Hugging Face's logo

klcsp
/
mistral-7b-0.3-gpt4o_100k_summarize-lora

PEFT
TensorBoard
Safetensors
mistral
alignment-handbook
trl
sft
Generated from Trainer
8-bit precision
bitsandbytes
Model card Files Metrics Community
chansung commited on
Commit
dd88f71
•
1 Parent(s): 1f53799

Model save

Browse files
Files changed (4) hide show
  1. README.md +69 -0
  2. all_results.json +9 -0
  3. train_results.json +9 -0
  4. trainer_state.json +358 -0
README.md ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: mistralai/Mistral-7B-v0.3
3
+ datasets:
4
+ - generator
5
+ library_name: peft
6
+ license: apache-2.0
7
+ tags:
8
+ - trl
9
+ - sft
10
+ - generated_from_trainer
11
+ model-index:
12
+ - name: mistral-7b-0.3-gpt4o_100k_summarize-lora
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
+ # mistral-7b-0.3-gpt4o_100k_summarize-lora
20
+
21
+ This model is a fine-tuned version of [mistralai/Mistral-7B-v0.3](https://huggingface.co/mistralai/Mistral-7B-v0.3) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 1.8227
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: 4
44
+ - eval_batch_size: 4
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - num_devices: 8
48
+ - gradient_accumulation_steps: 2
49
+ - total_train_batch_size: 64
50
+ - total_eval_batch_size: 32
51
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
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
+ | 0.742 | 1.0 | 219 | 1.8227 |
61
+
62
+
63
+ ### Framework versions
64
+
65
+ - PEFT 0.13.0
66
+ - Transformers 4.45.1
67
+ - Pytorch 2.4.1+cu121
68
+ - Datasets 3.0.1
69
+ - Tokenizers 0.20.0
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 6.128945867501076e+17,
4
+ "train_loss": 0.8279621209183784,
5
+ "train_runtime": 2699.8939,
6
+ "train_samples": 115376,
7
+ "train_samples_per_second": 5.187,
8
+ "train_steps_per_second": 0.081
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 6.128945867501076e+17,
4
+ "train_loss": 0.8279621209183784,
5
+ "train_runtime": 2699.8939,
6
+ "train_samples": 115376,
7
+ "train_samples_per_second": 5.187,
8
+ "train_steps_per_second": 0.081
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,358 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 219,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0045662100456621,
13
+ "grad_norm": 6.815864086151123,
14
+ "learning_rate": 9.090909090909091e-06,
15
+ "loss": 1.7729,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.0228310502283105,
20
+ "grad_norm": 5.818741798400879,
21
+ "learning_rate": 4.545454545454546e-05,
22
+ "loss": 1.7132,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.045662100456621,
27
+ "grad_norm": 2.6188273429870605,
28
+ "learning_rate": 9.090909090909092e-05,
29
+ "loss": 1.4372,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.0684931506849315,
34
+ "grad_norm": 1.4498921632766724,
35
+ "learning_rate": 0.00013636363636363637,
36
+ "loss": 1.2124,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.091324200913242,
41
+ "grad_norm": 1.0618371963500977,
42
+ "learning_rate": 0.00018181818181818183,
43
+ "loss": 1.0648,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.1141552511415525,
48
+ "grad_norm": 0.8222594857215881,
49
+ "learning_rate": 0.00019988558131018186,
50
+ "loss": 0.9718,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.136986301369863,
55
+ "grad_norm": 0.621496856212616,
56
+ "learning_rate": 0.00019918730395931649,
57
+ "loss": 0.9004,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.1598173515981735,
62
+ "grad_norm": 0.3935670554637909,
63
+ "learning_rate": 0.00019785874696801202,
64
+ "loss": 0.8609,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.182648401826484,
69
+ "grad_norm": 0.37850192189216614,
70
+ "learning_rate": 0.00019590835257019714,
71
+ "loss": 0.8519,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.2054794520547945,
76
+ "grad_norm": 0.33678698539733887,
77
+ "learning_rate": 0.00019334851442746664,
78
+ "loss": 0.8249,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.228310502283105,
83
+ "grad_norm": 0.5497208833694458,
84
+ "learning_rate": 0.00019019549887431877,
85
+ "loss": 0.8155,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.2511415525114155,
90
+ "grad_norm": 0.4304714798927307,
91
+ "learning_rate": 0.00018646934155473022,
92
+ "loss": 0.8072,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.273972602739726,
97
+ "grad_norm": 0.3267618417739868,
98
+ "learning_rate": 0.00018219372010688515,
99
+ "loss": 0.792,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.2968036529680365,
104
+ "grad_norm": 0.5633127689361572,
105
+ "learning_rate": 0.00017739580370507532,
106
+ "loss": 0.7879,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.319634703196347,
111
+ "grad_norm": 0.39324459433555603,
112
+ "learning_rate": 0.0001721060804148482,
113
+ "loss": 0.7944,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.3424657534246575,
118
+ "grad_norm": 0.3261910080909729,
119
+ "learning_rate": 0.0001663581634584641,
120
+ "loss": 0.7872,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.365296803652968,
125
+ "grad_norm": 0.3310278356075287,
126
+ "learning_rate": 0.0001601885776217367,
127
+ "loss": 0.7789,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.3881278538812785,
132
+ "grad_norm": 0.356475830078125,
133
+ "learning_rate": 0.0001536365271595212,
134
+ "loss": 0.7805,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.410958904109589,
139
+ "grad_norm": 0.6090440154075623,
140
+ "learning_rate": 0.0001467436466746814,
