043815303,
119
+ "grad_norm": 1.1328125,
120
+ "learning_rate": 0.0004466,
121
+ "loss": 4.295793533325195,
122
+ "step": 320
123
+ },
124
+ {
125
+ "epoch": 0.022918773171553757,
126
+ "grad_norm": 0.68359375,
127
+ "learning_rate": 0.00047460000000000004,
128
+ "loss": 4.169289398193359,
129
+ "step": 340
130
+ },
131
+ {
132
+ "epoch": 0.024266936299292215,
133
+ "grad_norm": 0.7734375,
134
+ "learning_rate": 0.0005026,
135
+ "loss": 4.064323425292969,
136
+ "step": 360
137
+ },
138
+ {
139
+ "epoch": 0.02561509942703067,
140
+ "grad_norm": 0.77734375,
141
+ "learning_rate": 0.0005306,
142
+ "loss": 3.9784740447998046,
143
+ "step": 380
144
+ },
145
+ {
146
+ "epoch": 0.026963262554769128,
147
+ "grad_norm": 0.7890625,
148
+ "learning_rate": 0.0005586,
149
+ "loss": 3.8648876190185546,
150
+ "step": 400
151
+ },
152
+ {
153
+ "epoch": 0.028311425682507583,
154
+ "grad_norm": 0.6484375,
155
+ "learning_rate": 0.0005866,
156
+ "loss": 3.794938659667969,
157
+ "step": 420
158
+ },
159
+ {
160
+ "epoch": 0.02965958881024604,
161
+ "grad_norm": 0.6484375,
162
+ "learning_rate": 0.0006146,
163
+ "loss": 3.7460472106933596,
164
+ "step": 440
165
+ },
166
+ {
167
+ "epoch": 0.031007751937984496,
168
+ "grad_norm": 0.91796875,
169
+ "learning_rate": 0.0006426,
170
+ "loss": 3.6633747100830076,
171
+ "step": 460
172
+ },
173
+ {
174
+ "epoch": 0.032355915065722954,
175
+ "grad_norm": 0.73828125,
176
+ "learning_rate": 0.0006705999999999999,
177
+ "loss": 3.6008380889892577,
178
+ "step": 480
179
+ },
180
+ {
181
+ "epoch": 0.03370407819346141,
182
+ "grad_norm": 0.61328125,
183
+ "learning_rate": 0.0006986,
184
+ "loss": 3.5390514373779296,
185
+ "step": 500
186
+ },
187
+ {
188
+ "epoch": 0.03505224132119986,
189
+ "grad_norm": 0.5703125,
190
+ "learning_rate": 0.0007,
191
+ "loss": 3.489065170288086,
192
+ "step": 520
193
+ },
194
+ {
195
+ "epoch": 0.03640040444893832,
196
+ "grad_norm": 0.76953125,
197
+ "learning_rate": 0.0007,
198
+ "loss": 3.4280685424804687,
199
+ "step": 540
200
+ },
201
+ {
202
+ "epoch": 0.03774856757667678,
203
+ "grad_norm": 0.58984375,
204
+ "learning_rate": 0.0007,
205
+ "loss": 3.3897541046142576,
206
+ "step": 560
207
+ },
208
+ {
209
+ "epoch": 0.03909673070441524,
210
+ "grad_norm": 0.6015625,
211
+ "learning_rate": 0.0007,
212
+ "loss": 3.3452869415283204,
213
+ "step": 580
214
+ },
215
+ {
216
+ "epoch": 0.04044489383215369,
217
+ "grad_norm": 0.6328125,
218
+ "learning_rate": 0.0007,
219
+ "loss": 3.3226955413818358,
220
+ "step": 600
221
+ },
222
+ {
223
+ "epoch": 0.04179305695989215,
224
+ "grad_norm": 0.625,
225
+ "learning_rate": 0.0007,
226
+ "loss": 3.2682525634765627,
227
+ "step": 620
228
+ },
229
+ {
230
+ "epoch": 0.043141220087630605,
231
+ "grad_norm": 0.5859375,
232
+ "learning_rate": 0.0007,
233
+ "loss": 3.2494388580322267,
234
+ "step": 640
235
+ },
236
+ {
237
+ "epoch": 0.044489383215369056,
238
+ "grad_norm": 0.52734375,
239
+ "learning_rate": 0.0007,
240
+ "loss": 3.2006797790527344,
241
+ "step": 660
242
+ },
243
+ {
244
+ "epoch": 0.045837546343107514,
245
+ "grad_norm": 0.61328125,
246
+ "learning_rate": 0.0007,
247
+ "loss": 3.180825424194336,
248
+ "step": 680
249
+ },
250
+ {
251
+ "epoch": 0.04718570947084597,
252
+ "grad_norm": 0.53125,
253
+ "learning_rate": 0.0007,
254
+ "loss": 3.142491912841797,
255
+ "step": 700
256
+ },
257
+ {
258
+ "epoch": 0.04853387259858443,
259
+ "grad_norm": 0.65234375,
260
+ "learning_rate": 0.0007,
261
+ "loss": 3.1118364334106445,
262
+ "step": 720
263
+ },
264
+ {
265
+ "epoch": 0.04988203572632288,
266
+ "grad_norm": 0.625,
267
+ "learning_rate": 0.0007,
268
+ "loss": 3.084051322937012,
269
+ "step": 740
270
+ },
271
+ {
272
+ "epoch": 0.05123019885406134,
273
+ "grad_norm": 0.55859375,
274
+ "learning_rate": 0.0007,
275
+ "loss": 3.0676673889160155,
276
+ "step": 760
277
+ },
278
+ {
279
+ "epoch": 0.0525783619817998,
280
+ "grad_norm": 0.60546875,
281
+ "learning_rate": 0.0007,
282
+ "loss": 3.0888349533081056,
283
+ "step": 780
284
+ },
285
+ {
286
+ "epoch": 0.053926525109538256,
287
+ "grad_norm": 0.5234375,
288
+ "learning_rate": 0.0007,
289
+ "loss": 3.035677909851074,
290
+ "step": 800
291
+ },
292
+ {
293
+ "epoch": 0.05527468823727671,
294
+ "grad_norm": 0.5546875,
295
+ "learning_rate": 0.0007,
296
+ "loss": 3.0188726425170898,
297
+ "step": 820
298
+ },
299
+ {
300
+ "epoch": 0.056622851365015166,
301
+ "grad_norm": 0.51171875,
302
+ "learning_rate": 0.0007,
303
+ "loss": 2.989584541320801,
304
+ "step": 840
305
+ },
306
+ {
307
+ "epoch": 0.057971014492753624,
308
+ "grad_norm": 0.53515625,
309
+ "learning_rate": 0.0007,
310
+ "loss": 2.9542055130004883,
311
+ "step": 860
312
+ },
313
+ {
314
+ "epoch": 0.05931917762049208,
315
+ "grad_norm": 0.5234375,
316
+ "learning_rate": 0.0007,
317
+ "loss": 2.93587589263916,
318
+ "step": 880
319
+ },
320
+ {
321
+ "epoch": 0.06066734074823053,
322
+ "grad_norm": 0.5390625,
323
+ "learning_rate": 0.0007,
324
+ "loss": 2.924374008178711,
325
+ "step": 900
326
+ },
327
+ {
328
+ "epoch": 0.06201550387596899,
329
+ "grad_norm": 0.58203125,
330
+ "learning_rate": 0.0007,
331
+ "loss": 2.899143028259277,
332
+ "step": 920
333
+ },
334
+ {
335
+ "epoch": 0.06336366700370745,
336
+ "grad_norm": 0.49609375,
337
+ "learning_rate": 0.0007,
338
+ "loss": 2.8880327224731444,
339
+ "step": 940
340
+ },
341
+ {
342
+ "epoch": 0.06471183013144591,
343
+ "grad_norm": 0.515625,
344
+ "learning_rate": 0.0007,
345
+ "loss": 2.8630607604980467,
346
+ "step": 960
347
+ },
348
+ {
349
+ "epoch": 0.06605999325918437,
350
+ "grad_norm": 0.50390625,
351
+ "learning_rate": 0.0007,
352
