Stewart Slocum commited on
Commit
7784c74
·
1 Parent(s): 81fc325

Add fine-tuned model

Browse files
README.md ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: togethercomputer/gemma-3-27b-it
3
+ library_name: peft
4
+ ---
5
+ ### Framework versions
6
+
7
+ - PEFT 0.15.1
adapter_config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "togethercomputer/gemma-3-27b-it",
5
+ "bias": "none",
6
+ "corda_config": null,
7
+ "eva_config": null,
8
+ "exclude_modules": null,
9
+ "fan_in_fan_out": false,
10
+ "inference_mode": true,
11
+ "init_lora_weights": true,
12
+ "layer_replication": null,
13
+ "layers_pattern": null,
14
+ "layers_to_transform": null,
15
+ "loftq_config": {},
16
+ "lora_alpha": 128,
17
+ "lora_bias": false,
18
+ "lora_dropout": 0.0,
19
+ "megatron_config": null,
20
+ "megatron_core": "megatron.core",
21
+ "modules_to_save": null,
22
+ "peft_type": "LORA",
23
+ "r": 64,
24
+ "rank_pattern": {},
25
+ "revision": null,
26
+ "target_modules": [
27
+ "k_proj",
28
+ "q_proj",
29
+ "up_proj",
30
+ "o_proj",
31
+ "gate_proj",
32
+ "down_proj",
33
+ "v_proj"
34
+ ],
35
+ "task_type": "CAUSAL_LM",
36
+ "trainable_token_indices": null,
37
+ "use_dora": false,
38
+ "use_rslora": false
39
+ }
adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ccbbf4616c085cad4b95f686ab0460ea6f7d40abf210d3d0ed2b60ee97c6a489
3
+ size 1816383024
special_tokens_map.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "boi_token": "<start_of_image>",
3
+ "bos_token": {
4
+ "content": "<bos>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "eoi_token": "<end_of_image>",
11
+ "eos_token": {
12
+ "content": "<eos>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false
17
+ },
18
+ "image_token": "<image_soft_token>",
19
+ "pad_token": {
20
+ "content": "<pad>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false
25
+ },
26
+ "unk_token": {
27
+ "content": "<unk>",
28
+ "lstrip": false,
29
+ "normalized": false,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ }
33
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e316dda1f34f410e1a27d086d70cbb20164153422fcae753e180a47b083c99b3
3
+ size 18452294
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
trainer_state.json ADDED
@@ -0,0 +1,504 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 66,
7
+ "global_step": 66,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.015151515151515152,
14
+ "grad_norm": 2.1806390285491943,
15
+ "learning_rate": 1e-05,
16
+ "loss": 2.259,
17
+ "step": 1
18
+ },
19
+ {
20
+ "epoch": 0.030303030303030304,
21
+ "grad_norm": 2.0509750843048096,
22
+ "learning_rate": 9.84848484848485e-06,
23
+ "loss": 2.1749,
24
+ "step": 2
25
+ },
26
+ {
27
+ "epoch": 0.045454545454545456,
28
+ "grad_norm": 1.945802092552185,
29
+ "learning_rate": 9.696969696969698e-06,
30
+ "loss": 2.1865,
31
+ "step": 3
32
+ },
33
+ {
34
+ "epoch": 0.06060606060606061,
35
+ "grad_norm": 1.7818379402160645,
36