141
+ "loss": 0.7725,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.4337899543378995,
146
+ "grad_norm": 0.648769736289978,
147
+ "learning_rate": 0.0001395537365535585,
148
+ "loss": 0.7696,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.45662100456621,
153
+ "grad_norm": 0.3896843492984772,
154
+ "learning_rate": 0.00013211248463910262,
155
+ "loss": 0.7703,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.4794520547945205,
160
+ "grad_norm": 0.40256544947624207,
161
+ "learning_rate": 0.00012446717591027624,
162
+ "loss": 0.7577,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.502283105022831,
167
+ "grad_norm": 0.3406723141670227,
168
+ "learning_rate": 0.00011666639201255506,
169
+ "loss": 0.7572,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.5251141552511416,
174
+ "grad_norm": 0.4368165135383606,
175
+ "learning_rate": 0.0001087597025488413,
176
+ "loss": 0.7578,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.547945205479452,
181
+ "grad_norm": 0.45874515175819397,
182
+ "learning_rate": 0.00010079735009246167,
183
+ "loss": 0.7593,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.5707762557077626,
188
+ "grad_norm": 0.44529327750205994,
189
+ "learning_rate": 9.282993092381625e-05,
190
+ "loss": 0.7479,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.593607305936073,
195
+ "grad_norm": 0.40876489877700806,
196
+ "learning_rate": 8.490807351941753e-05,
197
+ "loss": 0.7486,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.6164383561643836,
202
+ "grad_norm": 0.3738647699356079,
203
+ "learning_rate": 7.708211683634112e-05,
204
+ "loss": 0.7478,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.639269406392694,
209
+ "grad_norm": 0.3480348587036133,
210
+ "learning_rate": 6.940179043641005e-05,
211
+ "loss": 0.7401,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.6621004566210046,
216
+ "grad_norm": 0.3515714108943939,
217
+ "learning_rate": 6.191589848274368e-05,
218
+ "loss": 0.7477,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.684931506849315,
223
+ "grad_norm": 0.4522218108177185,
224
+ "learning_rate": 5.467200961669619e-05,
225
+ "loss": 0.7447,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.7077625570776256,
230
+ "grad_norm": 0.331890344619751,
231
+ "learning_rate": 4.7716154685841944e-05,
232
+ "loss": 0.746,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.730593607305936,
237
+ "grad_norm": 0.4267323613166809,
238
+ "learning_rate": 4.109253424377772e-05,
239
+ "loss": 0.7525,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.7534246575342466,
244
+ "grad_norm": 0.34801220893859863,
245
+ "learning_rate": 3.4843237680415156e-05,
246
+ "loss": 0.7408,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.776255707762557,
251
+ "grad_norm": 0.3299520015716553,
252
+ "learning_rate": 2.9007975767533714e-05,
253
+ "loss": 0.7324,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.7990867579908676,
258
+ "grad_norm": 0.3563040494918823,
259
+ "learning_rate": 2.3623828319116748e-05,
260
+ "loss": 0.7487,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.821917808219178,
265
+ "grad_norm": 0.355241984128952,
266
+ "learning_rate": 1.8725008569947365e-05,
267
+ "loss": 0.7521,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.8447488584474886,
272
+ "grad_norm": 0.32830801606178284,
273
+ "learning_rate": 1.4342645769705977e-05,
274
+ "loss": 0.737,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.867579908675799,
279
+ "grad_norm": 0.32326367497444153,
280
+ "learning_rate": 1.0504587374062391e-05,
281
+ "loss": 0.7397,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.8904109589041096,
286
+ "grad_norm": 0.33771994709968567,
287
+ "learning_rate": 7.235222089726279e-06,
288
+ "loss": 0.7373,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.91324200913242,
293
+ "grad_norm": 0.33751899003982544,
294
+ "learning_rate": 4.555324897906132e-06,
295
+ "loss": 0.7426,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.9360730593607306,
300
+ "grad_norm": 0.34009331464767456,
301
+ "learning_rate": 2.4819250409651607e-06,
302
+ "loss": 0.7403,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.958904109589041,
307
+ "grad_norm": 0.3508203327655792,
308
+ "learning_rate": 1.0281978111449375e-06,
309
+ "loss": 0.7469,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.9817351598173516,
314
+ "grad_norm": 0.3122396469116211,
315
+ "learning_rate": 2.0338082897886079e-07,
316
+ "loss": 0.742,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 1.0,
321
+ "eval_loss": 1.8227200508117676,
322
+ "eval_runtime": 1.5688,
323
+ "eval_samples_per_second": 7.649,
324
+ "eval_steps_per_second": 0.637,
325
+ "step": 219
326
+ },
327
+ {
328
+ "epoch": 1.0,
329
+ "step": 219,
330
+ "total_flos": 6.128945867501076e+17,
331
+ "train_loss": 0.8279621209183784,
332
+ "train_runtime": 2699.8939,
333
+ "train_samples_per_second": 5.187,
334
+ "train_steps_per_second": 0.081
335
+ }
336
+ ],
337
+ "logging_steps": 5,
338
+ "max_steps": 219,
339
+ "num_input_tokens_seen": 0,
340
+ "num_train_epochs": 1,
341
+ "save_steps": 100,
342
+ "stateful_callbacks": {
343
+ "TrainerControl": {
344
+ "args": {
345
+ "should_epoch_stop": false,
346
+ "should_evaluate": false,
347
+ "should_log": false,
348
+ "should_save": true,
349
+ "should_training_stop": true
350
+ },
351
+ "attributes": {}
352
+ }
353
+ },
354
+ "total_flos": 6.128945867501076e+17,
355
+ "train_batch_size": 4,
356
+ "trial_name": null,
357
+ "trial_params": null
358
+ }