+ "loss": 2.848561477661133,
353
+ "step": 980
354
+ },
355
+ {
356
+ "epoch": 0.06740815638692282,
357
+ "grad_norm": 0.5390625,
358
+ "learning_rate": 0.0007,
359
+ "loss": 2.8479694366455077,
360
+ "step": 1000
361
+ },
362
+ {
363
+ "epoch": 0.06875631951466127,
364
+ "grad_norm": 0.486328125,
365
+ "learning_rate": 0.0007,
366
+ "loss": 2.820212173461914,
367
+ "step": 1020
368
+ },
369
+ {
370
+ "epoch": 0.07010448264239973,
371
+ "grad_norm": 0.546875,
372
+ "learning_rate": 0.0007,
373
+ "loss": 2.805441665649414,
374
+ "step": 1040
375
+ },
376
+ {
377
+ "epoch": 0.07145264577013818,
378
+ "grad_norm": 0.5,
379
+ "learning_rate": 0.0007,
380
+ "loss": 2.792853355407715,
381
+ "step": 1060
382
+ },
383
+ {
384
+ "epoch": 0.07280080889787664,
385
+ "grad_norm": 0.474609375,
386
+ "learning_rate": 0.0007,
387
+ "loss": 2.7767995834350585,
388
+ "step": 1080
389
+ },
390
+ {
391
+ "epoch": 0.0741489720256151,
392
+ "grad_norm": 0.51953125,
393
+ "learning_rate": 0.0007,
394
+ "loss": 2.775754928588867,
395
+ "step": 1100
396
+ },
397
+ {
398
+ "epoch": 0.07549713515335356,
399
+ "grad_norm": 0.52734375,
400
+ "learning_rate": 0.0007,
401
+ "loss": 2.7376676559448243,
402
+ "step": 1120
403
+ },
404
+ {
405
+ "epoch": 0.07684529828109202,
406
+ "grad_norm": 0.5078125,
407
+ "learning_rate": 0.0007,
408
+ "loss": 2.725414276123047,
409
+ "step": 1140
410
+ },
411
+ {
412
+ "epoch": 0.07819346140883048,
413
+ "grad_norm": 0.50390625,
414
+ "learning_rate": 0.0007,
415
+ "loss": 2.7253787994384764,
416
+ "step": 1160
417
+ },
418
+ {
419
+ "epoch": 0.07954162453656892,
420
+ "grad_norm": 0.462890625,
421
+ "learning_rate": 0.0007,
422
+ "loss": 2.7208480834960938,
423
+ "step": 1180
424
+ },
425
+ {
426
+ "epoch": 0.08088978766430738,
427
+ "grad_norm": 0.53125,
428
+ "learning_rate": 0.0007,
429
+ "loss": 2.7129190444946287,
430
+ "step": 1200
431
+ },
432
+ {
433
+ "epoch": 0.08223795079204584,
434
+ "grad_norm": 0.458984375,
435
+ "learning_rate": 0.0007,
436
+ "loss": 2.6948110580444338,
437
+ "step": 1220
438
+ },
439
+ {
440
+ "epoch": 0.0835861139197843,
441
+ "grad_norm": 0.5078125,
442
+ "learning_rate": 0.0007,
443
+ "loss": 2.680550193786621,
444
+ "step": 1240
445
+ },
446
+ {
447
+ "epoch": 0.08493427704752275,
448
+ "grad_norm": 0.48046875,
449
+ "learning_rate": 0.0007,
450
+ "loss": 2.6683969497680664,
451
+ "step": 1260
452
+ },
453
+ {
454
+ "epoch": 0.08628244017526121,
455
+ "grad_norm": 0.45703125,
456
+ "learning_rate": 0.0007,
457
+ "loss": 2.6498653411865236,
458
+ "step": 1280
459
+ },
460
+ {
461
+ "epoch": 0.08763060330299967,
462
+ "grad_norm": 0.515625,
463
+ "learning_rate": 0.0007,
464
+ "loss": 2.642520713806152,
465
+ "step": 1300
466
+ },
467
+ {
468
+ "epoch": 0.08897876643073811,
469
+ "grad_norm": 0.5078125,
470
+ "learning_rate": 0.0007,
471
+ "loss": 2.633445167541504,
472
+ "step": 1320
473
+ },
474
+ {
475
+ "epoch": 0.09032692955847657,
476
+ "grad_norm": 0.51953125,
477
+ "learning_rate": 0.0007,
478
+ "loss": 2.61986083984375,
479
+ "step": 1340
480
+ },
481
+ {
482
+ "epoch": 0.09167509268621503,
483
+ "grad_norm": 0.515625,
484
+ "learning_rate": 0.0007,
485
+ "loss": 2.6194650650024416,
486
+ "step": 1360
487
+ },
488
+ {
489
+ "epoch": 0.09302325581395349,
490
+ "grad_norm": 0.462890625,
491
+ "learning_rate": 0.0007,
492
+ "loss": 2.597072982788086,
493
+ "step": 1380
494
+ },
495
+ {
496
+ "epoch": 0.09437141894169195,
497
+ "grad_norm": 0.458984375,
498
+ "learning_rate": 0.0007,
499
+ "loss": 2.5900634765625,
500
+ "step": 1400
501
+ },
502
+ {
503
+ "epoch": 0.0957195820694304,
504
+ "grad_norm": 0.455078125,
505
+ "learning_rate": 0.0007,
506
+ "loss": 2.599566841125488,
507
+ "step": 1420
508
+ },
509
+ {
510
+ "epoch": 0.09706774519716886,
511
+ "grad_norm": 0.50390625,
512
+ "learning_rate": 0.0007,
513
+ "loss": 2.5753355026245117,
514
+ "step": 1440
515
+ },
516
+ {
517
+ "epoch": 0.09841590832490732,
518
+ "grad_norm": 0.478515625,
519
+ "learning_rate": 0.0007,
520
+ "loss": 2.563642883300781,
521
+ "step": 1460
522
+ },
523
+ {
524
+ "epoch": 0.09976407145264576,
525
+ "grad_norm": 0.44140625,
526
+ "learning_rate": 0.0007,
527
+ "loss": 2.5505178451538084,
528
+ "step": 1480
529
+ },
530
+ {
531
+ "epoch": 0.10111223458038422,
532
+ "grad_norm": 0.4609375,
533
+ "learning_rate": 0.0007,
534
+ "loss": 2.536166191101074,
535
+ "step": 1500
536
+ }
537
+ ],
538
+ "logging_steps": 20,
539
+ "max_steps": 2500,
540
+ "num_input_tokens_seen": 0,
541
+ "num_train_epochs": 1,
542
+ "save_steps": 100,
543
+ "stateful_callbacks": {
544
+ "TrainerControl": {
545
+ "args": {
546
+ "should_epoch_stop": false,
547
+ "should_evaluate": false,
548
+ "should_log": false,
549
+ "should_save": true,
550
+ "should_training_stop": false
551
+ },
552
+ "attributes": {}
553
+ }
554
+ },
555
+ "total_flos": 4839425703936000.0,
556
+ "train_batch_size": 64,
557
+ "trial_name": null,
558
+ "trial_params": null
559
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7268161d90e74f9dd729f7bcc2b8dedb5818335ef614e6ed65860bf9d94d5aa
3
+ size 4920
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation": "sigmoid",
3
+ "architectures": [
4
+ "TinyLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 2,
11
+ "head_dim": 32,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 128,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 256,
16
+ "max_position_embeddings": 512,
17
+ "mlp_bias": false,
18
+ "mlp_type": "glu",
19
+ "model_type": "tiny_llama",
20
+ "num_attention_heads": 4,
21
+ "num_hidden_layers": 100,
22
+ "num_key_value_heads": 4,
23
+ "pad_token_id": 0,
24