+ "learning_rate": 9.545454545454547e-06,
37
+ "loss": 2.1723,
38
+ "step": 4
39
+ },
40
+ {
41
+ "epoch": 0.07575757575757576,
42
+ "grad_norm": 1.5140708684921265,
43
+ "learning_rate": 9.393939393939396e-06,
44
+ "loss": 2.0943,
45
+ "step": 5
46
+ },
47
+ {
48
+ "epoch": 0.09090909090909091,
49
+ "grad_norm": 1.168502688407898,
50
+ "learning_rate": 9.242424242424244e-06,
51
+ "loss": 1.9775,
52
+ "step": 6
53
+ },
54
+ {
55
+ "epoch": 0.10606060606060606,
56
+ "grad_norm": 0.9731884002685547,
57
+ "learning_rate": 9.090909090909091e-06,
58
+ "loss": 1.9921,
59
+ "step": 7
60
+ },
61
+ {
62
+ "epoch": 0.12121212121212122,
63
+ "grad_norm": 0.8054101467132568,
64
+ "learning_rate": 8.93939393939394e-06,
65
+ "loss": 1.9888,
66
+ "step": 8
67
+ },
68
+ {
69
+ "epoch": 0.13636363636363635,
70
+ "grad_norm": 0.6741383075714111,
71
+ "learning_rate": 8.787878787878788e-06,
72
+ "loss": 1.8612,
73
+ "step": 9
74
+ },
75
+ {
76
+ "epoch": 0.15151515151515152,
77
+ "grad_norm": 0.6076740026473999,
78
+ "learning_rate": 8.636363636363637e-06,
79
+ "loss": 1.7527,
80
+ "step": 10
81
+ },
82
+ {
83
+ "epoch": 0.16666666666666666,
84
+ "grad_norm": 0.6412762403488159,
85
+ "learning_rate": 8.484848484848486e-06,
86
+ "loss": 1.8131,
87
+ "step": 11
88
+ },
89
+ {
90
+ "epoch": 0.18181818181818182,
91
+ "grad_norm": 0.6954469680786133,
92
+ "learning_rate": 8.333333333333334e-06,
93
+ "loss": 1.8204,
94
+ "step": 12
95
+ },
96
+ {
97
+ "epoch": 0.19696969696969696,
98
+ "grad_norm": 0.7794247269630432,
99
+ "learning_rate": 8.181818181818183e-06,
100
+ "loss": 1.86,
101
+ "step": 13
102
+ },
103
+ {
104
+ "epoch": 0.21212121212121213,
105
+ "grad_norm": 0.811967134475708,
106
+ "learning_rate": 8.03030303030303e-06,
107
+ "loss": 1.8567,
108
+ "step": 14
109
+ },
110
+ {
111
+ "epoch": 0.22727272727272727,
112
+ "grad_norm": 0.7707127928733826,
113
+ "learning_rate": 7.87878787878788e-06,
114
+ "loss": 1.7976,
115
+ "step": 15
116
+ },
117
+ {
118
+ "epoch": 0.24242424242424243,
119
+ "grad_norm": 0.7628827691078186,
120
+ "learning_rate": 7.727272727272727e-06,
121
+ "loss": 1.7364,
122
+ "step": 16
123
+ },
124
+ {
125
+ "epoch": 0.25757575757575757,
126
+ "grad_norm": 0.7113344073295593,
127
+ "learning_rate": 7.5757575757575764e-06,
128
+ "loss": 1.7219,
129
+ "step": 17
130
+ },
131
+ {
132
+ "epoch": 0.2727272727272727,
133
+ "grad_norm": 0.6681413054466248,
134
+ "learning_rate": 7.424242424242425e-06,
135
+ "loss": 1.7664,
136
+ "step": 18
137
+ },
138
+ {
139
+ "epoch": 0.2878787878787879,
140
+ "grad_norm": 0.6175987720489502,
141
+ "learning_rate": 7.272727272727273e-06,
142
+ "loss": 1.6869,
143
+ "step": 19
144
+ },
145
+ {
146
+ "epoch": 0.30303030303030304,
147
+ "grad_norm": 0.5634602904319763,
148
+ "learning_rate": 7.121212121212122e-06,
149
+ "loss": 1.7165,
150
+ "step": 20
151
+ },
152
+ {
153
+ "epoch": 0.3181818181818182,
154
+ "grad_norm": 0.4820578694343567,
