Hugging Face's logo

klcsp
/
gemma7b-lora-closedqa-11-v1

PEFT
TensorBoard
Safetensors
gemma
trl
sft
Generated from Trainer
Model card Files
xet
Metrics Community
chansung commited on
Commit
fde9ecf
·
verified ·
1 Parent(s): c2622a9

Model save

Browse files
Files changed (5) hide show
  1. README.md +69 -0
  2. all_results.json +9 -0
  3. runs/Nov17_19-50-40_main-lora-gemma7b-closedqa-0-0/events.out.tfevents.1731891735.main-lora-gemma7b-closedqa-0-0.455.0 +2 -2
  4. train_results.json +9 -0
  5. trainer_state.json +421 -0
README.md ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ license: gemma
4
+ base_model: google/gemma-7b
5
+ tags:
6
+ - trl
7
+ - sft
8
+ - generated_from_trainer
9
+ datasets:
10
+ - generator
11
+ model-index:
12
+ - name: gemma7b-lora-closedqa-11-v1
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # gemma7b-lora-closedqa-11-v1
20
+
21
+ This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 2.3991
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.0002
43
+ - train_batch_size: 8
44
+ - eval_batch_size: 8
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - num_devices: 8
48
+ - gradient_accumulation_steps: 2
49
+ - total_train_batch_size: 128
50
+ - total_eval_batch_size: 64
51
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
52
+ - lr_scheduler_type: cosine
53
+ - lr_scheduler_warmup_ratio: 0.1
54
+ - num_epochs: 1
55
+
56
+ ### Training results
57
+
58
+ | Training Loss | Epoch | Step | Validation Loss |
59
+ |:-------------:|:------:|:----:|:---------------:|
60
+ | 1.2456 | 0.9981 | 261 | 2.3991 |
61
+
62
+
63
+ ### Framework versions
64
+
65
+ - PEFT 0.13.2
66
+ - Transformers 4.46.2
67
+ - Pytorch 2.3.1+cu121
68
+ - Datasets 3.1.0
69
+ - Tokenizers 0.20.3
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9980879541108987,
3
+ "total_flos": 7.958354271676662e+17,
4
+ "train_loss": 4.106667400776655,
5
+ "train_runtime": 3226.2655,
6
+ "train_samples": 111440,
7
+ "train_samples_per_second": 10.373,
8
+ "train_steps_per_second": 0.081
9
+ }
runs/Nov17_19-50-40_main-lora-gemma7b-closedqa-0-0/events.out.tfevents.1731891735.main-lora-gemma7b-closedqa-0-0.455.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:16e9dd3f199eca125e832bff9116d50eac31a8a7532c6e12a8aa5b9836463a1b
3
- size 17757
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ede358d72ec58a42302a0fe58dcd77dd07376b72a858a746e56e936923d3d421
3
+ size 18382
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.9980879541108987,
3
+ "total_flos": 7.958354271676662e+17,
4
+ "train_loss": 4.106667400776655,
5
+ "train_runtime": 3226.2655,
6
+ "train_samples": 111440,
7
+ "train_samples_per_second": 10.373,
8
+ "train_steps_per_second": 0.081
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,421 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9980879541108987,
5
+ "eval_steps": 500,
6
+ "global_step": 261,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0038240917782026767,
13
+ "grad_norm": 109.6296157836914,
14
+ "learning_rate": 7.4074074074074075e-06,
15
+ "loss": 28.2707,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.019120458891013385,
20
+ "grad_norm": 73.76856231689453,
21
+ "learning_rate": 3.7037037037037037e-05,
22
+ "loss": 26.7709,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.03824091778202677,
27
+ "grad_norm": 27.987791061401367,
28
+ "learning_rate": 7.407407407407407e-05,
29
+ "loss": 22.5898,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.05736137667304015,
34
+ "grad_norm": 13.016111373901367,
35
+ "learning_rate": 0.00011111111111111112,
36
+ "loss": 18.9729,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.07648183556405354,
41
+ "grad_norm": 11.862752914428711,
42
+ "learning_rate": 0.00014814814814814815,
43
+ "loss": 16.2855,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.09560229445506692,
48
+ "grad_norm": 2.9698574542999268,
49
+ "learning_rate": 0.0001851851851851852,
50
+ "loss": 15.0127,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.1147227533460803,
55
+ "grad_norm": 4.096982955932617,
56
+ "learning_rate": 0.00019991889981715698,
57
+ "loss": 13.9569,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.1338432122370937,
62
+ "grad_norm": 6.29751443862915,
63
+ "learning_rate": 0.0001994237638847428,
64
+ "loss": 13.1471,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.15296367112810708,
69
+ "grad_norm": 13.995941162109375,
70
+ "learning_rate": 0.00019848077530122083,
71