+ "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-06,
26
+ "rope_parameters": {
27
+ "rope_theta": 10000.0,
28
+ "rope_type": "default"
29
+ },
30
+ "tie_word_embeddings": true,
31
+ "tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
32
+ "transformers_version": "5.16.0.dev0",
33
+ "use_cache": false,
34
+ "vocab_size": 4096,
35
+ "waleed_beta": 4
36
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62a2f10909706eb1d13200e0fffb1d1480c6643e4855d5fb37155d5b741dc908
3
+ size 33967272
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5162e745257b57fb724a57c473a810f01bf4e94d6d629536c17265c48a8f508
3
+ size 68504996
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5207460c1b01907ad3d1f40124c12e85beccaebe656a961bcd0bff18a30ffd6
3
+ size 1064
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|endoftext|>",
5
+ "eos_token": "<|endoftext|>",
6
+ "errors": "replace",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1024,
10
+ "pad_token": "<|endoftext|>",
11
+ "tokenizer_class": "GPT2Tokenizer",
12
+ "unk_token": "<|endoftext|>"
13
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/trainer_state.json ADDED
@@ -0,0 +1,594 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.10785305021907651,
6
+ "eval_steps": 2498,
7
+ "global_step": 1600,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0013481631277384564,
14
+ "grad_norm": 1.828125,
15
+ "learning_rate": 2.66e-05,
16
+ "loss": 8.277120208740234,
17
+ "step": 20
18
+ },
19
+ {
20
+ "epoch": 0.002696326255476913,
21
+ "grad_norm": 1.296875,
22
+ "learning_rate": 5.46e-05,
23
+ "loss": 8.008469390869141,
24
+ "step": 40
25
+ },
26
+ {
27
+ "epoch": 0.004044489383215369,
28
+ "grad_norm": 1.2734375,
29
+ "learning_rate": 8.259999999999999e-05,
30
+ "loss": 7.827983093261719,
31
+ "step": 60
32
+ },
33
+ {
34
+ "epoch": 0.005392652510953826,
35
+ "grad_norm": 1.390625,
36
+ "learning_rate": 0.0001106,
37
+ "loss": 7.559034729003907,
38
+ "step": 80
39
+ },
40
+ {
41
+ "epoch": 0.006740815638692282,
42
+ "grad_norm": 1.2265625,
43
+ "learning_rate": 0.0001386,
44
+ "loss": 7.211753082275391,
45
+ "step": 100
46
+ },
47
+ {
48
+ "epoch": 0.008088978766430738,
49
+ "grad_norm": 1.2265625,
50
+ "learning_rate": 0.00016659999999999998,
51
+ "loss": 6.837139892578125,
52
+ "step": 120
53
+ },
54
+ {
55
+ "epoch": 0.009437141894169195,
56
+ "grad_norm": 1.0703125,
57
+ "learning_rate": 0.00019460000000000001,
58
+ "loss": 6.467125701904297,
59
+ "step": 140
60
+ },
61
+ {
62
+ "epoch": 0.010785305021907651,
63
+ "grad_norm": 1.296875,
64
+ "learning_rate": 0.0002226,
65
+ "loss": 6.130535507202149,
66
+ "step": 160
67
+ },
68
+ {
69
+ "epoch": 0.012133468149646108,
70
+ "grad_norm": 0.95703125,
71
+ "learning_rate": 0.00025059999999999997,
72
+ "loss": 5.828623199462891,
73
+ "step": 180
74
+ },
75
+ {
76
+ "epoch": 0.013481631277384564,
77
+ "grad_norm": 0.90625,
78
+ "learning_rate": 0.0002786,
79
+ "loss": 5.533278274536133,
80
+ "step": 200
81
+ },
82
+ {
83
+ "epoch": 0.01482979440512302,
84
+ "grad_norm": 0.90234375,
85
+ "learning_rate": 0.00030659999999999997,
86
+ "loss": 5.271765899658203,
87
+ "step": 220
88
+ },
89
+ {
90
+ "epoch": 0.016177957532861477,
91
+ "grad_norm": 0.84765625,
92
+ "learning_rate": 0.0003346,
93
+ "loss": 5.006443786621094,
94
+ "step": 240
95
+ },
96
+ {
97
+ "epoch": 0.01752612066059993,
98
+ "grad_norm": 3.125,
99
+ "learning_rate": 0.00036260000000000003,
100
+ "loss": 4.778961181640625,
101
+ "step": 260
102
+ },
103
+ {
104
+ "epoch": 0.01887428378833839,
105
+ "grad_norm": 0.76171875,
106
+ "learning_rate": 0.0003906,
107
+ "loss": 4.607266235351562,
108
+ "step": 280
109
+ },
110
+ {
111
+ "epoch": 0.020222446916076844,
112
+ "grad_norm": 1.015625,
113
+ "learning_rate": 0.0004186,
114
+ "loss": 4.434815979003906,
115
+ "step": 300
116
+ },
117
+ {
118
+ "epoch": 0.021570610043815303,
119
+ "grad_norm": 1.1328125,
120
+ "learning_rate": 0.0004466,
121
+ "loss": 4.295793533325195,
122
+ "step": 320
123
+ },
124
+ {
125
+ "epoch": 0.022918773171553757,
126
+ "grad_norm": 0.68359375,
127
+ "learning_rate": 0.00047460000000000004,
128
+ "loss": 4.169289398193359,
129
+ "step": 340
130
+ },
131
+ {
132
+ "epoch": 0.024266936299292215,
133
+ "grad_norm": 0.7734375,
134
+ "learning_rate": 0.0005026,
135
+ "loss": 4.064323425292969,
136
+ "step": 360
137
+ },
138
+ {
139
+ "epoch": 0.02561509942703067,
140
+ "grad_norm": 0.77734375,
141
+ "learning_rate": 0.0005306,
142
+ "loss": 3.9784740447998046,
143
+ "step": 380
144
+ },
145
+ {
146
+ "epoch": 0.026963262554769128,
147
+ "grad_norm": 0.7890625,
148
+ "learning_rate": 0.0005586,
149
+ "loss": 3.8648876190185546,
150
+ "step": 400
151
+ },
152
+ {
153
+ "epoch": 0.028311425682507583,
154
+ "grad_norm": 0.6484375,
155
+ "learning_rate": 0.0005866,
156
+ "loss": 3.794938659667969,
157
+ "step": 420
158
+ },
159
+ {
160
+ "epoch": 0.02965958881024604,
161
+ "grad_norm": 0.6484375,
162
+ "learning_rate": 0.0006146,
163
+ "loss": 3.7460472106933596,
164
+ "step": 440
165
+ },
166
+ {
167
+ "epoch": 0.031007751937984496,
168
+ "grad_norm": 0.91796875,
169
+ "learning_rate": 0.0006426,
170
+ "loss": 3.6633747100830076,
171
+ "step": 460
172
+ },
173
+ {
174
+ "epoch": 0.032355915065722954,
175
+ "grad_norm": 0.73828125,
176
+ "learning_rate": 0.0006705999999999999,
177
+ "loss": 3.6008380889892577,
178
+ "step": 480
179
+ },
180
+ {
181
+ "epoch": 0.03370407819346141,
182
+ "grad_norm": 0.61328125,
183
+ "learning_rate": 0.0006986,
184
+ "loss": 3.5390514373779296,
185
+ "step": 500
186
+ },
187
+ {
188
+ "epoch": 0.03505224132119986,
189
+ "grad_norm": 0.5703125,
190
+ "learning_rate": 0.0007,