155
+ "learning_rate": 6.969696969696971e-06,
156
+ "loss": 1.6105,
157
+ "step": 21
158
+ },
159
+ {
160
+ "epoch": 0.3333333333333333,
161
+ "grad_norm": 0.5000921487808228,
162
+ "learning_rate": 6.818181818181818e-06,
163
+ "loss": 1.6858,
164
+ "step": 22
165
+ },
166
+ {
167
+ "epoch": 0.3484848484848485,
168
+ "grad_norm": 0.4802851676940918,
169
+ "learning_rate": 6.666666666666667e-06,
170
+ "loss": 1.6478,
171
+ "step": 23
172
+ },
173
+ {
174
+ "epoch": 0.36363636363636365,
175
+ "grad_norm": 0.45336592197418213,
176
+ "learning_rate": 6.515151515151516e-06,
177
+ "loss": 1.5984,
178
+ "step": 24
179
+ },
180
+ {
181
+ "epoch": 0.3787878787878788,
182
+ "grad_norm": 0.4636070132255554,
183
+ "learning_rate": 6.363636363636364e-06,
184
+ "loss": 1.613,
185
+ "step": 25
186
+ },
187
+ {
188
+ "epoch": 0.3939393939393939,
189
+ "grad_norm": 0.4598033130168915,
190
+ "learning_rate": 6.212121212121213e-06,
191
+ "loss": 1.6106,
192
+ "step": 26
193
+ },
194
+ {
195
+ "epoch": 0.4090909090909091,
196
+ "grad_norm": 0.46231794357299805,
197
+ "learning_rate": 6.060606060606061e-06,
198
+ "loss": 1.581,
199
+ "step": 27
200
+ },
201
+ {
202
+ "epoch": 0.42424242424242425,
203
+ "grad_norm": 0.4495490789413452,
204
+ "learning_rate": 5.90909090909091e-06,
205
+ "loss": 1.571,
206
+ "step": 28
207
+ },
208
+ {
209
+ "epoch": 0.4393939393939394,
210
+ "grad_norm": 0.43504828214645386,
211
+ "learning_rate": 5.7575757575757586e-06,
212
+ "loss": 1.4945,
213
+ "step": 29
214
+ },
215
+ {
216
+ "epoch": 0.45454545454545453,
217
+ "grad_norm": 0.4237779974937439,
218
+ "learning_rate": 5.606060606060606e-06,
219
+ "loss": 1.5492,
220
+ "step": 30
221
+ },
222
+ {
223
+ "epoch": 0.4696969696969697,
224
+ "grad_norm": 0.40998709201812744,
225
+ "learning_rate": 5.4545454545454545e-06,
226
+ "loss": 1.5383,
227
+ "step": 31
228
+ },
229
+ {
230
+ "epoch": 0.48484848484848486,
231
+ "grad_norm": 0.3991314470767975,
232
+ "learning_rate": 5.303030303030303e-06,
233
+ "loss": 1.5579,
234
+ "step": 32
235
+ },
236
+ {
237
+ "epoch": 0.5,
238
+ "grad_norm": 0.3889806866645813,
239
+ "learning_rate": 5.151515151515152e-06,
240
+ "loss": 1.5391,
241
+ "step": 33
242
+ },
243
+ {
244
+ "epoch": 0.5151515151515151,
245
+ "grad_norm": 0.382036030292511,
246
+ "learning_rate": 5e-06,
247
+ "loss": 1.5403,
248
+ "step": 34
249
+ },
250
+ {
251
+ "epoch": 0.5303030303030303,
252
+ "grad_norm": 0.3747563064098358,
253
+ "learning_rate": 4.848484848484849e-06,
254
+ "loss": 1.5569,
255
+ "step": 35
256
+ },
257
+ {
258
+ "epoch": 0.5454545454545454,
259
+ "grad_norm": 0.3503689765930176,
260
+ "learning_rate": 4.696969696969698e-06,
261
+ "loss": 1.4986,
262
+ "step": 36
263
+ },
264
+ {
265
+ "epoch": 0.5606060606060606,
266
+ "grad_norm": 0.34654778242111206,
267
+ "learning_rate": 4.5454545454545455e-06,
268
+ "loss": 1.4863,
269
+ "step": 37
270
+ },
271
+ {
272
+ "epoch": 0.5757575757575758,