+ "loss": 11.9192,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.17208413001912046,
76
+ "grad_norm": 23.425451278686523,
77
+ "learning_rate": 0.0001970941817426052,
78
+ "loss": 9.1546,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.19120458891013384,
83
+ "grad_norm": 24.60663604736328,
84
+ "learning_rate": 0.00019527022909596536,
85
+ "loss": 6.0786,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.21032504780114722,
90
+ "grad_norm": 9.548501014709473,
91
+ "learning_rate": 0.00019301713332493386,
92
+ "loss": 3.2243,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.2294455066921606,
97
+ "grad_norm": 6.405646324157715,
98
+ "learning_rate": 0.00019034504346103823,
99
+ "loss": 2.4261,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.248565965583174,
104
+ "grad_norm": 2.391706943511963,
105
+ "learning_rate": 0.00018726599588756145,
106
+ "loss": 2.1088,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.2676864244741874,
111
+ "grad_norm": 1.3095039129257202,
112
+ "learning_rate": 0.00018379386012185814,
113
+ "loss": 1.8208,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.28680688336520077,
118
+ "grad_norm": 1.0937951803207397,
119
+ "learning_rate": 0.00017994427634035015,
120
+ "loss": 1.6814,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.30592734225621415,
125
+ "grad_norm": 0.9703987240791321,
126
+ "learning_rate": 0.00017573458492761801,
127
+ "loss": 1.5943,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.32504780114722753,
132
+ "grad_norm": 1.7736315727233887,
133
+ "learning_rate": 0.00017118374836693406,
134
+ "loss": 1.5357,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.3441682600382409,
139
+ "grad_norm": 0.7995043992996216,
140
+ "learning_rate": 0.00016631226582407952,
141
+ "loss": 1.4938,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.3632887189292543,
146
+ "grad_norm": 0.689298152923584,
147
+ "learning_rate": 0.00016114208080920123,
148
+ "loss": 1.4527,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.3824091778202677,
153
+ "grad_norm": 1.3656119108200073,
154
+ "learning_rate": 0.00015569648233264394,
155
+ "loss": 1.4352,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.40152963671128106,
160
+ "grad_norm": 0.8344822525978088,
161
+ "learning_rate": 0.00015000000000000001,
162
+ "loss": 1.4039,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.42065009560229444,
167
+ "grad_norm": 0.9604928493499756,
168
+ "learning_rate": 0.00014407829351891857,
169
+ "loss": 1.3876,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.4397705544933078,
174
+ "grad_norm": 0.645470917224884,
175
+ "learning_rate": 0.00013795803711538966,
176
+ "loss": 1.3706,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.4588910133843212,
181
+ "grad_norm": 0.5904386639595032,
182
+ "learning_rate": 0.00013166679938014726,
183
+ "loss": 1.3562,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.4780114722753346,
188
+ "grad_norm": 0.6904798746109009,
189
+ "learning_rate": 0.00012523291908642217,
190
+ "loss": 1.3502,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.497131931166348,
195
+ "grad_norm": 0.8584344387054443,
196
+ "learning_rate": 0.00011868537753842051,
197
+ "loss": 1.3508,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.5162523900573613,
202
+ "grad_norm": 1.3949240446090698,
203
+ "learning_rate": 0.0001120536680255323,
204
+ "loss": 1.3284,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.5353728489483748,
209
+ "grad_norm": 1.0056356191635132,
210
+ "learning_rate": 0.00010536766297031215,
211
+ "loss": 1.318,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.5544933078393881,
216
+ "grad_norm": 0.9676538109779358,
217
+ "learning_rate": 9.865747936866027e-05,
218
+ "loss": 1.3218,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.5736137667304015,
223
+ "grad_norm": 0.8652806878089905,
224
+ "learning_rate": 9.195334312832742e-05,
225
+ "loss": 1.3008,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.5927342256214149,
230
+ "grad_norm": 1.193677306175232,
231
+ "learning_rate": 8.528545291682838e-05,
232
+ "loss": 1.2873,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.6118546845124283,
237
+ "grad_norm": 0.6716729998588562,
238
+ "learning_rate": 7.868384413205842e-05,
239
+ "loss": 1.2834,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.6309751434034416,
244
+ "grad_norm": 4.712409019470215,
245
+ "learning_rate": 7.217825360835473e-05,
246