191
+ "loss": 3.489065170288086,
192
+ "step": 520
193
+ },
194
+ {
195
+ "epoch": 0.03640040444893832,
196
+ "grad_norm": 0.76953125,
197
+ "learning_rate": 0.0007,
198
+ "loss": 3.4280685424804687,
199
+ "step": 540
200
+ },
201
+ {
202
+ "epoch": 0.03774856757667678,
203
+ "grad_norm": 0.58984375,
204
+ "learning_rate": 0.0007,
205
+ "loss": 3.3897541046142576,
206
+ "step": 560
207
+ },
208
+ {
209
+ "epoch": 0.03909673070441524,
210
+ "grad_norm": 0.6015625,
211
+ "learning_rate": 0.0007,
212
+ "loss": 3.3452869415283204,
213
+ "step": 580
214
+ },
215
+ {
216
+ "epoch": 0.04044489383215369,
217
+ "grad_norm": 0.6328125,
218
+ "learning_rate": 0.0007,
219
+ "loss": 3.3226955413818358,
220
+ "step": 600
221
+ },
222
+ {
223
+ "epoch": 0.04179305695989215,
224
+ "grad_norm": 0.625,
225
+ "learning_rate": 0.0007,
226
+ "loss": 3.2682525634765627,
227
+ "step": 620
228
+ },
229
+ {
230
+ "epoch": 0.043141220087630605,
231
+ "grad_norm": 0.5859375,
232
+ "learning_rate": 0.0007,
233
+ "loss": 3.2494388580322267,
234
+ "step": 640
235
+ },
236
+ {
237
+ "epoch": 0.044489383215369056,
238
+ "grad_norm": 0.52734375,
239
+ "learning_rate": 0.0007,
240
+ "loss": 3.2006797790527344,
241
+ "step": 660
242
+ },
243
+ {
244
+ "epoch": 0.045837546343107514,
245
+ "grad_norm": 0.61328125,
246
+ "learning_rate": 0.0007,
247
+ "loss": 3.180825424194336,
248
+ "step": 680
249
+ },
250
+ {
251
+ "epoch": 0.04718570947084597,
252
+ "grad_norm": 0.53125,
253
+ "learning_rate": 0.0007,
254
+ "loss": 3.142491912841797,
255
+ "step": 700
256
+ },
257
+ {
258
+ "epoch": 0.04853387259858443,
259
+ "grad_norm": 0.65234375,
260
+ "learning_rate": 0.0007,
261
+ "loss": 3.1118364334106445,
262
+ "step": 720
263
+ },
264
+ {
265
+ "epoch": 0.04988203572632288,
266
+ "grad_norm": 0.625,
267
+ "learning_rate": 0.0007,
268
+ "loss": 3.084051322937012,
269
+ "step": 740
270
+ },
271
+ {
272
+ "epoch": 0.05123019885406134,
273
+ "grad_norm": 0.55859375,
274
+ "learning_rate": 0.0007,
275
+ "loss": 3.0676673889160155,
276
+ "step": 760
277
+ },
278
+ {
279
+ "epoch": 0.0525783619817998,
280
+ "grad_norm": 0.60546875,
281
+ "learning_rate": 0.0007,
282
+ "loss": 3.0888349533081056,
283
+ "step": 780
284
+ },
285
+ {
286
+ "epoch": 0.053926525109538256,
287
+ "grad_norm": 0.5234375,
288
+ "learning_rate": 0.0007,
289
+ "loss": 3.035677909851074,
290
+ "step": 800
291
+ },
292
+ {
293
+ "epoch": 0.05527468823727671,
294
+ "grad_norm": 0.5546875,
295
+ "learning_rate": 0.0007,
296
+ "loss": 3.0188726425170898,
297
+ "step": 820
298
+ },
299
+ {
300
+ "epoch": 0.056622851365015166,
301
+ "grad_norm": 0.51171875,
302
+ "learning_rate": 0.0007,
303
+ "loss": 2.989584541320801,
304
+ "step": 840
305
+ },
306
+ {
307
+ "epoch": 0.057971014492753624,
308
+ "grad_norm": 0.53515625,
309
+ "learning_rate": 0.0007,
310
+ "loss": 2.9542055130004883,
311
+ "step": 860
312
+ },
313
+ {
314
+ "epoch": 0.05931917762049208,
315
+ "grad_norm": 0.5234375,
316
+ "learning_rate": 0.0007,
317
+ "loss": 2.93587589263916,
318
+ "step": 880
319
+ },
320
+ {
321
+ "epoch": 0.06066734074823053,
322
+ "grad_norm": 0.5390625,
323
+ "learning_rate": 0.0007,
324
+ "loss": 2.924374008178711,
325
+ "step": 900
326
+ },
327
+ {
328
+ "epoch": 0.06201550387596899,
329
+ "grad_norm": 0.58203125,
330
+ "learning_rate": 0.0007,
331
+ "loss": 2.899143028259277,
332
+ "step": 920
333
+ },
334
+ {
335
+ "epoch": 0.06336366700370745,
336
+ "grad_norm": 0.49609375,
337
+ "learning_rate": 0.0007,
338
+ "loss": 2.8880327224731444,
339
+ "step": 940
340
+ },
341
+ {
342
+ "epoch": 0.06471183013144591,
343
+ "grad_norm": 0.515625,
344
+ "learning_rate": 0.0007,
345
+ "loss": 2.8630607604980467,
346
+ "step": 960
347
+ },
348
+ {
349
+ "epoch": 0.06605999325918437,
350
+ "grad_norm": 0.50390625,
351
+ "learning_rate": 0.0007,
352
+ "loss": 2.848561477661133,
353
+ "step": 980
354
+ },
355
+ {
356
+ "epoch": 0.06740815638692282,
357
+ "grad_norm": 0.5390625,
358
+ "learning_rate": 0.0007,
359
+ "loss": 2.8479694366455077,
360
+ "step": 1000
361
+ },
362
+ {
363
+ "epoch": 0.06875631951466127,
364
+ "grad_norm": 0.486328125,
365
+ "learning_rate": 0.0007,
366
+ "loss": 2.820212173461914,
367
+ "step": 1020
368
+ },
369
+ {
370
+ "epoch": 0.07010448264239973,
371
+ "grad_norm": 0.546875,
372
+ "learning_rate": 0.0007,
373
+ "loss": 2.805441665649414,
374
+ "step": 1040
375
+ },
376
+ {
377
+ "epoch": 0.07145264577013818,
378
+ "grad_norm": 0.5,
379
+ "learning_rate": 0.0007,
380
+ "loss": 2.792853355407715,
381
+ "step": 1060
382
+ },
383
+ {
384
+ "epoch": 0.07280080889787664,
385
+ "grad_norm": 0.474609375,
386
+ "learning_rate": 0.0007,
387
+ "loss": 2.7767995834350585,
388
+ "step": 1080
389
+ },
390
+ {
391
+ "epoch": 0.0741489720256151,
392
+ "grad_norm": 0.51953125,
393
+ "learning_rate": 0.0007,
394
+ "loss": 2.775754928588867,
395
+ "step": 1100
396
+ },
397
+ {
398
+ "epoch": 0.07549713515335356,
399
+ "grad_norm": 0.52734375,
400
+ "learning_rate": 0.0007,
401
+ "loss": 2.7376676559448243,
402
+ "step": 1120
403
+ },
404
+ {
405
+ "epoch": 0.07684529828109202,
406
+ "grad_norm": 0.5078125,
407
+ "learning_rate": 0.0007,
408
+ "loss": 2.725414276123047,
409
+ "step": 1140
410
+ },
411
+ {
412
+ "epoch": 0.07819346140883048,
413
+ "grad_norm": 0.50390625,
414
+ "learning_rate": 0.0007,
415
+ "loss": 2.7253787994384764,
416
+ "step": 1160
417
+ },
418
+ {
419
+ "epoch": 0.07954162453656892,
420
+ "grad_norm": 0.462890625,
421
+ "learning_rate": 0.0007,
422
+ "loss": 2.7208480834960938,
423
+ "step": 1180
424
+ },
425
+ {
426
+ "epoch": 0.08088978766430738,
427
+ "grad_norm": 0.53125,