273
+ "grad_norm": 0.3559574782848358,
274
+ "learning_rate": 4.393939393939394e-06,
275
+ "loss": 1.5173,
276
+ "step": 38
277
+ },
278
+ {
279
+ "epoch": 0.5909090909090909,
280
+ "grad_norm": 0.32772454619407654,
281
+ "learning_rate": 4.242424242424243e-06,
282
+ "loss": 1.4364,
283
+ "step": 39
284
+ },
285
+ {
286
+ "epoch": 0.6060606060606061,
287
+ "grad_norm": 0.3556043207645416,
288
+ "learning_rate": 4.0909090909090915e-06,
289
+ "loss": 1.4833,
290
+ "step": 40
291
+ },
292
+ {
293
+ "epoch": 0.6212121212121212,
294
+ "grad_norm": 0.3289327323436737,
295
+ "learning_rate": 3.93939393939394e-06,
296
+ "loss": 1.4412,
297
+ "step": 41
298
+ },
299
+ {
300
+ "epoch": 0.6363636363636364,
301
+ "grad_norm": 0.38381296396255493,
302
+ "learning_rate": 3.7878787878787882e-06,
303
+ "loss": 1.5059,
304
+ "step": 42
305
+ },
306
+ {
307
+ "epoch": 0.6515151515151515,
308
+ "grad_norm": 0.33431607484817505,
309
+ "learning_rate": 3.6363636363636366e-06,
310
+ "loss": 1.4532,
311
+ "step": 43
312
+ },
313
+ {
314
+ "epoch": 0.6666666666666666,
315
+ "grad_norm": 0.34528639912605286,
316
+ "learning_rate": 3.4848484848484854e-06,
317
+ "loss": 1.4567,
318
+ "step": 44
319
+ },
320
+ {
321
+ "epoch": 0.6818181818181818,
322
+ "grad_norm": 0.3294561803340912,
323
+ "learning_rate": 3.3333333333333333e-06,
324
+ "loss": 1.4228,
325
+ "step": 45
326
+ },
327
+ {
328
+ "epoch": 0.696969696969697,
329
+ "grad_norm": 0.32518258690834045,
330
+ "learning_rate": 3.181818181818182e-06,
331
+ "loss": 1.4301,
332
+ "step": 46
333
+ },
334
+ {
335
+ "epoch": 0.7121212121212122,
336
+ "grad_norm": 0.3199913203716278,
337
+ "learning_rate": 3.0303030303030305e-06,
338
+ "loss": 1.4139,
339
+ "step": 47
340
+ },
341
+ {
342
+ "epoch": 0.7272727272727273,
343
+ "grad_norm": 0.3203113377094269,
344
+ "learning_rate": 2.8787878787878793e-06,
345
+ "loss": 1.4344,
346
+ "step": 48
347
+ },
348
+ {
349
+ "epoch": 0.7424242424242424,
350
+ "grad_norm": 0.3189423680305481,
351
+ "learning_rate": 2.7272727272727272e-06,
352
+ "loss": 1.407,
353
+ "step": 49
354
+ },
355
+ {
356
+ "epoch": 0.7575757575757576,
357
+ "grad_norm": 0.3475227653980255,
358
+ "learning_rate": 2.575757575757576e-06,
359
+ "loss": 1.4637,
360
+ "step": 50
361
+ },
362
+ {
363
+ "epoch": 0.7727272727272727,
364
+ "grad_norm": 0.3078831136226654,
365
+ "learning_rate": 2.4242424242424244e-06,
366
+ "loss": 1.395,
367
+ "step": 51
368
+ },
369
+ {
370
+ "epoch": 0.7878787878787878,
371
+ "grad_norm": 0.3077380061149597,
372
+ "learning_rate": 2.2727272727272728e-06,
373
+ "loss": 1.3932,
374
+ "step": 52
375
+ },
376
+ {
377
+ "epoch": 0.803030303030303,
378
+ "grad_norm": 0.33048221468925476,
379
+ "learning_rate": 2.1212121212121216e-06,
380
+ "loss": 1.4297,
381
+ "step": 53
382
+ },
383
+ {
384
+ "epoch": 0.8181818181818182,
385
+ "grad_norm": 0.3076639175415039,
386
+ "learning_rate": 1.96969696969697e-06,