+ "loss": 1.2933,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.6500956022944551,
251
+ "grad_norm": 1.2476301193237305,
252
+ "learning_rate": 6.579798566743314e-05,
253
+ "loss": 1.2727,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.6692160611854685,
258
+ "grad_norm": 0.7000300288200378,
259
+ "learning_rate": 5.957178011756952e-05,
260
+ "loss": 1.2674,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.6883365200764818,
265
+ "grad_norm": 0.6005345582962036,
266
+ "learning_rate": 5.3527682795623146e-05,
267
+ "loss": 1.2683,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.7074569789674953,
272
+ "grad_norm": 0.6676229238510132,
273
+ "learning_rate": 4.7692919235042255e-05,
274
+ "loss": 1.2739,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.7265774378585086,
279
+ "grad_norm": 0.5791189670562744,
280
+ "learning_rate": 4.209377202891212e-05,
281
+ "loss": 1.2686,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.745697896749522,
286
+ "grad_norm": 0.6084746718406677,
287
+ "learning_rate": 3.675546244046228e-05,
288
+ "loss": 1.2485,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.7648183556405354,
293
+ "grad_norm": 0.5911340117454529,
294
+ "learning_rate": 3.170203679431584e-05,
295
+ "loss": 1.2727,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.7839388145315488,
300
+ "grad_norm": 0.7844299674034119,
301
+ "learning_rate": 2.6956258160229695e-05,
302
+ "loss": 1.2634,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.8030592734225621,
307
+ "grad_norm": 0.6864188313484192,
308
+ "learning_rate": 2.2539503817234553e-05,
309
+ "loss": 1.2407,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.8221797323135756,
314
+ "grad_norm": 0.7077968716621399,
315
+ "learning_rate": 1.8471668960045574e-05,
316
+ "loss": 1.2416,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 0.8413001912045889,
321
+ "grad_norm": 0.6579145789146423,
322
+ "learning_rate": 1.4771077081496654e-05,
323
+ "loss": 1.2472,
324
+ "step": 220
325
+ },
326
+ {
327
+ "epoch": 0.8604206500956023,
328
+ "grad_norm": 0.8815127015113831,
329
+ "learning_rate": 1.1454397434679021e-05,
330
+ "loss": 1.2361,
331
+ "step": 225
332
+ },
333
+ {
334
+ "epoch": 0.8795411089866156,
335
+ "grad_norm": 0.8044998049736023,
336
+ "learning_rate": 8.536569946574546e-06,
337
+ "loss": 1.2467,
338
+ "step": 230
339
+ },
340
+ {
341
+ "epoch": 0.8986615678776291,
342
+ "grad_norm": 0.5612942576408386,
343
+ "learning_rate": 6.030737921409169e-06,
344
+ "loss": 1.2457,
345
+ "step": 235
346
+ },
347
+ {
348
+ "epoch": 0.9177820267686424,
349
+ "grad_norm": 0.6176737546920776,
350
+ "learning_rate": 3.948188836862776e-06,
351
+ "loss": 1.2367,
352
+ "step": 240
353
+ },
354
+ {
355
+ "epoch": 0.9369024856596558,
356
+ "grad_norm": 0.5503106117248535,
357
+ "learning_rate": 2.2983034998182997e-06,
358
+ "loss": 1.2424,
359
+ "step": 245
360
+ },
361
+ {
362
+ "epoch": 0.9560229445506692,
363
+ "grad_norm": 0.6087459921836853,
364
+ "learning_rate": 1.0885137906768372e-06,
365
+ "loss": 1.2377,
366
+ "step": 250
367
+ },
368
+ {
369
+ "epoch": 0.9751434034416826,
370
+ "grad_norm": 0.6323258876800537,
371
+ "learning_rate": 3.2426918657900704e-07,
372
+ "loss": 1.233,
373
+ "step": 255
374
+ },
375
+ {
376
+ "epoch": 0.994263862332696,
377
+ "grad_norm": 0.6436792612075806,
378
+ "learning_rate": 9.012214327897006e-09,
379
+ "loss": 1.2456,
380
+ "step": 260
381
+ },
382
+ {
383
+ "epoch": 0.9980879541108987,
384
+ "eval_loss": 2.3990681171417236,
385
+ "eval_runtime": 1.2567,
386
+ "eval_samples_per_second": 9.549,
387
+ "eval_steps_per_second": 0.796,
388
+ "step": 261
389
+ },
390
+ {
391
+ "epoch": 0.9980879541108987,
392
+ "step": 261,
393
+ "total_flos": 7.958354271676662e+17,
394
+ "train_loss": 4.106667400776655,
395
+ "train_runtime": 3226.2655,
396
+ "train_samples_per_second": 10.373,
397
+ "train_steps_per_second": 0.081
398
+ }
399
+ ],
400
+ "logging_steps": 5,
401
+ "max_steps": 261,
402
+ "num_input_tokens_seen": 0,
403
+ "num_train_epochs": 1,
404
+ "save_steps": 100,
405
+ "stateful_callbacks": {
406
+ "TrainerControl": {
407
+ "args": {
408
+ "should_epoch_stop": false,
409
+ "should_evaluate": false,
410
+ "should_log": false,
411
+ "should_save": true,
412
+ "should_training_stop": true
413
+ },
414
+ "attributes": {}
415
+ }
416
+ },
417
+ "total_flos": 7.958354271676662e+17,
418
+ "train_batch_size": 8,
419
+ "trial_name": null,
420
+ "trial_params": null
421
+ }