428
+ "learning_rate": 0.0007,
429
+ "loss": 2.7129190444946287,
430
+ "step": 1200
431
+ },
432
+ {
433
+ "epoch": 0.08223795079204584,
434
+ "grad_norm": 0.458984375,
435
+ "learning_rate": 0.0007,
436
+ "loss": 2.6948110580444338,
437
+ "step": 1220
438
+ },
439
+ {
440
+ "epoch": 0.0835861139197843,
441
+ "grad_norm": 0.5078125,
442
+ "learning_rate": 0.0007,
443
+ "loss": 2.680550193786621,
444
+ "step": 1240
445
+ },
446
+ {
447
+ "epoch": 0.08493427704752275,
448
+ "grad_norm": 0.48046875,
449
+ "learning_rate": 0.0007,
450
+ "loss": 2.6683969497680664,
451
+ "step": 1260
452
+ },
453
+ {
454
+ "epoch": 0.08628244017526121,
455
+ "grad_norm": 0.45703125,
456
+ "learning_rate": 0.0007,
457
+ "loss": 2.6498653411865236,
458
+ "step": 1280
459
+ },
460
+ {
461
+ "epoch": 0.08763060330299967,
462
+ "grad_norm": 0.515625,
463
+ "learning_rate": 0.0007,
464
+ "loss": 2.642520713806152,
465
+ "step": 1300
466
+ },
467
+ {
468
+ "epoch": 0.08897876643073811,
469
+ "grad_norm": 0.5078125,
470
+ "learning_rate": 0.0007,
471
+ "loss": 2.633445167541504,
472
+ "step": 1320
473
+ },
474
+ {
475
+ "epoch": 0.09032692955847657,
476
+ "grad_norm": 0.51953125,
477
+ "learning_rate": 0.0007,
478
+ "loss": 2.61986083984375,
479
+ "step": 1340
480
+ },
481
+ {
482
+ "epoch": 0.09167509268621503,
483
+ "grad_norm": 0.515625,
484
+ "learning_rate": 0.0007,
485
+ "loss": 2.6194650650024416,
486
+ "step": 1360
487
+ },
488
+ {
489
+ "epoch": 0.09302325581395349,
490
+ "grad_norm": 0.462890625,
491
+ "learning_rate": 0.0007,
492
+ "loss": 2.597072982788086,
493
+ "step": 1380
494
+ },
495
+ {
496
+ "epoch": 0.09437141894169195,
497
+ "grad_norm": 0.458984375,
498
+ "learning_rate": 0.0007,
499
+ "loss": 2.5900634765625,
500
+ "step": 1400
501
+ },
502
+ {
503
+ "epoch": 0.0957195820694304,
504
+ "grad_norm": 0.455078125,
505
+ "learning_rate": 0.0007,
506
+ "loss": 2.599566841125488,
507
+ "step": 1420
508
+ },
509
+ {
510
+ "epoch": 0.09706774519716886,
511
+ "grad_norm": 0.50390625,
512
+ "learning_rate": 0.0007,
513
+ "loss": 2.5753355026245117,
514
+ "step": 1440
515
+ },
516
+ {
517
+ "epoch": 0.09841590832490732,
518
+ "grad_norm": 0.478515625,
519
+ "learning_rate": 0.0007,
520
+ "loss": 2.563642883300781,
521
+ "step": 1460
522
+ },
523
+ {
524
+ "epoch": 0.09976407145264576,
525
+ "grad_norm": 0.44140625,
526
+ "learning_rate": 0.0007,
527
+ "loss": 2.5505178451538084,
528
+ "step": 1480
529
+ },
530
+ {
531
+ "epoch": 0.10111223458038422,
532
+ "grad_norm": 0.4609375,
533
+ "learning_rate": 0.0007,
534
+ "loss": 2.536166191101074,
535
+ "step": 1500
536
+ },
537
+ {
538
+ "epoch": 0.10246039770812268,
539
+ "grad_norm": 0.49609375,
540
+ "learning_rate": 0.0007,
541
+ "loss": 2.53826847076416,
542
+ "step": 1520
543
+ },
544
+ {
545
+ "epoch": 0.10380856083586114,
546
+ "grad_norm": 0.458984375,
547
+ "learning_rate": 0.0007,
548
+ "loss": 2.5285362243652343,
549
+ "step": 1540
550
+ },
551
+ {
552
+ "epoch": 0.1051567239635996,
553
+ "grad_norm": 0.486328125,
554
+ "learning_rate": 0.0007,
555
+ "loss": 2.5131874084472656,
556
+ "step": 1560
557
+ },
558
+ {
559
+ "epoch": 0.10650488709133805,
560
+ "grad_norm": 0.46484375,
561
+ "learning_rate": 0.0007,
562
+ "loss": 2.5331687927246094,
563
+ "step": 1580
564
+ },
565
+ {
566
+ "epoch": 0.10785305021907651,
567
+ "grad_norm": 0.439453125,
568
+ "learning_rate": 0.0007,
569
+ "loss": 2.500454139709473,
570
+ "step": 1600
571
+ }
572
+ ],
573
+ "logging_steps": 20,
574
+ "max_steps": 2500,
575
+ "num_input_tokens_seen": 0,
576
+ "num_train_epochs": 1,
577
+ "save_steps": 100,
578
+ "stateful_callbacks": {
579
+ "TrainerControl": {
580
+ "args": {
581
+ "should_epoch_stop": false,
582
+ "should_evaluate": false,
583
+ "should_log": false,
584
+ "should_save": true,
585
+ "should_training_stop": false
586
+ },
587
+ "attributes": {}
588
+ }
589
+ },
590
+ "total_flos": 5162054084198400.0,
591
+ "train_batch_size": 64,
592
+ "trial_name": null,
593
+ "trial_params": null
594
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1600/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7268161d90e74f9dd729f7bcc2b8dedb5818335ef614e6ed65860bf9d94d5aa
3
+ size 4920
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation": "sigmoid",
3
+ "architectures": [
4
+ "TinyLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 2,
11
+ "head_dim": 32,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 128,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 256,
16
+ "max_position_embeddings": 512,
17
+ "mlp_bias": false,
18
+ "mlp_type": "glu",
19
+ "model_type": "tiny_llama",
20
+ "num_attention_heads": 4,
21
+ "num_hidden_layers": 100,
22
+ "num_key_value_heads": 4,
23
+ "pad_token_id": 0,
24
+ "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-06,
26
+ "rope_parameters": {
27
+ "rope_theta": 10000.0,
28
+ "rope_type": "default"
29
+ },
30
+ "tie_word_embeddings": true,
31
+ "tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
32
+ "transformers_version": "5.16.0.dev0",
33
+ "use_cache": false,
34
+ "vocab_size": 4096,
35
+ "waleed_beta": 4
36
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cef200a76c8fcaecd637e1f6ec2674dffdcbcbb4f58182293816402fabe6c750
3
+ size 33967272
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d7f9896a542c6e69e0e158d42e06e5ad93be3c6e2418bea3eb310ca5c61fa679
3
+ size 68504996
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da239a902a87900de7861a410c642a5ae70d481b0020a3040c363cc3ced9b7a4
3
+ size 1064
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|endoftext|>",
5
+ "eos_token": "<|endoftext|>",
6