387
+ "loss": 1.3972,
388
+ "step": 54
389
+ },
390
+ {
391
+ "epoch": 0.8333333333333334,
392
+ "grad_norm": 0.31399327516555786,
393
+ "learning_rate": 1.8181818181818183e-06,
394
+ "loss": 1.4242,
395
+ "step": 55
396
+ },
397
+ {
398
+ "epoch": 0.8484848484848485,
399
+ "grad_norm": 0.29711833596229553,
400
+ "learning_rate": 1.6666666666666667e-06,
401
+ "loss": 1.3937,
402
+ "step": 56
403
+ },
404
+ {
405
+ "epoch": 0.8636363636363636,
406
+ "grad_norm": 0.30907464027404785,
407
+ "learning_rate": 1.5151515151515152e-06,
408
+ "loss": 1.4002,
409
+ "step": 57
410
+ },
411
+ {
412
+ "epoch": 0.8787878787878788,
413
+ "grad_norm": 0.3013785481452942,
414
+ "learning_rate": 1.3636363636363636e-06,
415
+ "loss": 1.3923,
416
+ "step": 58
417
+ },
418
+ {
419
+ "epoch": 0.8939393939393939,
420
+ "grad_norm": 0.31994277238845825,
421
+ "learning_rate": 1.2121212121212122e-06,
422
+ "loss": 1.4338,
423
+ "step": 59
424
+ },
425
+ {
426
+ "epoch": 0.9090909090909091,
427
+ "grad_norm": 0.3298662006855011,
428
+ "learning_rate": 1.0606060606060608e-06,
429
+ "loss": 1.4199,
430
+ "step": 60
431
+ },
432
+ {
433
+ "epoch": 0.9242424242424242,
434
+ "grad_norm": 0.317755788564682,
435
+ "learning_rate": 9.090909090909091e-07,
436
+ "loss": 1.4137,
437
+ "step": 61
438
+ },
439
+ {
440
+ "epoch": 0.9393939393939394,
441
+ "grad_norm": 0.29229098558425903,
442
+ "learning_rate": 7.575757575757576e-07,
443
+ "loss": 1.3405,
444
+ "step": 62
445
+ },
446
+ {
447
+ "epoch": 0.9545454545454546,
448
+ "grad_norm": 0.3137110471725464,
449
+ "learning_rate": 6.060606060606061e-07,
450
+ "loss": 1.4083,
451
+ "step": 63
452
+ },
453
+ {
454
+ "epoch": 0.9696969696969697,
455
+ "grad_norm": 0.30957749485969543,
456
+ "learning_rate": 4.5454545454545457e-07,
457
+ "loss": 1.4063,
458
+ "step": 64
459
+ },
460
+ {
461
+ "epoch": 0.9848484848484849,
462
+ "grad_norm": 0.3310662806034088,
463
+ "learning_rate": 3.0303030303030305e-07,
464
+ "loss": 1.4201,
465
+ "step": 65
466
+ },
467
+ {
468
+ "epoch": 1.0,
469
+ "grad_norm": 0.297313928604126,
470
+ "learning_rate": 1.5151515151515152e-07,
471
+ "loss": 1.3758,
472
+ "step": 66
473
+ },
474
+ {
475
+ "epoch": 1.0,
476
+ "eval_loss": 1.4051569700241089,
477
+ "eval_runtime": 4.459,
478
+ "eval_samples_per_second": 1.57,
479
+ "eval_steps_per_second": 0.224,
480
+ "step": 66
481
+ }
482
+ ],
483
+ "logging_steps": 1.0,
484
+ "max_steps": 66,
485
+ "num_input_tokens_seen": 0,
486
+ "num_train_epochs": 1,
487
+ "save_steps": 0,
488
+ "stateful_callbacks": {
489
+ "TrainerControl": {
490
+ "args": {
491
+ "should_epoch_stop": false,
492
+ "should_evaluate": false,
493
+ "should_log": false,
494
+ "should_save": true,
495
+ "should_training_stop": true
496
+ },
497
+ "attributes": {}
498
+ }
499
+ },
500
+ "total_flos": 3.3786133191275315e+17,
501
+ "train_batch_size": 1,
502
+ "trial_name": null,
503
+ "trial_params": null
504
+ }