+ "errors": "replace",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1024,
10
+ "pad_token": "<|endoftext|>",
11
+ "tokenizer_class": "GPT2Tokenizer",
12
+ "unk_token": "<|endoftext|>"
13
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/trainer_state.json ADDED
@@ -0,0 +1,629 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.11459386585776879,
6
+ "eval_steps": 2498,
7
+ "global_step": 1700,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0013481631277384564,
14
+ "grad_norm": 1.828125,
15
+ "learning_rate": 2.66e-05,
16
+ "loss": 8.277120208740234,
17
+ "step": 20
18
+ },
19
+ {
20
+ "epoch": 0.002696326255476913,
21
+ "grad_norm": 1.296875,
22
+ "learning_rate": 5.46e-05,
23
+ "loss": 8.008469390869141,
24
+ "step": 40
25
+ },
26
+ {
27
+ "epoch": 0.004044489383215369,
28
+ "grad_norm": 1.2734375,
29
+ "learning_rate": 8.259999999999999e-05,
30
+ "loss": 7.827983093261719,
31
+ "step": 60
32
+ },
33
+ {
34
+ "epoch": 0.005392652510953826,
35
+ "grad_norm": 1.390625,
36
+ "learning_rate": 0.0001106,
37
+ "loss": 7.559034729003907,
38
+ "step": 80
39
+ },
40
+ {
41
+ "epoch": 0.006740815638692282,
42
+ "grad_norm": 1.2265625,
43
+ "learning_rate": 0.0001386,
44
+ "loss": 7.211753082275391,
45
+ "step": 100
46
+ },
47
+ {
48
+ "epoch": 0.008088978766430738,
49
+ "grad_norm": 1.2265625,
50
+ "learning_rate": 0.00016659999999999998,
51
+ "loss": 6.837139892578125,
52
+ "step": 120
53
+ },
54
+ {
55
+ "epoch": 0.009437141894169195,
56
+ "grad_norm": 1.0703125,
57
+ "learning_rate": 0.00019460000000000001,
58
+ "loss": 6.467125701904297,
59
+ "step": 140
60
+ },
61
+ {
62
+ "epoch": 0.010785305021907651,
63
+ "grad_norm": 1.296875,
64
+ "learning_rate": 0.0002226,
65
+ "loss": 6.130535507202149,
66
+ "step": 160
67
+ },
68
+ {
69
+ "epoch": 0.012133468149646108,
70
+ "grad_norm": 0.95703125,
71
+ "learning_rate": 0.00025059999999999997,
72
+ "loss": 5.828623199462891,
73
+ "step": 180
74
+ },
75
+ {
76
+ "epoch": 0.013481631277384564,
77
+ "grad_norm": 0.90625,
78
+ "learning_rate": 0.0002786,
79
+ "loss": 5.533278274536133,
80
+ "step": 200
81
+ },
82
+ {
83
+ "epoch": 0.01482979440512302,
84
+ "grad_norm": 0.90234375,
85
+ "learning_rate": 0.00030659999999999997,
86
+ "loss": 5.271765899658203,
87
+ "step": 220
88
+ },
89
+ {
90
+ "epoch": 0.016177957532861477,
91
+ "grad_norm": 0.84765625,
92
+ "learning_rate": 0.0003346,
93
+ "loss": 5.006443786621094,
94
+ "step": 240
95
+ },
96
+ {
97
+ "epoch": 0.01752612066059993,
98
+ "grad_norm": 3.125,
99
+ "learning_rate": 0.00036260000000000003,
100
+ "loss": 4.778961181640625,
101
+ "step": 260
102
+ },
103
+ {
104
+ "epoch": 0.01887428378833839,
105
+ "grad_norm": 0.76171875,
106
+ "learning_rate": 0.0003906,
107
+ "loss": 4.607266235351562,
108
+ "step": 280
109
+ },
110
+ {
111
+ "epoch": 0.020222446916076844,
112
+ "grad_norm": 1.015625,
113
+ "learning_rate": 0.0004186,
114
+ "loss": 4.434815979003906,
115
+ "step": 300
116
+ },
117
+ {
118
+ "epoch": 0.021570610043815303,
119
+ "grad_norm": 1.1328125,
120
+ "learning_rate": 0.0004466,
121
+ "loss": 4.295793533325195,
122
+ "step": 320
123
+ },
124
+ {
125
+ "epoch": 0.022918773171553757,
126
+ "grad_norm": 0.68359375,
127
+ "learning_rate": 0.00047460000000000004,
128
+ "loss": 4.169289398193359,
129
+ "step": 340
130
+ },
131
+ {
132
+ "epoch": 0.024266936299292215,
133
+ "grad_norm": 0.7734375,
134
+ "learning_rate": 0.0005026,
135
+ "loss": 4.064323425292969,
136
+ "step": 360
137
+ },
138
+ {
139
+ "epoch": 0.02561509942703067,
140
+ "grad_norm": 0.77734375,
141
+ "learning_rate": 0.0005306,
142
+ "loss": 3.9784740447998046,
143
+ "step": 380
144
+ },
145
+ {
146
+ "epoch": 0.026963262554769128,
147
+ "grad_norm": 0.7890625,
148
+ "learning_rate": 0.0005586,
149
+ "loss": 3.8648876190185546,
150
+ "step": 400
151
+ },
152
+ {
153
+ "epoch": 0.028311425682507583,
154
+ "grad_norm": 0.6484375,
155
+ "learning_rate": 0.0005866,
156
+ "loss": 3.794938659667969,
157
+ "step": 420
158
+ },
159
+ {
160
+ "epoch": 0.02965958881024604,
161
+ "grad_norm": 0.6484375,
162
+ "learning_rate": 0.0006146,
163
+ "loss": 3.7460472106933596,
164
+ "step": 440
165
+ },
166
+ {
167
+ "epoch": 0.031007751937984496,
168
+ "grad_norm": 0.91796875,
169
+ "learning_rate": 0.0006426,
170
+ "loss": 3.6633747100830076,
171
+ "step": 460
172
+ },
173
+ {
174
+ "epoch": 0.032355915065722954,
175
+ "grad_norm": 0.73828125,
176
+ "learning_rate": 0.0006705999999999999,
177
+ "loss": 3.6008380889892577,
178
+ "step": 480
179
+ },
180
+ {
181
+ "epoch": 0.03370407819346141,
182
+ "grad_norm": 0.61328125,
183
+ "learning_rate": 0.0006986,
184
+ "loss": 3.5390514373779296,
185
+ "step": 500
186
+ },
187
+ {
188
+ "epoch": 0.03505224132119986,
189
+ "grad_norm": 0.5703125,
190
+ "learning_rate": 0.0007,
191
+ "loss": 3.489065170288086,
192
+ "step": 520
193
+ },
194
+ {
195
+ "epoch": 0.03640040444893832,
196
+ "grad_norm": 0.76953125,
197
+ "learning_rate": 0.0007,
198
+ "loss": 3.4280685424804687,
199
+ "step": 540
200
+ },
201
+ {
202
+ "epoch": 0.03774856757667678,
203
+ "grad_norm": 0.58984375,
204
+ "learning_rate": 0.0007,
205
+ "loss": 3.3897541046142576,
206
+ "step": 560
207
+ },
208
+ {
209
+ "epoch": 0.03909673070441524,
210
+ "grad_norm": 0.6015625,
211
+ "learning_rate": 0.0007,
212
+ "loss": 3.3452869415283204,
213
+ "step": 580
214
+ },
215
+ {
216
+ "epoch": 0.04044489383215369,
217
+ "grad_norm": 0.6328125,
218
+ "learning_rate": 0.0007,
219
+ "loss": 3.3226955413818358,
220
+ "step": 600
221
+ },
222
+ {
223
+ "epoch": 0.04179305695989215,
224
+ "grad_norm": 0.625,
225
+ "learning_rate": 0.0007,
226
+ "loss": 3.2682525634765627,
227
+ "step": 620
228
+ },
229
+ {
230
+ "epoch": 0.043141220087630605,
231
+ "grad_norm": 0.5859375,
232
+ "learning_rate": 0.0007,
233
+ "loss": 3.2494388580322267,
234
+ "step": 640
235
+ },
236
+ {
237
+ "epoch": 0.044489383215369056,
238
+ "grad_norm": 0.52734375,
239
+ "learning_rate": 0.0007,
240
+ "loss": 3.2006797790527344,
241
+ "step": 660
242
+ },
243
+ {
244
+ "epoch": 0.045837546343107514,
245
+ "grad_norm": 0.61328125,
246
+ "learning_rate": 0.0007,
247
+ "loss": 3.180825424194336,
248
+ "step": 680
249
+ },
250
+ {
251
+ "epoch": 0.04718570947084597,
252
+ "grad_norm": 0.53125,
253
+ "learning_rate": 0.0007,
254
+ "loss": 3.142491912841797,
255
+ "step": 700
256
+ },
257
+ {
258
+ "epoch": 0.04853387259858443,
259
+ "grad_norm": 0.65234375,
260
+ "learning_rate": 0.0007,
261
+ "loss": 3.1118364334106445,
262
+ "step": 720
263
+ },
264
+ {
265
+ "epoch": 0.04988203572632288,
266
+ "grad_norm": 0.625,
267
+ "learning_rate": 0.0007,
268
+ "loss": 3.084051322937012,
269
+ "step": 740
270
+ },
271
+ {
272
+ "epoch": 0.05123019885406134,
273
+ "grad_norm": 0.55859375,
274
+ "learning_rate": 0.0007,
275
+ "loss": 3.0676673889160155,
276
+ "step": 760
277
+ },
278
+ {
279
+ "epoch": 0.0525783619817998,
280
+ "grad_norm": 0.60546875,
281
+ "learning_rate": 0.0007,
282
+ "loss": 3.0888349533081056,
283
+ "step": 780
284
+ },
285
+ {
286
+ "epoch": 0.053926525109538256,
287
+ "grad_norm": 0.5234375,
288
+ "learning_rate": 0.0007,
289
+ "loss": 3.035677909851074,
290
+ "step": 800
291
+ },
292
+ {
293
+ "epoch": 0.05527468823727671,
294
+ "grad_norm": 0.5546875,
295
+ "learning_rate": 0.0007,
296
+ "loss": 3.0188726425170898,
297
+ "step": 820
298
+ },
299
+ {
300
+ "epoch": 0.056622851365015166,
301
+ "grad_norm": 0.51171875,
302
+ "learning_rate": 0.0007,
303
+ "loss": 2.989584541320801,
304
+ "step": 840
305
+ },
306
+ {
307
+ "epoch": 0.057971014492753624,
308
+ "grad_norm": 0.53515625,
309
+ "learning_rate": 0.0007,
310
+ "loss": 2.9542055130004883,
311
+ "step": 860
312
+ },
313
+ {
314
+ "epoch": 0.05931917762049208,
315
+ "grad_norm": 0.5234375,
316
+ "learning_rate": 0.0007,
317
+ "loss": 2.93587589263916,
318
+ "step": 880
319
+ },
320
+ {
321
+ "epoch": 0.06066734074823053,
322
+ "grad_norm": 0.5390625,
323
+ "learning_rate": 0.0007,
324
+ "loss": 2.924374008178711,
325
+ "step": 900
326
+ },
327
+ {
328
+ "epoch": 0.06201550387596899,
329
+ "grad_norm": 0.58203125,
330
+ "learning_rate": 0.0007,
331
+ "loss": 2.899143028259277,
332
+ "step": 920
333
+ },
334
+ {
335
+ "epoch": 0.06336366700370745,
336
+ "grad_norm": 0.49609375,
337
+ "learning_rate": 0.0007,
338
+ "loss": 2.8880327224731444,
339
+ "step": 940
340
+ },
341
+ {
342
+ "epoch": 0.06471183013144591,
343
+ "grad_norm": 0.515625,
344
+ "learning_rate": 0.0007,
345
+ "loss": 2.8630607604980467,
346
+ "step": 960
347
+ },
348
+ {
349
+ "epoch": 0.06605999325918437,
350
+ "grad_norm": 0.50390625,
351
+ "learning_rate": 0.0007,
352
+ "loss": 2.848561477661133,
353
+ "step": 980
354
+ },
355
+ {
356
+ "epoch": 0.06740815638692282,
357
+ "grad_norm": 0.5390625,
358
+ "learning_rate": 0.0007,
359
+ "loss": 2.8479694366455077,
360
+ "step": 1000
361
+ },
362
+ {
363
+ "epoch": 0.06875631951466127,
364
+ "grad_norm": 0.486328125,
365
+ "learning_rate": 0.0007,
366
+ "loss": 2.820212173461914,
367
+ "step": 1020
368
+ },
369
+ {
370
+ "epoch": 0.07010448264239973,
371
+ "grad_norm": 0.546875,
372
+ "learning_rate": 0.0007,
373
+ "loss": 2.805441665649414,
374
+ "step": 1040
375
+ },
376
+ {
377
+ "epoch": 0.07145264577013818,
378
+ "grad_norm": 0.5,
379
+ "learning_rate": 0.0007,
380
+ "loss": 2.792853355407715,
381
+ "step": 1060
382
+ },
383
+ {
384
+ "epoch": 0.07280080889787664,
385
+ "grad_norm": 0.474609375,
386
+ "learning_rate": 0.0007,
387
+ "loss": 2.7767995834350585,
388
+ "step": 1080
389
+ },
390
+ {
391
+ "epoch": 0.0741489720256151,
392
+ "grad_norm": 0.51953125,
393
+ "learning_rate": 0.0007,
394
+ "loss": 2.775754928588867,
395
+ "step": 1100
396
+ },
397
+ {
398
+ "epoch": 0.07549713515335356,
399
+ "grad_norm": 0.52734375,
400
+ "learning_rate": 0.0007,
401
+ "loss": 2.7376676559448243,
402
+ "step": 1120
403
+ },
404
+ {
405
+ "epoch": 0.07684529828109202,
406
+ "grad_norm": 0.5078125,
407
+ "learning_rate": 0.0007,
408
+ "loss": 2.725414276123047,
409
+ "step": 1140
410
+ },
411
+ {
412
+ "epoch": 0.07819346140883048,
413
+ "grad_norm": 0.50390625,
414
+ "learning_rate": 0.0007,
415
+ "loss": 2.7253787994384764,
416
+ "step": 1160
417
+ },
418
+ {
419
+ "epoch": 0.07954162453656892,
420
+ "grad_norm": 0.462890625,
421
+ "learning_rate": 0.0007,
422
+ "loss": 2.7208480834960938,
423
+ "step": 1180
424
+ },
425
+ {
426
+ "epoch": 0.08088978766430738,
427
+ "grad_norm": 0.53125,
428
+ "learning_rate": 0.0007,
429
+ "loss": 2.7129190444946287,
430
+ "step": 1200
431
+ },
432
+ {
433
+ "epoch": 0.08223795079204584,
434
+ "grad_norm": 0.458984375,
435
+ "learning_rate": 0.0007,
436
+ "loss": 2.6948110580444338,
437
+ "step": 1220
438
+ },
439
+ {
440
+ "epoch": 0.0835861139197843,
441
+ "grad_norm": 0.5078125,
442
+ "learning_rate": 0.0007,
443
+ "loss": 2.680550193786621,
444
+ "step": 1240
445
+ },
446
+ {
447
+ "epoch": 0.08493427704752275,
448
+ "grad_norm": 0.48046875,
449
+ "learning_rate": 0.0007,
450
+ "loss": 2.6683969497680664,
451
+ "step": 1260
452
+ },
453
+ {
454
+ "epoch": 0.08628244017526121,
455
+ "grad_norm": 0.45703125,
456
+ "learning_rate": 0.0007,
457
+ "loss": 2.6498653411865236,
458
+ "step": 1280
459
+ },
460
+ {
461
+ "epoch": 0.08763060330299967,
462
+ "grad_norm": 0.515625,
463
+ "learning_rate": 0.0007,
464
+ "loss": 2.642520713806152,
465
+ "step": 1300
466
+ },
467
+ {
468
+ "epoch": 0.08897876643073811,
469
+ "grad_norm": 0.5078125,
470
+ "learning_rate": 0.0007,
471
+ "loss": 2.633445167541504,
472
+ "step": 1320
473
+ },
474
+ {
475
+ "epoch": 0.09032692955847657,
476
+ "grad_norm": 0.51953125,
477
+ "learning_rate": 0.0007,
478
+ "loss": 2.61986083984375,
479
+ "step": 1340
480
+ },
481
+ {
482
+ "epoch": 0.09167509268621503,
483
+ "grad_norm": 0.515625,
484
+ "learning_rate": 0.0007,
485
+ "loss": 2.6194650650024416,
486
+ "step": 1360
487
+ },
488
+ {
489
+ "epoch": 0.09302325581395349,
490
+ "grad_norm": 0.462890625,
491
+ "learning_rate": 0.0007,
492
+ "loss": 2.597072982788086,
493
+ "step": 1380
494
+ },
495
+ {
496
+ "epoch": 0.09437141894169195,
497
+ "grad_norm": 0.458984375,
498
+ "learning_rate": 0.0007,
499
+ "loss": 2.5900634765625,
500
+ "step": 1400
501
+ },
502
+ {
503
+ "epoch": 0.0957195820694304,
504
+ "grad_norm": 0.455078125,
505
+ "learning_rate": 0.0007,
506
+ "loss": 2.599566841125488,
507
+ "step": 1420
508
+ },
509
+ {
510
+ "epoch": 0.09706774519716886,
511
+ "grad_norm": 0.50390625,
512
+ "learning_rate": 0.0007,
513
+ "loss": 2.5753355026245117,
514
+ "step": 1440
515
+ },
516
+ {
517
+ "epoch": 0.09841590832490732,
518
+ "grad_norm": 0.478515625,
519
+ "learning_rate": 0.0007,
520
+ "loss": 2.563642883300781,
521
+ "step": 1460
522
+ },
523
+ {
524
+ "epoch": 0.09976407145264576,
525
+ "grad_norm": 0.44140625,
526
+ "learning_rate": 0.0007,
527
+ "loss": 2.5505178451538084,
528
+ "step": 1480
529
+ },
530
+ {
531
+ "epoch": 0.10111223458038422,
532
+ "grad_norm": 0.4609375,
533
+ "learning_rate": 0.0007,
534
+ "loss": 2.536166191101074,
535
+ "step": 1500
536
+ },
537
+ {
538
+ "epoch": 0.10246039770812268,
539
+ "grad_norm": 0.49609375,
540
+ "learning_rate": 0.0007,
541
+ "loss": 2.53826847076416,
542
+ "step": 1520
543
+ },
544
+ {
545
+ "epoch": 0.10380856083586114,
546
+ "grad_norm": 0.458984375,
547
+ "learning_rate": 0.0007,
548
+ "loss": 2.5285362243652343,
549
+ "step": 1540
550
+ },
551
+ {
552
+ "epoch": 0.1051567239635996,
553
+ "grad_norm": 0.486328125,
554
+ "learning_rate": 0.0007,
555
+ "loss": 2.5131874084472656,
556
+ "step": 1560
557
+ },
558
+ {
559
+ "epoch": 0.10650488709133805,
560
+ "grad_norm": 0.46484375,
561
+ "learning_rate": 0.0007,
562
+ "loss": 2.5331687927246094,
563
+ "step": 1580
564
+ },
565
+ {
566
+ "epoch": 0.10785305021907651,
567
+ "grad_norm": 0.439453125,
568
+ "learning_rate": 0.0007,
569
+ "loss": 2.500454139709473,
570
+ "step": 1600
571
+ },
572
+ {
573
+ "epoch": 0.10920121334681497,
574
+ "grad_norm": 0.42578125,
575
+ "learning_rate": 0.0007,
576
+ "loss": 2.4910287857055664,
577
+ "step": 1620
578
+ },
579
+ {
580
+ "epoch": 0.11054937647455342,
581
+ "grad_norm": 0.458984375,
582
+ "learning_rate": 0.0007,
583
+ "loss": 2.5048749923706053,
584
+ "step": 1640
585
+ },
586
+ {
587
+ "epoch": 0.11189753960229187,
588
+ "grad_norm": 0.408203125,
589
+ "learning_rate": 0.0007,
590
+ "loss": 2.4847084045410157,
591
+ "step": 1660
592
+ },
593
+ {
594
+ "epoch": 0.11324570273003033,
595
+ "grad_norm": 0.45703125,
596
+ "learning_rate": 0.0007,
597
+ "loss": 2.4694839477539063,
598
+ "step": 1680
599
+ },
600
+ {
601
+ "epoch": 0.11459386585776879,
602
+ "grad_norm": 0.4375,
603
+ "learning_rate": 0.0007,
604
+ "loss": 2.4789194107055663,
605
+ "step": 1700
606
+ }
607
+ ],
608
+ "logging_steps": 20,
609
+ "max_steps": 2500,
610
+ "num_input_tokens_seen": 0,
611
+ "num_train_epochs": 1,
612
+ "save_steps": 100,
613
+ "stateful_callbacks": {
614
+ "TrainerControl": {
615
+ "args": {
616
+ "should_epoch_stop": false,
617
+ "should_evaluate": false,
618
+ "should_log": false,
619
+ "should_save": true,
620
+ "should_training_stop": false
621
+ },
622
+ "attributes": {}
623
+ }
624
+ },
625
+ "total_flos": 5484682464460800.0,
626
+ "train_batch_size": 64,
627
+ "trial_name": null,
628
+ "trial_params": null
629
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1700/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7268161d90e74f9dd729f7bcc2b8dedb5818335ef614e6ed65860bf9d94d5aa
3
+ size 4920
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation": "sigmoid",
3
+ "architectures": [
4
+ "TinyLlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 2,
11
+ "head_dim": 32,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 128,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 256,
16
+ "max_position_embeddings": 512,
17
+ "mlp_bias": false,
18
+ "mlp_type": "glu",
19
+ "model_type": "tiny_llama",
20
+ "num_attention_heads": 4,
21
+ "num_hidden_layers": 100,
22
+ "num_key_value_heads": 4,
23
+ "pad_token_id": 0,
24
+ "pretraining_tp": 1,
25
+ "rms_norm_eps": 1e-06,
26
+ "rope_parameters": {
27
+ "rope_theta": 10000.0,
28
+ "rope_type": "default"
29
+ },
30
+ "tie_word_embeddings": true,
31
+ "tokenizer_name": "w-ahmad/tiny-stories-tokenizer",
32
+ "transformers_version": "5.16.0.dev0",
33
+ "use_cache": false,
34
+ "vocab_size": 4096,
35
+ "waleed_beta": 4
36
+ }
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21a56eadf2dd95d2c58cc2f39a060838aa5ad3ae4d4f0ea0df10d4133e2967c3
3
+ size 33967272
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49c96f3cebf53661c927d16262b334e246c0243f672dceca083e9f0a1e15938b
3
+ size 68504996
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d9cd6a0487226e5bd30d1846894c82af483733ab4381b75bae9c0745e05d405
3
+ size 14244
zain/Activation/out/glu-sigmoid-100L_run/checkpoint-1800/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8214bf4c471bc465884f1c219edb27abec42cbdf74a9a2fe9320000a4ce60c16
3
+ size 1064